GUI Bugfixes: 2bc018e96 CHANGE: (#29966) - An easier way to see the correct update history for the server at the about windows BUGFIX: (#29868) - Report "Client Backup Summary" shows wrong / incomplete information wip: (#19001) - Restore only NEWER or OLDER files than on the target exist NEW: (#29856) - Support of LIS v2 in restore wizard CHANGE: (#29780) - Restore users are unable to see any savesets in Jaglion, because the default ACL for all media pools is set to deny CHANGE: Improve the presentation of the standard groups in the user and groups settings panels and dialogs. BUGFIX: (#29706) - Configuring VM with more than one SCSI controller leads to 'VM Exception: [com.vmware.vim25.InvalidArgument] caused by invalid property=deviceChange[1].device.key' CHANGE: (#29669) - VADP backup must clean up correctly when backup is canceled during snapshot creation BUGFIX: (#29653) - Resolving a message in notification center causes NPE NEW: (#29602) - Mount vSphere datastores using NFS v4 CHANGE: (#29547) - Restore of a VMware VM does not restore all properties of original one CHANGE: (#29547) - Restore of a VMware VM does not restore all properties of original one BUGFIX: (#29584) - MEDIA selection is not displayed for a media_event when reopening the media_event window. BUGFIX: (#29548) - MEDIA selection is not displayed for a media_event when reopening the media_event window. BUGFIX: (#29579) - VADP backup could be marked as successful, also when error occurs at parsing SGM file BUGFIX: (#29579) - VADP backup could be marked as successful, also when error occurs at parsing SGM file CHANGE: (#29555) - Deactivate "Initial Seed" button for Si3-NG data stores BUGFIX: (#29447) - Display schedule properties sometimes throws JavaException NEW: (#21845) - Front Side calculation must cumulate the clients backup front side together with the VM backup size. Set pool for saveset (result) with max(data_size) NEW: (#21845) - Front Side calculation must cumulate the clients backup front side together with the VM backup size. Order by name,task,media_pool NEW: (#21845) - Front Side calculation must cumulate the clients backup front side together with the VM backup size. Take client_id from tasks_clients to avoid use of hypervisor client_id which must not appear in report NEW: (#21845) Front Side calculation must cumulate the clients backup front side together with the VM backup size. New rythm template client-report-virtual-size.html. User has to substitutes client-report.html with the new one to get cumulated front side sizes of all tasks for that client - hypervisor and local Client's tasks with location or client report BUGFIX: (#27549) - Create SSL certificate for GUI client authentication by UI or CLI BUGFIX: (#29435) - Reset password via CLI does not work as expected BUGFIX: (#29420) - Configuring VM on VMware restore fails with VM Exception "invalid property=deviceChange[3].device.key" BUGFIX: (#29360) - When importing interfaces via the CLI, the provided list of interfaces is not validated NEW: (#29383) - Add option to set data mover for migrations on which sbc copy will be executed NEW: (#27549) - Create SSL certificate for GUI client authentication by UI or CLI NEW: (#24351) - Add option to execute FULL/DIFF/INC snapshot within vSphere host BUGFIX: (#29352) - When constructing the CLI examples, sometime command and object are the same BUGFIX: (#29342) - Exporting the credentials via CLI does not work BUGFIX: (#29341) - Removing a credentials set via CLI throws SQL exception when credentials set is still in use BUGFIX: (#29264) - Add the help for commands for sm_cmd NEW: (#29250) - DB:media.capacity is not visible in UI BUGFIX: (#26406) - Log information read by guest API during backup and restore NEW: (#29283) - Show DB:data_stores.fsck_status/fsck_msg in UI BUGFIX: (#29226) - Customer tries to delete a corrupt entry in the media table BUGFIX: (#28866) - It is not possible to type in a password with more than 27 characters into several access-tabs BUGFIX: (#29281) - VMware backup fails with 'Final state: Internal script error' when VM has a disk of type RAW/independend BUGFIX: (#29035) - Save Sesam key files in DB:credentials BUGFIX: (#29035) - Save Sesam key files in DB:credentials BUGFIX: (#29206) - Add data import from JSON file for clients, tasks and interfaces BUGFIX: (#29206) - Add data import from JSON file for clients, tasks and interfaces NEW: (#29201) - Add filter in Components/Mediapools to distinguish between media pool types BUGFIX: (#29036) - GUI client shows ONP exception when authentication mode of RMI mode changes BUGFIX: (#28023) - Add final status log line in sbc_vadp CHANGE: (29174) - Remove Note: "Using single file restore is not possible, when compression or encryption is enabled!" when activating compression for a BSR backup CHANGE: (#29194) - Add in DOWNLOAD LOG ARCHIVES in the last window the PATH where the archive is saved NEW: (#28693) - Refresh VM cache during browsing on demand CHANGE: (#29137) - Enhance handling of attachments when sending mail via GUI BUGFIX (#29179 ) - Don't cut of longer text columns in 'list' operations BUGFIX: (#29164) - Restore Wizard expert options: start time, time frame, duration and follow up settings have no effect NEW: (#29172) - Add DB:newday_events.options field to Newday Event Dialog NEW: (#29155) - Automatic task generation for path backups CHANGE: (#29000) - Remove filter items that don't show up NEW: (#29151) - Add media pool EOL to "Events as list" BUGFIX: (#29068) - Toggling replication type filter while still rendering shows results not matching the filter NEW: With new versions of Java (Java 8 >= build 292 and Java 11 >= build 11), the TLS version 1.0 and 1.1 are disabled by default. BUGFIX: (#29069) - Disabling the refresh while applying filters in 'Job State/Migrations and Replications' leads to wrong results BUGFIX: (#29003) - Job status tables does not load automatically when changing the view BUGFIX: (#29058) - GUI-262: "Save View" in Media table does not show results correctly wip: (#19001) - Restore only NEWER or OLDER files than on the target exist BUGFIX: (#29031) - Execution off filter in "Tasks as list" shows only tasks where both client and task are disabled BUGFIX: (#29092) - Mapping of drive numbers is wrong when changing the data store selection in manual 'Purge' or 'Clean Up' BUGFIX: (#29030) - When copying a backup task assigned schedules are not copied BUGFIX: (#29086) - If the repeat flag of a schedule is active the GUI calculates wrong end time of repetition CHANGE: (#29083) - Add the same 'delete' button in all events BUGFIX: (#29040) - Send button has no function in migration properties for "Main log" NEW: (#19001) - Restore only NEWER or OLDER files than on the target exist BUGFIX: (#29014) - start disaster_recovery from task bar fails BUGFIX - Refactored CLI client FileRetriever. Added explicit command-object-parameter definition BUGFIX - Fixed CLI parser: Filter objects ignored boolean value CHANGE: (#28979) - Remove "Hide migrated" checkbox and filter attribute BUGFIX: (#28645) - GUI offers backups with state 'd' for restore CHANGE: (#28902) - The source media pool isn't select able, if backup multiple savesets already exist BUGFIX: (#28943) - REST API end point '.../v2/backups' does not work as documented NEW: (#24731) - Changes to the Access Mode in the New Clients dialog NEW: (#27476) - Lock tape media in backup view NEW: (#25295) - Setting of alternative VFS mount path by GUI BUGFIX: (#28434) - If a BootOrderSection is present in ovf file the creation of the VM fails with error CHANGE: (#26436) - GUI browser hangs forever, if 'sm_client dir' command on Sesam client does not finish NEW: (#28200) - Show schedules tab for migration and replication tasks CHANGE: (#21915) - Deletion of an interface in client properties doesn't work CHANGE: (#24730) - Insert a new checkbox to configure a Virtualization server in the New Clients dialog NEW: (#28888) - Add a filter or search possibility to the command configuration window NEW: (#28901) - Backup type is missing in detailed view of a backup CHANGE: (#28898) - Rename 'Lotus Notes' backup type to 'HCL Domino' NEW: (#28882) - Add service pack information to HELP->ABOUT output BUGFIX: (#28866) - It is not possible to type in a password with more than 27 characters into several access-tabs BUGFIX: (#28864) - Backup of a virtual machine from an ESXi with restricted license results in a deadlock CHANGE: (#28750) - Adapt help for filter in expert options of restore wizard NEW: (#23804) - Delete all data of a tape from Sesam DB. Realized via a setting in the media pool CHANGE: (#27755) - User without permission to access the SEP sesam RTS is able to access drive-group and drive properties of attached tape-library. BUGFIX: Fixed error while list schedule CHANGE: (#26018) - Change EOL or remove multiple media at once BUGFIX: (#28803) - Any form of .zip .ZIP .ZiP .ZIp ... should be accepted by the GUI when reading a LIC file CHANGE: (#27741) - REST post request changes OS of client to LINUX, if not set BUGFIX: (#28744) - Unable to deselect files in selective restore if complete restore is attempted CHANGE: (#28695) - Restore wizard must adapt settings of migrated saveset CHANGE: (#28691) - Target node selection list is empty in restore wizard for OpenNebula restore BUGFIX: (#28629) - In case of type NSS restore wizard must not show a dialog box to switch to another type BUGFIX: (#28622) - In an existing backup event the source side dedup flag cannot be switched BUGFIX: (#28621) - Results of a command event are not show if opened over context menu BUGFIX (#28546) - No option to set "restore as path" in sm_cmd restore call BUGFIX (#28607) - Default local directory for downloading files needs to be handled on client side BUGFIX: (#28608) - Adding write only drive of another drivegroup in the datastore dialog triggers NPE BUGFIX (#28546) - No option to set "restore as path" in sm_cmd restore call BUGFIX (#28563) - "sm_cmd show log ... -B YESTERDAY" keeps printing the same day and appears to be stuck after a while BUGFIX (#28563) - "sm_cmd show log ... -B YESTERDAY" keeps printing the same day and appears to be stuck after a while BUGFIX (#28546) - No option to set "restore as path" in sm_cmd restore call BUGFIX (#28546) - No option to set "restore as path" in sm_cmd restore call CHANGE: (#28570) - GUI doesn't show active backup results entry, if results are opened over task context menu CHANGE: (#28559) - RMI RSS FeedFetcher should follow HTTP redirects instead of interpreting a redirect info site NEW: (#27881) - General multi task generation by GUI CHANGE: (#28510) - Preserve file extension '.log' of sm_gui_server logfiles before compression CHANGE: (#24729) - Insert a new checkbox to configure a Virtual machine in the New Clients dialog CHANGE: (#25132) - Unify data mover handling in clients and task dialog CHANGE: (#16399) - Offer only usable clients for data mover CHANGE: (#23397) - Current password must be requested if something changes in task encryption properties CHANGE: (#22132) - vsphere task creation in UI: create sm_pre_event_group_ per default and add sm_cmd as sample BUGFIX: (#28439) - Restore to original VM shows current cluster status "clustered=yes" in target VM name BUGFIX: (#28453) - The "Failed jobs report" may ends with error: "Illegal unquoted character ((CTRL-CHAR, code 13)): has to be escaped using backslash to be included in string value" CHANGE: (#28449) - Include name of Sesam Server in Rythm report "backups last 24h" BUGFIX (#28438) - CLI hangs with option "-o -1" for sm_cmd download log sesam CHANGE: (#28440) - Set HTTP response header 'X-Frame-Options' for REST interface CHANGE: Fixed non-resolver vApp field in VSphereParams BUGFIX (#28442) - CLI error message for object drivegroup/parameter default interface not meaningful BUGFIX (#28438) - CLI hangs with option "-o -1" for sm_cmd download log sesam BUGFIX (#28418) - CLI hangs with option "-o -1" for sm_cmd download log sesam NEW: (#28332) - Show release notes button on first start [after update] of GUI/WebUI BUGFIX: (#28352) - GUI freezes after selection of "Broken Tasks" options in restore wizard NEW: (#27105) - Deletion of data store: Show more information about related DB entries BUGFIX: Fix NPE in MediaPoolsDaoRestImpl if a media pool without no name exist NEW: (#28227) - Setting of 'ignore_finderr' backup option for Linux path tasks CHANGE: (#28376) - Deselecting single items at restore wizard does not work BUGFIX: (#28371) - Can't create data store if data store name matches target directory name BUGFIX: (#28381) - Menu option "Create new medium" must not be visible at all for DataStore at media pool overview BUGFIX (#24840) - Import/export of calendar BUGFIX (#28272) - Enabling Tape encryption on already existing mediapool sets crypt_key_media to plaintext value CHANGE: (#15934) - Improve display of meta information in client browser NEW: (#28345) - Global setting for case sensitive sort order NEW: (#27030) - Show backup coverage in client browser of task dialog CHANGE: fixed bug when calendar events import only firs one from file. GUI version BUGFIX (#24840) - Import/export of calendar NEW: (#26274) - Browsing of vSphere with same tree structure as in vCenter BUGFIX (#28301) - CLI help displays examples for some commands twice BUGFIX (#28308) - Force parameter missing for remove in "help client" BUGFIX (#28311) - Parameter declared in super class are not printed in help BUGFIX: (#28299) - Restore of a newly created VM with nvme-controller fails (svp-1980: Error: Exception: [InvalidController]) BUGFIX (#28263) - It is possible to add a notification with an invalid notification type BUGFIX (#28122) - CLI help: For commands "sm_cmd help list log" and "sm_cmd help log" parameters are not shown BUGFIX: (#28293) - Can't select '/' for backup source, excluding directories not possible in client browser BUGFIX: (#28294) - Client browser throws exception if top level entry is expanded BUGFIX: (#28184) - Keep target settings for migration tasks BUGFIX: (#28253) - When changing the settings on the first drive of a data store, the credential set must be kept NEW: (#28205) - Add API V2 endpoint /mediapools//ssdd to get SSDD enablement information BUGFIX: (#28242) - vSphere backup fails with NullPointerException just after trying to enable CBT though it is already enabled BUGFIX: (#28041) - GUI shows in task configuration for HANA a (useless) tab for HANA credentials BUGFIX: (#28108) - Sorting the columns in the 'restore wizard' worked until 4.4.3.72, but not since 4.4.3.79. NEW: (#28203) - Enhance the /datastores/find API endpoint to filter by driveGroupIds CHANGE: (#28222) - GUI does not set check boxes at VMware Single File Restore for elements in sub directory BUGFIX: (#28204) - After update to 4.4.3.84 SP1 it is not possible to assign a drive group to a DISK_CHANGE drive NEW: (#28221) - Allow IPV6 addresses as client name CHANGE: (#23374) - vSphere restore of VM with independent VMDK does not give a hint that not all original disks have been restored BUGFIX: (#28188) - Restore of vSphere VM template fails to recover VMDK data because of "Error: Exception: [NotSupported]" NEW: (#27985) - Show effective time period in migration window during which backups are searched BUGFIX: (#27818) - GUI download fails with: Failed to determine file name from response BUGFIX: (#27181) - GUI remote installation sets wrong user name and password NEW: (#27985) - Show effective time period in migration window during which backups are searched NEW: (#28074) - New V2 API for Restart and Immediate Start of Migrations/Replications Revert "CHANGE: (#28179) - UI must only allow to set maximal tape block_size 512K" CHANGE: (#28179) - UI must only allow to set maximal tape block_size 512K BUGFIX: (#28172) - GUI add always "all" as source Kopano independent of selection in client browser BUGFIX: (#28159) - Field 'Service Pack Package' in the CLIENT view must be renamed to 'Package'. There the Sesam Install Package is displayed NEW: (#28100) - Add end points to add and modify drive groups NEW: (#27381) Add help button to "Task Name Template" field in "Defaults" - "Extras" BUGFIX: (#28133) - SFR via mount: sm_copy_files is called with relocation for "disk0" but files to restore are on "disk1" BUGFIX: (#28050) When importing a new license via the UI, an exception occurs. Problem: License file is locked form another process, stacktrace will be be shown. Solution: Filter exception. NEW: (#28127) - Add drive group to media table NEW: (#28094) - Add end points to manage drive associations for drive groups CHANGE: (#28091) - Add the possibility in the GUI to refresh datastore information CHANGE: (#27870) Changes for Cli object 'drive': Add new parameters (-a , -T , -P , -M , -U , -E ) with unittest; remove deprecated examples (drivetype: DISK_HARD, DLT) and add RTX example BUGFIX: (#28073) - RHEV: restore selection in wizard is missing target nodes and general RHV related information BUGFIX: (#28006) - GUI Server wont start with recent OpenJDK Versions: SEP sesam DB is empty NEW: (#28078) - Hyper-V task with Resilient Change Tracking (RCT) CHANGE: (#23224) - Advanced and expert mode settings in GUI BUGFIX: (#28079) - Not possible to change or delete manually created user groups BUGFIX: (#28083) - Selecting Oracle backup source, task is set to type Path NEW: (#28043) - REST API: Add end points to add and modify media pools BUGFIX: (#27943) - No mail send possible with legal official certificate chain changed on the REST Server NEW: (#28042) - REST API: Add end points for dealing with notifications BUGFIX: (#28051) - Column for EOL in newday event is named EOM NEW: (#27694) - REST-API needed for the "Size History" chart of a datastore bugfix: UI cannot be started after moving client and server version to 5.0.0. Server sends "INSTALL_NEW" update recommendation. NEW: (#27967) Set GLBV, if key in DB:defaults starts with 'gv_' CHANGE: (#28021) - Restore Wizard changes for clustered Hyper-V VM to Hyper-V cluster NEW: (#20612) - Show backup tasks which are not scheduled or a part of a scheduled group in Monitoring/Last Backup state/ as "Not scheduled" BUGFIX: (#291) no mailer appears, when send mail from component 'Logging'. Problem: no log file will be found, mailer action will be canceled. Solution: set right file filter NEW: (#27936) - Add a tooltip for Linux backups in the GUI in TASK-PROPERTIES-ENCRYPTION&COMPRESSION why compression is not possible for CPIO BUGFIX: (#291) no mailer appears, when send mail at migrationbystatus dialog. Problem: no log file will be found, mailer action will be canceled. Solution: set right file filter with file type '.mig' CHANGE: (#291) show messagebox 'manually attach not/prt file' as attachment, when send mail at backup/restore results. NEW: (#27938) - Adds an info line in the task browser, that you can search by clicking in the expanded tree (also with wildcards) NEW: (#27895) - Optional popup window to edit DB:defaults in GUI defaults table BUGFIX: (#27947) - Time frame selection in restore wizard doesn't work CHANGE: (#27949) - Network must be selected during VMware restore to another vCenter BUGFIX: (#27957) - Deduplication ratio may could contain wrong values when result of calculation is very low BUGFIX: (#291) send mail show messagebox 'manually attach license file'. Problem: will be shown when send mail for all cases. Solution: show message only, when send mail from license dialog NEW: (#27935) - REST API endpoint "count" needed for findTasks BUGFIX: (#27764) - Automatic creation of backup tasks for RHEV virtual machines BUGFIX: (#27930) - Restoring a single file from a Hyper-V backup fails NEW (#26922) - Sending of 'CLI' command line to UI server using pipe or socket NEW: (#27730) - Add 'Create LOG Archiv' to the context menu in JOB STATE - BACKUPS / RESTORES / MIGRATIONS BUGFIX: (#27691) - New restore directory option doesn't work in GUI 4.4.3.79 BUGFIX: (#27839) - GUI dialog freezes when no drive is selected at task properties CHANGE: (#27841) - Adjusts all vm-task settings when the VMware type is selected BUGFIX: (#27748) - Do not allow to select a DS drive group for an exiting tape device BUGFIX: (#27746) - Multi selection in Components/Media is no longer possible BUGFIX: (#27751) - In COMPONENT - DATASTORES - SAVESETS the date range in the view is not saved BUGFIX: (#27752) - There are different messages when I try do delete a data store BUGFIX: (#27878) - Relocation of MSSQL database does not work via GUI, because it does not provide logical names NEW: (#27899) - Add DB:result_lbls.creation_time to 'Storage Locations' list BUGFIX: (#26777) If system load is very high, 'sm_cmd' may hang and blocks other processes. BUGFIX: (#26923) Remove VADP temporary files after backup and restore. Problem: remove action will not be done at end of 'try' operation, after exception will be thrown. Solution: change code, shift remove action to the end as finally action. BUGFIX: (#27869) - GUI: Hide "Create Media Pool" button in "New Data Store" Dialog NEW: (#27861) Add handler DB:migration_events.options BUGFIX: (#27876) - Media action properties are displayed with 'Mediensicherung' BUGFIX: When creating a new data store and switching between the HPE data store types, the HPE credentials tab is shown twice. NEW: (#27862) - Migation Tasks: Add "delete" flag column to migration task tree view NEW (#26922) - Sending of 'CLI' command line to UI server using pipe or socket BUGFIX: (#27866) - Master: after fresh installation sesam user is missing in database, sm_cmd operations fail if GUI authentication is enabled BUGFIX: (#27802) - "Auto update" Button in "Calendar Sheet" view disappears, when selecting another date BUGFIX: (#27800) - Shortcut "Alt + C" does not open "Cancel Current Running Activities" Dialog NEW (#26922) - Sending of 'CLI' command line to UI server using pipe or socket BUGFIX: (#27682) - Column 'locked' in saveset view of datastore shows locked state of wrong saveset if migration lays in datastore BUGFIX: (#291) Write license info to server failed. Problem: write file denied. Solution: allow write file to gv path 'gv_rw_tmp' at server side CHANGE: (#291) Some preparation for external mailer are incorrect. 'mail to' field is not set (extract from license info token 'eMail'); Change eMail message, remove message and set title as message, write whole licence info to external file (gv_rw_tmp) and inform user, that sm.ini and gui_license_info.txt must be added as attachment to mail. NEW (#26922) - Sending of 'CLI' command line to UI server using pipe or socket CHANGE: (#27464) Show information message in the first row. BUGFIX: (#27838) - Drives with access mode 'read only' must not be shown at task properties NEW: (#23210) - More information in the GUI - EOL information in overviews/select menus BUGFIX: (#18096) option 'SEP sesam for Recovery Pro' for type 'Exchange' is hidden at restore wizard, although type is 'Exchange Server'. Problem: option will not be shown. Solution: show as disabled option. NEW (#26922) - Sending of 'CLI' command line to UI server using pipe or socket NEW: (#27812) - Set DB:migration_tasks.options by GUI BUGFIX: (#27813) - sm_cmd start migration set DB:migrated_flag='0' BUGFIX: (#27806) - Setting VM server type to RHV-Management-Server sets the Access mode to SM_SSH BUGFIX: (#27804) - Adding vcenter fails with Password Invalid Exception BUGFIX: (#27745) - NPE when trying to immediate start a media action and no loaders are configured on the Sesam Server BUGFIX: add tooltip and improve text for backup lock state checkbox BUGFIX: (#27783) - NPE comes up if you click on an empty field in the Notification Center, in the Resolved Messages BUGFIX: (#27765) - It is not possible to create a schedule with start date in past BUGFIX: (#27722) - Hide S3 tab for Si3-NG as long it is not supported BUGFIX: (#27724) Add selective restore task not possible, always full. Problem: sel file will not be transferred to server. Solution: Overwrite method 'create', read the content of the sel file and create it on server side. NEW (#26922) - Sending of 'CLI' command line to UI server using pipe or socket CHANGE: (#291) Auto create or update "Sales" account if missing (External Mailer); Get the "Sales" email address from the "eMail" line in the body; Select "Sales" account (External Mailer); Add Attachment sm.ini. BUGFIX: (#27761) - GUI shows too small throughput value for Si3 replications because of mixed up units BUGFIX: (#27607) - The environment variable (gv_prot/gv_dayfile) handling is no longer correct in 4.4.3.79 with sm_smtp. CHANGE: (#27667) Set correct information message (german) for datastore action 'purge' CHANGE: (#27464) Add information, what purge means to the purge-window (as it is realized in 'Clean Up'-window) BUGFIX: (#27718) - Setting media write protection throws exception NEW (#27359) - Add task parameter for 'list restart' and 'restart' CHANGE: (#27667) Hide mount restore functionality for BSR Windows NEW (#27568) - Allow CLI commands and parameter to be defined as 'hidden' BUGFIX: Rearranged labels in migrationResultsPanel for proper layout BUGFIX: (#27682) - Column 'locked' in saveset view of datastore shows locked state of wrong saveset if migration lays in datastore BUGFIX: (#18096) handling of option 'SEP sesam Recovery pro for Exchange' is incorrect. Problem: incorrect browser url for help and option is enabled, when backup is only available on tape. Solution: use right url and set right logic for option BUGFIX: (#27639) - Limit maximal streams to be configured for a RDS NEW: (#25521) - Extend information drives view BUGFIX: (#27642) - Selecting the relocation for MSSQL restore leads to a NullPointerException BUGFIX: (25711) - Remove 'Auto Unload' in OPTIONS in drive properties or make it only visible with expert options CHANGE: (#26405) - Speed up the 'vm host' selection - in larger environments this may take >5 minutes until the selection is complete CHANGE: (#27625) - GUI server takes too long to spin up during installation, resulting in incomplete configuration and failed test-backup BUGFIX: (#27621) - Option 'Monitoring > Drives' fails with 'Missing parameters' when selecting a particular drive BUGFIX: (#27605) - "sm_update_client" not work if a Location is selected as base BUGFIX: (#27609) - sm_cmd unload drive 39 is not possible because of 2 a digits drive number NEW: (#15393) - Meta task group in GUI with all jobs not in any group CHANGE: (#26002) - Manual write protection setting at properties of DataStore/Si3 must also change setting for child savesets BUGFIX: (#27607) - The environment variable (gv_prot/gv_dayfile) handling is no longer correct in 4.4.3.79 with sm_smtp. BUGFIX: (#27604) - It's not possible to switch from a custom update source to the default package source BUGFIX: (#27466) - Hide "Exchange Access" tab completely and show "restore granularity" options in expert mode only BUGFIX: (#15298) selection of drive at loader content table will not be taken as initialization for loader action dialog. Problem: only PK of drive is set. Solution: use deep object of drive. NEW: (#27520) - New V2 API for media pools and interfaces BUGFIX: (#27555) - After creating and deleting OS-Access data in a job, you can make changes in the job, but you can no longer accept or save them without the exception CHANGE: With DB authentication enabled, the REST server logs clear text passwords BUGFIX: (#26675) - Check for exiting path during creation of a new data store isn't done on RDS BUGFIX: (#27543) - Drives for loader actions aren't updated, if loader has been changed BUGFIX: (#27532) - Private keys generated by GUI are world readable NEW: (#24614) - Add a column in schedules & show the right information for migrations (followUp) CHANGE: (#26626) Misleading note ''Single file restore (Disabled, because backup format '...' isn't supported)''; Improve the single file restore panel at restore wizard by tasktype. Show only 'Single file restore' panel for task types which are not bases on Netware. CHANGE: (#25781) - Handling of 'COPY' backup level for vSphere without IMG option set BUGFIX: (#27492) - When I open a medium (tape) with an expired EOL, DELETE MEDIA is grayed out. Only when I click on Properties 2 and again on Properties 1, then the deletion works. CHANGE: (#27454) - Replace certificate and key for WebUI access CHANGE: (#27471) - RW: Show "Save as Template" only when explicitly enabled CHANGE: (#26626) Improve note ''Single file restore (Disabled, because backup format '...' isn't supported), add highlighting for note (blue font); Show note for all task types based on 'Netware' BUGFIX: (#27458) - Restore of vSphere VM using VxRail could fail with 'Size of file .ovf is 0 byte' CHANGE: (#27373) - Adds an hour glass when a data store is deleted BUGFIX: (#27409) - Datastore saveset overview view stops at 99% BUGFIX: (#27448) - Immediate of backup does not work anymore due to ONPE BUGFIX: (#27439) - Context menu of external backups must not contain option to immediate start it NEW: (#27362) - Preset values for new drive, if called by context menu of existing drive or drive group NEW: (#27211) - Restart failed backup directly in backup results dialog CHANGE: (#27054) Redesign archive adjustment dialog; show only required fields at dialog by loader ctrl type, hide field 'storage pool' for real tape loaders CHANGE: (#27420) - Immediate start of schedule leads to "duplicate key value" DB error BUGFIX: (#27398) - Not all / wrong interfaces selectable during migration task / event creation in case of shared drives NEW: (#27029) Task by clients add the column "Backup Group" CHANGE: (#27253) Show shared drives in loader content dialog; The drive that loaded the tape is displayed in bold. BUGFIX: (#27408) - If you look at the properties of a backup and its e.g. 2 migrations in COMPONENT DATASTORES-SAVESETS, they are identical for all of them CHANGE: (#27367) - Restrict name of drive groups BUGFIX: (#27403) - Excel export in DATASTORE-SAVESETS is not correct BUGFIX: (#27406) - When starting an event or schedule, sometime the events below a schedule disappear BUGFIX: (#27397) - It's not possible to "Immediate Start" a schedule CHANGE: (#27404) - Access mode not set to PROXY when switching VM server type to 'vcenter' BUGFIX: (#26701) - Warning message, if last saveset of a backup will be expired BUGFIX: (#25970) - DS Creation: path should be mandatory for all datastore types accept HPE Store NEW: (#26209) - sm_cmd list restore/migration/media results BUGFIX: (#27217) - If a tape medium is to be deleted, you must not check whether the init drive still exists BUGFIX: (#27390) - GUI does not allow to set the default day range to '-1' for unlimited BUGFIX: (#27383) - The combo boxes 'VM host' and 'VM name' in 'New Client' window are not correctly initialized NEW: (#27360) - Create migration task name by media pool names NEW: (#27385) - Add REST V2 API count end points when searching for backups/restores/migrations NEW: (#27137) - Add REST API endpoints for backup, restore and migration tasks - added new API endpoints: BUGFIX: (#27378) - If GUI is opened in German the sort order of the migration task delete flag combobox is wrong BUGFIX: (#27346) - Selection of source interface during immediate start of migration for a backup BUGFIX: (#27326) - It isn't possible to do a follow-up event of a sm_cmd start command BUGFIX: (#27331) - GUI must not allow expiration of saveset as long as it's in active or queued state BUGFIX: (#25067) - sm_cmd start inventory - parameter -n (barcode adjustment) and -f (fast mode) are disfunctional CHANGE: (#16661) - Rename all referring command events if command is renamed BUGFIX: (#26926) - Refactoring of CLI help: extra option parameters and ressource texts are missing NEW: (#26209) - sm_cmd list restore/migration/media results CHANGE: (#24784) - Adjust all dependencies if a client entered as data mover is to be deleted or renamed. CHANGE: (#27229) - Show checkbox 'HPE Bandwidth Optimized Data Transfer' with grey color and tool tip to inform customer BUGFIX: (#27323) - NPE when opening properties of a Hyper-V restore task without existing savesets CHANGE: (#27277) - Change default time range for migration tasks BUGFIX: (#27307) - RW: Show disaster recovery checkbox in Advanced mode BUGFIX: (#27234) - When opening tab 'Actions' at DataStore just actions of current sesam day are shown, though filter is set different BUGFIX: (#27280) - The GUI displays incorrect information after uninstalling a client via the GUI CHANGE: (#27309) - New data store: Create one drive with write and 2nd with read allocation CHANGE: #(26884) Missing field to enter DB:schedules.user_comment. You could not enter a user comment and it's also not visible, even the DB field exists. BUGFIX: (#26926) - Refactoring of CLI help: extra option parameters and ressource texts are missing BUGFIX: (#27302) - List vSphere snapshots via CLI works correctly, but error message is written to exception log of GUI server CHANGE: (#27156) BUGFIX: (#27299) - Do not use default date range filter for backup results in media properties CHANGE: (#27266) - Remove task from task group in group tree by context menu doesn't update tree CHANGE: (#26628) - Improve command event function CHANGE: (#26826) - Use DB:media_capacity to show size of medium CHANGE: (#27268) swap position of tables port and changer in loader content dialog CHANGE (#27267) - Extent delete flag for migration/events to switch immediate purge BUGFIX: (#26926) - Refactoring of CLI help: extra option parameters and ressource texts are missing BUGFIX: (#15298) change: Improve Loader Content Visualization; add square icons at legend dialog; remove header line of tables; use font bolt for first column header; set media action to 'export', when only slots are selected; CHANGE: (#27195) - Disable 'Decent other FS' and 'Save NFS' options in task dialog, if backup over LVM is selected BUGFIX: (#26938) Infinite attempts to create VM snapshots even though the user has no rights to do this and the number of attempts is 2. Problem: Internal exception 'NoPermission' will not be handled. Solution: stop snapshot creation with clean error and terminate whole vm process. CHANGE: (#27171) - Remove "do you want to create new media pool" popup during DS creation BUGFIX: (#27202) - Print human readable error message in case the host name resolution fails BUGFIX: (#24844) set correct initialization time for spinner 'task canceled after' at tab 'execution time' for run immediately dialogues BUGFIX: Reports are empty if "-L " is specified on command line. BUGFIX: (#27173) - KVM restore fails with "Unable to define VM: "XML error: name hda:/vm/maxdb01.img cannot contain / " BUGFIX: (#15298) wrong initialization for loader action dialog. Problem: when first select a port, than a slot, port and slot fields take same value slot. Solution: get value directly from port/slot table without listener. BUGFIX: (#27155) Restore Wizard don't submit date restrictions in search function (sm_search_lis). Problem: convert String from date combobox 'from/to' to date 'Feb 25, 2020)' failed. Solution: add date pattern 'MMM dd, yyyy' for string->date converter. BUGFIX: (#27121) - Backup of vSphere VM fails due to NPE when export of OVF is not permitted BUGFIX: (#27125) - 27125 - Column "Last Successfull run" and "Last Full Backup" don't use the correct date-time format BUGFIX: (#27134) - Restart Exchange Server restore fails with NPE CHANGE: (#15298) Improve Loader Content Visualization, add column 'barcode' to loadercontent table, rework table/model structure BUGFIX: (#27131) - NPE in RW Exchange try to browse path for RDB BUGFIX: (#27127) General restore of vSphere VM fails with 'VM Exception: [Hostsystem does not exist]'. Problem: vcenter server name will be used for hostsystem check. Solution: use parameter -a server= NEW: (#27102) - No option to set the data mover for a backup task BUGFIX: (#27111) - Backup task exclude maximum length not in sync with DB BUGFIX: (#23988) - Error create command because of length mismatch "clients" -> "name" and "commands" -> "host" fields BUGFIX: (#14086) Port table at loader content dialog show wrong data. Problem: filled data of the tables port, transport are interchanged. Solution: fill data to corresponding table BUGFIX: (#26875) add mount points to all mountable task type types with mount option is set (param -d ymountpath> for command 'sm_copy_files' CHANGE: (#22562) Filter options in RW: Allow only one pattern per line and disable ', ' separated items CHANGE: (#27077) - After opening the day log, the GUI doesn't respond for minutes BUGFIX: (#27075) - It is not possible to select all media pools, when creating a 'new media event' in a schedule BUGFIX: (#27043) - GUI does not show whole restore log if it is compressed BUGFIX: (#27027) - IllegalArgumentException at Java restore wizard, if source contains a percent sign CHANGE: (#27007) - Immediate Start: Replication and Command - show right dialog and hide fields "Cancel execution after/at" BUGFIX: (#27006) - Immediate Start: Migration - Do not show replication tasks in task selection BUGFIX: (#27005) - RW: Show buttons "Expert Options" and "Save as Template" only for UI mode "Expert" BUGFIX: (#27021) - NPE when folder without content is expanded at restore wizard BUGFIX: (#26973) - Wrong message "Virtual disk successfully saved" is sent to Sesam DB, VMware VMDK backup failed BUGFIX: (#26964) - Longer values than allowed by sesam DB can be entered in credential dialogs BUGFIX: (#26968) - If 'Automatic introduction' in media event for archive adjustment is selected, some graphical control elements don't work correctly BUGFIX: (#26954) - NPE during delete Action of restore task BUGFIX: (#26942) - Removing restore option "-o verify=list" fails BUGFIX: (#26946) - The restore item window does not show the media pool value BUGFIX: (#22036) - Make the restore hooks in the current restore-agent consistent BUGFIX: (#26916) - Media pool tree table sorts EOL after lexical order instead of number value BUGFIX: (#26914) - An exception comes up, when the customer tries to change expert options in the restore wizard BUGFIX: (#26878) - Changing count of streams of DISK_STORE drive does not start reconfiguration NEW - (#26847): Add data store name to 'Drives' view CHANGE: Execute the "Initial Seed" in the background and show a infinite progress bar during the process. BUGFIX: Next execution date calculated wrong when a schedule has "Repeat with end time" set. CHANGE: (#26864) - Clarify the selection of sesam dates in restore wizard BUGFIX: (#26894) - Source prefix "ONE:" not stripped from backup source, if an OpenNebula task is created BUGFIX: (#26677) - When adding a new client HTTP and HTTPS interface is not created automatically anymore BUGFIX: (#26872) - Filter on vApp doesn't work for sm_cmd dir vsphere BUGFIX: (#26870) - Automatic creation of vSphere tasks fails with 'Missing Field de.sep.sesam.model.Clients.operSystem.' BUGFIX: (#26804) - Quitting Data-Store Dialog without any changes results in misleading error message BUGFIX: (#26850) - CLI ignores option '-o' at 'generate vsphere' NEW: (#26832) - Add the possibility to delete/uninstall a client by GUI BUGFIX: (#26840) - VMware task creation error if task for VM already exist BUGFIX: (#26042) - In some cases the labels of the 'Refresh', 'Properties' and some other buttons are cut off BUGFIX: (#26752) - Type filter for table "All Results" - correction and rework NEW: (#26613) - Filter to hide all clients, which are backup clients only CHANGE: (#26648) - Column "Total size" does not work in view mode "Group by Task" BUGFIX: (#26825) - Could not clear loader and slot in Media Properties BUGFIX: (#26815) - An exception comes up, when an old systemstate backupevent/task exists with typ F, another db field is changed (interface) and the task/event is attempted to save BUGFIX: (#26824) - Tab "Savesets" in data store properties shows savesets which are no longer in the data store BUGFIX: (#26791) - CLI option to filter VMware-DataStores just shows VMs located on one VMware-DataStores BUGFIX: (#26810) - VMware DataCenter must be enclosed by double quotes BUGFIX: (#26811) - Exception in data store dialog when creating a data store without any drives CHANGE: (#26557) - Limit number of backups shown in GUI NEW: (#26785) - Add in the 'Events as List' the column -> 'Migration Task' BUGFIX: (#26726) - Changing the saveset EOL in JOB STATE - BACKUPS of a saveset to the future does not change the Backup EOL in the GUI BUGFIX: (#26774) - Saveset EOL is not changed CHANGE: (#26682) - Hide SEP sesam Install option on Location BUGFIX: (#26748) - Excel export from media table - column write protection is empty BUGFIX: (#26754) - Restore from mounted saveset with relocation always uses the original path as relocation target BUGFIX: (#26755) - Remove leading double quote from relocation source and target (mounted saveset only) BUGFIX: (#26725) - Changes in data store properties are only visible after reopening GUI BUGFIX: (#26742) - UI (client side) crash if try to delete tape media and drive no more exists BUGFIX: (#26743) - Operation Not Possible when try to delete a external group in "Permission Management" CHANGE: (#26737) - Shortcut "Alt+G" for GUI update does not work BUGFIX: (#26738) - Import Sesam DB: File browser does not allow selection of DB file BUGFIX: (#26642) - Single item restore from mounted or attached vSphere saveset fails with: "Data center is not set" CHANGE: (#26721) - Change text in Configuration->Defaults->Install/Update drop down menu update mode CHANGE: (#26722) - Change "Update UI Server" to "Install UI Server Service Pack" BUGFIX: (#26716) - Mask 'Manage VM tasks' does not generate Group flag option BUGFIX: (#26691) - LDAP User Authentication does not work when no group search base is specified BUGFIX: (#26674) - Set a filter (e.g. loader 2) in tab MEDIA and press REFRESH then all displayed lines are deleted BUGFIX: (#26672) - Creating new user with group operator or restore failed. BUGFIX: (#26698) - Mask at 'Manage VM tasks' does not apply filters 'DataStore' and 'ESX Server' for CLI command CHANGE: (#26686) - Date range configuration for "Restart Backups" or "Restart Migrations" BUGFIX: (#26702) - Deleting a mediapool or media inside datastore properties -> media results in a NPE BUGFIX: (#26706) - Drive number in loader cannot be edited in the "Manage multiple drives" window BUGFIX: (#26705) - Filter "Backup Execution Off" has no function in "Components" - "Clients" tab CHANGE: (#26708) - Remove the two options in the BSR task, because both are no more necessary in Beefalo V2 and newer BUGFIX: (#26683) - Immediate start of media action "Archive Adjustment" causes NPE BUGFIX: (#26697) - Saving main.log from tab "Migrations and Replications" fails without any error message BUGFIX: (#26700) - Help functions inside data store properties redirects to non existing web site BUGFIX: (#26679) - VMDK is missing on target VM after vSphere IR BUGFIX: (#26684) - Changing a value in Configuration - Defaults - Settings not applied BUGFIX: (#26688) - NPE during RMI reload on externally opened GUI BUGFIX: (#26676) - When clicking at "Create command" at "Manage VM tasks" an NPE occurs BUGFIX: (#26647) - Column "Saveset EOL" in "Job State -> Backups" cannot be hidden CHANGE: (#26646) - Wrong summary - Jobstate Backups - Overall duration - in view mode - Group by task BUGFIX: (#26662) - Java RW shows no backups, because of wrong SQL statement NEW: (#25407) Add a new combobox to "VM Restore Options" for adding virtual machine to cluster after restore BUGFIX: (#26650) - Add an alternate text to status images, in case the HTML clients not support SVG tag wip: Add support for VMDK and EXCHANGE mount/umount to V2 mount service BUGFIX: (#26632) - Duplicate function in context menu for properties of sub element of task event CHANGE: (#18979) - Exception adding a client, if client name already exists in DB: interfaces BUGFIX: (#26572) - Filter panel for Job State/Migrations and Replications disfunctional for section Type BUGFIX: (#26588) - When browsing for user defined source folder the client file system is browsed instead of the required SEP sesam server file system BUGFIX: (#26587) - User defined source for update packages can only be set the first time - the source cannot be modified BUGFIX: (#26513) - Restore fails if top tree item selected BUGFIX: (#26567) - Schedules->Media Event->Init offers wrong labels CHANGE: (#26562) - All update/install options must be switchable in the GUI CHANGE: Improve "By Status" view performance BUGFIX: (#26558) - Taskgroup->Schedules: Exception, if media pool is changed CHANGE: Allow to switch from "DB:count_results" to "DB:results" via "gui.force_results_table" switch in "DB:defaults" BUGFIX: (#25941) - For role restore and operator: prevent access to results without task or client connection CHANGE: (#26512) - GUI results/savesets .. has to show the complete size of a VM tasks not only the KB size of the config BUGFIX: (#26462) - "html" folder missing when upgrading SEP sesam BUGFIX: (#26429) - Restore wizard must show short information when backup chain is inconsistent BUGFIX: (#26412) - SFR Hyper-V does not work BUGFIX: (#26495) - Restore of VMware machine fails with "Invalid configuration for device" when video size of VM was 0 CHANGE: (#24114) - SEP sesam RMI does not make use of an HTTP proxy to retrieve RSS notifications CHANGE: (#23565) - Add tool tip to multi-drive-tool, if the mouse hovers over the editable columns and multi-change is possible NEW: (#26492) - Feature request: Selection of "Update Mode" in "Topology" tab with drop down menu BUGFIX: (#26497) - Adding or changing drive properties does not start reconfiguration of drives BUGFIX: (#26479) vSphere IR to other vCenter fails with "Datacenter 'null' does not exist". Problem: Datastore does not exist at target vcenter or is not set. Solution: extract datastore at correct vmdk path <[ds]source==::==[>ds<]target); Prevent start VM restore, when no datacenter/datastore is selected. BUGFIX: (#26493) - Subsequent opening of task group events slows down the UI CHANGE: (#25399) - Don't ask ".. create a task for the client" during add new client in case of access mode "virtual" BUGFIX: (#26487) - NPE if mount of vmware backup is attempted CHANGE: (#24696) - Micro Focus settings are set to NULL when clicking OK at client options CHANGE: (#26488) hide recover option 'rest.conf.attach-ro-nfs.snap.start' for restore VM at restore wizard, until implementation of VM sandbox restore has been finished. BUGFIX: (#26117) - The results window of a backup job forgets its entries if it's open while a new sesam day begins BUGFIX: (#25409) - The pulldown menu in the gui seems to show the same values several times BUGFIX: (#24102) - Exception after entering ESX vSphere password CHANGE: (#25910) - Filter for "Last Backup State" for inactive clients not work as expected CHANGE: (#26394) - Show message box when customer tries to create a migration task between two Si3 stores BUGFIX: (#26212) - Exception 'Unable to load Results' if the protocol is going to be reloaded but it has already been removed CHANGE: (#26428) - Enhance description of search pattern of restore wizard BUGFIX: (#26461) restore VM to other vcenter with VMDK relocation failed. Problem: Datastore does not exist at target vcenter. Solution: extract datastore at correct vmdk path <[ds]source==::==[>ds<]target) BUGFIX: (#26424) - GUI view of "locked until" at media overview must contain correct value BUGFIX: (#26452) - GUI does not show the hook of the selected drive BUGFIX: (#26454) - NPE if adding a new drive to a drive group that does not yet have a drive assigned BUGFIX: (#26420) - GUI shows wrong date/time format in job state backup/all results tables BUGFIX: (#26450) - Configuration of shared drives not possible via GUI anymore BUGFIX: (#26443) Set VMWare vSphere authentification at browser will not applied to DB. Problem: client will not be updated with new authentification data, before the request is send again. Solution: update client with changed user/password in DB. SRC Bugfixes: 5bb754bc1e CHANGE: (#29892) - Database of users should be checked and cleaned before an update is carried out. NEW: Shutdown Sesam server if free space is below 'min_free' and 'gv_conf_shutdown_minfree' is set BUGFIX: (#29796) use correct DB statement with numbers equal to 0 instead of NULL in sm_update_client update statement to avoid DB error 'Value not a number' in clients table. BUGFIX: (#29451) Do not execute sm_main start during update in sm_setup before executing sm_startup -f. In sm_startup continue with startup process after sm_main start since sm_main only starts sm_startup if it is not already running. NEW: (#21848) Delete savesets parallel in 'sm_remove_savesets' as default BUGFIX: (#29451) remove the newly inserted semaphor for sm_startup since there was already a semaphor present in the sm_startup. NEW: (#29824) SFR fails with 'Warning: Saveset not created' because 'tmp/mnt' directory is excluded. Solution: Add '-o skip_sbc_exclude' to SBC command line to skip reading default excludes from 'sm.ini' BUGFIX: (#29464) eDirectory restore fails with 'Cannot find matching LIS line for item'. Solution: If LIS file has ISO-8859 encoding then con BUGFIX: (29464) eDirectory restore fails with 'Cannot find matching LIS line for item'. Solution: If LIS file has ISO-8859 encoding then convert SEL file from UTF-8 to ISO-8859 NEW: Set pattern for VMware and Nutanix in 'gv_conf_pattern_lisv2' on RDS and client too NEW: (#21142) Add new license FDL needed for GNU cpio to list of used licenses. NEW: (#29460) Add extra sbc message number for VADP transport mode used NEW: Skip LIS file generation for restore using mount of saveset or VD CHANGE: (#24735) Use 'sbc_vadp.jar' as default for VMware backups CHANGE: (#24735) Add 'sbc_vadp.jar' to Sesam server kit BUGFIX: (#29798) Exchange Server Backup - Log truncation after VSS_S_ASYNC_PENDING. Problem: Exchange Server Backup cause log truncate. Solution: Avoid BackupComplete call except case then we are sure backup succeed NEW: (#25064) New BSR DiskImage 17.1. Problem: JobPath may end without path delimiter. Use os.path.join(dir,file) to get proper path NEW: (#25064) New BSR DiskImage 17.1. Problem: JobPath may end without path delimiter. Use os.path.join(dir,file) to get proper path NEW: (#29122) Add 'pidof' to dependencies for SLES15 SP3 because REAR backup requires it bugfix: Do not restart STPD during drive configuration if 'ADD-DS-DRIVE' or 'CHANGE-DS-DRIVE' option is set BUGFIX: (#29451) 'sm_qm_main start' will now call 'sm_main startup' instead of 'sm_main start'. This ensures that the sm_startup is only called if the qm is started and it is only called once. BUGFIX: (#29451) insert a semophor for sm_startup to ensure that it can only be started one at a time. Calling sm_main start for Windows will call sm_startup only if qm is not running and if sm_startup is not already running. NEW: (#25064) New BSR registry key for DiskImage 17.1.x is missing. Add registry key detection for version 17.1.x to executable sbc_bsr.exe and sm_client.exe (remove 17.0.x) NEW: (#27855) Client update: Do not rebuild 'index.txt' file in Sesam package directory, if '.htaccess' file exists too. Assuming SEP repository in this case. NEW: (#21142) Backup of sparse files. New: Now a sparse file is detected during backup, a special cpio mode is set but data is completely stored. Restore of a sparse file is done by detecting zero areas (>=512 Bytes) and creating a sparsely allocated file (with 'holes'). Additional: Retry write loop improved. Note: Functions sparse_write() and buf_all_zeros() are fetched from Free Software Foundation, Inc. GNU cpio 2.13 src/util.c. BUGFIX: (#29772) If an exception occurred during a XenServer restore, the final state is wrongly successful. Added a general exception handler to sbc_proxy that catches any unhandled exception from any sbc_proxy module and writes SBC error. That way unhandled exceptions will always cause failed operations. CHANGE: Citrix XEN restore: Avoid restore of vmdk if 'vmdk' is missing in recover option BUGFIX: (#28822) EOM doesn't work on RDX SayFUSE because SetTapeBlockSize() is called. Solution: Retry GetPrivateProfileString() if device type is empty NEW: (#28761) Only check AVX support needed for HTTPS on sesam server and rds. NEW: (#28761) Insert a check for the support of AVX during update and installation. AVX as well as FMA3/FMA4 are needed for HTTPS to work. A warning message is displayed in case HTTPS is not supported. BUGFIX: (#29559) - SBC Exchange database backups are running sequential due to the use of a global semaphore. Solution: Unlock the global semaphore after StartSnapshotSet..DoSnapshotSet CHANGE: (#28761) Add CPU check to sm_setup. The parameter 'check_https_support' returns information if HTTPS is supported on the system or not. NEW: (#29742) Add Java 17 to list of valid Java Versions. BUGFIX: (#29638) Migration to HPE StoreOnce with stpd_http sometimes fails - 'Checksum of saveset copy does not match'. Problem: In case of STPD_MAIN detect connection error (MHD_REQUEST_TERMINATED_CLIENT_ABORT) STPD_HTTP is Terminated without chance to properly close LIS/SGM. Solution: Send soft kill signal 10 sec prior Termination to allow STPD_HTTP finish regularly BUGFIX: (#29739) Force a new full backup when INC of MS Exchange/MSSQL fails at parsing SGM file. Solution: Change logic: Enforce full if some data was processed even if no SGM file was parsed BUGFIX: (#29718) SESAM_BACKUP fails with 'VSS API error: Volume is not supported by the VSS provider' as long as SM_VFS virtual file system for a saveset mount is active. Solution: Add SBC exclude: sm.ini [SBC_EXCLUDE] ExcludePattern916=gv_rw_tmp:mnt during update BUGFIX: (#29718) SESAM_BACKUP fails with 'VSS API error: Volume is not supported by the VSS provider' as long as SM_VFS virtual file system for a saveset mount is active. Solution: Add SBC exclude: sm.ini [SBC_EXCLUDE] ExcludePattern916="gv_rw_tmp:mnt" NEW: Set correct pattern for VMware and Nutanix in 'gv_conf_pattern_lisv2' BUGFIX: (#29713) Blocking Date without end_time do not work with 'suppress_with_timerange'. Solution: Get surrounding newdays then check with ((start<='{now}' and end_time >='{now}') or ((end_time='' or end_time is null) and start>='{recent_newday}' and start<='{next_newday}') NEW: Set 'gv_conf_use_sm_remove_saveset' and 'gv_conf_pattern_lisv2' during installation and update BUGFIX: Set overwrite flag during restore via 'mount' BUGFIX: (#29705) Full restore doesn't work, if LIS is missing. Omit check for full restore BUGFIX: (#25781) Restore of vSphere backups done with single saveset option doesn't work because of wrong restore type prefix NEW: (#25064) New BSR registry key for DiskImage 17.x is missing. Add registry key detection for version 17.xx to executable sbc_bsr.exe and sm_client.exe BUGFIX: (#29638) Migration to HPE StoreOnce with stpd_http sometimes fails - 'Checksum of saveset copy does not match'. Problem: Backup could be successful even if "Operation failed." or none response received from server. Solution: Backup should be marked as successful only if server respond with "Operation succeeded." NEW: (#23442) Custom sort order for migrations started by one migration task. Set DB:migation_tasks.option like 'order=data_size desc' BUGFIX: (#29680) Do not set state of backup or migration to cancelled, if task has already finished during sm_break is active CHANGE: Set FSCK status for data stores to '1' (Warning), if data files without DB entry are found. This isn't a severe error in opposite to a backup in DB, without data file NEW: (#29665) Reserve executing streams for tasks with priority: DB:hw_drives.qm_options: 'prio=_'. Example: update hw_drives set qm_options='prio=9_6,prio=15_4 BUGFIX: (#29638) Migration to HPE StoreOnce with stpd_http sometimes fails - 'Checksum of saveset copy does not match'. Problem: Backup could be successful even if "Operation failed." or none response received from server. Solution: Backup should be marked as successful only if server respond with "Operation succeeded." NEW: (#28940) Start external restore requests with higher default priority (15) then backups (10) BUGFIX: (#29638) Migration to HPE StoreOnce with stpd_http sometimes fails - 'Checksum of saveset copy does not match'. Problem: At backup end XBSA wait some time to get data from STPD, and if now data in time - hard abort connection. Solution: Increase wait timeout and add additional logs BUGFIX: (29654) Full restore reads data from the beginning of the tape instead of the correct segment, because STPD requests segment 0 from SMS. Solution: Parsing old segment file header without 'BLOCKSIZE=... and VERSION=...' lead to consuming 2 additonal lines so real segment could not be detected NEW: openebula: allow to configure snapshot creation timeout via additional arguments BUGFIX: (#29553) Fixed mount of Hyper-V VM VHD on Windows RDS. Powershell execution script was set with path of the Sesam server instead path of the RDS client NEW: Improve Java memory calculation for Si3-NG change: Do not restart 'stpd' for every drive during drive configuration, but do it only once CHANGE: (#29559) SBC Exchange database backups are running sequential due to the use of a global semaphore. Change: Message log level revised NEW: (#23064) sm_data_store fsck: Do not count saveset with 'DB:result_lbls.segement=-1' as error NEW: (#23064) Remove EOL free savesets with 'DB:result_lbls.segement=-1', which usually happens, if saveset hasn't been created BUGFIX: (#29627): - Sesam ships expired Root CA Certificate for letsencrypt: update certificates BUGFIX: (#29641) Core dump in 'sbc_vmware' during backup of '.vmx' file with very long lines (>32766=MAX_PATH_LEN) new: (#29560) SBC treated HPE store failure as successful and truncated Exchange Server transaction Logs. NEW: If stpd.ini [CI] END_WITH_FAILURE=1 is set then end backup operation but not a COM_ operation new: (#29560) SBC treated HPE store failure as successful and truncated Exchange Server transaction Logs. NEW: If stpd.ini [CI] END_WITH_FAILURE=1 is set then end backup operation ftp/http[s] with 553 STOR Failed. CI test: stpd.ini [CI] END_WITH_FAILURE=1 is set! (0) BUGFIX: (#29560) SBC treated HPE store failure as successful and truncated Exchange Server transaction Logs. Solution: Set BackupComplete only in case of success (STATUS=0) CHANGE: (#29437) Set DB:results.sub_type='_CBT' for existing VMware full backups made without '_IMG' option CHANGE: (#29437) Set DB:results.sub_type='_CBT' for existing VMware full backups made without '_IMG' option BUGFIX: (#29560) SBC treated HPE store failure as successful and truncated Exchange Server transaction Logs. Solution: Set BackupComplete only in case of success (STATUS=0) CHANGE: (#29368) SBC_COM: iSELECT_FROM_RESULTS() Do not return saveset with failed state BUGFIX: (#29588) Execute sm_break2 to stop all tasks during Sesam shutdown BUGFIX: (#29579) VADP backup could be marked as successful, also when error occurs at parsing SGM file. Solution: A failed state of a subtask now results in a failed state of the main task NEW: (#28708) Stop all running tasks before starting full drive configuration BUGFIX: (#28887) Blocking date with suppress with 'suppress_with_timerange' set does not work. Solution: In this case only the start and end_time together with priority must be checked, newday event does not matter, seek for events with prio > my_prio and start <= current_time <= end_time NEW: Sesam release v5_0_0_1 branch BUGFIX: (#29039) Exeption in 'sm_replicate' during initial Si3 seed operation NEW: (#16099) Improve log messages for drive configuration: Add hostname. BUGFIX: (#29450) Hana Restore waited for specific drive while other 'read' drive was available. Solution: If external restore with multiple read drives then let sm_qm_main decide to use available drive NEW: (#28169) Block drive queues of SDS drives during Sesam startup until SDS service is up and running" NEW: (#28169) Write overall status of drive configuration into Sesam main log NEW: (#28169) Write DB:params_fix.sts/msg for Sesam startup NEW: (#28169) Set 'gv_server_not_available' to block start of new tasks during drive configuration NEW: (#28169) Wait for for complete Sesam startup NEW: (#24995) Keep 'gv_server_not_available' variable set if 'sm_break' is called with '-u ' paramter set, so no new external backup jobs will be started BUGFIX: (#29495) Adopt 'SAP ASE' module for Python 3 on Linux BUGFIX: (#29451) sm_main start is only executed if the service is not running at all. BUGFIX: (#29484) Since python3 migration several zfs module actions failed with 'a bytes-like object is required, not str' BUGFIX: (#22150) Fix SQL statement to stop all running tasks of a certain task group BUGFIX: (#28952) sbc.exe crashed on a MSSQL task, if more then 4 dbs are selected. Increase source length to 1024 BUGFIX: (#29475) Migrations with active state in DB but without any process are not set to cancelled by sm_break. Handle main task too BUGFIX: (#29475) Migrations with active state in DB but without any active process are not set to cancelled by sm_break BUGFIX: (#28282) After an active migration gets cancelled the state remains active. TODO: Update is done in signal handler, which is not safe BUGFIX: (#22042) Submitted sub-migration in state pending are not stopped by 'sm_break' NEW: (#27166) Set 'gv_server_not_available' to block start of external backups/restores during termination of all active Sesam jobs BUGFIX: (#29412) create python3 subfolder before copying files in use there. BUGFIX: (#29413) Restore from tape fails with "Data synchronization lost" when first tape is not involved. Use given tape in case of FTF NEW: (#28862) Data store FSCK: Do not complain of missing saveset file if DB:results.data_size IS NULL. Before only '0' has been checked NEW: (#28862) Data store FSCK: Do not complain about wrong saveset file size, if DB:result_lbls.segment number=-1 (usually brocken backup or migration) CHANGE: sbc_vmware_mount: Set log file path to Sesam LGC directory NEW: (#28357) Add backup and saveset removal information to DB:sesam_events NEW: Scheduled file restore using VMDK mount NEW: (#25522) Execute command during VMware backup by guest tools: Set OS user/password from Sesam DB, if 'run=' action is detected in backup options BUGFIX: (#29402) OpenNebula restore: use NETWORK_ID instead of NETWORK during vm creation, attach all existant network cards to vm BUGFIX: (#29359) Do not show Hyper-V on a Windows 10 system since it is not supported. NEW: (#28357) Add backup and saveset removal information to DB:sesam_events. Return OLD for delete from 'result_lbls', because else delete isn't done NEW: (#28357) Add backup and saveset removal information to DB:sesam_events BUGFIX: (#25884) Copy saveset information into target store after migration using correct interface. Already made change in 8b003931f89 has been overwritten NEW: (#28981) Add compression option to MS-SQL backup. Solution: If backup option compression is enabled then sbc option -o compress is used to start Microsoft SQL Server Backup 'WITH COMPRESSION' and avoid packdata compression. Note: Microsoft SQLEXPRESS does not support backup option WITH COMPRESSION BUGFIX: (#27344) Insert spaces in the warning messages in case of service restart to easen readability. BUGFIX: (#28433) Check for Windows platform on chosen RDS when mounting Hyper-V image not on sesam server. BUGFIX: (#28941) do not raise an error but rather return an error so the process can finish with unmounting the disk. NEW: (#29383) Allow setting of data mover for migration. Default is source RDS BUGFIX: (#28941) do not raise an error but rather return an error so the process can finish with unmounting the disk. NEW: (#28862) Remove saveset with DB:result_lbls=-1 by 'sm_pur_status' only, if saveset size is 0. Other ones will be removed by data store purge operation, if EOL is gone NEW: (#28357) Data store FSCK: Do not complain about saveset files without DB entry, if files are on S3 store only, because they will only be deleted, if complete snapshot is removed BUGFIX: (#29367) STPD results in an endless loop when OSCLT_ERR_PAST_EXTENT occurs. Solution: Check EOF flag and buffer empty in case of complete restore BUGFIX: (#29368) Do not return savesets with invalid DB:result_lbls.segement (-1) to external caller CHANGE: (#25781) Set 'DB:results.sub_type=_CBT' for vSphere tasks with backup with single stream during option set and not level COPY BUGFIX: (#15264) Re-creation of LIS files from saveset on datastore. Saveset name has been cut of by wrong use of 'M_X_SNPRINTF()' macro CHANGE: (#28397) Handling of 'database locked' errors. Solution: Call sm_alarm to inform customer about fatal database errors. Add A001-DATABAS message to daily log BUGFIX: 'sm_stop': Last character of 'DB:results.stop_time' has been cut off BUGFIX: (#28688) - Object attributes of Windows volume must be backed up as well. Implemented for vss. Fix after failed CI BDA tests. NEW: (#29329 ) Add OS user and password for 'run' action during vSphere sandbox restore NEW: (#21848) sm_remove_saveset: Use only threads on Windows NEW: (#29314) Add information about proceeding full/incr backup to saveset info NEW: (#21848) sm_remove_saveset: Use only threads on Windows and log with thread id CHANGE: (#23844) sm_recover: Offer tape drives only, because sm_recover doesn't work with datastores BUGFIX: (#23844) sm_recover: Set correct exit code and message BUGFIX: (#29297) Sesam DB export duplicates line feeds in values. Solution: Use binary mode for out file too, check line break of import file and write same line break to outfile BUGFIX: (#23844) sm_recover does not provide saveset and segment number for sbc command NEW: (#15892) Saveset recover from tape: Use 'backup id' in DB:result_lbls to support migrated savesets BUGFIX: (#29302) Saveset info couldn't be read by SMS dir function any more. Cut 'savesetname' string at correct place BUGFIX: (#27775) Data transfer to HPE StoreOnce could fail with 'Failed sending data to the peer' when Ctl_OpenObject failed with OSCLT_ERR_INVALID_OBJECTSTORE_KEY. Solution: Avoid race condition during initialization NEW: (#28470) Check SQLite DB export for migration to Postgresql: Check foreign keys NEW: Use 'sm_remove_saveset' instead of 'sm_data_store purge' to speed up backup removal NEW: (#28470) Check SQLite DB export for migration to Postgresql: Use 'getopt' to set parameters NEW: (#28470) Check SQLite DB export for migration to Postgresql: Execute SQL statement to fix problems NEW: (#28470) Check SQLite DB export for migration to Postgresql: Print out SQL statement to fix problems BUGFIX: (#29296) Also update the access state of the Hyper-V cluster resource not only the cluster nodes after successful access check. NEW: (#17561) Issue an error if number of files saved according to sbc and LIS don't match and count from LIS is smaller then reported by SBC. BUGFIX: (#28389) - sm_reformat_lis does not consider the exclude of a vhd. Cause: The following format -x ",, ... " is not handled in the correct way. CHANGE: (#29272) Changing Media EOL does not fill DB fields eol_changedby_user_name and eol_changedby_msg. Additional: Update medai for single saveset, update results.eol via result_lbls.backup_id NEW: Use LIS file version 2 for vSphere and Nutanix backups for faster virtual disk mount BUGFIX: (#28822) EOM doesn't work on RDX SayFUSE because SetTapeBlockSize() is called. Problem: Function INI_GetString() was not thread safe BUGFIX: (#29262) An empty update_message= entry in the sm.ini leads to douplicate entries of the update_message. Use function ReplaceProfString to insert or replace the update_message instead of inserting a new string using AddProfString. bugfix: (#29258) Archive adjustment for DISK_HARD/DISK_CHANGE does not write the tape_header/_trailer. Solution: Call sm_arch with option '-i over' BUGFIX: (#29217) Reinsert the update options for all setup types which were accidentally moved to the server part. BUGFIX: (#29217) Only check for db export file for setup type server. Correct the path of the sm_setup.exe using the full qualified path. bugfix: (#29258) Archive adjustment for DISK_HARD/DISK_CHANGE does not write the tape_header/_trailer. Solution: Call sm_arch with option '-i over' BUGFIX: (#29217) Reinsert the update options for all setup types which were accidentally moved to the server part. BUGFIX: (#29217) Only check for db export file for setup type server. Correct the path of the sm_setup.exe using the full qualified path. "BUGFIX: (#29245): set restore options based on opennebula version BUGFIX: (#29228) NDMP restore from FDI chain fails: NDMP error: ERR-CONN NDMP3_DATA_START_RECOVER exchange-failed. Solution: Set relocation for every restore not only for the last NEW: Add 'gv_rw_db:log' (default PostgreSQL trace log directory) to directories which are purged by 'sm_pur_status' CHANGE: (#24783) 'sm_rmi_main' invokes Java directly instead of using 'sm_java' to start UI server BUGFIX: (#28360) Remote installation fails with Python error, if client isn't reachable. Check if username and password is given if client is not reachable before trying to connect to the client via WMI. CHANGE: (#28916) Limit VM Essential socket license to 50 VMs NEW: (#18515) Import/recover of backups from data stores. Copy meta data file from RDS to Sesam server BUGFIX: (#5884) Copy saveset information into target store after migration using correct interface BUGFIX: (#28688) - Object attributes of Windows volume must be backed up as well. CHANGE: (#29033) Remove 'gv_rw_glbv' from backup source of 'SESAM_BACKUP'. Most GLBVs are Sesam internal only and user configuration should be done over Sesam DB only ('gv_conf*' variables NEW: (#29169) Execute Sesam alarm/notification during 'newday' in background, if 'DB:newday_events.options=submit_notify' is set. Use flag for 'ALL_JOBS_DONE' notification too NEW: (#29169) Execute Sesam alarm/notification during 'newday' in background, if 'DB:newday_events.options=submit_notify' is set BUGFIX: (#17233) 'sm_db' analyze of error output isn't correct and leading to false import error. If 'Error:' is at begin of line of a multi-line value, parsing fails. Adopt for sqlite BUGFIX: (#17233) sm_db analyze of error output incorrectly leading to import error. Solution: Adapt sqlite to print 'sqlite: Error: ' in front of every error message BUGFIX: (#17233) 'sm_db' analyze of error output isn't correct and leading to false import error. If 'Error:' is at begin of line of a multi-line value, parsing fails. For Postgresql error messages start with 'psql:' and this is used now. TODO: Sqlite BUGFIX: (#28064) Re-start media init for backup, if current medium is in use by another drive. Fix SQL statement, so init isn't done after every backup all the time NEW: (#28953) Add support for Windows reparse tags: IO_REPARSE_TAG_STORAGE_SYNC, IO_REPARSE_TAG_WCI_TOMBSTONE, IO_REPARSE_TAG_UNHANDLED and IO_REPARSE_TAG_PROJFS_TOMBSTONE BUGFIX: (#28688) - Object attributes of Windows volume must be backed up as well NEW: (#29175) Add build host information during build time to Sesam executable NEW: (#29175) Log build host information added during build, instead of reading it from 'sm.ini' NEW: (#29176) Do not vacuum DB using Postgres during Sesam newday, because usually 'autovacuum' is active on. Vacuum can be enforced by setting 'GLBV:gv_conf_vacuum_db=1' NEW: (#29175) Add build host information during build time to Sesam executable NEW: (#29122) Remove dependency for 'sysvinit-tools' on SLES 11 NEW: (#29170) Execute 'sm_alarm' and 'sm_notify' in background during Sesam newday, if 'DB:newday_events.options=submit_notify' is set NEW: (#29169) Submit (optional) execution of 'sm_alarm' and 'sm_notify' into queue, so script execution could not block caller NEW: (#27868) Reduce function calls in 'sm_newday' protected by 'sepuler' semaphore NEW: (#26959) Add default exclude for '/run' directory on Linux. Directory is saved, but without any content NEW: (#29133) Restart of failed task group: Better log message, if no backups could be found. which could be restarted CHANGE: (#28213) Support for HCL Domino 12. Change: Add 'events4.nsf' to the list of ignored database files which must not be logged. This will avoid warning state with FULL backup BUGFIX: (#29145) FSCK for 'Path' datastore doesn't detect difference in size between DB and file system any more, because size reported by 'sm_dedup_interface dir' is always 0. Solution: Call 'dir full' command NEW: (#29122) Add 'sysvinit-tools' for SLES required by REAR ('pidof') wip: (#26177) repair function iChangeValueFromFile to also accept CHANGE_LINE_NR: again as deprecated function did. BUGFIX: (#29107) Mount of virtual disks on Windows: Do not try to import Python 'sh' module, because it's available on Linux only CHANGE: (#29072) SAP Hana delete backup: Encode backup source in 'hdbbackint' on client side NEW: (#29097) - Add timeout option to 'sm_cmd' NEW: (#29094) HPE Catalyst replication fails, if FC interface is used for backups. Use source side interface from DB:migration_tasks to access source Catalyst store NEW: (#29094) Do not use 'DB:credentials.store_name' for Catalyst replication, if empty NEW: (#29094) Do not use 'DB:credentials.store_name' to set Catalyst store name NEW: (#29094) HPE Catalyst replication fails, if FC interface is used for backups. For replication another interface has to be used to access source side store. Not used 'DB:credentials.store_name' field can be used, to set this interface. TODO: Use another table to enter this information. BUGFIX: (#29090) Do not report cancelled Catalyst replications as successful CHANGE: sbc_vcb is obsolete CHANGE: sbc_xensrv is obsolete BUGFIX: (#29041) Occupancy of datastore media is unrealistically high. Using Py3 for Windows the DB statement works fine. CHANGE: (#29072) Deletion of SAP Hana backups: Simplify function CHANGE: (#29072) Deletion of SAP Hana backups: Delete backup using given (and unique) saveset id only. Comparing backup source doesn't make sense CHANGE: (#29072) Deletion of SAP Hana backups: Decode backup source before comparing it with DB:results.source BUGFIX: (#28454) Saveset migration: Handle errors during post processing (copy LIS/SGM, updat DB:results) correctly CHANGE: (#28010) Merge remote-tracking branch 'origin/python3_kernel'. All kernel module build with Python 3 BUGFIX: (#29050) Do not use start time of migrated backups to calculate next FULL backup execution using 'MaxSinceFull' option in backup event wip: (#19001) Restore only NEWER or OLDER files from saveset, than exiting on the target NEW: (#27197) Support LVM backup of multiple volumes in one task NEW: (#28839) Backup of thin provisioned LVM volumes by creating thing provisioned volume snapshots NEW: Write all entries from DB:defaults, where key starts with 'gv_' as Sesam GLBV by Server 'sm_sms_watch'. Remove function from 'sm_start_all' NEW: Writes all entries from DB:defaults, where key starts with 'gv_' as Sesam GLBV by Server 'sm_sms_watch' NEW: (#27550) Set owner of 'CERT' files to 'sesam' BUGFIX: (#28431) - Windows all backup fails with "553 STOR Failed. MTF analyze error [Error: Invalid MTF stream block checksum. (58)]". Improvement - to ensure that break of stream routine loop is done when the new block is detected BUGFIX: (#28431) - Windows all backup fails with "553 STOR Failed. MTF analyze error [Error: Invalid MTF stream block checksum. (58)]". Description block is interpreter as stream in some rare case. BUGFIX: (#28394) Return error to external backup caller. Additional: In case of DB error set return message to a value with error pattern 'E...' to allow correct error forwarding BUGFIX: (#28957) No reasonable message during restore, if correct tape is not present in loader. Check if tape is available in another loader was wrong. Now correct error message should be set during further steps in 'sm_restore' CHANGE: (#27389) Check DB triggers regarding performance and use case. Improve performance for statement e.g. "update results set eol_changedby_msg='0',eol_changedby_user_name='SEPsesam' where eol='2021-06-30 04:04:02' and eol_changedby='RI20210531155708057@398OpY25zyF'\ by adding index for results.eol CHANGE: (#27389) Check DB triggers regarding performance and use case. Improve performance by adding UPDATE OF column especially for 'update_tasks_results_sts/update_taskgroup_results_sts AFTER UPDATE OF state' BUGFIX: (#28394) Return error to external backup caller, if update of 'DB:results' fails duing 'openjob' operation new: (#28921) Data throughput during restore from tape drive is too low. NEW: Windows: Added non interactive mode for sm_sms_tapeutil -d [-b blocks] [-f FMs] [-m 1|2|3|4|5|6] [-w] [-v loglevel] [-B |0] [-C read_count] [-R R|E] NEW: (#28959) Improve error handling during 'closejob' operation from external backups NEW: (#27415) Optimize delete object from datastore by avoid use of wildcards. Remove backup from DB, if at least one file has been deleted BUGFIX: (#28952) sbc.exe crashed on a MSSQL task, if more then 4 dbs are selected. Problem: Saveset INFO with full command line exceeds trace output line length. Solution: Avoid overrun and print message about truncated string CHANGE: (#27527) Set stricter permissions for PostgreSQL socket file and set group for it to 'sesam' during update BUGFIX: (#23339) Disaster Restore via Bootstrap database and restore type disaster. Windows: Function file_copy( source, target ) failed if target is a directory BUGFIX: Windows: sm_client function filecopy may fail due to path with spaces "C:\Program Files\SEPsesam\bin\sesam\sbc.exe" which lead to 'C:\Program' not found NEW: (#26529) Support for Oracle-VM BUGFIX: (#28936) Parts of output from 'Monitoring > Drives' appears twice. Flush 'STDOUT' between 'sm_qm show' commands NEW: (#28950) Allow setting of 'sbc' options for external backups NEW: (#28940) Submit migrations with lower priority then restores to allow faster start of restore NEW: (#28922) Add complete command line to follow up event, which has triggered event NEW: (#28900) - Script for automatic MSSQL task generation BUGFIX: (#28248) Umount of Sayfuse drive does not work because sayFUSE4SESAM_OFF.exe is called with NULL instead of drive letter. Solution: If hw_drives.device_path is not set to the drive letter then use hw_drives.path NEW: (#17110) Add 'DB:task_types.display_name' for showing other/new name for a task type, without changing all places in source code NEW: (#28388) Remove orphaned meta data files from Sesam LIS directory NEW: (#28892) Support Sesam server installation on Debian 11 (Bullseye) using build for Debian 10 NEW: Remove meta files from task groups after all according backups have been deleted durint next 'sm_pur_status' run BUGFIX: (#28894) Remove main log files (NOT) for command events correctly. Filename has been changed but function in 'sm_pur_status' not been adopted NEW: (#28736) Add PID file for Sesam start over 'systemd', to identify Sesam main (sm_qm_main) process correctly BUGFIX: (#28454) Saveset migration: Handle errors during post processing (copy LIS/SGM, updat DB:results) correctly BUGFIX: (#28867) Update fails because of buggy copy DB file command. Problem: On some operating systems (e.g. SLES12) the behavior of overlapping sting operations is undefined. Solution: Do not use same string as source and target in string operation NEW: (#28026) Retry connection to PostgreSQL server in case of 'connection refused' error NEW: (#28026) Retry connection to PostgreSQL server in case of 'the database system is shutting down' error. (Restart of DB) NEW: Added new BSR version 16.1.192 to build. NEW: (#28843) Log Sesam DB problems with severity 'ALERT' NEW: (#28854) 'Alert' severity log level for Sesam main log. Message prefix is 'AXX-' (A001-DATABASE) CHANGE: (#28026) Set max. DB connection retry to 30 NEW: (#28026) Retry connection to PostgreSQL server, if connection limit is hit. Rise exception, if connect still fails after retires NEW: (#28026) Retry connection to PostgreSQL server, if connection limit is hit (Python modules) BUGFIX: (#28026) Write message to Sesam error log, if max connection limit for Postres is reached NEW: (#23804) Remove all meta data for tape medium: If DB:media_pools.purge_flag='2' is set, then try to initialize tape, so data couldn't be read from it any more NEW: (#23804) Remove all meta data for tape medium: Fix SQL statement in 'sm_sms_watch' NEW: (#23804) Remove all meta data for tape medium: Set DB:media.purge_flag='1' to avoid calling remove function again and again NEW: (#23804) Remove all meta data for tape medium: If 'sm_arch remove_journal' is called with drive only, then DB:media_pools.purge_flag=1 has to set NEW: (#23804) Remove all meta data for tape medium: If 'sm_arch remove_journal' is called with media pool (-m) parameter or with drive (-d) parameter only, then meta data of all EOL free tape media will be deleted BUGFIX: (#28822) EOM doesn't work on RDX SayFUSE because SetTapeBlockSize() is called. Solution: Call only for real tape device BUGFIX: (#28810) Replication (Si3 to Si3) fails because to many savesets are selected. Solution: Upgrade to SQLite version 3.35.5 to retrieve correct rows NEW: (#28817) check if virtual machine has architecture set and restore accordingly BUGFIX: (#28813): use --format parameter for oneimage create on opennbula 6 BUGFIX: (#26958) SMS remembers EOM IO ERROR status during following normal EOM handling BUGFIX: (#28813) Restored opennebula virtual machine fails to boot: add fallback to changed xml definition, cause warning if not valid xml entry is found during restore BUGFIX: (#27797) Kopano backup: Public folder backup does not delete dump files. Wait longer time (200 sec.) for termination of backup thread BUGFIX: (#26860) Do not initialize tape, if 'getlabel' does not return a valid label and drive is attached to a loader BUGFIX: (#28512, #28523) To be able to use Windows ssh client to access sm_sshd the acls of the key file were restricted. Now only SYSTEM and group Administrators have full access. NEW: (#20798) Remove of tape media and remove of data on tape media: If '-i over' option is set for sm_arch, then medium will be erased or initialzied BUGFIX: (#28650) - Windows Backup of mounted folder - volume assigned to a directory. wcsstr_path() function is bugged. BUGFIX: (#28518) Backup to tape during IO Error is sucessful but restore fails with 'Error: Cannot read from remote archive'. Solution: Accept message '1014: Failed to read from tape. System error.' also coz '1044: End of data reached.' must not appear. Note: Logic for pTapeCS->iSeenTapeTrailer now correct BUGFIX: (#27808) Consistent export of Sesam DB SQLITE. Solution: Use of sqlite3.exe requires correct setting of PATH within i_setenv_sesam(...) NEW: (#28780) Stop backup operation in case of IO error reported by SMS, if 'gv_conf_no_eom_ioerror' is set NEW: Set DB:media_results.state to warning for EOM because of IO error NEW: (#11472) Trigger alarm notification for EOM because of IO error NEW: (#26631) Add error message returned by QM, if submit of external job fails. QM error message will be in English BUGFIX: (#28779) sbc cores on Windows sporadically during adaption of exclude items. Solution: i_AdaptPattern() must call i_ResolvePathInPattern() for output parameter BUGFIX: (#28518) Backup to tape during IO Error is sucessful but restore fails with 'Error: Cannot read from remote archive'. Solution: The restore with EOM due to IO error was only implemented for selective restore. Now complete restore is also implemented BUGFIX: (#26009) Reformat third party license to easen reading of single licenses. NEW: (#28463) Faster selective NDMP restore. Solution: New sbc_ndmp allows to specify only selected directories or only selected files. This requires switch '-a seldir=1'. GLBV 'gv_conf_use_ndmp_only_dirs_or_files' controls this behavior. The new behavior is the default with DataMover version >= 5.0.0.1 NEW: (#26981) Saveset migration: Do not check correct target drive/pool connections if not drive is set BUGFIX: (#28153) use NUL instead of null to get ride of stdout output. BUGFIX: (#28153) Powershell path is now determined using a different registry key, hence service pack update does not fail anymore. Powershell path is obtained checking 3 different registry keys and taking the correct one. NEW: (#26981) Saveset migration: Check for correct target drive/pool connections and reset target drive to usable value, if necessary NEW: (#23779) Replication based on source pool selection only, without setting drive number (which restrict source side to one data store) NEW: Replication from media pools on different data stores NEW: (#22616) Add '-j ' parameter to 'sm_sbc_com_ext' call, so task name will be used to create log file name NEW: (#28637) Source side deduplication: Add 'DB:clients.stpd_port' to sbc server parameter (-S) to allow using another listen port then 11001 on client for STPD BUGFIX: (#28752) Restore exclude filter in 'sm_restore' doesn't work correctly, because file name from LIS file isn't extracted correctly (leading / hasn't been removed) NEW: (#28661) Remove server side PRE/POST logfiles at once, if nothing has been done by script. Script has to return with exit code '20' in this case BUGFIX: (#18629) Continue to start saveset migrations, even if one sub-migration failed BUGFIX: (#28282) Migrations stay active in Sesam DB, if submit of sub migration task fails BUGFIX: (#27413) Restore exclude filter isn't reset after first match, so all following items are excluded too BUGFIX: (#27287) Selective restore doesn't work using replicated saveset, which was migrated before replication NEW: (#28732) Adopt saveset selection to work without 'DB:results' for migration of already migrated savesets BUGFIX: (#28335) Unversioned files like *.jar, *.cmd, *.ps1, *.config, *.txt, etc. have to be removed manually to enforce being overwritten during update. For some reason some of these files have a changed modification date and hence are skipped during update. Some string declarations were wrong. BUGFIX: (#28335) Unversioned files like *.jar, *.cmd, *.ps1, *.config, *.txt, etc. have to be removed manually to enforce being overwritten during update. For some reason some of these files have a changed modification date and hence are skipped during update. NEW: (#28636) DIFF/INCR backups for Nutanix-AHV BUGFIX: (#28650) - Windows Backup of mounted folder - volume assigned to a directory. Backing mount points without following them BUGFIX: (#28650) - Windows Backup of mounted folder - volume assigned to a directory. Skip directory feature improved. BUGFIX: (#28650) - Windows Backup of mounted folder - volume assigned to a directory. LIS file generation issue, TEST-BACKUP issue, "following" logic is improved CHANGE: (#28722) Print clear error message, if 'sbc' command line becomes too long. Do not add command string to error message, because this will lead to truncation of error reason in backup message BUGFIX: (#28672) Core dump sm_sms_copy with SP3. Solution: License check with SHARED_DRIVES lead to crash dump when looking for sm_lic_demo_*.ini files. Use resolve_path() instead of resolve_glbv(). CHANGE: (#23784) Saveset migration without 'DB:results': Do not create DB:results entries for default. Old behviour can be switched on by DB:migration_tasks.options=use-results BUGFIX: (#25911) Do not copy msi installer file to obelix. BUGFIX: (#28650) - Windows Backup of mounted folder - volume assigned to a directory. Introduce the new implementation of file system traverse BUGFIX: (#28694) Restore from migrated saveset must not switch RDS interface from http[s] to ftp protocol. Solution: CHECK_AND_SET_INTERFACE(): Check protocol of given interface and use appropriate drive node interface BUGFIX: (#27808) Consistent export of Sesam DB SQLITE. Solution: Call sm_db export -c -o use_dump -f {file} for SQLITE too BUGFIX: (#27808) Consistent export of Sesam DB SQLITE. Solution: Since commit e268ae6 sqlite3 supports .export command to dump without 'CREATE TABLE/INDEX' statements BUGFIX: (#28699) Failed to delete VDI snapshot after FDI backup. Catch XENAPI exception VDI_IN_USE after a call to xenapi.VDI.data_destroy. If exception is thrown from server, wait 10 seconds and retry to execute data_destroy in 10 retries loop BUGFIX: (#21769) Erase ERRORs and WARNINGs which were determined during Certification Test Tool evaluation needed to get Windows Certification. BUGFIX: (#28628) Restore of vSphere configuration creates files with empty content on specific systems. Problem: Windows use case-insensitive folders may cause issue if VM name differs only in case. Solution: Use case-insensitive compare NEW: (#28661) Remove server side PRE/POST logfiles at once, if nothing has been done by script. Script has to return with exit code '20' in this case BUGFIX: (#28664) Restore of a migrated XenServer VM backup does not work BUGFIX: (#28669) sm_restore core dumps if lis file contains file named 'Total:'. Solution: Improve logic and create specific error messages BUGFIX: (#28195) If exclude string at VMware backup is too long, the drive string may be overwritten, e.g. 'Drive: .vmdk'. Additional: Increase size for XBSA Descriptor saveset specifier. Also increase size for RCMD.szMsg BUGFIX: (#21332) Insert more log lines. BUGFIX: (#21332) Insert log lines. Try to identify the problem why modified file sm_disaster.ps1 is replaced without asking the user for it. NEW: (#28660) Remove all meta data (DB/LIS...) of EOL free tape media durint Sesam newday, if 'gv_conf_purge_metadata_tape' is set BUGFIX: (#28663) 'sm_loader' doesn't take care about shared drives, if it tries to load medium in use by another drive. Change SQL select statement to return drive with DB:hw_drives.label!='' first BUGFIX: (#28480) sbc_com_interface hangs, after reboot of Sesam server. Problem: sbc_com_interface hangs if switch off network of VM where server is running. Solution: Make sbc_com_interface expect KeepAlive messages if it was started with "-O keep-alive" CHANGE: (#24710) use logged on user for desktop link using kha. BUGFIX: (#28588) Correct typo of gr_ro to gv_ro to obtain the installation path for SEP sesam BSR Pro. Now installation works as expected. BUGFIX: (#28631) Inquire backup by external backups (Oracle, SAP...): Do not return saveset, if saveset has failed state, but return 'does not exist' error instead BUGFIX: (#28454) Oracle backup: Handle errors set by Sesam server during sbc_com 'close' correctly. Backup will be set to failed state for Oracle side BUGFIX: (#28560) NDMP restore from tape failed with NDMP: 1 info: LOG_MESSAGE: 'RESTORE: Mangled directory'. Solution: Add offset low/high for every buffer. Old sequential optimization becomes obsolete. Reset l_blocks_fetched in case of rewind. Check restore with offset before loop and if buffer found then set pBCK_MESSAGE_r->un.ReadWrite.SegmentNumber = 0 to avoid repositioning during read_data() in SMS NEW: (#25064) New BSR registry key for DiskImage 16.x is missing. Add registry key detection for version 16.xx to executable sbc_bsr.exe and sm_client.exe NEW: (#28567) Set DB:hw_drives.clean_bit=1', if drive reports that cleaning is required BUGFIX: (#28612) Enclose backup type 'Dovecot IMAP' with quotes during restore NEW: (#28567) If cleaning tape is loaded by 'sm_loader': Wait until cleaning has finished and move cleaning tape back to slot afterwards NEW: (#28584) sbc_com_interface has to verify to get correct answer. Check if answer returned from correct sm_sbc_com BUGFIX: (#28568) sm_sms_watch core dumps during Sesam startup and newday. Problem: SM_QM status output during drive configuration lead to overrun in stack. Solution: Increase char array and improve char array processing BUGFIX: (#28588) Do not insert double backslashes in the registry for SEP sesam BSR Pro. NEW: (#28584) Return id build from PID+uuid send by sbc_com_interface back to it, so correct answer process could be checked NEW: merge all changes from sm_servicepack.cmd v4_4_3_5 to master. BUGFIX: (#28573) Restore of vSphere backup with level copy and CBT enabled BUGFIX: (#28586) Update default for FORMAT_LOGICAL_BLK_SIZE and do not overwrite on backup when larger than default. Solution: Check if [SBC_MTF] [FORMAT_LOGICAL_BLK_SIZE] < 2048 before updating of the value NEW: (#17860) Cleaning tape barcode prefix can be set by GLBV 'gv_conf_cleaning_tape_barcode_prefix'. Default is 'CLN' NEW: (#17860) Do not load cleaning tapes during archive adjustment. Check for cleaning tape is done by barcode prefix 'CLN' BUGFIX: (#28577) - VSS backup of disk with mounted subvolume ends in endless loop if -o follow specified BUGFIX: (#16983) - Diff/Incr. backup using NTFS journal. Compilation error under VS2010 BUGFIX: (#28474) Oracle module must correctly verify com requests and answers. Do not compare saveset from reply for 'inquire' because saveset id from migration could be returned BUGFIX: (#16983) - Diff/Incr. backup using NTFS journal. Improve logic around "renamed/moved" folders for SBC BUGFIX: (#16983) - Diff/Incr. backup using NTFS journal. Improve logic around "renamed/moved" folders for VSS NEW: (#26177) Wrong call format of Windows_postgresql section. NEW: (#26177) Unzip postgresql zip file only if installed version is smaller than the one located in the zip file. NEW: (#26177) New common.ps1 containing all common sesam relevant functions which can be reused. New set_postgresql.ps1 extracts the postgresql zip file located in skel directory to bin\db folder. BUGFIX: (#28554) Insert the uninstall string needed for SEP sesam BSR Pro version 15. BUGFIX: (#28554) inserted several error messages to easen error analysis of installation/update/uninstall. BUGFIX: (#28554) Client update fails if BSR has been installed on the system due to a syntax error in the cmd file. NEW: Update 'DB:loader_contents' during mount media action. DB:loader_contents is usually updated by 'sm_loader' only, but information can be lost, if a loader action fails. Then only next archive adjustment would correct this again BUGFIX: (#27175) Block execution of jobs started by 'sepuler' during drive configuration NEW: Extra license copyright file PostgreSQL_license.txt for PostgreSQL and libpgeasy NEW: (#22831) erase function i_READ_SLU_OUTPUT_NEW and put its content in directly in the original function i_READ_SLU_OUTPUT. NEW: (#22831) use new slu -J parameter to get encryption relevant information in i_READ_SLU_OUTPUT. Use new function only if client version is larger or equal to 5.0.0.1, otherwise use old slu commands. NEW: (#22831) show all output when calling -J all not usage error. NEW: (#22831) Inserted a new parameter -J [all|short] which executes 4 functions Log sense, mode sense, read MAM attributes, and encryption specific in a sequence. If all is inserted all output is presented, if short is inserted only filtered output is presented which is needed for encryption check. Revert "NEW: (#22831) To reduce number of slu calls during drive options it is now possible to call multiple slu operations at once using a ; as a separator for the commands e.g. 'slu Tape0 -s; -o; -R get -t all'. In addition a new parameter is introduced called -z short/encryption/extended which will be used to limit the stdout output to easen the log file analysis." Revert "BUGFIX: (#22831) start looping the input parameters from index 0 not 1 to ensure that also commands with only one parameter like 'topology' still work as expected." Revert "NEW: (#22831) Filter stdout output of slu commands using new parameter -z. e.g. slu Tape0 -s -z cartridge_loaded,cleaning_bit will only show the two given fields in the output." NEW: (#22831) Filter stdout output of slu commands using new parameter -z. e.g. slu Tape0 -s -z cartridge_loaded,cleaning_bit will only show the two given fields in the output. CHANGE: (#28492) RPM installation of Sesam server: Use new RPM spec syntax to specify pre install requirements BUGFIX: (#22831) start looping the input parameters from index 0 not 1 to ensure that also commands with only one parameter like 'topology' still work as expected. CHANGE: (#28492) RPM installation of Sesam server: Add 'Java' as pre-requirement to enforce full installtion of Java package, before Sesam setup is started BUGFIX: (#28522) SP3 for Windows contains older CBFS driver package, which makes VFS not usable anymore. Solution: Add CBFS_VERSION v'6' to path to locate recent version BUGFIX: (#28490) Do not write STDERR into target file during remote file copy. Solution: Set bStderr and improve logic to forward stderr output to process stderr NEW: (#22831) To reduce number of slu calls during drive options it is now possible to call multiple slu operations at once using a ; as a separator for the commands e.g. 'slu Tape0 -s; -o; -R get -t all'. In addition a new parameter is introduced called -z short/encryption/extended which will be used to limit the stdout output to easen the log file analysis. BUGFIX: (#27175) Block execution of other jobs like external backups until drive configuration has been finished during Sesam startup BUGFIX: (#28456) Error during restore of data split on multiple tapes: The archive could not be resynchronized. Exiting. Solution: EOM was enforced via command 'sm_sms_interface force_eom -d ' this lead to last block written twice. NEW: If 'DB:clients.access_options' starts with 'use_ssh' and access mode is SMSSH: Executcommand using 'ssh' instead of 'sm_ssh'. Because 'ssh' is usually written in C, startup is much faster then Python 'sm_ssh'. Port to Windows NEW: If 'DB:clients.access_options' starts with 'use_ssh' and access mode is SMSSH: Executcommand using 'ssh' instead of 'sm_ssh'. Because 'ssh' is usually written in C, startup is much faster then Python 'sm_ssh'. Works currently on Linux only BUGFIX: (#28497) Enforce a clean for building the SP in case something has changed in git, independent of the switch -c (meaning no clean should be executed). BUGFIX: (#28474) Oracle module must correctly verify com requests and answers. Additional: Call sbtclose2() after failure to end SBC_COM session BUGFIX: (#27597) Backup session gets wrong COM answers during backup with multiple channels. Problem: STPD connection slot may be detected as free for new connection, but still be in use by previous. Solution: Connection slot should be detected as free only after connection ended BUGFIX: (#28474) Oracle module must correctly verify com requests and answers. Solution: Verify if SM_SBC_COM retrieved saveset ID matches to origin saveset ID BUGFIX: (#22831) skip call 'slu -R get -t all' in case no cartridge is loaded since the call leads to an DEVICE_BUSY error. NEW: (#28475) Add description of 'gv_conf_com_return_source' NEW: (#28475) SAP Hana backup: Check if backup source send to Sesam server is the same then returned from server. Workaround for 'sbc_com' issue, where answer from another call is returned. Backup will be in failed state BUGFIX: (#27999) Sporadically all write operations to a tape drive hang or fail after medium change, except the one that was responsible for the EOM process. Solution: Since commit 3f20473 "(#27422) One data block isn't written to next tape after EOM" writing a tape mark into the EOM area immediately sets the status to 'CHANGING_MEDIA', but write_data() then occupies the mutex twice and releases it only once BUGFIX: (#22831) reduce the log messages during encryption check. BUGFIX: SAP Hana: Set correct backup type for query oldest backup in Sesam DB BUGFIX: (#27597) Backup session gets wrong COM answers during backup with multiple channels. Problem: On high load systems CreateProcess may fail to start process correctly, but we still give sub-process time to init. Solution: If CreateProcess fails - return with error immediately. BUGFIX: (#28347) Return error in case BSR uninstall or install fails not warning. Enhance the error message in case BSR uninstall or install fails. BUGFIX: (#28347) BSR uninstall routine returned error even though no error occured. If the installed BSR version is smaller or equal to the BSR version in the installation package, success has to returned. BUGFIX: (#28459) if index.txt is located in packages folder it is being deleted during update process. The packages folder is being cleaned from all files which are not present in the index.txt during update. Solution: Only remove files which have file endings .deb, .rpm, .sh, and .exe and leave all others untouched. NEW: (#26177) PostgreSQL for Sesam DB on Windows. Windows: DB_ACCESS() does now check for gv_db_type 'postgres'. Then necessary postgres functions are dynamically loaded from libpq.dll and pgeasy.dll. Static link to libpq.lib and pgeasy.lib removed from VS project files. BUGFIX: (#28247) SAP Hana backup: Do not report successful backup to Hana DB, if backup post processing on Sesam server fails (could not get SGM file, checksums form LIS and sbc don't match...) BUGFIX: (#28446) 'sm_dedup_interface dir' for 'Path' data store returns successfully, even data store directory isn't accessable. This can lead to removal of backup information from Sesam DB during data store 'fsck' operation NEW: (#28327) Mount tape by 'sm_loader' directly, if called during restore, so 'mount' process is protected by loader semaphre CHANGE: (#27389) Check DB triggers regarding performance and use case. Improve performance by avoiding mtime triggers for 'migration_results' CHANGE: Do not create trigger to update 'DB:mtimes' for 'DB:*results' tables NEW: (#26177) PostgreSQL for Sesam DB on Windows. Add PestgreSQL libs as dependencies CHANGE: (#28399) Use global variable gv_sqlite_busy_timeout if present, otherwise use default timeout 180, for SQLITE busy timeout in Python DB access module. NEW: (#28332) - Show release notes button on first start [after update] of GUI/WebUI CHANGE: (#28399) Increase SQLITE busy timeout in Python DB access module from 60 to 180 sec. TODO: Use 'glbv' to make timeout configurable NEW: (#28396) Log saveset inquire calls by Oracle as 'INFO' instead of 'ERROR' in Sesam main log, if saveset couldn't be found, because Oracle searches for last control file backup using 'trail and error' BUGFIX: (#28174) Clear update_message in case update_state is empty or NULL. Do not exit if check_user_password fails, only return the errror and continue. NEW: (#28388) Remove all migration log files for a Sesam day, if no DB:migration_results exist for it NEW: (#28406) Set 'DB:results.data_size=1', if data size is '0' but a valid SGM file exists. This can happen, if STPD couldn't write total line into LIS BUGFIX: (#28408) Set drive mount status correctly, even if tape trailer could not be read, but header is correct BUGFIX: (#27805) FDI_CHAIN adjust may take very long on systems where many backups exist. Solution: If GLBV gv_conf_fdi_chain_recursive_select is set then use 'recursive select' SQL statement to retrieve chain of FDI savesets BUGFIX: (#28406) Try to copy and analyze SGM and LIS file created by an external backup, even backup has failed BUGFIX: (#28189) EOM doesn't work, because other STPD reset EOM status. Solution: Old implementation sets glbv gv_sms_eom_{device} to 'READY' after write loop - so finishing backup may interrupt EOM handling. Now moved to correct place - SEP_sync() BUGFIX: (#28394) Return error to external backup, if insert into 'DB:result_lbls' fails CHANGE: (#27891) Read next label required for restore: Cut label from current label 'TAPE00038 mounted' returned by drive handler NEW: (#28327) Occupy loader semaphore again, after submit of external job returned BUGFIX: (#28384) SMS does not stop if medium error is recognized, backup state is set to success. Problem: In case of medium error retry write loop started. Solution: Break retry loop and set EOM NEW: (#28070) Load of mounted tape from another drive during EOM at restore is started with higher priority then backup tasks NEW: (#28327) Avoid dead lock by 'sm_loader', if following tape for restore is mounted in other drive then used for restore BUGFIX: (#28383) Retry dismount in case dismount fails because target is busy BUGFIX: (#28382): sbc_proxy: LVM backup fails with 'TypeError: a bytes-like object is required, not 'str' NEW: 'sm_sesam': Log with PID, because 'sm_sesam' can be executed parallel, but all use same log file BUGFIX: (#28123) - Restore fails with 'At current offset 144 (0x90) no valid MTF Stream ID was found'. Source changes after coder review BUGFIX: (#28123) - Restore fails with 'At current offset 144 (0x90) no valid MTF Stream ID was found'. Cause: Wrong set SPAR header length in case of zero-ranged file after previously processed SPAR non zero-ranged file. BUGFIX: (#27796) Core dump of sm_sms_copy after LOAD_VOLUME_INTO_DRIVE. Solution: If gv_prot is not set properly (e.g. during day change) then stdout file handle may be closed BUGFIX: (#28309) NDMP restore via FTP fails with 'Found a checksum error in header' during data transfer, but works via HTTP. Solution: Sequential read only implemented for tape devices (and old DISK_HARD) CHANGE: (#17427) Set stop time for external to at least one day for restores only BUGFIX: (#28174) Do not overwrite the update_state by the update_message. BUGFIX: (#28174) For Winodws installation/update add update_state and update_message sbc output and update these values during alive test if update_date in DB is different from update_date given by alive test. BUGFIX: (#28174) For Winodws installation/update add update_state and update_message to sm.ini and display it using sbc. BUGFIX: (#28174) show cleaning log message only in higher log level. BUGFIX: (#28174) Clean client version information about available versions from DB to ensure obsolete entries are removed before checking for new available versions. BUGFIX: (#28309) NDMP restore via FTP fails with 'Found a checksum error in header' during data transfer, but works via HTTP. Solution: FTP must not rewind in case of a DataStore. Keep position only for tapes BUGFIX: (#28337): sbc_one: Restore failed with AttributeError: Cause: vm xml includes empty node, check for empty string BUGFIX: (#28337): Restore failed with AttributeError: Cause: vm xml includes empty node BUGFIX: (#28233) HCL Domino 11 backup fails with 'Problem while loading dynamic link library: [dlopen() returned: libgsk8iccs_64.so'. Solution: Link without gsk8iccs_64 and use option --allow-shlib-undefined BUGFIX: (#28316) STPD doesn't flush LIS and SGM file in case of backup termination. Problem: STPD signal handler on Windows perform Hard termination. Solution: Add graceful termination similar as on Linux BUGFIX: (#28316) STPD doesn't flush LIS and SGM file in case of backup termination. Problem: LIS and SGM files flushed and closed only if backup finished normally. Solution: Call LIS/SGM flush/close if termination signal handled NEW: (#28329) Add new operating system 'DELL EMC Unity' BUGFIX: (#28316) STPD doesn't flush LIS and SGM file in case of backup termination. Problem: LIS and SGM files flushed and closed only if backup finished normally. Solution: Call LIS/SGM flush/close if termination signal handled NEW: (#28329) Do not show task type 'NetApp' for DELL EMC Unity NDMP BUGFIX: (#28309) NDMP restore via FTP fails with 'Found a checksum error in header' during data transfer, but works via HTTP. Solution: During FTP there is one data read session where every portion is read without repositioning but this is not supported for Si3 - reset memorized position BUGFIX: (#28189) EOM is not recognized for unknown reason. Retry writing gv_sms_eom_... value into registry. Add trace() and PROT() if writing of GLBV failed BUGFIX: (#28302, #22454) Add firewall info and firewall rule for SEP sesam BSR Pro and show firewall info dialog with all consequences also during update. BUGFIX: (#28290) sbc_smdr: The last item is not counted in the total data size and number of restored items. Solution: Logic for 'sbc-3008 Processing ...', 'sbc-3108 Item processed ...' and 'sbc-3021 Item [...] already existing' revised NEW: (#16997) Delete backup by external caller. Remove entry from Sesam DB, instead of setting status='6'. This will fix problems, if for example Oracle uses same saveset id again for control file backup BUGFIX: (#28298) use SHA-256 for signing all executables in the Windows kit. This is possible since we do not support Windows Vista SP2 and Windows Server 2008 SP2 anymore which had problems with executing files being signed with SHA-256 NEW: Use LISv2 if 'gv_conf_pattern_lisv2=' is found in connect string send by SBC NEW: Use LISv2 if 'gv_conf_pattern_lisv2=' is found in connect string send by SBC CHANGE: (#28284) Increase maximum data timeout of STPD. Set MAX_DATA_TIMEOUT to 655350, which is one week. Timeout is defined as an 'int' but this shouldn't be a problem on currently supported operating systems BUGFIX: (#28288) Restore from StoreOnce fails RETR Failed. Command error with HPE StoreOnce server [COFC-CZ3935RB01]: OSCLT_ERR_INVALID_CREDENTIALS. Problem: Password decryption failed. Solution: Initialize variables before usage BUGFIX: DI failed to connect to HPE SO store using encrypted password. Problem: Password decryption failed. Solution: Initialize variables before usage NEW: (#17427) Set stop time for external backups/restores to at least one day and if a longer submit timeout is set then use submit 'timeout*2' NEW: (#17427) Default queue submit timeout for external backups/restores can be changed by setting 'gv_conf_extern_submit_timeout= BUGFIX: (#28280) Do not enable NBD on all network interfaces every time a backup is started. Enabling it an all interfaces is wrong as it may override user settings on some interfaces. Users must take care that their Citrix XEN servers have NBD enabled. This is expected to be the default state on the servers. BUGFIX (#28260) Backup trace log isn't shown in UI, if Sesam newday has been executed between submit and execution of backup job. Now Sesam date from DB is used to build log file name instead of current day BUGFIX: (#28264) support disks which are configured using pool/volume notation BUGFIX: VFS must set szStoreKey to get HPE Catalyst Store configuration BUGFIX: (#23813) Termination of running backup: Kill STPD process on client, if source side deduplication is active CHANGE: Use restore task name of main restore task for sub tasks too (vSphere, Xen...) BUGFIX: (#28181) Possible data lost when STPD timeout occurs at closing saveset operation. Problem: On timeout the same value returned as on EOD. Solution: Return different value on Timeout occurred to make distinguish between EOD and Timeout BUGFIX: (#28030) DB dead lock during update of DB:media.readcheck_state. Solution: PostgreSQL: Use semaphore to avoid concurrent update of table media. Additional: In case of new media update only given label BUGFIX: (#28256) Connection to HPE StoreOnce is not closed in case of restore. Solution: Call osCltData_CloseObjectDataSession() for restore too NEW: (#16233) Set backup state to failed, if checksum from client side doesn't match checksum from LIS (STPD) during 'closejob' operation. TODO: sbc_com_interface doesn't check error in resturn message NEW: (#17561) Write warning message to Sesam main log if number of saved items reported by SBC and in LIS file don't match. Warning is used, to find out all kinds of backups, where number doesn't match, but backup is successful BUGFIX: (#26665) Backup META data files couldn't be uploaded, because 'sbc -p (put)' uses same temporary fileas input. Solution: Add PID to filename BUGFIX: (#28150) 'sm_config_drives' adds Si3 crypt key twice to STPD config file, if write of INI file is retried (RDS not available). Do not append crypt key to drive options during every loop, but empty option string before BUGFIX: (#27564) Sesam wrongly counts Proxmox CPU cores instead of CPU sockets in license calculation. Solution: Proxmox does not provide socket count. Now 16 CPUs are treated as 1 socket NEW: Data store purge: Add new option '-o sem_wait=' to allow waiting for a longer time, if another purge is already running. BUGFIX: Remove '*_SG.lis' files generated by restore older then 14 days NEW: SAP Hana backup: EOL for hdbbackint trace file can be set by 'SSB_LOG_REMOVE_TIME=' in 'UTL' file NEW: Remove 'mig_all_finished_*' semaphore files older 14 days NEW: (#28035) Rewrite CLI client part (sm_cmd) in pure 'C' NEW: (#17561) Compare number of saved items from SBC and LIS file and return with error, if not match. Error return can be configured by 'gv_conf_backup_checks'. BUGFIX: (#28224) SQL error during generation saveset migration, if several drives can be used on source side NEW: (#16233) Compare checksum from LIS file and SBC for sub-savesets (vSphere, Xen...) and set backup status to error, if checksums do not match NEW: (#16233) Update 'DB:results.checksum' by checksum read from LIS file for external backups during 'closejob' operation BUGFIX: (#28181) Compare backup data size reported by sbc with written size to disk/SMS by STPD and set backup status to error, if they don't match BUGFIX: (#28181) Possible data lost when STPD timeout occurs at closing saveset operation. Problem: STPD count data checksum after data received but before data sent for write, which may cause wrong checksum if write failed. Solution: STPD must count data checksum only after data sent for write CHANGE: (#17768) Include Domino backup module in Sesam Linux kits BUGFIX: (#20299) STPD log shows cpio errors lost / junk bytes processing, but backup is successful. Problem: SBC read soft link size using stat command, which may return wrong size. Solution: Replace stat size using readlink size similar to cpio tool NEW: (#27088) Remove the entry 'LD_LIBRARY_PATH=/usr/lib/vmware-vix-disklib/lib64:$LD_LIBRARY_PATH' from 'sm.ini' during update, but only if entry matches exactly BUGFIX: (#27475) BSR Pro Windows: Source side dedplication backup doesn't work. Return interface with protocoll and port (http://client:11000), because parser in BSR library could not handle client name only BUGFIX: Source side deduplication for external backups: Set correct Si3 server listen port in device parameter CHANGE: (#23374) - vSphere restore of VM with independent VMDK does not give a hint that not all original disks have been restored NEW: (#27278) Write Sesam client version into backup main log (NOT). Because information is updated by every alive test it can be done in general without adding it to 'sbc_bsr' output CHANGE: (#20299) Abort backup, if STPD encounters problems analyzing CPIO or MTF stream. Defaul has been changed and old behaviour can be enforced by 'GLBV:gv_abort_after_CPIO|MTF_stream_problem=0' NEW: (#27410) Set index size to 'DB:hw_drives.block_size*100' for Si3-NG. Value from DB will be used only, if higher then calculated one BUGFIX: (#28192) Installation from UI don't work because username is not converted correctly. Solution: Replace all occurences of slashes (/) in the username with backslashes (\) before trying to connect to the client. BUGFIX: (#2070) New powershell script including all functions of sbc_oracle_rman.sh. BUGFIX: (#28195) Long exclude list overwrites backup STRUCT in 'sm_sbc_com'. Adopt definition to length of 'DB:tasks.exclude BUGFIX: (#28117) Also execute update routine when executing major update. BUGFIX: (#28162) Hyper-V VM complete restore finishes successful but VM is not recovered. Solution: Set MailMode for Hyper-V complete restore to set recover options BUGFIX: (#28034) Path restore of a Hyper-V vhdx disk ends with 'No entries found in LIS for...'. Problem: UI allows to select vhdx in Hyper-V VM selection view. Solution: Expand selected line to file item line and go on with file restore BUGFIX: (#26694) Set estimated data size of savesets to be migrated, if FDI chain function has been used to select savesets BUGFIX: (#23336) Fixed VFS mount for sm_vmdk_interface with pool argument. When mounting to a migrated saveset with -m pool has to be set in function calls to get_migrated_saveset and get_saveset_path NEW: (#28179) Use default size 1 MB for STPD Shared Memory Size and Tape Device Block Size BUGFIX: (#28178) Oracle runs into error during backup with 2 channels. Problem: Global dictionary for saveset information. Solution: Thread specific dictionary BUGFIX: (#23585) Do not show messsage 'E014-DRIVES Encryption analysis in drive XXX failed' in case encryption has already been turned on anymore. This is not an error. BUGFIX: (#28154) Mount of VMDK with same file name but on different data stores. Add data store name to path of VMDK NEW: sbc_iscsi: iscsiadm wrapper for sesam BUGFIX: (#28071) - Sporadically stpd crashes after MSSQL job because of ERROR: read_DBLK: Found unknown block. Cause: STPD is trying to analyze encrypted SQL data that is no sense as entire stream is encrypted. Solution: skip processing encrypted data (etc. TAPE block is not detected). BUGFIX: (#27987) Backup source contains '+' sign and multiple sources for path backup are set CHANGE: (#28109) Do not retry initialization of tape medium, if media load timeout has been reached. This has lead to an useless EOL extension for 4 tapes BUGFIX: (#28129) Changed the timeout of the zip file extraction from 300 sec to 600 second. Moved the logging to the log file to the call functions, since the logging in the subfunction does not work. Added some log messages to easen the log file analysis. BUGFIX: (#27835) Link 'libcurl' statically into 'sm_dedup_cli' to avoid addtional dependencies BUGFIX: (#25462) Add milliseconds to trace of SBC modules if gv_log_milliseconds is set. Problem: ANALYSE_BACKUP_LOG() did not take care about milliseconds NEW: (#27833) Enforce 'check_space' to update data store status after 'fsck' has been finished BUGFIX: (#28111) Oracle: Restore of controlfile from AUTOBACKUP fails: 'no AUTOBACKUP found or...'. Reset sbc-com message before every query, because not found message isn't reset in case of success NEW: (#28091) Set enforce check space flag after data store 'purge' has been executed NEW: (#28064) Re-start media init for backup, if current medium is in use by another drive. 'gv_conf_init_mounted_other_drive' has to set to enable function BUGFIX: (#28103) Check for active restore using 'DB:restore_results.saveset' and 'DB:restore_results.saveset_used' for right detection, if saveset is migrated NEW: (#17427) SAP Hana restore: Allow to set waiting time for restore medium by 'COM_RESTORE_TIMEOUT' in UTL file NEW: (#25757) Send backup data size to Sesam server for VMDK, so UI can show backup progress BUGFIX: (#28099) For automatic client and gui SP extraction sm_glbv has been used which is not present on client and gui systems. Solution: Use sm_client to read glbv instead. BUGFIX: (#28099) For automatic client and gui SP extraction sm_glbv has been used which is not present on client and gui systems. Solution: Use sm_client to read glbv instead. NEW: (#26578) Use new VDDK 7.0.1. NEW: (#27629) DB2 restore: Allow redirected restore into new database (DB name does not match name from backup) BUGFIX: (#27456) HCL Domino 11 COPY backup ends with warning. Solution: Now in case of COPY backup level the warnings about not logged databases are ignored NEW: (#28088) DB2 backup: Enforce match of DB:task.source and source set by DB2 to avoid using same backup task for different DB2 databases. Check has to be enabled by 'gv_conf_enforce_db2_source_match=1' BUGFIX: (#24043) Set 'DB:restore_resuls.size' correctly for restores using migrated savesets NEW: (#28081) move functions and type definitions for rhev stream backup to own file and class NEW: add example script how to reconstruct mountable qemu image from rhev backup stream format NEW: (#27541) Allow incremental backup for RHEV. Last change has been overwritten later in SQL script again BUGFIX: (#28066) Data restored from BF64 encrypted savesets does not decrypt files. Problem: Wrong encrypt password used while backup. Solution: Revise BF64 password usage while backup BUGFIX: (#27797) Kopano backup: Public folder backup does not delete dump files. Wait longer time for termination of backup thread NEW: (#27541) Allow incremental backup for RHEV. There is currently no check for older versions, not supporting incr. backups NEW: (#28039) insert switch SHOWINFOBOXES to allow installation/update of Sesam on Windows without user input. Call update as follows: sesam-srv-5.0.0.1-windows.x64.exe /v"SHOWINFOBOXES=0" BUGFIX: (#25327) Open only outbound firewall port 11401 for GUI installation. BUGFIX: (#25327) Remove unused ports 11402 and 11403 from the list of ports to be opened. NEW: (#27554) Do not allow Sesam server update from version < 4.4.3.42. NEW: (#27846) Encrypt S3 access key in 'stpd conf' file, if 'gv_conf_s3_secretkey_crypt' is set NEW: (#27846) Encrypt S3 secret key in 'stpd conf' file, if 'gv_conf_s3_secretkey_crypt' is set NEW: (#28045) Support of SAP Hana backint API 1.5 NEW: (#27541): CBT based Full backup for RHV with version 4.4. BUGFIX: (#27988) - File restore fails when Windows volume is mounted as folder and when source was 'all'. Cause: double slashes in the folder path during the restore operation. BUGFIX: (#28025) Remote installation of SP did not wait until extraction finished, hence ended with error. Now it should wait. BUGFIX: (#26823): sbc_imap2: local variable 'imap' referenced before assignment exception, rework imports BUGFIX: (#28000) Remote command hangs, if one line is longer than 16382 bytes. Increase read buffer to 32768 NEW: (#28011) "Sesam Java version check: Regex to used for check can be set in 'sm.ini' by 'java_version_check='. (java_version_check='version.*"1\.[89]\|version.*"1[012345][-.]'). default is '.....version.*"1[012]....' NEW: (#27554) Do not allow Sesam server update from version < 4.4.3.42. Correct version number to compare NEW: (#27554) Do not allow Sesam server update from version < 4.4.3.44 to 5.0 (Debian) NEW: sesam2000.profile: alias 'cat_err' to display current Sesam error protocol NEW: (#27554) Do not allow Sesam server update from version < 4.4.3.44 to 5.0 BUGFIX: (#26638) German umlaute in email subject are broken: change wording in sm_notify, do not use umlauts CHANGE: Disable login flag for MySQL and PostgreSQL until kernel will be correctly adopted NEW: (#27990) Add data store trash size to DB:data_stores.trash BUGFIX: (#27981) - MSSQL backup does not work anymore due to core of sbc. Access array's item violation. NEW: (#24753) Use 'sbc_ssl102' for vSphere backups as default NEW: (#24753) Allow to set 'sbc' directory to be used for vSphere backups. (-a sbc_ssl=_ssl102). TODO: Switch automatically, if default 'sbc' is linked agains OpenSSL 1.1.1 BUGFIX: (#27729) VFS mount of virtual disks on Linux was changed so that single guestmount command is executed for all virtual disks of the mounted Sesam saveset. It solves issues of mounting virtial disks without OS partitions NEW: (#16233) SAP Hana backup: Send saveset checksum from client side to Sesam server during 'closejob' call CHANGE: (#27649) - Restore a VMware VM into a file is offered by the GUI, but does not work BUGFIX: Backup: Update 'DB:results' in case of problems during argument parsing (for example not existing migration task). BUGFIX: (#27942) Restore fails if SMS data server does not reset block size to variable after write operation with LTO-7/LTO-8. Solution: Reset to variable block size for Linear Tape-Open drives which exceeds SEP sesam maximum SMS block size 1MB, e.g. LTO-7 or LTO-8 BUGFIX: (#26605) Backup to tape fails because of wrong (unsupported) read block_size - errno EBUSY. Solution: Decrease read block size until EBUSY is gone BUGFIX: (#27780) Rephrase the db insert statement to avoid quotes in update client STATUS and DB message. NEW: (#25289) Add check for correct Sesam date as Postgresql 'CHECK CONSTRAINT' BUGFIX: (#27933) VM backup fails, if VM name consists only of numbers (especially on Proxmox and OpenNebula) NEW: sm_exception: errno can be passed as parameter, defaults to ERROR_WARNING BUGFIX: (#27931) 'sm_dedup_interface' for HPE Catalyst doesn't work any. Adopt 'clean' rule to delete object files correctly Revert "BUGFIX: (#27361) Migrations to to tape fail with 550 CWD Failed. :3009: Unable to open first free detected pipe (0). Problem: Linux: Data thread terminated due to 'file lock with fcntl() failed: (35) Resource deadlock avoided' in x_Semaphore(). Solution: Avoid deadlock by calling GET_EXCLUSIVE_ACCESS [==: i_SEMA( '+', &sctExclBusySema )] always after GET_WORKING_ACCESS [==: i_SEMA( '+', &sctSetWorkingState )] - never before" Revert "BUGFIX: (#27361) Migrations to to tape fail with 550 CWD Failed. :3009: Unable to open first free detected pipe (0). Problem: Linux: Deadlock may occur if process requires exclusive semaphore while other has not yet released it. Logic changed - do not set working access semaphore before releaseing excluxive semaphore", because EOM does not work anymore NEW: Add Sesam meta data datastore, holding information of used disk space by Sesam meta data (DB, LIS..) NEW: (#25814) Implement function to delete obsolete GLBV from Windows Registry CHANGE: (#27593) - Skip call of 'sm_data_store write_saveset_info', if no saveset has been created NEW: (#21987) Simpler sm_db_update process. Solution: Windows: Add triggers from sesam_sqlite.sql to sesam_insert_start.sql to avoid active triggers during database import. Database creation sequense is sesam_db.sql, sesam_db_old(new).sql (export), sesam_insert_start.sql NEW (#27904) Set backup to failed, if vSphere specific information read from log file becomes bigger then 'DB:results.source' BUGFIX: (#27904) vSphere restore fails, because VM information is cut off. Do not cut off messages read from backup log after 512 bytes, but read up to 32 Kib BUGFIX: (#27915) Selective restore of Windows saveset may restore wrong files or it may fail with MTF error due to ESET after new segment. Solution: SSET generates wrong offset2 for ESET when using gci_curr_begin_dir_offset after ESET (must use gci_curr_begin_file_offset) NEW: (#17628) Check minimum free disk space of 5 GB on 'var' partition during Sesam server DEB installation BUGFIX: (#27361) Migrations to to tape fail with 550 CWD Failed. :3009: Unable to open first free detected pipe (0). Problem: Linux: Deadlock may occur if process requires exclusive semaphore while other has not yet released it. Logic changed - do not set working access semaphore before releaseing excluxive semaphore BUGFIX: (#27361) Migrations to to tape fail with 550 CWD Failed. :3009: Unable to open first free detected pipe (0). Problem: Linux: Data thread terminated due to 'file lock with fcntl() failed: (35) Resource deadlock avoided' in x_Semaphore(). Solution: Avoid deadlock by calling GET_EXCLUSIVE_ACCESS [==: i_SEMA( '+', &sctExclBusySema )] always after GET_WORKING_ACCESS [==: i_SEMA( '+', &sctSetWorkingState )] - never before BUGFIX: (#27907) do not enode configuration data during restore process BUGFIX: (#27907) fix Created configuration file does not have valid XML format during backup: do not attempt to encode configuration data NEW: Add saveset creation time to output from 'saveset tree' function CHANGE: (#27863) Do not log backups wit status warning in Sesam error log BUGFIX: (#26777) If system load is very high, 'sm_cmd' may hang and blocks other processes. Solution: Add idle timeout to RCMD() and set it to 1800 (1/2 hour) - override possible with GLBV 'gv_sm_cmd_idle_timeout' NEW: (#27852) sm_remote_installer: use exceptions for status messages and error handling where possible NEW: (#27852) sm_update_client: use exceptions for status messages and error handling where possible NEW: (#27852) sm_config_hw: use exceptions for status messages and error handling NEW: (#27852) sm_copy_files: use exceptions for status messages and error handling NEW: (#27852) sm_main: use exceptions for status messages and error handling instead of hardcoded STATUS=..MSG= printing NEW: (#27852) sm_replicate: use new exception handler from sm_exception NEW: (#27852) sm_remove_saveset: use new exception handler from sm_exception NEW: (#27852) sm_event: use new exception handler from sm_exception NEW: (#27852) sm_reformat_lis: use new exception handler from sm_exception NEW: (#27852) sm_cross_check: use new exception handler from sm_exception CHANGE: (#27891) Read next label required for restore (also migration read side) over EOM from DB:media.next_label only. This is much more relyable then searching over DB:restore_results (which will not work for migrations at all) NEW: (#27883) Automatic tape device configuration has to be switched on by setting 'sm_glbv w gv_conf_auto_hw_device 1', so it can be disabled when it doesn't work correctly NEW: (#17628) Check minimum free disk space of 5 GB on 'var' partition during Sesam server RPM installation NEW: (#27852) sm_data_store: use new exception handler from sm_exception NEW: (#27825) Check for all 'sbc-2xxx' message lines in backup log and set state to warning. Set warning status if current status is still undefined BUGFIX: (#26902) SESAM_BACKUP may end with warning on Windows system due to thread exit code STATUS_THREAD_IS_TERMINATING with disaster interface. Solution: Thread exit code STATUS_THREAD_IS_TERMINATING must be handled like ERROR_SUCCESS BUGFIX: (#27473) Hyper-V SFR from RDS fails with "The system cannot find the path specified". Problem: Function get_lisfile() did not return the correct .lsl file name which must contain '_SG_' generation restore prefix NEW: (#17628) Check minimum free disk space of 5 GB on 'var' partition during Sesam server RPM installation BUGFIX: (#27887) catch possible exception if event creation fails, it is not relevant to backup functionality BUGFIX: (#27887) random integer should not exceed maximum integer length BUGFIX: (#27887) fix External Event with this custom id already exists error: create event ID by using random function instead of time BUGFIX: (#27882) Do not remove DB:results entries for running backups after 7 days NEW: (#27852) introduce SesamArgumentError exception, on raise, program exits with ERROR_ARGUMENTS exitcode. Provide example code. NEW: Compress complete DB export at once, if compress flag is set BUGFIX: (#27877) Restore from RDX disk drive stalled ... stpd generates log files in endless loop. Solution: Since commit c865fa7 [(#27078) Restore of Linux saveset fails with 'Data Synchronization lost' when no tree structure should be created] the segment position was only calculated if saveset was splitted on further tapes NEW: (#27852) exception handler can now be used with python2 too, move example code and create build targets NEW: (#23990) Remove all savesets and meta data of a backup task, client or location BUGFIX: (#27877) Restore from RDX disk drive stalled ... stpd generates log files in endless loop. Solution: Do not set buffer filled flag after moving to new position BUGFIX: (#27874) Introduce media for DISK_HARD (e.g. RDX) ends with 'E001-DATABAS Failure during access to database: update media set encryption_capable='- Solution: Do not check for encryption capable in case of DISK_HARD. Additional: Add error from sm_arch to media action log (PROT). If label in DISK_HARD directory cannot be read than use folder (barcode) as label BUGFIX: (#27875) 'sm_data_store check_space' empties trash folder without day offset. NEW: (#21198) Remove all savesets and meta data of a backup task NEW: (#27852): add example how to catch exceptions globally using sys.excepthook, allow to define exit and errno options on decorator level during excepthook setup NEW: (#23784) Refactoring remove of backups and savesets NEW: (#27852): if exception log target is a regular file handle, log full traceback to file NEW: (#27852) allow output to both logfiles and stdout/stderr based on decorator or exception definition NEW: (#27852) add example for logfile writer defined on decorator level NEW: (#27852) add example for common exception handling BUGFIX: (#20414) Exit 'sm_reformat_lis' with 'STATUS= MSG=' in all cases. TODO: Check for correct error messages in case of failure CHANGE: (#27741) Analyze backup log: Log output from 'sm_reformat_lis' call only in trace level 2, to avoid big 'bck..log' files. Information is written to 'sm_reformat_lis' log file BUGFIX: (#27786) Concurrent processes created same unique ID. Solution: Avoid problem with DST (double hour) by using current time without DST for mode 'S' (SEP sesam old time stamps) BUGFIX: (#27786) Concurrent processes created same unique ID. Problem: When seconds changed timestamp may be generated with old value 'SS' in 'YYYYmmDDHHMMSS...', e.g. '20200812124531999' followed by '20200812124531000'. Solution: Avoid to call a 2nd timestamp generating function BUGFIX: (#27808) Consistent export of Sesam DB on Postgresql. Use 'pg_dump', instead of exporting every table in a loop BUGFIX: backint_sapdb: fix ImportError: No module named sbccom BUGFIX: fix backint_sapdb build: symlink required sm_common/sm_version files to right build directory Revert "BUGFIX: fix backint_sapdb build, add sm_common to build rules" BUGFIX: fix backint_sapdb build, add sm_common to build rules BUGFIX: (#27786) Concurrent processes created same unique ID. Solution: Add repeat logic also for type 'R' (Sesam old timestamp with random) BUGFIX: (#27786) Concurrent processes created same unique ID. Solution: Waiting for 1 ms does not assure that timestamp is changed. Now get timestamp before and repeat timestamp generation until timestamp changed BUGFIX: (#27786) Concurrent processes created same unique ID. Problem: When seconds changed timestamp may be generated with old value 'SS' in 'YYYYmmDDHHMMSS...', e.g. '20200812124531999' followed by '20200812124531000'. Solution: Avoid to call a 2nd timestamp generating function BUGFIX: (#27808) Consistent export of Sesam DB on Postgresql. Use 'pg_dump', instead of exporting every table in a loop BUGFIX: backint_sapdb: fix ImportError: No module named sbccom BUGFIX: fix backint_sapdb build: symlink required sm_common/sm_version files to right build directory Revert "BUGFIX: fix backint_sapdb build, add sm_common to build rules" BUGFIX: fix backint_sapdb build, add sm_common to build rules NEW: (#27825) Check for all 'sbc-2xxx' message lines in backup log and set state to warning BUGFIX: (#27786) Concurrent processes created same unique ID. Solution: Add repeat logic also for type 'R' (Sesam old timestamp with random) BUGFIX: (#27786) Concurrent processes created same unique ID. Solution: Waiting for 1 ms does not assure that timestamp is changed. Now get timestamp before and repeat timestamp generation until timestamp changed NEW: (#23784) Migration without DB:results entry. Adopt DB:result_lbls NEW: (#23784) Migration without DB:results entry. Do not use 'DB:results' of migrated saveset for restore BUGFIX: (#27716) Automatic hardware configuration: Set tape device names in DB doesn't work, because of DB errors. Declare UNIQUE constraint as DEFERRABLE Revert "BUGFIX: (#27716) Automatic hardware configuration: Set tape device names in DB doesn't work, because of DB errors. Declare UNIQUE constraint as DEFERRABLE" BUGFIX: (#27716) Automatic hardware configuration: Set tape device names in DB doesn't work, because of DB errors. Declare UNIQUE constraint as DEFERRABLE NEW: (#27716) HW configuration: Print out result of SQL statements correctly BUGFIX: (#27827) do not return function to early if removing multiple disks from data-mover CHANGE: (#23784) Migration without DB:results entry. Adaption: Now we have to use tuple [('{saveset}',(select backup_id from result_lbls where saveset='{saveset}'))] instead of [(select saveset from results where saveset='{saveset}' or original_saveset='{saveset}') NEW: (#27781) Remove old media logs BUGFIX (#27821) do not reset disk bus to VIRTIO_SCSI if IDE disk is restored BUGFIX: (#27821) sbc_rhev: fix restore with disk interface SATA. Set correct disk interface after disk attachment BUGFIX: (#27820) sbc_rhev: fix selection of default cluster if specified by user NEW: (#25798) Support major version >= 5.0 in 'sm_setup' NEW: (#20435) Write errors during DB access into Sesam day log. Solution: Add error message to daily log for PostgreSQL CHANGE: (#27786) 'iCREATE_UNIQUE_ID(): Wait 2 ms (instead of 1ms) before release lock, because function returned same value sometimes. The reason for this is unknown. Perhaps it only happens on Windows, but according to documentation 'Sleep()' should not return to early CHANGE: 'iCREATE_UNIQUE_ID(): Wait 2 ms (instead of 1ms) before release lock, because function returned same value sometimes. The reason for this is unknown. Perhaps it only happens on Windows, but according to documentation 'Sleep()' should not return to early CHANGE: (#27762) Add 'sesam-novell-client' to RPM 'Conflict' list for SLES12, because 'sbc_smdr' binary is already included in Sesam client (server/RDS) package BUGFIX: (#27726) - BSR Windows backup fails, when migration of same task is active at this time. Exclude migrations entries in DB:results during check BUGFIX: (#27473) - Hyper-V SFR from RDS fails with "The system cannot find the path specified. Correct error handling according to last comment CHANGE: (#27779) Re-activate backup type 'MaxDB' during new Sesam installation BUGFIX: (#27778) SQL query fails during real start of BSR backup by 'sm_sbc_com'. Query is for backup types with one main backup and serveral sub-backups like KVM, Citrix Xen. Fix: Do not execute query for BSR Windows BUGFIX: (#27473) - Hyper-V SFR from RDS fails with "The system cannot find the path specified". Support list directory with merged disks on the RDS BUGFIX: (#27772) Problem: Wrong SQL update statement of Hyper-V backup. Solution: erase wrong single quote from SQL update statement to fix the SQL update statement. BUGFIX: (#27769) Correct handling of SQL statements using backup source longer then 1024 or 2048 characters BUGFIX: (#27769) Update statement for 'DB:results.source' fails, because size of SQL command string is to long. Solution: Use 'SQLLEN' for command. TODO: Return with error, if source becomes still to long BUGFIX: (#27646) 'sm_java server_debug' does not start the REST server in debug mode. This only worked using the SM_GUI_SERVER entry in the debug.ini until now. Solution: also enable functionality via command line. BUGFIX: (#27714) Multiplex backup to tape writes corrupt data. Problem: Oracle saveset name may be longer than 32 BUGFIX: (#24919) use echo F in front of xcopy to tell xcopy that a file is being copied. BUGFIX: (#27714) Multiplex backup to tape writes corrupt data. Problem: Race condition could lead to invalid thread_ID > 99 BUGFIX: include /Y to xcopy. BUGFIX: (#24919) use xcopy /X instead of copy to rename sm.ini etc. Copies file audit settings (implies /O which copies file ownership and ACL information.) CHANGE: (#27669) Remove obsolete Exchange Server 2003 ex2k.dll library NEW: Set default Java heap space to 4GiB for Si3-NG BUGFIX: (#27175) Failed backups during drive configuration, because drive queue isn't created yet: Retry submit of executable to a queue for 2 min, if queue doesn't exist. BUGFIX: (#27675) Dismount medium from source drive, if 'sm_loader' loads a tape from one drive into another. Because of missing dismount further actions like restore could fail, because they assume, that requested medium is still in drive NEW: Set Sesam client: Retry getting Sesam server SSH public key using 'https' transfer CHANGE: (#27583) Write tape status 'STATUS= MSG=' message from 'sm_client' to STDOUT instead of 'STDERR' because message on STDERR is read by caller after another 'STATUS=' message written to STDOUT (Wrong order). But caller uses latest 'STATUS=' line for further processing, which is wrong then BUGFIX (#27686) Restore VM over RDS with transport mode https fails: pass tapeserver to sbccom openjob call during restore too BUGFIX: (#27686): Restore VM over RDS with transport mode https fails: pass tapeserver to sbccom openjob call during restore too BUGFIX: (#27679) STPD_Main cores if MAX_CONNECTIONS exceeds. Problem: Close handle without check may cause core. Solution: Add handle check before close BUGFIX: (#23064) Remove savesets with 'DB:result_lbls.segment=-1' during purge by 'sm_pur_status' older 7 days BUGFIX: (#27679) STPD_Main cores if MAX_CONNECTIONS exceeds. Problem: Sub process PID does not saved, which cause incorrect check is process alive. Solution: Save PID for further checks BUGFIX: Oracle module: Ignore unknown keys returned by Sesam server, instead of stop with error. This allows Sesam server to send more information, without checking client version before CHANGE: (#27654) HPE StoreOnce compresses data with the Catalyst profile 'High Bandwidth'. CHANGE: Only in cse of Low Bandwidth Mode the Payload Compression and the Payload Checksums are enabled BUGFIX: (#27650) Start of http process blocks STPD main too long. Problem: STPD_MAIN wait for STPD_HTTP to be started, and this take a long because of lot stuff (HPE communication, Authorization). Solution: Remove long running operations from start process lock Revert "BUGFIX: (#27650) Start of http process blocks STPD main too long. Problem: STPD_MAIN wait for STPD_HTTP to be started, and this take a long because of lot stuff (HPE communication, Authorization). Solution: Remove long running operations from start process lock" BUGFIX: (#27650) Start of http process blocks STPD main too long. Problem: STPD_MAIN wait for STPD_HTTP to be started, and this take a long because of lot stuff (HPE communication, Authorization). Solution: Remove long running operations from start process lock BUGFIX: (#27595) oracle.exe crashes during backup with 2 channels. Problem: Global array allocated and use from different threads. Solution: Use TLS to make array thread safe. BUGFIX: (#19512) Do not add savesets to 'DB:result_lbls' during EOM, which are not really active (there is no active data connection to STPD). CHANGE: (#27598) Set 'LC_ALL' in sbc_vadp to 'C.UTF-8' instead of 'de_DE.UTF-8' because German language pack isn't available by default on all Linux systems BUGFIX: (#27630) Do not continue with post installation steps if an error occured during installation. CHANGE: (#24989) Extent command line length for 'commands' to 2048 CHANGE: (#27614): 'gv_conf_wait_reload_sms' to set sleep time in seconds between stop and start by 'sm_main reload sms' CHANGE: (#27614) Longer wait (5 sec) between stop and start during 'sm_main reload sms' BUGFIX (#27584): proxmox: wait for other vzdump tasks on the same node before attempting to transfer data BUGFIX: (#27502) - sbc cores during HCL Domino COPY backup. Possible reason: Access to memory that was free before. NEW: Show front side license volume used for level 2 and 3 backup tasks in normal license output and not in extended only NEW: (#27577) Avoid deadlock by 2 external restores requesting tape medium mounted in other drive, then restore has started on NEW: (#27586) Insert release field to enforce SP update if version and SP version differ. Release field has to be adjusted to the version of the SP than SP update is possible even if installed version and SP version differ. BUGFIX: (#27617) orasbt.dll throws application fault while sbttest. Problem: orasbt.dll core while function with unknown signature called. Solution: Add missing include file NEW: (#27569) Add support for Kopano 8.7.12 (kopano-admin output has been changed) BUGFIX: (#27615): deactivate disk on data-mover before de-attaching BUGFIX: (#27561) Remove cvs_revision and cvs_id from installshield since it is deprecated. BUGFIX: (#27459) During automatic update BSR has not been uninstalled since version comparison was evaluted wrong. Now evalution has been fixed. BUGFIX: (#27595) oracle.exe crashes during backup with 2 channels Problem: XBSA BSAGetLastError function is not thread safe. Solution: Use thread-local storage variable for XBSA LastErrorMessage NEW: proxmox: backup via zfs: allow to disable zfs and fallback to vzdump if -a trans=vzdump is specified BUGFIX: (#27377) - Backup or browsing of Hyper-V VMs fails if Windows was configured to use hungarian language. Logic around get list of VMs is changed without using literals and constant strings BUGFIX: (#27586) Insert an enforce switch for SP update. If release field is present in sm.ini than this version is used to compare with the SP version, otherwise the version field is used. BUGFIX: (#27459) Adjust the sm_execute_update.cmd which is needed for the automatic update of windows clients. BUGFIX (#27564): sbc_pxmx: use cpu_count instead of cpu_sockets in dir output CHANGE: (#26518) DB2 inquire: Sort savesets by 'start_time' instead of 'saveset' because 'saveset' isn't a teim stamp any more CHANGE: (#27557) Increase maximum allowed connection for STPD in 'stpd.ini' to 1024 during update CHANGE: (#27557) Increase maximum allowed connection for STPD in 'stpd.ini' to 1024 during installation BUGFIX: (#27560) 2 STPD process are running on the same stream number at the same time. Problem: Race condition between different STPD processes for the running stream. Solution: Use lock guard for new STPD process creation until process started BUGFIX: (#27563) Kopano backup removes own full index files needed for incr. backups after 7 days CHANGE: (#27389) Drop trigger to set mtime for DB:migration_results because it can lock DB for a very long time during Sesam newday CHANGE: (#27527) Set stricter permissions for PostgreSQL socket file and set group for it to 'sesam' BUGFIX: (#27556) DB2 rollforward db failed with "SQL2062" while retrieving log file "S0000187.LOG" for database "RTD" on database partition "0" and log stream "0". Problem: The results.source was set with sequence number 0. Solution: Retry in case of sequence number 1 with 0 NEW: If SQL statment failed with SQLITE then add message to SEP sesam daily log - {day}.prot file NEW: (#26912) Remove 'glbv' used by 'sm_sbc_com_ext' after 'del_entry' by caller to avoid thousands of not necessary 'glbvs' BUGFIX: (#27502) sbc cores during HCL Domino COPY backup. Problem: Concurrent thread may access same string array. Solution: Access to string array only in critical section BUGFIX: (#27293) SAP Hana doesn't recognize successful backup correctly. Problem: SAP Hana backup ends with 'backint did not respond for 600 seconds'. Solution: Avoid waiting for semaphore NEW: Purge of old SAP Hana Sesam trace files: Do not walk over directory tree for performance reasons NEW: (#27521) Set stricter permissions for files in 'gv_rw_ini:ssl' during Sesam start CHANGE: REAR: check for multipath configurationon all linux variants, not only SLES4SAP NEW: (#27518) Submit long running purge of some Sesam trace files like 'gv_rw_lgc:' and 'gv_rw_stpd:' into queue, so 'sm_pure_status' returns quick BUGFIX: (#27506) sm_config_drives crashes when configuring HPE StoreOnce Data Stores. Solution: Avoid PROT() with arbitrary data in format specifier. Adapt scanf for HPE answer with 'Physical_Storage_Quota_Byte' or 'Logical_Storage_Quota_Bytes' BUGFIX: (#27422) One data block isn't written to next tape after EOM. Solution: Write data block first and then the file mark with segment header but only if EOM area was not yet reached BUGFIX: (#23232) sm_alarm message is sent to the customer every 4 minutes when waiting for requested media in a tape drive without loader. Solution: Call sm_alarm only in the beginning and after another wrong label is detected. Call sm_notify in the beginning and after 360 minutes (override with GLBV gv_notify_single_drive_timeout) BUGFIX: (#27508) SAP Hana restore/inquire: Do not return migrated savesets, if SAP Hana queries backups done by Sesam NEW: (#27453) Build vSphere module using VDDK 7.0.0 BUGFIX: (#27389) Drop trigger to set mtime for DB:sesam_events because it can lock DB for a very long time during Sesam newday BUGFIX: (#27498) Core dump of sm_data_server in 'x_DumpEx()' function. Solution: Avoid to call trace function with arbitrary data in format argument CHANGE: (#21987) Update of Sesam server: Exclude Sesam database backup directory from 'sbc' backup created before update starts, because in can become to big BUGFIX: (#25064) New BSR registry key for DiskImage 15.0 is missing. Problem: Since 15.0 the DefaultPath doe not end with path delimiter. Solution: Use os.path.join(directory,filename) to generate proper file name NEW: (#27489) Include O&O version 15.4.196 - in Sesam kit build BUGFIX: (#27478) All child processes started by sm_startup hang on Windows 2012R2. Problem: New improved read output function is stalled on some Windows Server 2012 R2 systems. Solution: Use old pipe reader if operating system is Windows Server 2012 R2 or older. Allow to override with GLBV gv_new_read or gv_old_read BUGFIX: (#27480) 'sm_config_drives' removes loader assignment of disk change (SayFuse) drives NEW: Enforce check space for data stores at once after (re) start of 'sm_sms_watch' BUGFIX: (#27478) All child processes started by sm_startup hang on Windows 2012R2. Problem: New improved read output function is stalled on some Windows Server 2012 R2 systems. Solution: Use new function only if GLBV gv_new_read is set BUGFIX: (#26605) Backup to tape fails because of wrong (unsupported) read block_size - errno EBUSY. Solution: Set Linux default block sizes for LTO-3/LTO-4/LTO-5/LTO-6 reduced to 128KB BUGFIX: (#26214) Purge of DataStore does not work anymore when vSphere VM is part of VM-folder with mutated vowels. Solution: Decode UTF-8 with ignore errors to skip non utf8 chars BUGFIX: (#25968) Adjusted the text of the servicepack specific popup, for a better understanding. BUGFIX: (#27447) User defined location of SEP sesam packages does not work when path does not end with slash sign. Solution: Add slash at the end if missing. BUGFIX: (#27440) DB2 restore: Do not return migrated savesets, if DB2 queries backups done by Sesam BUGFIX: (#27400) Oracle backup must not start when communication between STPD and libobk fails. Solution: Check com_intf() return state and message and check if XBSA host is set. BUGFIX: (#27433) Check interface (http://...) during vSphere restore against data mover and not client name, because on vCenter no Sesam client is (mandatory) be installed BUGFIX: (#27425) Oracle RMAN multi channel restore does not proper end the SBC_COM session. Solution: Move COM state structure to session to assure proper logic BUGFIX: (#27426) Changed the german text of the error message to better german. BUGFIX: (#27432) Added missing MsiCloseHandle called for all calls of MsiRecordSetInteger and replaced CloseHandle calls with MsiCloseHandle due to possible some memory leak issues with CloseHandle. Also made a small code review and removed all unneeded comments. NEW: (#27315) Options needed for SAP ASE to backup several databases instances on one ASE server. Solution: With new option -a SYBASE={path_to_SYBASE.env} a specific SYBASE path can be set to locate SYBASE.env file BUGFIX: (#26797) insert missing quotes. BUGFIX: (#26797) Execute submodules build using STPD_STATIC=1. BUGFIX: (#26797) Execute submodules build using STPD_STATIC=1. BUGFIX: (#26797) remove the gmp.dll, gnutls.dll, libhogweed.dll, and libnettle.dll from the msi build since executables are now build statically. BUGFIX: (#26797) sm_stpd_main.exe dumps core if custom specified SSL certificates are used. Problem: Using customer CERT cause STPD_MAIN core in nettle lib calling gmp function. Solution: Allow STPD_MAIN build with static dependencies to gnutls/nettle/gmp (STPD_STATIC=1) BUGFIX: (#26797) sm_stpd_main.exe dumps core if custom specified SSL certificates are used. Problem: Using customer CERT cause STPD_MAIN core in nettle lib calling gmp function. Solution: Allow STPD_MAIN build with static dependencies to gnutls/nettle/gmp (STATIC=1) NEW: (#27288) Detection of broken 'sbc_com openjob' connections by using 'keep-alive' packets. sbc and sbc_com_interface have to be updated on client side in addition to 'hdbbackint' BUGFIX: (#27354) Small migration as well as external backup/restore jobs take longer than needed due to signal handing of 'sm_sbc_com_ext'. Solution: Sleep() on Windows does not abort when signal is catched therefore loop with sleep(1) and while ( ... && gcInterrupt!='y' ) must be used with macro interruptable_sleep(A) BUGFIX: (#19381) Extend sm_java command length as well as the command length of i_cm_execvp... commands to ensure that the command works. In case the command exceeds the command length an error is returned. BUGFIX: (#27413) Restore pattern filter in sbc Linux is ignored CHANGE: EOM during restore: Use 'DB:media:next_label' to fetch next medium (if label is invalid from tape trailer) before trying to get it by active restore DB entry. That's much simpler way to do it BUGFIX: (#27386) Update of Sesam GUI RPM package: Ignore error, that Sesam service not running, because this is OK for a GUI only installation NEW: (#4804) Remove sm_system_tray.exe and sm_gui_access.exe Components from the msi package. NEW: (#4804) Remove all system tray specific called from msi installer and from sm_setup. NEW: (#27388) Inserted the new sesam_db_pre.sql in the build process. It is appended at the beginning of the sesam_insert_start.sql. NEW: (#4804) Removed the sm_system_tray and the sm_gui_access from build process. BUGFIX: (#27353) Oracle RMAN multi channel backup failed with error ORA-19506: failed to create sequential file. Solution: Restore requires spSMS->sSMS_XBSA.szTape BUGFIX: (#27353) Oracle RMAN multi channel backup failed with error ORA-19506: failed to create sequential file. Solution: Adapted for new com_int() with structure 'struct _sSMS' to set sXBSA_Descriptor.cpSavesetSpec to proper STOR or RETR argument BUGFIX: (#27353) Oracle RMAN multi channel backup failed with error ORA-19506: failed to create sequential file. Solution: Adapted for new com_int() with structure 'struct _sSMS' to set sXBSA_Descriptor.cpSavesetSpec to proper STOR or RETR argument NEW: (#26907) Start 'sm_data_store purge/fsck' process in drive queue (qu_trans) by 'sm_sms_watch'. Do start in first drive of data store and do not start, if already running. Do not execute 'purge' and 'fsck' during same time NEW: (#4804) remove SEP sesam system tray from the Windows installer. NEW: Use Postgresql 12 for on Ubuntu 20.04 NEW: Allow Postgresql 12 for Debian 10 to support Ubuntu 20.04 NEW: (#27396) Enforce faster check space for data stores, if Si3 service changes from unavailable to running NEW: (#25968) Insert a checkbox during installation and update which asks the user for servicepack update. The answer is passed to the sm_setup who inserts the information into defaults value gui.enable.action.install.update as well as in the sm.ini using 'install_servicepack' NEW: Update 'DB:data_store.msg' during 'check_space' operation too BUGFIX: (#26906) Reduce use of blocking semaphores. Solution: int_INSERT_INTO_RESULTS(): Avoid blocking semaphore 'get_unique_savesetid' use task specific 'get_unique_cnt_{task}' instead NEW: (#24690) Update 'DB:result_lbls.backup_id' during Sesam update and add index for it NEW: (#24690) Add 'DB:result_lbls.backup_id' referencing 'DB:results.saveset' of original backup bugfix: Correct date format for path to migration log file NEW: (#24690) Add 'DB:result_lbls.backup_id' referencing 'DB:results.saveset' of original backup CHANGE: Use 'DB:restore_tasks.client_id' instead of of 'DB:restore_tasks.client'. BUGFIX: (#27353) Oracle Backup failed with error ORA-19506: failed to create sequential file. Solution: XBSA function BSAHTTPHeaderCallback() was not thread safe, now only thread specific data is used BUGFIX: (#22814) Missing some symbols in NOT log. Problem: sm_sshd may miss some symbols in stderr output due to buffering based on '\n' parser. Solution: disable buffering, as it cause much troubles and should be revised or removed BUGFIX: (#26605) Backup to tape fails because of wrong (unsupported) read block_size - errno EBUSY. Solution: Set Linux default block sizes for LTO-6/LTO-7 and LTO-M8 to 256KB NEW: (#27368) Execute post processing of migration task in other non drive queue, if 'gv_conf_que_move_copy_post=' is set NEW: 'que_search()' to search a QM entry using 'sm_qm shocsv' output. Add queue name to return values, because search can be done for all queues BUGFIX: sm_break: Add 'sm_sms_copy' to signaled modules NEW: 'que_search()' to search a QM entry using 'sm_qm shocsv' output BUGFIX: (#27353) Oracle Backup failed with error ORA-19506: failed to create sequential file. Solution: XBSA library must be compiled with /MD flag to use the multithread-specific and DLL-specific version of the run-time library. BUGFIX: (#27351) Program sm_restore or sm_sm_copy may core dump if prot file is not accessible. Solution: Use function RESTORE_PROT1() to ensure proper processing CHANGE: Remove not necessary call of 'i_CHECK_MIGRATED_SAVESET()' in 'sm_sms_copy' BUGFIX: (#27347) Enlarge size of 'szCnt' (used for DB:results.cnt) to 16 charaters to allow 'cnt>9999'. Has caused core dump in 'sm_sms_copy' BUGFIX: (#27342) Access check should not be executed after new installation of a server since no access is possible to server after installation. NEW: (#11752) Put migration main log files into 'gv_rw_lis:/migration/' sub-directory NEW: (#27327) qm_main supports new command move_entry that moves job from one queue to another NEW: (#19241) Improve message text, if external backup task is stopped in pending state NEW: (#27306) Option to reduce default log messages in sbc Linux. If '-o shortlog' option is set, information like Sesam kit name or default exclude list isn't logged. Useful for backup types like SAP Hana, where 'sbc' is called 1000 times an hour, to reduce redundant information BUGFIX: (#19315) Backup fails with 'Medium is mounted' if the current medium of the tape drive is in mount state in another drive. Solution: Try to occupy other drive (sm_sbc_com_ext) and if successful then load it into required tape drive. NEW: (#27288) Correct termination of SAP Hana backups on Sesam server, if 'hdbbackint' gets termination signal from SAP Hana NEW: (#27288) Improve termination of running SAP Hana restore jobs by stopping pending requests first. This will avoid, that pending jobs will become active, if currently active ones terminate. Add stricter check for 'sbc' output during restore, if action was successful. NEW: (#27288) Improve termination of running SAP Hana backup jobs by stopping pending requests first. This will avoid, that pending jobs will become active, if currently active ones terminate CHANGE (#27294) Do not call 'alive_test_sbc()' for during migration. This was a left over from 'sm_restore.c' CHANGE: (#27324): sm_sshd needs up to 10 sec to establish connection. Problem: 'gethostbyaddr() could cost a lot of time. Solution: Remove 'gethostbyaddr()' used to print host name in log file only BUGFIX: Correct 'http' port for client to server access check BUGFIX: (#27250) Error during read operation doesn't return error correctly to Oracle and Sesam server. Solution: All BSA errors are now passed to SOB BUGFIX: (#27251) Oracle session list is not implemented correctly. Solution: xbsa_open must use Thread-local storage (TLS) for static variables BUGFIX: (#27251) Oracle session list is not implemented correctly. Solution: Now an array of sob sessions is used to allow direct access of sob session via handle (index) Revert "Revert "BUGFIX: (#27067) - Data of submounted volumes on Windows is not backed up"" NEW: (#27026) Correct path relocation for restore over mount saveset on Windows. Check if mount point is a dirctory or a Windows drive letter NEW: (#27000) SAP Hana multi stream backup and restore: Do not continue, if a backup or restore fails, but return with error for all remaining items BUGFIX: Si3 sm_data_store fsck: Do not call 's3 status' command, if no S3 connetion is configured NEW: (#26900) Avoid using of wild cards to find backup meta files for upload to data store, but do it by checking all possible file names directly (try/error). That's much faster then using wild cards, if many files are in LIS directory CHANGE: (#27294) 'i_FIND_MEDIUM()': Use segment number from DB:result_lbls, instead of searching it in SGM file NEW: (#27179) Include O&O version 15.3.179 in kit build. BUGFIX: (#27263) FSCK for data stores may report inconsistence, when backup starts before 'FSCK' and ends before DB select for active backups. CHANGE: URL encode saveset info in 'sbc -i' parameter CHANGE: SAP Hana module: Reduce default pipe read/write timeout for 'sbc' to 60 sec. Can be adopted by 'UTL' file: 'SBC_TIMEOUT=' NEW: SAP Hana backup: Extend check of backup pipe name to detect DIFF or INCR backup level CHANGE: (#27074) Reduce number of license checks. Solution: Instead of checking full license for Server and timeout only timeout value is read from database BUGFIX: (#27067) - Data of submounted volumes on Windows is not backed up CHANGE: backint SAP Hana: do not enclose backup id by quotes for '#SAVED' line in output file Revert "BUGFIX: (#27067) - Data of submounted volumes on Windows is not backed up" BUGFIX: (#27282) SAP ASE backups end with error: Failed to load library. Problem: Sourcing the LD_LIBRARY_PATH failed. Solution: If LD_LIBRARY_PATH is not set then subsitute "$LD_LIBRARY_PATH" with empty string CHANGE: (#27074) Reduce number of license checks. In case of CHECK_LICENSE for a license key fetch value from SEP sesam database NEW: SAP/Oracle backint: Write 'sbc' and 'sbc_com' command line to STDERR NEW: (#25064) New BSR registry key for DiskImage 15.0 is missing. Add registry key detection for version 15.0.xx to executable sbc_bsr.exe and sm_client.exe BUGFIX: (#27196) Inserted several database inserts to easen the error analysis using SEP sesam GUI. BUGFIX: (#27272) Restore of external backup does not work, if saveset is migrated to RDS BUGFIX: (#27067) - Data of submounted volumes on Windows is not backed up NEW: (#20346) Source side deduplication for external backups without mandatory SESAM_TAPESERVER setting NEW: (#20346) Set interface for sbc (-S) to client name in case of source side deduplication NEW: (#20346) Simplify configuration of source side deduplication for external backups NEW: (#20346) Source side deduplication for external backups: Set tape server (-S) to client name automatically. SESAM_TAPESERVER variable is not needed to be set any more NEW: (#27274) Set source side deduplication device, if sbc is called with 'com' option. (Start of sbc on client side BUGFIX: Calculate index size for SI3-NG correctly BUGFIX: (#27263) FSCK for data stores may report inconsistence, when backup starts before 'FSCK' and ends before DB select for active backups. NEW: If 'DB:migration_tasks/events.delete_flag=2': Do not start purge of saveset after migration at once, but set EOL into past only NEW: If 'gv_conf_no_purge_after_migration' is set: Do not start purge of saveset after migration at once, but set EOL into past only CHANGE: (#18629) Do not start migration for savesets, if EOL of it is already gone NEW: (#20346) Source side deduplication fro external backups: Set tape server (-S) to client name automatically. SESAM_TAPESERVER variable is not needed to be set any more BUGFIX: (#27196) Determination of the BSR version failed due to missing quotes. Enhanced the error logging during BSR uninstallation and installation. NEW: (#27230) Check if 'HPE Bandwidth Optimized Data Transfer' is available on backup client BUGFIX: (#27259) Set EOL for savesets imported from S3 snaphot correctly by adding EOL from media pool to current date BUGFIX: (#27260) Avoid DB errors during calculation of DB:cal_sheets. End time for calculation of a day has overlaped because of daylight saving time (one hour added: 2020-03-29 00:00:00 - 2020-03-30 01:00:00) BUGFIX: (#27196) Some directory separator were still missing during the installation of BSR Pro. Should now work as expected. NEW: (#27255) SAP on Oracle backup: Support for HPE bandwidth optimized transfer BUGFIX: (#27196) Used the wrong registry key for SEP sesam BSR Pro, no use the right one. BUGFIX: (#27196) Added registry insert statements to adjust the app data folder in the registry of the SEP sesam BSR Pro installation. In case of the installation this is done before the SEP sesam BSR Pro installation in case of the update using sm_execute_update.cmd it is done after the SEP sesam BSR Pro installation. BUGFIX: (#26304) Removed the program sm_java.exe from the firewall rule. Now port 11401 is open for all programs. NEW: (#27252) Log level for Oracle backup: Set XBSA log level after getting it over 'SEND' command again to have same XBSA logs as if set by 'ENV' command BUGFIX: (#27196) Pass installation var and bin directory to BSR Pro installation. Workaround: Even with installation under 'D:' the registry pathes may refer to volume 'C:' so adapt registry keys appropriate BUGFIX: (#27196) The path for the subdirectories OO_SOFTWARE_APPDATA may not include the subdirectory 'SEP sesam BSR Pro' since it is automatically added by O&O CHANGE: (#27172) - Citrix XenServer can be CBT based backed up by using certicate signed by CA CHANGE: Retry 'sm_dedup_interface df' command, if format of output isn't correct CHANGE: Try to update recent failed and matching 'DB:migration_results' entry as successful, if a saveset has been successfully added to 'DB:result_lbls'. NEW: No sleep after submit of 'sm_sms_watch' in case of data store drive BUGFIX: (#27182) - Exchange DAG backup fails if the sesam service user have no administrative access to the entire Exchange organization CHANGE: Restore with verify option set: Set status to failed, if no checksum could be found in 'sbc' log BUGFIX: Compare of checksum hasn't been done any more during restore with 'verify' option set, because of wrong 'sscanf()' format specification BUGFIX: (#27196) Additionally use the parameters OO_SOFTWARE for the bin directory and OO_SOFTWARE_APPDATA for the var directory of BSR Pro. BUGFIX: #(19312) - Backup with source 'VSS:/Performance Counters Writer' failed with wrong item [\\?\GLOBALROOT\Device\HarddiskVolumeShadowCopy4\Windows\System32\Prfc????.dat] BUGFIX: SAP Hana backup script: Do not execute housekeeping, if backup has failed CHANGE: (#24196) Do not copy 'sm_ui.jar' to 'sm_gui.jar', but remove 'sm_gui.jar' during update BUGFIX: (#27225) 'sm_restore' creates wrong SGM file name, if restore should use migrated saveset. 'cnt' and 'sesam_date' will be read from 'DB:results' again, after saveset id to be used has been set BUGFIX: (#27223) Restore fails, because restore tries to use a 'write' only drive BUGFIX: (#27222) Set correct queue allocation (media pool) for external backups and migration write sied again BUGFIX: (#25670) - Browsing Exchange Server 2019 source does not longer set the 'not_selectable' for logs and files BUGFIX: (#26649) The package list 'index.txt' is automatically refreshed if the modification time of the list is older than the modification time of at least one file in the directory or subdirectory where the package list is located. NEW: (#27208) Remove temp files under 'gv_rw_work:templates/rythm' by 'sm_pur_status' BUGFIX (#27209) Correct SQL statement in for deleting orphaned task group results, so it works on SQLITE correctly too NEW: Calculate 'DB:restore_results.duration' at end of restore for external backups. Fill 'DB:restore_results.data_size' with exact byte value NEW: Calculate 'DB:restore_results.duration' at end of restore CHANGE: (#27205) Set status of restore to 'WARNING' instead of 'ERROR', if ACLs couldn't be restored NEW: (#26998) Restore performance from tape is 2-3 times slower than the backup performance. Solution: Fill 2nd buffer while processing 1st buffer, EOF handling adapted BUGFIX: (#26392) - 'sm_reformat_lis' complains about missing Exchange log files from excluded database BUGFIX: (#27196) Pass installation var and bin directory to BSR Pro installation. BUGFIX: (#27190) Restore of a replicated saveset with preferred media may use origin saveset. Solution: Use only the preferred media pool in the SQL where clause NEW: (#26649) if parameter -R is given index.txt is newly generated. Subdirectories are also considered for generating index.txt now. BUGFIX: (#27177) - Restore of a selected Exchange database with import to original database via RDB fails with 'Failed to match string from msl file' NEW: Do not show 'language plpgsql already exists' as error during DB import NEW: (#27179) Include O&O version 14.5.415 in kit build. BUGFIX: (#27175) Initialize all queues at the very beginning of config drive process to avoid errors, if backups are already been started during sm_config_drives is still active BUGFIX: (#27149) - Hyper-V Cluster VM backup fails with 'file is not found in the LSL file'. Cause: mess with lower/upper characters handling in the LSL file parser. BUGFIX: (#27163) - Restore of Exchange database via generated RDB fails with '.COMException (0x80004005): Database::Dismount:ExchangeManagementShell::DismountDatabase: Couldn't dismount the database that you specified.' NEW: (#27003) Set DB:restore_results.media_pool for external restores BUGFIX: (#26841) Always execute log on as a service policy if runas different user has been selected. BUGFIX: (#26841) Give inserted service user 'Log on as a service' policy during installation as well as if service user is changed via services. Use newly inserted parameter logon_as_service of sm_qm_main.exe during installation. BUGFIX: (#27163) - Restore of Exchange database via generated RDB fails with '.COMException (0x80004005): Database::Dismount:ExchangeManagementShell::DismountDatabase: Couldn't dismount the database that you specified. BUGFIX: (#27161) Installation of new client failed because no fitting package could be found. Solution: Architecture for a new client is None. Use a default architecture x64 in this case. NEW: (#16513) Start archive adjustment in 'qu_all', if 'over barcode only' is selected, so it is not blocked by running backups BUGFIX: (#27158) Restore of Exchange database via generated RDB leads to overwriting of the original database and data loss. Solution: MTF meta data stream must be closed with ESET BUGFIX: (#26431) Do not set error message to 'Checksum of saveset copy does not match' for failed backups, but continue to check log file for real error BUGFIX: (#22663) Failed backups with message 'Operation successful'. This usually happened for backup types using sub tasks (like vSphere), because a 'sbc' success message has been found in log output before error message. Now message will be overwritten with error CHANGE: (#27147) Remove '_pg' (for PostgreSQL) postfix from Linux Sesam server Debian package names BUGFIX: (#27103) Windows: qm_lib uses PeekNamedPipe and burns too much CPU power. Solution: If ReadFile() failed with error ERROR_PIPE_LISTENING then avoid WaitForSingleObject() NEW: (#27132) Break for external backups: Send signal to sbc on client side for SAP Hana and SAP Oracle, because in this case we have a running sbc CHANGE: (#27147) Remove '_pg' (for PostgreSQL) postfix from Linux Sesam server package names BUGFIX: (#27129) Restore of Exchange database ends with 'Empty Saveset' error. Solution: If mailbox database is selected in the Mailbox view then restore whole mailbox database CHANGE: (#27138) Allow to set timeout for suspending VM during cold Citrix XenServer backup BUGFIX: (#27132) Termination for external backups/restores: Add 'sm_sbc_com_ext' to list of signalled modules BUGFIX: (#27123) Oracle Restore Validate with more then 64 Pieces failed. Problem: When using HTTP interface the session was not released in BSATerminate(). Solution: Release XBSA and SOB session. NEW: Check 'DB:tasks.backup_srv_options' for 'use_ts_db' and 'use_ts_file' to switch between timestamp from DB or file system for INCR/DIFF backups BUGFIX: (#17924) Exclude with umlaut in exclude list for Windows doesn't work17924 - Exclude with umlaut in exclude list for Windows doesn't work. Problem: If exclude list is given with a file beginning with BOM then first pattern is corrupt. Solution: Use ReadSourceFromHandle() function which is used for include list file and handles BOM encoding properly NEW: (#27126) Use 'tar' format for uploading Sesam meta files into data store, because 'tar' module doesn't cause such a high CPU load as usinf uncompressed ZIP files BUGFIX: (#26484) Domino backup will stop if the last item is excluded. Solution: Process item thread must call WakeConditionVariable() in case of excluded items too BUGFIX: (#17924) Exclude with umlaut in exclude list for Windows doesn't work. Solution: Exclude uses UTF-8 therefor pattern must be converted to UTF-8 if not yet given with UTF_8 encoding. Implemented for -x {exclude} and -X {exclude_file} BUGFIX: (#27109) sm_sshd main process consumes 100% CPU during command execution. Problem: To many attempts to read from queue cause high CPU usage. Solution: Read from queue using 15 sec timeout to decrease CPU load NEW: (#27098) Citrix Xen: Set system encoding to UTF-8, so item names with UTF characters can been shown without causing Python encoding exception BUGFIX: Media readcheck: Avoid SQL error during 'select from result_lbls' becaus of missing cast to integer CHANGE: (#26628) Command events: Better main log message CHANGE: (#26628) Command events: Remove functions, not necessary any more BUGFIX: (#26474) Kopano backup: Correct parsing of folder names including comma (',') BUGFIX: (#26474) Kopano backup: Replace all double quotes by underscore in mail subjects to avoid problems duing parsing in restore wizzard BUGFIX: (#26101) Notes backup does not consider 'tmp_dir' in section [SBC_NOTES] from sm.ini. Solution: New installation will not create this obsolete entry. Note: The temporary SEP sesam path is given with 'gv_rw_tmp' in [PATHES] section in the sm.ini file. BUGFIX: (#26474) Fix creation of LIS file entries for Kopano single item restore, where 'Subject' log from 'sbc' spread over several lines CHANGE: (#24595) 'sm_client' shows wrong error message during command event execution. Do not interpret exit code > 128 as system 'errno' NEW: (#16624) Write checksum of backup into restore main log, if restore verify option is set BUGFIX: (#27089) A Exchange RPEX restore created icon is added to skel directory and used for the RPEX link. BUGFIX: (#27086) Get saveset tree returns wrong media availability for a FDI saveset chain. For a backup saveset with highest 'availability' can be used, but for FDI chain lowest one CHANGE: (#19488) Lotus Notes INC must backup all transaction logs with mode Archive and Active, also the transaction control file (nlogctrl.lfh). SOlution: Add LOGCTRL file nlogctrl.lfh to backup items BUGFIX: (#22891) Notes backup of Transaction logs must not perform NSFDoneArchivingLog if backup failed. Solution: Create list of processed transaction logs and call NSFDoneArchivingLog() in DB_CloseOperation() only in case of STATUS==0 CHANGE: (#27069) Retry 'df/du' command, if message text isn't set. Workaround for problem in 'PIPE_READ()' function BUGFIX: (#25285) Do not call 'sm_dedup_interface s3 list alive', if no S3 connection is configured CHANGE: (#24688) Enable 'CTRLD' service again, if a GroupWise installation is detected BUGFIX: (#27070) - Hyper-V VMs are no longer listed when browsing with sbc. Cause: The specified source "/Hyper-V" unlike to "Hyper-V" is not handled in the correct way. CHANGE: Use new 'sm_qm sho/status' function, to check, if queue is idle. Avoids reading much data from QM, because of many pending jobs NEW: Use new data store type 'SEP Si3 NG Deduplication Store' instead of 'DB:hw_drives.options NEW: Add new data store type 'SEP Si3 NG Deduplication Store' BUGFIX (#27028) - Parallel Hyper-V backups not possible. Now it's possible with option -a "force_unlock" but it's done just as workaround. The main focus is "HYPER-V RCT" solution. NEW: (#27048) Make 5 retries for remote command in 'e_TAPE_STATUS_REMOTE()' function BUGFIX: (#27065) Problem: The update of a strech server fails because the package name in the linux_srv... file name misses the ~pg after the distribution. Hence the wrong package_name is inserted in the sm.ini during update. Solution: insert the correct package_name in the linux_srv file during build process. BUGFIX: (#22126) Add possibility to install/update/uninstall BSR during update. BUGFIX: (#26505) Cancelling a running Domino backup via Sesam GUI leads to a crash of the Domino server. Solution: If sbc is aborted after already retrieving last item (SBC_BCK_NO_MORE_ITEMS) then the MTF end-of-set padding block 'ESPB' writting failed. Now ESPB is not written in this case. BUGFIX: (#26651) Also catch URLError and return an adequate error message. BUGFIX: (#26988) Update of a client by GUI does not work. Problem: BSR uninstall key was not set in the right file during kit build. Solution: Set BSR uninstall key in the sm_execute_update.cmd of the src\skel folder during kit build. BUGFIX: (#23851) - Since the last Windows 10 update, the system state restore fails to set standard attributes for Windows defender files BUGFIX: (#26651,#26765) sm_update_client can use HTTPs for downloading index.txt now. The CA certificate is contained in the skel directory of the installation package. The default is still http until the function has been verified. CHANGE: SAP Hana backup: Use 'sys.exit()' in backup sub function, because it's started as process from main thread NEW: SAP Hana backup: Set 'DB:results.fdi_type' to 'D' for Hana INCR backups, because we need 'I' exclusively for LOG backups NEW: (#26859) Restore of SAP Hana LOG backups: Check backup level in Sesam DB and do not start parallel restores, if level is 'I' (incr=LOG) BUGFIX: (#26505) Cancelling a running Domino backup via Sesam GUI leads to a crash of the Domino server. Solution: The NotesTerm() must be executed by the same thread which executed NotesInitExtended() therefor the signal processing thread must inform the main thread to terminate further processing BUGFIX: (#26967) SAP Hana: Enhance checks, if 'sbc' has saved all data from pipe NEW: (#27000): Compress older Sesam main log files ('.prt') CHANGE: (#27001) Remove obsolete data store type 'SEP EasyArchive' BUGFIX: (#26996) S3 replication is started although the S3 mediapool is disabled. Correct SQL statement checking 'inactive' value NEW: (#27004) Optimize selection of tape drives used for restore. External restores will run parallel, if savesets are written to different media and multiple drives are available BUGFIX: (#26980) SAP Hana restore does not work if savesets are only available in migrated form. Problem: The migrated saveset was retrieved without pool therefore the saveset's media_pool must be used instead of saveset_pool NEW: (#26974) Compress and remove Sesam trace files in sub-directories of LGC directory too BUGFIX: (#26972) - Restore of VMware virtual machine fails, because just config folder itself is restored, but no files BUGFIX: (#26989) Clean up of a data store removes data files for backups started after DB select has been done. Check for new 'DB:results' entries again, after list of saveset has been fetched from data store CHANGE: (#26803) Write 'sm_sbc_com' trace files into 'lgc/com/' directories BUGFIX: (#26157) No restore possible from migrated saveset, if backup is encrypted. Solution: Fetch encryption settings from original saveset BUGFIX: (#26772) Sesam Postgresql instance reports many errors 'deadlock detected'. Solution: Avoid deadlock with [update media set filled=(SELECT ....) where media.label in (select label from media where location='...')] by using table media in select clause CHANGE: (#26967) Do not report SAP Hana restores as successful, if 'sbc' returns a warning BUGFIX: (#26963) - Exclude via regular expression is ignored on Windows backup via VSS BUGFIX: (#26980) SAP Hana restore does not work if savesets are only available in migrated form. Problem: The database VIEW count_results was changed - now with adjusted database column names BUGFIX: (#26980) SAP Hana restore does not work if savesets are only available in migrated form. Problem: Preferred media pool was set to original media pool. Solution: Always use appropriate media pool from selected saveset BUGFIX: (#26958) sm_sms_watch remembers IO error for regular EOM cases. Reset flag during next EOM processing BUGFIX: (#26816) Backup to Si3 and migration to tape gives an (Checksum of saveset copy does not match). Solution: Since all Data Store types are avoiding padding the default in DS_IsStripped() is set to 'stripped'. Now if INFO cannot be retrieved stripping of data is avoided. BUGFIX: (#26505) Cancelling a running Domino backup via Sesam GUI leads to a crash of the Domino server. Solution: Only the main thread is connecting with the Domino Lotus Notes server so no additional NotesTerm() or NotesTermThread() required when process is signaled BUGFIX: (#26941) Set correct option to transfer LIS file in compressed mode from RDS to Sesam server BUGFIX: (#24792) Inserted an additional key to the reboot required check (HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\Component Based Servicing\RebootPending). Also the key to detect PendingFileRenameOperations is checked and printed to the log file but not evaluated. NEW: (#26900) Upload of backup meta info file to data store after backup can be suppressed by setting 'GLBV:gv_conf_no_info_upload_ds=1 NEW: (#26661) Add VMDK META files to INFO zip file for main saveset, because these files are created after saveset info for VMKD sub-savesets are copied into data store Replace RAW and COMPRESSED chunk type in log with NEW BUGFIX: (#22771) Si3 replication of some savesets fails with java.lang.AssertionError in background. Problem: Refs offsets are based on bytesInData but not on bytesOnPage. Solution: Use bytesInData for replication Refs offsets NEW: (#22739) Add snapshot feature for S3 BUGFIX: Fix failing PageRemover on unknown page size issue. Fix failing EncryptKeyMismatch exception creation on invalid input hash BUGFIX: (#22550) Fix stopping gc during ddl rewrite BUGFIX: Fix hanging GC startup if write actor failed the snapshot NEW: (#22319) Separate "trash" directory for Si3 store NEW: (#22319) Separate "trash" directory for Si3 store NEW: (#21827) Amazon S3 storage backend for Si3 BUGFIX: (#21811) Do not merge locked pages in page tracker BUGFIX: (#21811) Do not fail backup if page cannot be added to index since it is locked by GC NEW: (#18298) Add snapshot start/stop commands - add newly created file to print output NEW: (#18298) Add snapshot start/stop commands NEW: Add JVM args into thrift `getServerStatus` command NEW: (#21815) Make fsck to recognize whereas incr. fsck went into full check NEW: (#21815) Add time of last object check to object info NEW: (#21831) Introduce WriteActor bounded PageTracker. During each backup new PageTracker is created and will be merged into general GC's PageTracker when GC is started BUGFIX: (#20877) Fix race condition between Si3 GC and Restore and abort Restore when exception has been thrown by server BUGFIX: Problem: "Missing chunks during FSCK and Restore". Solution: "WriteActor does not create links to chunks in page which is locked by GC. When it creates link to chunk, it also reference that page in PageTracker" NEW: (#21775) Make Si3 to not accepts new connection during shutdown NEW: Si3 version 4.4.3 to match Sesam main version NEW: Upgrade to scala 2.11 and latest available libraries for scala 2.11/Java7 BUGFIX: (#21712) Fix exceptions during server shutdown NEW: (#21710) Add start time to server status response NEW: (#21713) add last full successful fsck run to fsck status response NEW: Change URL to 'www.sep.de' an increase minor version BUGFIX: Avoid SI3 index corruption, if dedup server crashes during GC