GUI Bugfixes: change: Fix profile is deleted when closing the confirmation dialog via the red 'X' change: Fix NPE on UI startup if DB authentication is enabled and local full access is true change: (#22314) prevent change backup eol cb, after change saveset eol, (listener problems) change: (merge:7fdc28b)(#22314) eol warning dialog appears two times, when change saveset eol and afterwards change the backup eol. bugfix: (#22340) Default view of "status->results" couldn't be adopted any more; ComponentLoader will not appear at gui, when initialize layout BUGFIX: (#22329) Default view of "status->results" couldn't be adopted any more bugfix: (#22314) set correct label for radiobuttons option 'full saveset chain', 'selected saveset' at eol warning dialog, which appears, when change eol change: (#22314) eol changes will not be applies, when edit eol from result dialog under datastore.saveset table. bugfix: build failed; missing class for last commit. change: show and exchange options 'for selected backup', 'complete backup chain (set as default)', when edit 'backup eol' only for backups with fdi type diff, incr at result dialog; bugfix: method 'adjusteol' will be invoked with wrong param 'isSingleSaveset=false', when change saveset eol of backup bugfix: show options 'for selected backup', 'complete backup chain', when edit 'saveset eol' only for backups with fdi type diff, incr at result dialog bugfix: options 'for selected backup', 'complete backup chain' not more visible, when change saveset eol at datastore dialog' change: (#22314) show eol warning dialog, when change backup eol and close dialog with button 'ok' or when change saveset eol at table CHANGE: (#22314) GUI must inform about EOL adjustment in case of reduce or increase of save set or backup EOL. bugfix: (#22221) cli action 'list vsphere -m snapshot' cause NPE; .Problem: set vm param -V to object does not work change: Fix param info fails to convert values into correct class bugfix: (#22285) - Harmonize the icon for deleting the search/entry-fields BUGFIX: (#22299) - UI views not saved with authenticated user BUGFIX: (#22310) - Tape encryption is not set, but encryption flags are set correctly (UI part) BUGFIX: (#22247) - Protocol views loose current scroll position if switching tabs or refreshing the view CHANGE: (#22162) set correct eol label at backup property dialog; improve message, when choose eol < current time (same like datastore.saveset eol change); BUGFIX: (#22279) eol warning dialog does not appear, when set eol to the past at field 'locked until'. BUGFIX: (#22279) eol warning dialog does not appear, when set eol to the past at eol cell of backup with edit table. change: translate label 'Interface' to 'Schnittstelle' at german gui bugfix: (#22294) Restore at customer system wasn't possible - ERROR: Cannot find matching LIS. Problem: GUI write LIS file without UTF8, umlauts get lost. bugfix: (#22289) docking data for gui -bystatus components will not be saved, when save view. bugfix: (#22290) Enter-key does not work anymore when authentication is enabled change: Fix NPE if the UI is launched with a 32bit Java environment bugfix: change backup eol for datastore saveset show message with html tokens, when set point of time to history bugfix: (#22278) Restore wizard hangs with NullPointerException change: Fix order and formating of translation properties change: Improve output on downloading a new UI version BUGFIX: (#21909) - Add build/commit time to the UI update info box BUGFIX: (#21955) - Confirmation dialog missing when removing a view bugfix: (#22251) NPE, when change saveset eol at datastore.saveset table. Problem: no mediapool is found for command 'adjust_eol'. Solution: check media of all mediaPools, which belongs to datastore to get the right one. change: clean code bugfix: column saveset eol at datastore dialog under saveset table is empty. bugfix: (#22251) Changing EOL of replicated savesets fails. Problem: wrong mediapool will be used for command 'adjust eol', when datastore is type 'SI3-EXTERN' change: disable VMDK of VM rows at restore wizard and prevent continue restore process change: Adjust version string in splash to be "Tigon" change: Update splash screen for Tigon V2 BUGFIX: Data store can be deleted without safety check when "Delete" button is clicked before all data is loaded in the data store properties dialog bugfix: (#22221) list vsphere cause NPE change: Fix branch match token for sep_build_version.properties to match the v4_4_3_2 branch with and without 'origin/' change: Replication type label 'Si3 Rep' to 'Si3 Replication' new: Makefile to copy UI components into 'common' for release build change: Move splash screen to separate directory and deploy splash screen to release location to be picked up by kit build change: (#22038) set VM cbt mode at task dialog as read only change: (#22199) prevent offer date formatter menu items 'localized time', 'iso time' for columns, which has only date without time change: (#22199) prevent offer date formatter menu items with time for columns, which has only date without time bugfix: (#22185) Date format of column "Sesam day" can't be changed in GUI->Media table; saved view of table 'media' will never be used, when reopen it. BUGFIX: (#22194) - The lock icon is missing in column "Media EOL" in "Components" - "Media" bugfix: (#22193) "taskEvent.getFdiType" exception when restarting a backup group with canceled backups. Problem: fditype will not be set for taskevent, when parse data from result of failed backup bugfix: (#22195) A NullPointerException is thrown when trying to open schedule or task event properties change: set help link of backup option 'Backup as image' to '.*/4_4_3_Tigon:VMware_Single_File_Restore_and_Instant_Recovery_Support_Matrix' bugfix: url help link for restore option 'mount saveset' is wrong bugfix: no calendar will be set, when open schedule a second time. bugfix: (#22182) Exception "Operation Not Possible" when saving a schedule with user defined calendar. Problem: NPE, when switch to calendar without 'nextexec'. BUGFIX: (#22160) - SQL exception thrown by PostgreSQL when updating the media table from the UI BUGFIX: (#22174) - Restore Wizard authentication does not accept restore user BUGFIX: (#22136) - GUI crashes when changing write protection bugfix: (#22165) GUI starts selective restore as full bugfix: (#22166) Check box for restore without tree structure is reseted, when first deselect combobox 'original tree structure' and set 'to new restore target' bugfix: (#22130) - UI update loads both sm_gui.jar and sm_ui.jar when a newer sm_ui.jar is available on the server BUGFIX: (#22140) - When the notification center is open while the newday happens, an IllegalParamterException may occur BUGFIX: (#22154) - Master GUI does not recognize primary server if specified with different case at command line change: Fix label in newday event dialog BUGFIX: (#22109) - Occupancy value in media overview is calculated wrong bugfix: (#22156) Restore Wizard just offers original drives, if saveset exists on two media pools change: offer option 'immediate start', when selected row is a taskgroup at taskbystatus bugfix: crypt task password dialog cannot be closed with 'cancel' button at taskdialog, after select 'save encryption key' BUGFIX: (#22115) - It is not possible to save and use a layout when one of the tabs has a custom table view configuration set change: improve some labels for bystatus column 'fdi_type_set' and property 'FDI Type', 'FDI Type Set' (merge: 45c71a5) bugfix: (#22084) - Rework: Saveset EOL field is empty within the saveset tab of the media pool properties dialog bugfix: Fix retry handling on opening the UI while server startup is still in progress change: Filter Results.getSavesetEOL calls from UI server log file change: (#22105) Restart of a backup task in window 'Restart Backups' does not use the origin CFDI type (fdi_type_set) BUGFIX: (#22056) Comment field was not visible bugfix: (#22092) - Show number of selected save sets for which EOL will be changed change: (#22101) fill overwrite mode combobox with default entries and without red font of combobox entry 'overwrite existing files' for backuptype 'NDMP' with 'mount saveset' at restore wizard bugfix: (#22095) - "Restart backup" window doesn't close after backup has been restarted bugfix: wrong target clients will be offered, when restore backup from client with windows os and server use linux os and with mode 'mount saveset' BUGFIX: (#22084) - Saveset EOL field is empty within the saveset tab of the media pool properties dialog new: (#22090) rename header of column 'fdi type set' to 'Original Type' at table taskbystatus change: (#21591) improve note for deactivated 'mount' due to invalid interpreter at restore wizard. change: (#21591) disable instead of hide mount/attach option with reason. change: show option 'mount saveset', when select result entry with interpreter=dump at restore wizard bugfix: build error at method 'customInitialization' new: (#22090) column 'fdi type set' for table 'taskbystatus'; change: use results.fdi_type_set instead of fdi_type as init for immediate start dialog from taskbystatus bugfix: (#22082) - Warning on creating a new data store is not shown correctly on Linux BUGFIX (#22056) New data store-> Comment field was not visible .Solution: put the properties panel in a scroll pane BUGFIX: (#22077) - Exception on client side when creating new sub-location change: Add proper ignore rule for translation properties backup files change: Remove translation properties backups bugfix: (#22050) VMware Single File via Attach shows Windows partitions twice change: (#22038) replace char ':' with '.' at the vm group toolbar infobox change: Fix media.filled is in KiB and not in Byte BUGFIX: (#22056) New data store->Comment field isn't visible. Problem: Dialog Size for new DS is too small (920x695). Solution: set size to 920x850 change: Fix UI needs to wait a bit longer for the server startup if still in progress instead of failing right away bugfix: (#22025) - Performance window shows exponential values at y-axis change: (#22038) Remove Checkbox 'VM config only' and replace hyperlink text at tooltip with 'here' bugfix: (#22032) - Change the format of SESAM DATE -> remove the timestamp - and harmonize the name bugfix: (#22039) A restore from a mounted saveset stays active in db when it is canceled before start with button 'X' bugfix: (#22039) A restore from a mounted saveset stays active in db when it is canceled before start bugfix: (#22053) - Linux style GUI (-llinux) doesn't work under windows any longer change: Fix ClassCastException in restore wizard expert options BUGFIX: (#22016) The spaces between the fields and the label in the results dialog tabs did not fit .Solution: Use jgoodies.forms.layout instead of javax.swing.GroupLayout change: (#22038) Change layout, labeling and tool tip of check boxes in vm task properties and vm browser bugfix: help url link for option attach/mount at restore wizard is wrong. Problem: tagname ends with '$2' at the end; Solution: cut '$.*' bugfix: cli test 'testMediaPool' failed(2) change: (#22024) param "-Z ", "-z duration" will not be shown at cli help bugfix: cli test 'testMediaPool' failed BUGFIX: (#22028) - Slot assignment for SayFuse loader does not work anymore change: (#22024) add param "-Z ", "-z duration" for cli operation 'start mediapoolevent' change: (#22024) add param "-Z ", "-z duration" for cli operation 'start taskgroup' change: Fix failing command event "createStart" test change: (#22024) add param "-Z ", "-z duration" for cli operation 'start command' change: Fix build error in VMRecoverOption bugfix: vm attach/mount restore options are not more visible. Solution: replace logic with older release, where no reason for disable option will be shown. bugfix: (#22037) - Fehlermeldung beim Löschen eines Data Store ist zum Teil nicht lesbar bugfix: (#22026) - When creating a new data store, the listed existing drive group names contains duplicate entries bugfix: (#22021) - When I export the 'backup test' view in the GUI the filename will be 'backup -test' - the space between 'backup' and '-test' change: improve code, use annotation '@Parameter(names = {"-z", "-w"})' for allow set duration with both params. bugfix: clitest 'testschedules' failed; allow param -w; -z for set duration of schedule. (when both are set: -z > -w) bugfix:(#22024) set only one param -z or -Z cause NPE change: (#22024) add param -z ; -Z to cli migrationevent with action start/migrate change: (#22024) add param -z ; -Z to cli restore actions start/restore change: (#22024) reset cli start/backup task param to -z , -Z change: (#22024) set correct help token for cli help backup param -f change: (#22024) add param "-f " for cli operation 'add schedule, backup' change: Fix potential dead lock on UI startup and NPE while checking the license change: Adjust schedule CLI test to test for old "-w" and new "-z" parameter for duration change: Fix Unittest and replace all "new (...)" with the corresponding ".valueOf(...)" calls bugfix: (#22023,#22033) - ClassCastExceptions happens for some spinner controls in the UI BUGFIX: (#22031) - Get "Parameter media.pool is missing" exception if trying to introduce a new media with a custom label specified to a media pool bugfix: (#22020) - GUI-DATA STORES-SAVESETS-STOP TIME-RIGHT MOUSECLICK the last menu item (date format) is there two times change: MAVEN - Added jgoodies-forms library change: Remove mybatis generator change: Improve initializing of static file delivery restlet to avoid occasionally NPEs bugfix: (#22009) - GUI - JOB STATE - BACKUPS - TASK PROPERTIES - INFO2 -> All time/date-infos are ranged right formated bugfix: (#22018) - Change wording (window title, text and message) for the current window Import Initial SI3 Seed change: Fix data size of restore result record shown with wrong unit in restore result properties dialog change: Fix findbugs issues in UI bugfix: (#21591) hide option 'mount saveset' at restorewizard, when restore VM task; use same visibility logic of option VM attach/mount like 'mount saveset' change: Move to HyperSQL 2.4.0 change: Fix an NPE change: remove old commons-dbcp dependency bugfix: (#21874) - Error updating database. Cause: java.sql.SQLException: statement is not executing change: Move to sqlite-jdbc 3.18 bugfix: (#22008) VMware Single File options must not be visible for regular backups change: Upgrade to MyBatis 3.4.4 change: Clicking on 'X' button of connection lost message shall exit the UI as well change: Make the maximum document count configurable change: Fix occasionally dead lock on starting up the UI bugfix: (#21951) vcenter server without tasks at table 'vm with task' or 'VM task without existing VM' show all VM entries, after expand node although no tasks defined bugfix: set correct params at examples of cli 'help vspheretask' bugfix: (#21951) count of removed/added VMs at status message is wrong; handle count of duplicate VMs with same name bugfix: (#21951) vm cannot be selected, if another virtual server has already an task with same vm;. solution: set used tasks, where task.clients equals vm.server change: Improve connection lost handling change: Fix NPE on refreshing system settings from the defaults table when the server connection is suspended bugfix: (#21977) - Restore Wizards sets restore_tasks.tree_type always to 'd'eep for 'System Recovery' backup type even if 'tree structure' is not set change: (#21951) init backup/restore options with data from tasktypes for generated vmtasks; bugfix: remove vm tasks cause ObjectNotFoundException bugfix: (#20825) apply backup/restoreoptions to all generated VM tasks bugfix: (#21994) vm with password use cryptkey instead of saveset for log 'Task is encrypted, decrypt and encrypt password again with saveset ''"' bugfix: (#21996) When select all VM without task entries, no tasks will be added with reason: 'empty selection' change: Fix icon node trees disappear when clicking on an node in the tree change: Fix several remaining issues with profiles navigation node handling bugfix: (#21997) - Message "no restore access" appears when calling restore wizard with admin permission change: Fix text and color of mount/attach warning label when compression or encryption is enabled in the task bugfix: (#21951) wrong VM selection, when VM table has several entries with same name and same datacenter, but the hostsystem differs bugfix: (#21867) - Restore Wizard allows to select 'active' backup jobs when 'broken tasks' is set change: (#21591) improve messages and set correct helptag for operation 'mount'. bugfix: when acticate encryption at task dialog, no note will be shown. change: (#21591) add warning message at taskdialog 'mount not possible', when active compress or encoding. bugfix: (#21976) - Bei Neuaufnahme von Bänder kann das LW geändert werden, aber das Gerät wird nicht angepasst bugfix:(#21939) Select path with umlauts at gui file browser cause MalFormedInputExcepion with stacktrace at log; (Exception will be handled) change: Fix log file naming for exception log for rolling policy change: (#21591) set error message as visible text at its component and not as tooltip change: (#21591) improve tooltip for invalid state of saveset tree change: (#21591) set correct tooltip as reason, why option'mount saveset' not available due to invalid status of saveset tree change: Fix job status views initialized with wrong default layout change: Fix "New Media" dialog format and texts bugfix: (#21958) - GUI view handling is inconsistent change: Fix NPE opening the schedules view bugfix: (#21955) - Saved job status view is not removed from navigation pane when removed via the view selection toolbar bugfix: (#21962) - Help of migration task redirects to wrong link bugfix: (#21968) - Task as list: Backup state has no renderer and no filter bugfix: (#21975) vSphere authentication window pops up on immediate start of a vSphere restore. Problem: Restoretask.type is mount/attach. Solution: disable option 'immediately run' with reason as tooltip. add: (#21951) combobox for sub type _IMG, _CONFIG or _NONE at VM taskgroup panel bugfix: (#21956) - NPE when I try to save default_profile in WINDOWS-LAYOUTS change: Fix profiles manager formatting change: Fix failing unit test change: Fix merge error from cherry-pick bugfix: (#21954) - RESTORE-PROPERTIES: The throughput-value (right property window) changes, when I change the text of the note bugfix: (#21929) - Strange error message if trying to restart a backup without selecting any change: If (interface) template does not exist, open editor with empty content instead of failing bugfix: (#21947) - Clicking to REFRESH activates the disabled AUTO UPDATE bugfix: (#21950) - Show 'Results' windows remains empty for BSR Windows backup tasks change: Remove "bayrisch" localization from UI bugfix: (#21906) - Pre-script for task groups change: (#20825) improve title for vm tables 'VM_TASK', 'VM_TASKS_WITHOUT_VM', add note, that selected tasks will be deleted remove: param -D for cli 'add vspheretask' bugfix: (#21943) - GUI -> SCHEDULES -> NEW-MIGRATION TASK: After changing the task name a change from ExecPara->Schedule->ExecPara gives a strange message bugfix: (#21944) - GUI-SCHEDULE-NEW -> if the element needs a name, it would be fine that the focus would be in the NAME-field bugfix: (#21946) - Adjust the menu item order in the schedule view "New" menus bugfix: (#21945) - Wrong message is shown when adding a media task to a schedule containing a backup task change: Fix new day event panel to allow suppressing specific events, but about only all nor none events. change: (#21934) Set default priority to '1' for scheduled restore change: cli argument 'dir vsphere vc' to 'dir vsphere dc' (list datacenter) bugfix: duplicate param -T for cli 'add vspheretask'; set folder param '-T' to '-D'; set tasks.subtype=null to possible for cli 'add vspheretask' add: (#21591) tooltip for mount saveset, why option is not available. (compressed,crypted) bugfix: (#21905) - Edit Sesam server scripts by UI server directly change: Fix "Listen Port" text field is visible when it should not bugfix: (#21923) - user role "restore" is able to export the sesam DB bugfix: (#21926) - The end of the "event" column is cut off when printing calendar sheets change: A refresh to the "Events as list" table does not change the column layout anymore bugfix: (#21927) Use the same name for main selection entry "Calendar sheet" and tab "Calendar" bugfix: delete event does not use selected row under component 'Events as List' remove: item 'delete' at mouse context menu with focus 'taskgroup.task' under component 'Events (Next Run)' bugfix: taskgroupevents under component 'Events (Next Run)' cannot be deleted change: Increase width of client dialogs to show 80 characters in the interface box change: (#21853) improve error message for failed CBT reset with VM, which contain snapshots with powerstate=suspended bugfix: (#21914) - Increase the interface-window size in the GUI bugfix: vm with umlauts will not be added to task group for cli 'add vspheretask' .Problem: parse inputstream/file set wrong VM name (clientside) bugfix: (#21912) GUI doesn't if the time zone of the server is configured using a three letter id, like GMT+01:00 BUGFIX: (#21912) - GUI doesn't if the time zone of the server is configured using a three letter id, like GMT+01:00 change: Improve client side debug logging a little bit change: set correct valid samples for cli action 'help dir' and 'help dir vsphere' cleanup code; remove unused code; define as private where possible bugfix: (#21869) - Adding new restore event to a schedule takes up to four minutes change: add new cli action 'help dir vsphere', which describe the filter params; remove cli filter vm help for action 'help dir' and add link to new help change: add removed/added task objects as retval for cli action '"add" "vspheretask"' bugfix: wrong cli error message format for exception, which will be thrown from client side change: Fallback to system look and feel if not set otherwise change: Make sure that clients less than 4.4.3 will have to perform a full upgrade first bugfix: (#20825) fill VM table cause NPE bugfix: add vm via file with cli remove VMs, after create it. bugfix: (#20594) UNC restore with relocation not work correctly; UI cut off the leading slash in reloc options change: Show data size values in the properties panel using the selected default data size bugfix: (#21902) default date format will not set for column 'access time', when no view data exist. new: (#21902) set date format for client table column 'access time' change; improve error message at task dialog, if VM does not exist bugfix: param '-a ssmoref' will not be refreshed, after recreate a VM snapshot. bugfix: vm diff backup (VSPHERE:) failed; VM cbt file not found. change: Only allow all or no events to be aborted at "newday" events change: Remove BETA ticker change: Set the UI icon to "Release" and use "Tigon" as branch name SRC Bugfixes: BUGFIX: (#22753) Backup of Windows directory fails because of directory block size is larger than 1024 bytes. Solution: The directory stream name size must be checked together with OS Info size to determine if PNAM stream must be used when Format Logical Block size is too small Merge remote-tracking branch 'origin/master' into v4_4_3_4 bugfix: Correct SQL statement for last successful client access time new: (#22091) Client configuration: Do not check remote access for access mode 'VIRTUAL' change: Comments revised bugfix: (#23189) MS-SQL restore fails: UNKNOWN error: 4294967295 -. Assembling BACKUP/RESTORE T-SQL command failed. Problem: Restore target was given with 'SEP-SQL2012/MSSQLSERVER' - in this case databse name must be retrieved from MTF stream (selected database) change: Problem: (Windows:) The select() in recv_with_timeout() timed out but the error message is not displayed. Now use WSASetLastError( WSAETIMEDOUT ) to return '10060: Connection timed out.' change: (#23643) Migration error - DS (on NetApp with deDup) to Si3 -> SBC-STATUS: E012-COPY Die Migration war nicht erfolgreich. 553 STOR Failed. NO ERROR: (0x0,0) (0). Problem: (Windows:) The select() in SEP_Recv() timed out but the error message is not displayed. Now use WSASetLastError( WSAETIMEDOUT ) to return '10060: Connection timed out.' change: sbc_ndmp - print error cannot open catalog to stderr CHANGE: To indicate 'BETA' version: update defaults set value='1' where key='beta' Merge remote-tracking branch 'origin/master' into v4_4_3_4 bugfix: FSCK for data stores: Do not report missing snapshots for non snapshot type data stores bugfix: (#23716) - Netapp Snap Store: in case sbc is killed, snapshot is left on netapp system... change: sbc_ndmp .snapshot volumes not displayed correctly if DB_VM_NO_CLEAN (NOCLEAN parameter) not set, default value is 0 instead of 1 BUGFIX: () Problem: the minor update 4.4.3.48 did not work correctly since the component to spadmin.dll was removed by accident. Solution: Reinsert the component to the msi project and only set it to 0 meaning empty component without executable. In addition the hyper_access.dll is removed since it is not needed in the x86 package. Merge remote-tracking branch 'origin/master' into v4_4_3_4 new: (#23749) Si3/S3 sync: Write main log file into media log directory change: sbc_ndmp - display snapshot folders as volumes bugfix: (#23713) Hyper-V restore without overwrite continues even though VM exists. Solution: Since commit 5b1e3d2 the restore was not aborted BUGFIX: Saveset count in VIEW 'count_results'. Failed migrations have been counted as successful new: Update 'DB:result_lbls.pool/id' automatically using 'label' during Sesam update. Solution: Implement SQLITE triggers update_add_pool_result_lbls/insert_add_pool_result_lbls new: Update 'DB:result_lbls.pool/id' automatically using 'label' during Sesam update. This will make many DB selects much more simple new: DB trigger to fill 'DB:result_lbls.pool/id' automatically using 'label'. This will make many DB selects much more simple Add log message CHANGE: (#23647) License check during newday can block complete Sesam. Solution: Submit license check with 'sm_info C'. Note: Now the startup finishes after 1 minute. The license check needs a few minutes therefore the license output appears in the daily log after approximately 100-400 lines bugfix: (#23470) - additional fix. Semicolon is no longer used as delimiter in sbc_ndmp internal processing bugfix: (#23530) - Hyper-V - Inc - Restore - Restored VM was not started. Implemented the "merge" operation on Hyper-V 2016 new: Do not start S3 sync, if S3 target pool is deactivated Bug 23734 - system_recovery backup is too small. For some mysterious circumstances this flag is not set on some systems. NEW: New OO_VERSION=12_1_159 BUGFIX: (#23443) At update (.48->.53) on shutdown a 10-minute wait time comes up. Solution: Avoid possible race condition between checking sm_qm_main LOCK and connecting to qm_server pipe by using 3 retries with 10seconds timeout as long as pipe is busy NEW: New OO_VERSION=12_1_159 bugfix: (#23746) Add 'DB:media_results' entry set DB:data_stores.clone_status='X' for failed Si3/S3 sync BUGFIX: (#23744) SAP ASE backup failed if environment is not set properly with Error: SAP ASE command returned with error: . Solution: Set environment via SYBASE.env file with '' substitution for Linux bugfix: (#23714) STPD sub process cores at backup to tape. Problem: Current sms support only 64K block size. Solution: if Saveset block size different then 64K exit with error BUGFIX: (#23744) SAP ASE backup failed if environment is not set properly with Error: SAP ASE command returned with error: . Solution: Set environment via SYBASE.env file bugfix: (#23740) Fix core from 'sm_arch' during 'adjust_eol'. In this case drive properties struct hasn't been set, which is used for 'info()' calls bugfix: (#20345) Make BSR bootimage usable in a network environment. Problem: On some systems the DNS suffix is added twice. Solution: Remove domain name from hostname bugfix: inserted a missing variable in a log message. new: NetApp SNAP store: Call 'delete snapshot' always with complete snapshot name, instead of building it on client side. This allows to delete system NetApp snapshots too bugfix: (#23719) Do not purge saveset on SNAP store from DB, if snapshot couldn't be removed change: Use isql with option '-X': Initiates the login connection to the server with client-side password encryption. -X enables both extended password encrypted connections and password encrypted connections without plain text password reconnection new: Remote command: Report error in case RCMD() function returns with '-1' (internal error) new: Remote command: Add command to be excuted to error message in case of error change: Do not try to delete files on clients with virtual snapshot drives bugfix: (#23652) - sbc_netapp backup doesn't stop, if snapshot couldn't be created Merge remote-tracking branch 'origin/master' into v4_4_3_4 bugfix: NDMP do not set ENHANCED_DAR_ENABLED during restore. it is causing backup failures on some NetApp systems. Not having it is causing warnings ons ome other systems. This should be further investigated for the best possible fix. change: workaround for sbc_ndmp core - don't call destructor - memory freed on process exit bugfix: NetApp9 mount to vserver address also with option copyfromsnap BUGFIX (#23683) Import of medium from mail slot: Exit with error, if destination slot is filled TODO: Do not call SLU move command in this case, because it doesn't make sense change: Added TODO comment for ADJUST_EOLS() in case of DIFF/INCR backups change: Added trace line with SEL file name in case of restore change: (#17369) STPD abort connection if tape in "CHANGING_MEDIA" state. Problem: STPD should check tape state and wait for correct one if possible. Solution: Run open_saveset_in_SMS in loop while tape state is different to correct one or timeout appears (default: retry 120 times every 30 sec) Do not reset restore command to 'sbc' for operating system 'NETAPP'. new: sm_event execute: Command line of original command (sm_backup...) can be set as parameter by setting '{}' as option Merge remote-tracking branch 'origin/master' into v4_4_3_4 NEW: (#22871) Delete automatically generated migration task created by immediate start migrations after execution bugfix: (#23681) No interfaces update during sm_setup change_servername. Solution: Use insert command with column list and add value for column uuid CHANGE: show actual error in case connection to netapp system fails change: In case of generation restore add log lines about saveset names and saveset counter into restore protocol bugfix: (#23530) - Hyper-V - Inc - Restore - Restored VM was not started. Tracing improvement. new: Set backup level to 'copy' for NetApp backup going to 'snapshot' pool new: struct DRV_PROP: Add 'iDn' to avoid 'atoi(szDn)' where int is necessary change: Avoid multiple trace lines 'PIPE_READ: line mode' change: Do not write message 'sbc-3174: Info: Restore without ACL due to option -o skip_acl' if without ACL is set via HYPERV: bugfix: compilation error "error C2039: 'CVirtualDiskV2' : is not a member of 'NHyperV'" bugfix: (#23663) - sbc backs up same data twice for task type Exchange bugfilx: compilation error bugfix: (#23656) Zombie process for unknown reason via HTTP request. Problem: since commit edb6244 sm_stpd_http_conn become zombie on Linux. Solution: partially remove commit edb6244 bugfix: cleanup temporary folders in case of abnormal Hyper-V termination - merge, import and etc. change: Linux: Compiler warning resolved BUGFIX: (#23660) sm_ctrlc communication is blocked after connect. Problem: The first recv() waits infinite for data. Solution: Use recv_with_timeout() with select() and 120 seconds timeout BUGFIX:(#23658) add sesam bin directory to LD_LIBRARY_PATH to fix backup error on centos6/7 and ubuntu change: sbc_ndmp sorts volumes before printing them out NEW: (#23531) Execute command by follow up event BUGFIX: (#23530) Hyper-V Incremental Restore - Restored VM was not started. Problem: Restore sequence must not be reverted in case of Hyper-V and overwrite not set change: Compiler warnings resolved BUGFIX: (#23530) Hyper-V Incremental Restore - Restored VM was not started. Problem: AVHDX must be merged to VHDX in correct order. Solution: Call sbc with '-a merge,delete' for all restores FULL, INC1, ... except the last INCx which must be called with '-a import,rename' to merge with original sequence and to import after last merge NEW: (#23531) Start Sesam command (DB:commands) by follow up event new: (#23653): 'SNAP drives': Do not try to call any command on RDS to dismount drive Add macro 'IS_SNAP_DRIVE' bugfix: Correct output of 'dir' command for NetApp filers change: Do not log check of directory in Sesam main log bugfix: Set 'volume' key during fsck of LVM snapshot data store CHANGE: Support for new Microsoft Windows Reparse Point tags change: Defines for windows reparse points (tags) updated change: Trace 'RootDir' with correct value BUGFIX: (#23005, #17781) Problem: Hyper-V Cluster nodes are not automatically configured correctly as clients according to the cluster resource. Solution: If the cluster nodes of a Hyper-V cluster have not yet been configured in the sesam they are tried to be configured automatically according to the client information found at the cluster resource client in the database. They are only added if 'sm_setup check_resolution ..' of the cluster node is successful. change: compile fix for vhdx_parser Revert "new: (#23588) Change SMS log level and segment size without restart of sm_data_server" bugfix: (#23630) do not append backslash bugfix: Set fsck start time to '', if 'NULL' is read from DB, because fsck will never be started after GC else Remove trace line used for testing change: (#23636) Re-enable start of Sesam GUI on Linux without '-S server' parameter new: (#23588) Change SMS log level and segment size without restart of sm_data_server NEW: 'sm_event': Add general exception handler and log error in Sesam main log BUGFIX: (#22978) sbc complains about 'Not supported reparse tag 0x00000000L' for some reparse points. Solution: If FindFirstFile() returns wrong reparse tag 0x00000000 then use DeviceIoControl( hReparseFile, FSCTL_GET_REPARSE_POINT, ...) ... to get the correct reparse tag NEW: Delete older 'DB:media_results' entries with 'action=sync' older then oldest saveset on this data store bugfix: (#23530) - Hyper-V - Inc - Restore - Restored VM was not started. Restore into original path is adapted Revert "bugfix: (#23592) Backup with option 'Descend directories on other file systems' failed due to reading from original volume. Solution: corrected exclude filter for relocated paths" change: (#17914) - Backup (Snapshot) netapp volumes with security style NTFS : create link directories if not exist' new: 'iRCMD()': Switch to set max. exit code, which will be seen as successful. So for example exit 1 from sbc will not be reported as error in Sesam main log BUGFIX: (#23628) Remove FDS DeDup Store from data_store_types. Solution: Remove 'SEP FDS Deduplication Store','SEP FDS DeDupStore' from data_store_types during update Merge remote-tracking branch 'origin/master' into v4_4_3_4 NEW: Function 'i_RCMD_EXEC_LOG()' with switch to avoid logging to Sesam main log. bugfix: (#23530) - Hyper-V - Inc - Restore - Restored VM was not started. Fix problems supplied by the new POC in the "ImportVM" bugfix: (#23530) - Hyper-V - Inc - Restore - Restored VM was not started. Made some adaption in according to new POC change: (#17914) - Backup (Snapshot) netapp volumes with security style NTFS - fixed Linux backup change: (#23588) Increase number of SMS streams without restart of sm_data_server. Code reformat, no business logic changes bugfix: Correct SQL statement, if saveset is set by command line bugfix: (#23592) Backup with option 'Descend directories on other file systems' failed due to reading from original volume. Solution: corrected exclude filter for relocated paths bugfix: (#23583) Restore switch back to drive used for backup. Problem: Since commit 7be7d04 '(#23191) File restore uses wrong media pool, if saveset was replicated ...' the logic may lead to a drive switch. Solution: Do not set restore drive pRT->szDn to original backup drive number in function i_SELECT_FROM_RESTORE_TASKS() change: (#17914) - Backup (Snapshot) netapp volumes with security style NTFS bugfix: (#23583) Restore switch back to drive used for backup. Problem: Since commit 7be7d04 '(#23191) File restore uses wrong media pool, if saveset was replicated ...' the logic may lead to a drive switch. Solution: Do not set restore drive pRT->szDn to original backup drive number in function i_SELECT_FROM_RESTORE_TASKS() change: (#17914) - Backup (Snapshot) netapp volumes with security style NTFS bugfix: (#23583) Restore switch back to drive used for backup. Problem: Since commit 7be7d04 '(#23191) File restore uses wrong media pool, if saveset was replicated ...' the logic may lead to a drive switch. WIP: Switch drive only in case of replicated saveset. bugfix: (#23189) MS-SQL restore fails: UNKNOWN error: 4294967295 -. Assembling BACKUP/RESTORE T-SQL command failed. Problem: The restore is set up with 'to original' but the module sets target to '//' (DB_OpenItem: restore to given place [//]... ). Solution: Set target to empty string to process restore to original. Note: Restore may fail if 'overwrite' option is not set! change: Message revised - AddMultiplyObjects() is also called for restore bugfix: (#23603) - Exchange backup fails with "Failed to get databases and its requred logs from not file". Incorrect iteration over vss components is done. change: Message lines reviewed bugfix: change indention's Revert "bugfix: Correct SQL statement, if saveset is set by command line" change: (#11852) Tape Trailer couldn't be repaired. Problem: Tape Trailer repair seams to be working, but no logs at all. Solution: Revise logging while Tape Trailer repair bugfix: 'sm_data_store mount': Fix exception in log file name creation change: Compilation failed with 'cm_common.lib(getopt.obj) : error LNK2005: optarg already defined in sesam.lib(sesam.dll)' NEW: Delete older 'DB:media_results' entries with 'action=fsck/empty-trash' older then oldest saveset on this data store change: (#17914) - Backup (Snapshot) netapp volumes with security style NTFS NEW: Delete older 'DB:media_results' entries with 'action=fsck' and status=OK Merge remote-tracking branch 'origin/master' into v4_4_3_4 BUGFIX: (#21537) Do not use 'SPARE' medium, if accept spare medium flag isn't set for media pool BUGFIX: (#20947) Drive configuration: Do not create drive queues twice bugfix: (#23171) Add possibility to activate compression in BSR-backup. Problem: May be called with '-o compress'. Solution: Use 'try:' for compress_option.split('=')[1] change: (#23407) No backup of mounted volumes although 'Descend directories on other file systems'. Change: Code reviewed NEW: (#23219) Do not restart Si3 server during 'sm_config_drives', if not really necessary Revert "NEW: (#23219) Do not restart Si3 server during 'sm_config_drives', if not really necessary" NEW: Delete older 'DB:media_results' entries with 'action=purge' Code cleanup and improve of log messages NEW: (#23219) Do not restart Si3 server during 'sm_config_drives', if not really necessary change: Use already read DB:hw_drives record in license check, instead of reading drive properties again change: iRCMD(): Do not log DB call to check for local server in default log level change: sm_vfs performance utility added bugfix: (#19991) Small backup to Windows tape device occasionally fails with "Setting the tape block size failed" when writing tape trailer. Problem: On small backups time between SetTapeBlockSize is also limited and API function SetTapeParameters may fail. Solution: Add retry loop for SetTapeParameters function - up to 8 retries after 15 sec sleep each new: Use 'DB:drive_properties.generic_type' bugfix: Correct SQL statement, if saveset is set by command line new: Use 'snprintf' to cut off message text BUGFIX: (#23576) SAP backup saves data into Sesam work directory instead of correct medium change: Windows build link sbc_proxy sbc_netapp module Merge remote-tracking branch 'origin/master' into v4_4_3_4 bugfix: Typo in column name: 'sesam_data'->'sesam_date' Merge remote-tracking branch 'origin/master' into v4_4_3_4 bugfix: (#23482) - Hyper-V INCR backup produces avhdx which refers to FULL vhdx - a differential saves. Adapt unit tests for "consolidate" function. change: (#21531) "WARNING: mtfana: readbuffer: SFMB block offset 0" occurs during path backup. Problem: log without severity Solution: Move log to Info level change: fixed sbc_ndmp build for Windows change: (#23473) - Adopt -x exclude option to NDMP ENV variables used by EMC VNX NEW: (#23171) Add possibility to activate compression in BSR-backup. Solution: if called with '-o compress[=LZNT1|BZIP2]' then add compression_type 'LZNT1 standard' (default) or 'BZIP2 4K' to the backup job's xml file change: (#11579) LIS file creation without saving data and backup of extended information only. Added log level checking for data size output NEW: (#23538) Start migration of saveset after backup: Read parallel execution and delete flag from migration task. If parallel execution is set, start migration at once for every task in a task group change: Si4-STPD integration bugfix: (#23553) sds failed to start after installation with var path set to 'C:\Program Files\SEPsesam\var'. Solution: Enclose configurationFile and gv_rw_stpd path in quotes bugfix: (#23352) STPD_MAIN process shows wrong information according active jobs, which results in error of drive handling. Problem: SYST command show already finished STPD-HTTP process. Solution: Set STPD-HTTP finished process PID's in structure to zero bugfix: (#23548) Single File Restore via HTTP-PIPE does not work on Windows RDS. Problem: STPD-HTTP semaphore locked till stpd_http_conn init finished. Solution: Unlock semaphore just before pipe creation to allow other STPD conect to STPD_MAIN change: (#23548) Single File Restore via HTTP-PIPE does not work on Windows RDS. Code reformat only, no business logic changes CHANGE: (#23157) BSR Windows: Add possibility to make excludes and put them into the xml-file. Solution: Implement functionality to parse '-x [,...]' and '-X ' and push the retrieved exclude patterns into .xml as NEW: Write size of files deleted by Si3 purge operation into 'DB:media_results.size new: Add 'DB:migration_tasks.delete_flag' to allow setting of 'delete saveset after migration' flag here and not only by event Revert "bugfix: (#23520) - qm_lib - arguments are no longer deargumentised/argumentised" Revert "bugfix: (#23520) - sm-qm - trim any whitespace at beginning and end of command" Revert "change: sm_qm_main replaces "--" with --" Adopt logging in 'i_GET_DRIVE_PROPERTIES()' change: Trace in i_ADJUST_EOLS() reviewed bugfix: COPY_SAVESETS() failed with 'Sesam COPY E001-DATABAS Failure during access to database: unrecognized token: ..'. Solution: Initialize szSl to empty string. bugfix: Remove single quote (') from Si3 status message, before writting it to Sesam DB, to avoid SQL syntax error Merge remote-tracking branch 'origin/master' into v4_4_3_4 change: (#11579) LIS file creation without saving data and backup of extended information only. Added windows part NEW: (#23181) Call 'ALL_JOBS_DONE' notification during Sesam newday, even if there a still jobs active. So this notification will be send in case of no backup interruption by newday too fixed compilation with NO_IPV6 flag change: (#11579) LIS file creation without saving data and backup of extended information only. Added output size of every stub (-v2) NEW: Purge orphand backup results, like migration backup results without original saveset change: Use commmon function to log command line NEW: (#23190) Socket based license for Hyper-V for SEP sesam edition 'VM Essential [Plus]'. Solution: Use the Hyper-V resource name to query for 'Cluster Node Name:' and 'CPU Sockets:'. Note: Customer has to set up all Hyper-V jobs via Hyper-V resource otherwise sockets may be counted twice on resource and on Hyper-V server change: sm_qm_main replaces "--" with -- CHANGE: (#23044) Inform customer about obsolete 'disk_info' backup task -> customer must switch to Linux BSR (REAR) backup. Solution: Changed Severity to 'strictly recommended' and module to 'KERNEL' CHANGE: (#23044) Inform customer about obsolete 'disk_info' backup task -> customer must switch to Linux BSR (REAR) backup. Solution: Add notification during update CHANGE: (#23529) License for Storage TB must show the underlying sizes. Solution: Details does now display the sizes occupied for DISK_HARD media and sizes of savesets with state in ('0','1','2') for the DataStores with data_store_type 'Path' CHANGE: (#17906) Write result of data store purge into 'DB:media_results', instead of result of 'check_space' made after purge new: Class 'DataStore': Add attribute 'isSi3' BUGFIX: (#21092) Do not start purge of data store twice at the same time, to avoid concurrent remove operations new: Delete orphaned 'DB:results' entries for non exisiting media bugfix: Windows: If command is executed on console and a child process is created then avoid Process Creation Flag CREATE_NO_WINDOW otherwise the stderr may be skipped. Example program call: sm_client.exe dir /Hyper-V: Merge remote-tracking branch 'origin/master' into v4_4_3_4 bugfix: (#17770) Informix 12 on Windows. Problem: onbar returned ERRORLEVEL=141. Solution: When compiled with I_API_PRELIMINARY then set XBSA_STANDARD to compile with XBSA_STANDARD BSA_MAX_OBJECTSPACENAME 1024 and BSA_MAX_PATHNAME 1024 new: Allow to start NDMP backup with or without '/NDMP:' prefix NEW: (#20811) Delete all 'DB:media_results' entries with action 'check_space' older than 7 days, except the one with the maximum filled value NEW: (#20811) Call 'check_space' function for data store using 1st drive only. So number of 'DB:media_results' entries will be reduced. bugfix: handle "/Hyper-V" keyword for any type of operation. bugfix: (#22553) Backup fails with "Current state is 'CHANGING_MEDIA'" when previous STPD handling EOM has been killed. Problem: STPD soft kill using "kill -2" while EOM processing cause all other backups to fail. Solution: STPD soft kill using "kill -2" should abort data transfer, finish SBC but keep running until EOM process finished bugfix: print out the CPU info in the new format and with using VMI calls. change: sbc_ndmp fix for crash at end of the backup bugfix: print out CPU information with setting blank between colon and value. change: (#11579) LIS file creation without saving data and backup of extended information only. Added output size of processed data change: sbc_ndmp - show /NDMP: prefix when displaying only volumes change: sbc_ndmp fix for non Cxx11 change: (#23246) - display message about empty stub; some minor fixes WORKAROUND: (#15766) Executing commands by Sesam: If command in 'DB:commnd_events.cmd' starts with 'DB:commands.cmd', then use command form events table change: (#23245) sbc_ndmp generates a filesystem catalog without mount ot making a data backup new: sm_copy logfile using PID in name bugfix: Since introduction of sbc browsing option '-a details' the 'Cluster Node Name: ...' was not present in the Hyper-V output. Solution: Add option '-a details' WIP: (#23190) Socket based license for Hyper-V for SEP sesam edition 'VM Essential [Plus]'. Solution: Add commands to get Hyper-V Sockets from Hyper-V nodes (= Hyper-V resources) with backup tasks. Additional: Now the license violation check for Hyper-V servers is done coz. backup tasks are now configured. Draft: The algorithm must use the cluster nodes not the resources! change: sbc_ndmp Create .lnd file with "w", release memory after completed backup. change: NetApp9 fixed mountsnap and umountsnap functions change: NetApp9 fixed listsnap function bugfix: (#23520) - sm-qm - trim any whitespace at beginning and end of command CHANGE: (#19815) Reduce log in 'i_CHECK_RESULTS_ROWS()' function, as it will usually add 3 lines for every backup job, which has runned at this Sesam day bugfix: (#23520) - qm_lib - arguments are no longer deargumentised/argumentised new: Add 'DB:capacities' entry for NetApp SnapShot drive Rename identifier of semaphore from 'restore' to 'sm_sms_copy' NEW: Saveset migration: Use drive with access mode 'READ' as source drive, if no source drive is set change: (#23515) BSR saveset-names and display-problem: Problem: If a 'BSR Windows' backup was started within a taskgroup then the saveset name starts without prefix 'BC'. Solution: Add 'BC' prefix bugfix: fixed crash on free() while "sbc -D /DB_INFO" running bugfix: possible fix to nested call for sbc -D bugfix:(#23470) - NDMP (netapp) restore fails with 'START RECOVER parse failure on nlist: missing leading slash on destination path bugfix: (#23468) - The error message should be improved if the source is missing. Print out just one error message instead of two. bugfix: (#23514) Uninstall previous VMware Vstor2 driver when updating VDDK. bugfix: (#23190) - Socket based license for Hyper-V for SEP sesam edition 'VM Essential [Plus]'. Print out the CPU information. change: sbc_ndmp allows only / when browsing change: sbc_ndmp requires /NDMP: at beginning of path in all operations change: Compiler warning resolved: variable ‘cpOrigSaveset’ set but not used bugfix: (#18092) sm_restore doesn't work if 'F' identifier is in LIS file. Problem: 'F' must be treated similar to 'f' bugfix: (#23342) Migration of DIFF/INC VMware backup works, but EOL adjustment fails in specified case. Problem: The logging of the adjusted savesets tried to use the original saveset but this saveset is only available as migrated saveset. Solution: Use migrated saveset in that case. Note: Now migration displays missing savesets when saveset adjustment failed due to missing savesets on target pool. change: (#23423) sbc on linux trys to use proxy set in http_proxy environment variable. Solution: added CURLOPT_PROXY, "" for xbsa change: (#23505) Updated aligning new: Add client name and OS to header of backup log new: 'sm_data_store': Adopt log file name for 'mount/umount' action, because there is no datastore name or drive number set CHANGE: (#22348) Allow task type 'NetApp:' for platform 'NDMP' change: sbc_ndmp fix some warnings change: sbc_ndmp - values not correctly reported by NDMP - math done differently change: (#23505) Aligned sbc key-value messages change: sbc_ndmp display dt instead of dm change: sbc_ndmp - display dt for volumes; fix for negative volume size - free size BUGFIX: (#22822) Restore wizard opens a MS exchange restore task with settings of a path restore. Problem: Granularity 'Databases' ( == '-a granularity=none' ) was used but GUI does not allow to select database rows with indicator 'D'. Solution: Generate LIS rows for Exchange server databases with indicator 'Db' to allow selection in the Restore Wizard. Note: Requires new backups to generate correct LIS file. bugfix: (#23165) vSphere config backup includes warning when specific excludes are defined. Solution: fixed work with -x option when it contains spaces (it always must be quoted), added escaping of '[', ']' for regexp bugfix: Do not use cm_strtok() in common functions coz the cm_strtok() nay be used by higher function and static buffere may be overwritten. Solution: Use cm_split() and FreeSplitList() new: (#23503) Sesam saveset migration from NetApp snapshots remove RCSID from sm_client change: Do not call 'sm_start' for virtual snapshot drives change: sbc_ndmp change d_ to dm for volumes Revert "change: sbc_ndmp browse - display dm for volumes when browsing" change: sbc_ndmp browse - display dm for volumes when browsing change: Purge of SnapStore: Use client name from 'DB:results' directly to delete snapshot BUGFIX: (#23179) EOL adjustment for failed vSphere backup uses own IMG sub-savesets. Solution: Check with 'AND session_id NOT IN (SELECT saveset FROM results WHERE ...)' to avoid selection of sub-savesets NEW: (#21260) Log the EOL adaption in Backup/Migration/Replication output. Solution: Add ADJUST_EOL output to replication log file too. change: Improved error messages if EOL adaption failed in case of replication NDMP Fixed windows build change: sbc_ndmp full tree browse support new: Drive type 'SNAP_NETAPP' bugfix: Do not overwrite backup sbc command for 'NetApp' add SEPzilla call number to TODO new: Use common 'get_credentials()' function to get login parameter for NetApp filer bugfix: (#23438) - check if file handle open before writing to qm_exit log; print stack trace on unpredictable crashes bugfix: (#22553) Backup fails with "Current state is 'CHANGING_MEDIA'" when previous STPD handling EOM has been killed. Problem: SMS do not know if EOM handling STPD process alive which can cause SMS hang in "CHANGING_MEDIA" state. Solution: Detect EOM handling STPD process pipe and check if its open till Resume called, if not - release SMS and finish concurrent backups with "1053: No further tape available" bugfix: Delete of obsolete 'DB:migration_results' doesn't work with SQLite DB, because of invalid SQL statement bugfix: (#23474) - Fixed Windows build bugfix: (#23474) - Allow deeper browsing of directories for backup source on NetApp filers change: Add Volume Professional/Ultimate and possibility to set NDMP license for volume edition BUGFIX: (#18092) sm_lis_items doesn't work if 'F' identifier is in LIS file. Problem: 'F' is used for stub files - indicate existence of a file but do not backup file data. These stubs are used to identify remaining files during DIFF or INCR (optional) and are used for NetApp snapshots too. Solution: Do not skip 'F' files. bugfix: Initial seed of Si3 store: Fix exception in check for running replications bugfix: (#23470) NDMP (netapp) restore fails with 'START RECOVER parse failure on nlist: missing leading slash on destination path bugfix: (#23485) sm_sesam was called with wrong media_pool. WIP: Add trace with PROT messages. Do not turn exit state to SUCCESS in case of called by GUI remove duplicate code bugfix: (#22872) Initial seed of Si3 store: Start replication processes, even, if there are less than 2000 objects in source store bugfix: Last change to use old submit syntax has doubled the executed command line sbc_ndmp Cleanup of unused variable change: NDMP excludes set with / instead of " bugfix: Since commit 12917e7 'new: Start Si3 service with restart flag, so QM will start it again, ...' the submit command executed on RDS could fail if RDS was not updated to 4.4.3.53. Solution: Use old sm_qm submit syntax if 'restart' flag is not set new: sm_backup: Add DRV_PROP to ARGUMENT Struct, because several drive related values are needed at many places in sm_backup NEW: (#23481) Submit drive configuratin task with unique identifier into queue, so only one 'sm_config_drives' will be active at same time bugfix: Si4 integration - backup from stdin and and restore to stdout fixed indexation issue in Hyper-V info browsing BUGFIX: (#23479) Avoid SQL error during insert into 'DB:media_results', caused by trying to write data store name into shorter defined label column NEW: (#23078) New BSR Pro 12.1. Add Uninstall keys to allow browsing and backup for new BSR Pro 3.1 (12.1) bugfix: Si4 integration - di get fix change: Function description GetStreamData() revised change: sbc_ndmp update doc change: sbc_nbdmp -o list=DIR not needed for browsing bugfix: Add '/' seperator for link rule bugfix: (#23015) fixed sbc's info call bugfix: (#19749) sm_ssh uses weak ciphers, could be security related risk. Problem: Weak ciphers/MAC available for connection to sm_sshd. Solution: Avoid using of weak ciphers by default and add possibility to change them using ini Merge remote-tracking branch 'origin/master' into v4_4_3_4 Correct allignment of message 'E009_MIGRATION_ABORT' change: (#23458) Disable VDDK cleanup routine. Solution: changed esx->szNoCleanup to bool from char*, and now TRUE is its default value bugfix: (#23467) Fix SQL error in 'INSERT_INTO_MEDIA_RESULTS()' because string for SQL command was limited to 512. Now 2048 characters are allowed NEW: (#23445) Show saveset tree: Add backup EOL to JSON output change: sbc_ndmp uses parameter -o list=Dir to list volume content change: Use def word for sbc_ndmp plain restore change: 'sm_sesam get_saveset_tree': Return better message, if saveset doesn't exist Revert "CHANGE: (#23433) Do not allow restore without directory strucuture for NDMP" This reverts commit 62df97f18a28b17c69270fa64bfd1d39f4789a65. Meanwhile the commit 2d145afea9627c12d2845de575bf17e6971a9886 Implemented option -o plain to work for flat restore. Note: In case of files with equal name the NDMP restore will only keep the last file due to restore mode 'overwrite' - other modes are not supported by NDMP. change: (#23433) - Implement NDMP flat restore fix merge conflict NEW: (#18007) Add new Sesam data store type 'NetApp Snap Store' NEW: Try to restart Sesam service, if terminated by operating system (segmentation fault for example) bugfix: 'i_GET_SAVESET_TREE()': Return with error, if saveset parameter is empty bugfix: (#23450) Set name of migration log file correctly. Name includes task name of saveset to be migrated, but it was't know at place, where it has been build change: display a message in sbc log if snapshot management extension is not enabled change: update sbc_ndmp documentation about how to use snap valult extension change: sbc_ndmp - clear some warnings new: (#23379) Re-write of Si3 deduplication server - new Si3 implementation integration into SMS/STPD change: added instructions to sbc_ndmp wiki - how to enable snapshot management extension bugfix: (#23382) - Hyper-V browsing issues. "Notes" are visible in dir output. bugfix: (#23215) - allow to specify multiple restart conditions bugfix: (#23438) - sm_qm_main crash location new: Si4 integration NetApp module - Added readme.wiki Merge remote-tracking branch 'origin/master' into v4_4_3_4 change: added references to ndmp source wiki change: sbc_ndmp readme.wiki introduced NEW: (#23368) Log any kind of EOL changes in Sesam main log. Solution: GUI now calls sm_sesam adjust_eol with new option '-o single_saveset' if user only wants to set the EOL for the selected savesets. Default: Set the EOL for the complete saveset chain new: Extend DB view 'drive_properties' by generic data store type bugfix: Fixed restore for Cluster mode NetApp 8 and 9 Merge remote-tracking branch 'origin/master' into v4_4_3_4 CHANGE: (#23433) Do not allow restore without directory strucuture for NDMP fix merge conflict bugfix: Correct define of macro 'IS_SNAP_STORE' change: Get 'DB:hw_drives' for data store: Return 1st drive (lowest drive number) as main drive, because many data store actions are only done using this one new: Use view 'DB:drive_properties' instead of 'DB:hw_drives' to get more information by one SQL call bugfix: (#23438) - sm_qm_main generates crash dump on every exception bugfix: (#23431) - NDMP selective restore failed with 'RESTORE: /smallfiles/1/file1: File does not exist in backup.' when meta file from backup is used NEW: (#23437) Set Sesam date for sub task results to Sesam date of main task bugfix: (#23193) sbc creates a file instead of folder at restore, when folder just has 1 character and restore path is UNC. Solution: fixed indexation of dir path bugfix: (#23326) copy console log from public folder to public old folder if it exists. change: WMI error description function is implemented. Blank operation TABS -> SPACES change: small fix of an access() macro change: small refactoring of cm_url_decode() and cm_url_encode() (migrated to int param instead of bool) bugfix: (#23420) - Change the Hyper-V restore path if new VM and new target path was inserted. Use new target path and add a folder called like the new VM name to restore the VM to. bugfix: (#23419) - Change the default Hyper-V restore path if only new VM name is given. Use Default Hyper-V Path + folder which is called like the new VM name that was inserted by the user (remove the "_restore" from the folder name) new: Si4 integration bugfix: (#18501) sm_sms_interface hangs if SMS isn't online. Problem: Pipe connection established but write operation stuck. Solution: Add timeout value to sendto_pipe() and read_from_pipe() to avoid such situation - wait up to 1800s bugfix: (#23382) - Hyper-V browsing issues. Make "not selectable" VM's files bugfix: (#23432) After update to 4.4.3.53 the NDMP backup with option for file index (-o DAR=1) failed with 'SBC logging without final state message'. Solution: Remove obsolete backup_option '-o DAR=1' bugfix: (#23414) declaration fix bugfix: (#23414) Error "A file was not found" at VMware backup when VM contains special characters. Solution: URI decrypting on vSphere source name now working only in case of "vsphere://" or similar prefix. Also we didnt touch '+' symbol in names anyway. change: Function SetNdmpVariable() now with trace sbc_log( 4000, ....) to trace the NDMP environment variables change: Function dump_settings() revised new: (#23418) Add the storage-free value to sm_dedup_interface, print output similar to "sm_client df /" bugfix: (#23382) - Hyper-V browsing issues. Make possible to select VM's files, display the size of VM's file, change item type of "Host Component" new: (#23418) Add the storage-free value to sm_dedup_interface, print output similar to "sm_client df /" add 'sob_clean' rule change class name change: Reduce trace log for remote copy file, if default log level (=1) is set new: (#23418) Add the storage-free value to sm_dedup_interface change: sm_qm_main - problem with lost log file handle change: sm_qm_main - don't print extra space bugfix: Purge of savesets from Si3 deduplication store. Set low water mark to '0', if set in 'DB:data_stores', because it is used as size of trash and not for low water mark here change: Macro LOG_CLOSE_LOGFILE with setting handle to NULL fixed linker error when vfs built with DEBUG flag bugfix: (#23403) Comment all BACKUP group relevant fields in the initial database insert. bugfix: (23382) - Hyper-V browsing issues. Problem: GUI doesn't shows VM files. Solution: Fixed it. change: (#22771) Si3 replication of some savesets fails with java.lang.AssertionError in background. Problem: We do not know with what software version current object was stored. Solution: Store and show current software GitId while "di objectinfo" bugfix: (23382) - Hyper-V browsing issues. Improved logic in the sbc. First steps for unit tests. bugfix: (#23413) Mount of saveset fails: Do not set -E parameter during queue submit, if 'end time' is 'now'. QM will set job to expired at once bugfix: (#23218) The process group somehow did not work properly. bugfix: (#23154) sm_sms_watch cores on diagnostix. Solution: before calling fclose() check if file handle != NULL Revert "bugfix: (#23361) Core dump of sm_stpd_main on Windows for unknown reason. Solution: reverted libmicrohttpd version to 0.9.48 until 0.9.59 fixing" change: Add EOL to message XX_BACKUP_EOL_SAVESET_ADJUSTED[_PREVIOUS] change: Windows: Compiler wanrings resolved: later include of cm_trace.h to avoid early include of cm_posix.h Revert "bugfix: (#23403) erase the BACKUP group from the list of groups. Set enable_gui_acl to true and set disable_gui_acl_full_permission_set to true. Hence ACL is enabled again and the Full permissions are enabled" CHANGE: (#23365): Replace 'Backup type' with 'Task type' in log files visible by user (no trace files) new: Add documentation file for Sesam global variables bugfix: (#23361) fixed libmicrohttpd 0.9.59 (also in sesamdev) change: Compiler warning resolved: format ‘%d’ expects argument of type ‘int’, but argument 5 has type ‘long unsigned int’ new: Add fucntion 'i_get_default()' to read value from 'DB:defaults' bugfix: (#23403) erase the BACKUP group from the list of groups. Set enable_gui_acl to true and set disable_gui_acl_full_permission_set to true. Hence ACL is enabled again and the Full permissions are enabled bugfix: (#23382) - Hyper-V browsing issues. Change the type of the item d_ => dv. bugfix: (#23382) - Hyper-V browsing issues. Speed up the VSS initialization and added the new option "-a details" specifying to print or not additional information. bugfix (#23399) - XEN diff backup with linux DM fails with Unhandled exception occurred bugfix: (#23388) - Citrix XEN restore to new VM overwrites original VM Merge remote-tracking branch 'remotes/origin/hyperv_cbt_branch' bugfix: (#23312) added handling of EAGAIN also bugfix: (#23361) Core dump of sm_stpd_main on Windows for unknown reason. Solution: reverted libmicrohttpd version to 0.9.48 until 0.9.59 fixing bugfix: (#23312) Operations with HTTP transport could fail with "Interrupted system call". Solution: reverted libmicrohttpd version to 0.9.48 until 0.9.59 fixing NEW: (#22015) Saveset migration: Set network interface for 'read' side too bugfix: (#23382) - Hyper-V browsing issues. Performance class is implemented to measure the function's execution time. change: sm_qm_main - log appended twice change: sm_qm_main - memory not freed bugfix: (#23154) Windows: sm_sms_watch cores on diagnostix. Problem: ASYNC_RCMD_THREAD() writes log message which refers to callers RCMD structure which may be released meanwhile. Solution: Use local RCMD structure change: sm_qm_main - fix compile time warnings change: sm_qm_main flush exit log for process on every occasion change: Use strlcpy() instead of strncpy() change: Increase RCMD_OUTLINE_LEN to avoid line break after 2047 characteres in RCMD output, e.g. sm_rexec ... bugfix: (#23390) sm_sms_watch for remote device cores with 0xc0000005 STATUS_ACCESS_VIOLATION. Solution: Avoid memory leak in function i_cm_exec() used to execute sm_data_store: release hThread handle change: Adapt call to READ_PERFORMANCE_FILE() - recently changed new: Purge of Si3 data store with active trash size limit: Make 'gv_dedup_purge_older_than' data store specific: 'gv_dedup_purge_older_than_' bugfix: (#23390) sm_sms_watch for remote device cores with 0xc0000005 STATUS_ACCESS_VIOLATION. Solution: Use scanf format string with 'maximum field width' to avoid overrun in string arrays bugfix: (#23312) Operations with HTTP transport could fail with "Interrupted system call". Solution: added handling of interrupted system call new: Add 'six.py' needed by recent 'XenAPI.py' change: sm_qm_main - append information about restarting job into job log change: sm_qm_main accepts 0 alongside now when specifying time new: Purge of Si3 data store: Make 'gv_dedup_purge_older_than' data store specific: 'gv_dedup_purge_older_than_' new: 'read_2_glbv_default(g1,g2,default)': If 'g1' does not start with 'gv_' construct name as _ CHANGE: (#15987) SAN shared drives are not displayed in license output. Solution: Calculation moved from sm_config_drives to license module NEW: Writes all entries from DB:defaults, where key starts with 'gv_' as Sesam GLBV CHANGE: (#23375) vSphere backup of VM using independent disk: Set backup status to 'Warning' and add a hint to exclude these VMDKs in task properties to get rid of warning message Fix merge conflict new: Start Si3 service with restart flag, so QM will start it again, after shutdown for getting DDLs from S3. Exit code from Si3 is '9' in this case NEW: sm_client client access check: Do not check locked clients change: Do not set dummy date/time values during browsing of KVM VMs change message for independent VMDK NEW: (#23375) add sbc_vadp message: '2971:Independent disk '%s' found, no snapshot of drive can be created.' (log level warn) change: sm_qm_main convert executable to full path name bugfix: (#23361) Core dump of sm_stpd_main on Windows for unknown reason. Potential solution: fixed libmicrohttpd struct (changed in v59 of lib) bugfix: GET_DRIVE_PROPERTIES(): Call DB_SEL_LOOP_RECORDS( &sctRL, 'd' ) before next call iDB_ACCESS( szSql, szBuf, &sctRL, strLog ) change: free() allocated memory in case of error change: Do not use uninitialized value in DB_ACCESS() BUGFIX: (#23380) Reading data from tape can fail with 'A tape access reached a filemark'. Problem: After reading a segment header the dir command failed. Solution: Since commit 46ddcac (#22283) Restore fails with ''Found interposed Segment Header after TapeMark.'' the segment header must be handled by the function caller (serv_fast_dir) BUGFIX: (#23380) Reading data from tape can fail with 'A tape access reached a filemark'. Problem: After reading a segment header the dir command failed. Solution: Since commit 46ddcac (#22283) Restore fails with ''Found interposed Segment Header after TapeMark.'' the segment header must be handled by the function caller change: License output revised: Customer related information now at beginning of license output change: (#23215) - restart jobs on processes terminated due to signal bugfix: (#23358() vSphere restore from full or incr fails with 'Datastore %s does not exist'. Problem: Since commit b0c74a8 (#21819) the message lines from NOT file are squeezed and therefore the offset to 'server=...' was changed. Solution: Avoid squeeze of message lines bugfix: sm_qm won't start if CLI > 1024. Solution: adjusted CLI size for sm_qm.exe bugfix: change: (#17909) Backup library for Sybase ASE. Problem: Restore must use saveset provided in LOAD command instead of saveset from ini file change: In case of SBC_COM the device must be checked and mapped to hw_drives.drive_num change: External task is called with -d {device} but i_GET_DRIVE_GROUP() handles only drive numbers. Solution: Check for prefix 'SMS:' and use hw_drives.device in such cases change: (#23215) support job restart on windows bugfix: (#21041) - Single item restore for Hyper-V. Optimized the PS script. bugfix: (#21041) - Single item restore for Hyper-V. Adjusted logic around avhd files. change: (#23161) License for SEP sesam Backup Client Si3S (Source Side Deduplication). Problem: If old Volume license is used then the use of SEP_DeDup_Si3S_Clients will reult in license violation. Solution: In case of a '... Volume ...' edition the SEP_DeDup_Si3S_Clients is set implicitly to unlimited bugfix: (#23354) sbc doesn't write mtime/ctime into LIS file during CBT backup. Solution: added creation of ctime/mtime for CBT change: (#23354) moved date mask to global header change: sm_vfs fetches ctime and mtime for every file in MTF stream and uses them if values in catalog are invalid bugfix: (#21041) - Single item restore for Hyper-V. Mount of INCR saveset for VM with and without installed Integration services. bugfix: (#21041) - Single item restore for Hyper-V. Improved logic around generating MAP file with involving additional AVHDX files from previous INCR. BUGFIX: (#23362) Sharepoint backup error 'sm_reformat_lis returned exit state 2' case sensitive error with drive letter 'd:'. Problem: Drive letter with lowercase. Solution: Set drive letter to uppercase BUGFIX: (#23347) Avoid creation of sesam-.sds.PID files in si3-root-directory. Problem: Windows: The rename function MoveFileEx() must be recalled if ERROR_ACCESS_DENIED raised up. change: Move macro '#define GET_LAST_ERROR GetLastError() | errno' to cm_common.h NEW: (#23324) Use low water mark of Si3 data store as maximal size of trash change: Disable ACL in UI by default via DB:defaults change: Static char array *gszCmdMode[] was not up to date with recent command arguments bugfix: Added wrong variable. Now gv_dayfile instead of just dayfile NEW: Read check for Si3 deduplication store: Read entries from 'recover.log' and show them as corrupted data files. If '-o remove_all|remove_recover' is set, then delete data files and Sesam DB entry bugfix: Hyper-V browse operation - print out VM's file descriptors. bugfix: (*23083) Hyper-V Host Component backup fails with "Backup finished. Status: ERROR Error: DB Module: [BackupProcessing: No virtual machine is found in the system]". Cause: Incremental backup of the host component has not been backed up. NEW: sbc_rhev: return false if attaching disks to data-mover fails bugfix: (#23348) Shutdown of SEP sesam service must stop Si3 daemon with softkill first. Problem: Sequence of daemons calls 'sm_qm shutdown' before calling 'sm_dedup_interface -d ... stop' for 'sds-#' process. Solution: Reverse sorted list of deamons after adding 'sds-#' daemons bugfix: Transfer Si3 database export file (for disaster recovery) URL encoded, because Windows coudn't handle line feeds change: Add DB:defaults entry to disable full permission set (key = 'disable_gui_acl_full_permission_set', user = 'sesam') BUGFIX:(#23345) sbc_rhev: utf8 encode vm comment to fix browse problems bugfix: Added global variable gv_dayfile BUGFIX: (#23334) SAP ASE failed with Error: The context allocation routine failed when it tried to load localization files!!. Problem: Locale LANG was set to 'POSIX'. Solution: Set default LANG 'en_US.UTF-8' instead of not supported 'POSIX' BUGFIX: (#23332) Strange files created under /var/opt/sesam/var/qm. Problem: sm_sepul_event uses NOT_LOG() but the file name may not be set. Solution: Declare file name empty to avoid arbitrary file names Fix log message fix merge conflicts bugfix: Read 'DB:drive_properties.generic_type' into DRV_PROP struct BUGFIX: (#23332) Strange files created under /var/opt/sesam/var/qm. Problem: sm_sepul_event uses NOT_LOG() but the file name may not be set. Solution: Declare file name empty to avoid arbitrary file names sbc_one: CHANGE: treat disk id as parameter to prevent TypeError Exception bugfix: (#23331) sm_prof for Powershell works on SEP sesam server only sbc_one: BUGFIX:(#23329) exclude disks with type FS from backup: not supported bugfix: (#21041) - Single item restore for Hyper-V. Implemented the INCR mount - generate LSL file and read entries from one instead of reading from LIS for creating SM_VFS "MAP" rules . change: (#23215) - added support for rolling time window bugfix: (#22238) Calling "sm_client dir \\?\pipe" leads to a coredump of sm_passd. Solution: added filter for GetFileInfoByName for pipes change: Common function to read 'DB:drive_properties' into 'DRV_PROP' struct change: Add 'DB:data_stores.uuid' as 'VIEW:drive_properties.ds_uuid' sbc_one: BUGFIX:(#23329) exclude disks with type FS from backup: not supported bugfix: (#21041) - Single item restore for Hyper-V. Implemented the INCR mount - generate LSL file and read entries from one instead of reading from LIS for creating SM_VFS "MAP" rules . change: (#23215) - added support for rolling time window bugfix: use ASCII format instead of Big-endian UTF-16. bugfix: (#22238) Calling "sm_client dir \\?\pipe" leads to a coredump of sm_passd. Solution: added filter for GetFileInfoByName for pipes new: changed version number from 4.4.3.52 to 4.4.3.53. BUGFIX: (#23311) Migration failed due to different checksum for encrypted Microsoft SQLServer backup. Problem:/Solution: The final block added for encryption was no summed up in the checksum new: (#23094) Set data mover for vSphere to vSphere server, if no data mover is defined and access mode is not 'PROXY' NEW: Script to download sm_ui.jar from server bugfix: enabled building x86 logic with old server libs bugfix: fixed compilation error (fixed patch utility path) NEW: (#22046) vSphere mount: Add 'INFO' message about mounted drive letters change: Add 'DB:hw_drives.previous_password' to allow re-encryption of Si3 data store with new password bugfix: (#23208) Selective Si3 restore is not able to skip in replicated savesets. Problem: Could not identify savesets with corrupted meta data. Solution: Add specific mode into FSCK to find corrupted meta data savesets bugfix: (#22950) Sesam uses sqlite3.exe in C:\Program Files (x86)\Intel\ASC, not the executable in \bin\db. Additional: Change directory back to gv_rw_work after execution of sqlite3 in gv_ro_db bugfix: (#23154) Windows: sm_sms_watch cores on diagnostix without logs with corresponding process ID. Problem: The RCMD_LL() function may lead to overrung in char array cpMsg. Solution: Use extra argument size_t iSize change: (#23300) Encrypt password used for Si3 encryption. Problem: showing information about unfinished DDL confusing. Solution: Hide last unfinished DDL in "di ddl status" output bugfix: If v_set_log_main() was called a second time (e.g. with 'sm_info.lgc' or 'sm_glbv.lgc') then the local log structure used in trace() was set to the new log file. Solution: Keep local log structure settings if already set new: (#23099): sm_vfs generates VHD on the fly new: add new chunkdir magic to di analyzer bugfix: (#23307) sbc restore q-pth_w2k8_none_dx loops with MTF_ERROR_STREAM_NOT_FOUND. Problem: Packdata does always retry the DBLK detection. Solution: Allow only 8192 retries change: TR message revised new: Set 'DB:task_types.short_name' for Kopano and Open Nebula bugfix: fixed x86 sesam building bugfix: improved the error-handling mechanism in the vmdk file to be able to get general/compilation errors with stack traces. bugfix: (#23208) Selective Si3 restore is not able to skip in replicated savesets. Problem: Replication link to existing metadata tree which could have wrong offsets. Solution: Additional FSCK mode to detect such issues bugfix: (#23289) Restore: Try to use preferred media pool change: (#20059) changes paths to libs and includes due to new sesamdev folders format sbc_one: BUGFIX:(#23294) ignore disks with type CDROM from backup BUGFIX: (#23285) Coredump in sm_passd during start of restore via ftp interface. Problem: IPv6 address with Zone ID lead to format error. Solution: Avoid using IPv6 as format specifier. Increase char arrays for client's IP to 51 chars (INET6_ADDRSTRLEN 46 plus zone ID, e.g. '%eth0') bugfix: (#23292) Warning message about EOL adaption will confuse customer. Solution: Avoid message 'ADJUST_FDI_MEDIA_EOLS failed with 'W014-BACKUP Die Anpassung der EOL für die Sicherung ... hat keine vorherige Sicherung mit Level 'C' gefunden.' NEW: Change message what todo, if Sesam server isn't allowed to connect to Sesam client, because we've SMSSH access mode now. (No 'Edit sm_ctrld.auth' message) bugfix: (#21041) - Single item restore for Hyper-V. Correct proceed the "unmount" operation under Linux. change: sm_vfs potential bug fixed bugfix: sm_vfs - use current time in case of 0000-00-00 time in LIS change: I005_REMOVE_SAVESET German text revised BUGFIX: (#23287) Problem: If old and new version are equal the db update is not executed. Solution: Insert a switch -F to force the db update even though versions are equal. new: (#23278) Exclude 'sbc_smdr' with correct path from REAR dependency check (remove doubled //). bugfix: AttributeError: 'NoneType' object has no attribute 'tr'. Possible cause: wrong logic around detecting defined 'log' variable. bugfix: global name 'log' is not defined. Possible cause is previous commit (SHA-1: 95407b3e1588668e64966b21cab8a1d89da5eb61) change: Initialize char szCmd[SL] bugfix: (#23154) Windows: sm_sms_watch cores on diagnostix without logs with corresponding process ID. Problem: sm_sms_watch.exe still core dumps in function qm_cmd_ll() with error STATUS_INVALID_CRUNTIME_PARAMETER (0xc0000417). Solution: Avoid to use overlapped buffer in strlcpy() sbc_proxy: CHANGE: Prettify log output sbc_rear: NEW: add excluded paths to COPY_AS_IS_EXCLUDE config variable too change: Add call to 'DestLL()' - will only be reached if while loop is exited CHANGE: (#23263) Sesam server installation: Use 'SMSSH' for remote access to Sesam server and do not try to check, if 'CTRLD' is available, so it can be switched off completely bugfix: (#23279) - Introduce an additional User Group, called BACKUP Update client: Change default to not checking inactive clients ÂCHANGE: Use 'SMSSH' as default remote access method BUGFIX: (#23279) Problem: sm_setup get_policy does not evaluate 'permission de.sep.sesam.gui.server.BackupPermission' correctly as backup permission. Solution: print backup in the get_policy of sm_setup if the permission is de.sep.sesam.gui.server.BackupPermission. bugfix: (#20059) fixed compilation error change: (#20059) modified makefiles to be possible to build stpd and sbc with newest verison of libraries (NEWLIB=1 flag when building) BUGFIX: (#23263) Problem: ctrl runs even if ctrld port has been set to 0. Solution: if ctrld port is set to 0 using -o ctrl=0 then the ctrl is set to off in the sm.ini. NEW: (#23263) Test of data connection during client configuration: Use common function 'i_CHECK_AND_SET_INTERFACE()' to get Sesam server interface. This will usually use 'http' transfer instead of 'ftp' sbc_rear: NEW: add files from exclude list to REAR COPY_AS_IS_EXCLUDE_SESAM sbc_rear: BUGFIX:(#23278) add sbc_smdr to exclude list if installed on system BUGFIX:(#23270) Use Image ID instead of image name to get disk information bugfix: (#23261) sm_data_store check_db: Remove '.info' files from data store, where no '.data' files exists any more bugfix: Fix exception in sm_data_store get_saveset_function BUGFIX: (#23123) sm_pur_status deletes used (of a backup) files after 4 days. Solution: Check if saveset is still active in table all_results change: (#23051) Useless source line log in Oracle trace. Solution: Use macro sbc_log() which calls sbc_log_f() with __FILE__ and __LINE__ new: Migration trace log: sm_copy and sm_sms_copy write into same trace file, if no submit of sub-tasks is requested change: vSphere restore: Do not enclose 'filter' argument with quotes, because it includes several arguments, which should been as seperate ones bugfix: (#21041) - Single item restore for Hyper-V. Print out assigned dirve letters in alphabetically sorted drive numbers. change: (#23051) Useless source line log in Oracle trace. Solution: Move trace line 'Parse answer: ...' to higher log level (XBSA_TRACE >= 3) change: Set 'DB:hw_drives.block_size=64' if 'block_size>16384'. GUI has set 'block_size=65536', but kernel uses 'block_size*1024'. GUI has been changed not to write 64 too. bugfix: (#23257) License timeout but external backup is executed. Problem: Last commit added result to resturn message BUGFIX: (#23257) License timeout but external backup is executed. Problem: External backup license does not check SERVER and TIMEOUT change: Trace messages revised which have been added with recent commit in v_set_log_all_close() CHANGE: Extend 'DB:restore_results.r_task' to VARCHAR(64), because external backups create longer automatic generated restore task names change: Call qm_lib_close_answer() to release file stream handle change: Use fclose() instead of close() when file stream is created with fdopen() new: (#19776) Check, if execution flag for a taskgroup is switched off and do not start group in this case bugfix: (#21041) - Single item restore for Hyper-V. Correct indentions in the python script. change: Last commit in cm_trace.h requires '{' and '}' to enclose TR#() macro when used in if ... else ... constucts bugfix: (#23154) Windows: sm_sms_watch cores on diagnostix without logs with corresponding process ID. Problem: Setting up the log structure for the sm_client performance thread lead to concurrent close of strLocalLog.fpLogFile - the trace() function log file handle. Solution: Function v_set_log_all_close() use CriticalSection to avoid concurrent use of strLocalLog.fpLogFile NEW: (#23209) Add information about EOL of media pool used for backup to backup main log change: sm-qm usage doc - added usage for -R change: sm_qm usage updated - example how to restart a job new: (#21556) Do not overwrite 'filter' (which includes VMDK names to be restored) for task type 'vSphere' new: Function: 'pDB_GETVALUE()' to read a single value from Sesam DB. Value is returned in a static allocated string. In case SQL statement doesn't return a value (not found or wrong SQL statement), default value is returned bugfix: (#21041) - Single item restore for Hyper-V. Changed logic around fetching assigned drive letters for VHDX files. bugfix: (#21041) - Single item restore for Hyper-V. Error: "Cannot bind argument to parameter 'ReferenceObject' because it is null." change: (#23215) - sm_qm_main restarts jobs on condition change: sm_vfs code cleanup bugfix: (#21041) - Single item restore for Hyper-V. Error: "Could not find executable [DiskPart.exe] in defined path(s)." BUGFIX: (#23225) sm_setup with options set_sm_ini fails with: 'set_sm_ini' failed: Error: '...'. Solution: Add logic to create sm.ini file when it does not exist or force switch '-f' is set. Get gv_rw/gv_ro from option '-o bin=...,var=...' or via current directory. Example: (Windows) C:\Program Files\SEPsesam\bin\sesam>sm_setup.exe set_sm_ini change: modified makefiles to be possible to build with newest libmicrohttpd, libcurl, libopenssl libraries (use NEWLIB=1) bugfix: (#20059) Backup over https:// is about three times slower than ftp:// or http:// Solution: client - standardization of TCP buffers used in project, server - increased memory allowed to libmicrohttpd to use bugfix: (#21041) - Single item restore for Hyper-V. Fix error "local variable 'osver' referenced before assignment" BUGFIX: (#23233) Problem: deactivation of ldap and ad did not work correctly via activate_uac. Solution: deactivate ldap/ad also using activate_uac, if they are set to 0. bugfix: (#21041) - Single item restore for Hyper-V. Fix error "global name 'DiskPart' is not defined" bugfix: Fix syntax error bugfix: (#21041) - Single item restore for Hyper-V. Impoved the error-hadnling mechanism of the PS and some other stuffs, format output for sm_vmdk_interface module. remove unused code line bugfix: (#23237) 'sm_sesam getpassword' function cuts off client name, because it calls 'i_SPLIT_LABEL()' function for -m parameter. Problem: If client name ends with 5 digits then it may be shortened. Solution: Check if function 'getpasswd' is called BUGFIX: (#23237) 'sm_sesam getpassword' function cuts off client name, because it calls 'i_SPLIT_LABEL()' function for -m parameter. But this parameter isn't used for media/pool related functions only bugfix: (#21041) - Single item restore for Hyper-V. Make able to run script on the RDS. new: Si3 new implementation integration new: Use new log rename function to write 'sm_sepul_event' log file change: S3 sync: Find first free label id for S3 pool change: Import common Sesam Python modules in main 'sm_vmdk_interface.py' again change: Import common Python modules in main 'sm_vmdk_interface.py' again bugfix: Include 'vhdx' module correctly during Linux build NEW: Sesam server on UCS: Check for PostgreSQL verion 9.6 NEW: sm_update_client: '-p 1' switch to check/update active clients only (DB:clients.permit='y') bugfix: (#23164) - check if lis file is compressed and uncompress it. Also check if lis file exists in the first place bugfix: inserted a missing indent. bugfix: (#21041) - Single item restore for Hyper-V. Compilation error under Linux bugfix: (#21041) - Single item restore for Hyper-V. Use diskpart instead of PS. new: Set 'DB:clients.access_success' to date, where Sesam server has reached client successfully last time bugfix: (#21041) - Single item restore for Hyper-V. Compilation error - "STATUS=ERROR MSG=local variable 'log' referenced before assignment" bugfix: (#23221) Fix exception in 'sm_data_store' caused by recent refactoring new: 'DB:clients.access_success' to store date of last successful remote access to client bugfix: (#21041) - Single item restore for Hyper-V. Compilation error. bugfix: (#21041) - Single item restore for Hyper-V. Support vhdx module and handle savesets with Hyper-V type of backup operation. bugfix: (#21041) - Single item restore for Hyper-V. Some adaptions in the main module to be albe to mount VHD(X) files BUGFIX: (#23218) start the remote process as process group. This enables the sm_remote_installer to wait until all child processes are finished. bugfix: (#21041) - Single item restore for Hyper-V. Implemented the mounting of the FULL savesets with image files inside. BUGFIX: (#23191) File restore uses wrong media pool, if saveset was replicated and already removed from original Si3 store. Problem: Generation restore starts with DIFF backup on a migrated tape media but FULL backup only exists as a replicated saveset. Solution: Add logic to check for replicated saveset new: new Si3 implementation client library change: sm_qm_main display remaining args for follow up job change: fixed some memory leaks sm_qm_main when job submitted to unexisting queue change: fixed some memory leaks in sm_qm_main change: thrift interface changes NEW: (#23201) 'sm_start': Do not send Sesam alarm message, if tape drive is busy, because backups are still running filter out cyclic Si3 message new: (#23102) Set 'DB:data_stores.status=4', if Si3 service is down change: Ask for TIMEOUTDATE in case of edition enum 'vm_essential' NEW: Allow bigger block size for tape devices change: SMS: Do not log block information with level 'WARNING' bugfix: (#23176) Analyze restore log file: Recognize 'no space left on device' error message correctly change: Linux: LIC_INSERT_ITEM() Use strncasecmp() instead of strnicmp() NEW: (#20426) Socket based license for VMware vSphere for SEP sesam edition 'VM Essential [Plus]'. Solution: Add new license key ESX_Sockets which is valid only together with SEP sesam edition 'VM Essential [Plus]'. Note: In this case the ESX_SERVER[_SEATS] are ignored. bugfix: (#15915) - Original Xen VM will be deleted, if VM is restored with new name bugfix: (#23163) fixed crash during vSphere config backup when exclude has 23 characters bugfix: (#23169) 'sm_sms_watch 0' call to 'sm_qm sho all' does not work anymore. Solution: Fix the call to qm_cmd_ll() with 1st argument 'sho' and 2nd argument 'all' new: (#23161) License for SEP sesam Backup Client Si3S (Source Side Deduplication). Solution: Add new keys to sesam_make_lic NEW: (#22179) Start 'sm_pur_status' asyncron by queue manager during Sesam day change, so it doesn't block SEPuler for longer time change: (#22504) Reverted change disabling printing of GitID in log. This change not needed anymore after moving v_set_log_mover() code to a separate function NEW: (#23161) License for SEP sesam Backup Client Si3S (Source Side Deduplication). Solution: The new key SEP_DeDup_Si3S_Clients licenses the amount of clients with backup tasks (task_events) using 'Source Side Deduplication' change: sbc_encrypt_password() Use memset(cpEncryptedPassword, 0, ulDestLen); new: Build sm_vmdk_interface using sub-modules NEW: sbc_rhev: save UUID of virtual machine during config file backup bugfix: inserted better error handling to easen error detection. BUGFIX: (#23126) If a wrong package is used for installation abort the procedure in the highest function, not in the subfuction. bugfix: (#23119) - Extend sm_vmdk_interface to mount/attach virtual disks from different hypervisors. workaround: (#23164) Restore: Remove check for LIS v2 lis file format, because it doesn't handle compressed LIS files BUGFIX: (#22978) sbc complains about 'Not supported reparse tag 0x00000000L' for some reparse points. Problem: The FindFirstFile() does return wrong reparse tag 0x00000000 for some IO_REPARSE_TAG_DEDUP files. Solution: Retry FindFirstFile() and backup reparse tag 0x00000000 as normal file CHANGE: (#17826) SQL module might overwrite database on restore if move option specified. Problem: This is the correct Microsoft SQL Server restore behavior. If restore to orignal database with move option and without 'REPLACE' then the original ldf and mdf files remains on the system bugfix: (#23132) MSSQL restore in to new db with move option overwrite the original existing db. Problem: Since last commit the backup and browsing of servers with default instance 'MSSQLSERVER' does not work. Solution: Default instance '(local)' or 'MSSQLSERVER' must not be used in SQLDriverConnect() BUGFIX:(#23155) add option -a skip_id to skip VM uuid and mac address regeneratoin new: Set 'restore' option for browsing Xen server to show available storage repositories CHANGE: sbc_one: do not call get_domaininfo twice NEW: sbc_one: show size of virtual machine during browse BUGFIX: (#23132) MSSQL restore in to new db with move option overwrite the original existing db. Problem: If target is given with '{server}/{database}' then the given database name is used as instance so the SQL Server connect must fail. Solution: Check restore target and convert to '{server}//{database}' to allow SQL Server connect to {server} Add pid to sm_arch log file change: (#22504) v_set_log_mover() now also recognizing tokens CHANGE: #23130: do not add saveset id to backupsrc bugfix: (#10258) fixed sm_arch warnings change: (#23130) RHEV restore: Correct search for VHD savesets over session id == main saveset id bugfix: (#10258) fixed struct initialization change: a little refactoring for MEDIA_RESULT struct filling, also fixed v_set_log_mover dup issue change: Copy shared libraries for SMS with 'cp -p' to preserve executable bit new: Set 'DB:media_resuls.pool' for action remove media NEW: (#23147) sbc_rhev: show VM size during browse CHANGE: #23130: adopt sbccom calls match with other task types like vSphere change: (#10258) Common function i_INSERT_INTO_MEDIA_RESULTS_ARCH() BUGFIX:(#23145) Fix restore of virtual machines with disk type Virtio_SCSI BUGFIX:(#23144) fix restore of disks with type virtio Add 'readdblk_vfs.c' to .gitignore Add 'sh.py' to .gitignore change: sm_lib - que_shocsv no longer accepts cpReturn as an argument change: fixed some indentation faults and erased commented code which is not needed anymore. BUGFIX: (#23140) Only execute the windows defender powershell script if powershell version 2.0 or higher is installed on the system. CHANGE: sbc_netapp: set volume_path if specified by option, do not default to 'vol'. Default keep_snap to False in order to prevent 'AttributeError: NetApp instance has no attribute 'keep_snap' exception during backup change: remove obsolete code bugfix: (#22748) - add support for shocsv to qmlib change: sm_qm_main fixed some memory leaks change: sm_qm_main displays complete command line for command with CSV printout is used BUGFIX: (#23132) MSSQL restore in to new db with move option overwrite the original existing db. Problem: Since commit fba32b6 (4.4.3.48) (#22758) Restore of a single db from ms-sql 'all' backup fails with 'SQL Server does not exist or access denied' a restore target given with {server}/{instance}/{database} is shortened to {server}/{instance}. Solution: Do not shorten restore target name. NEW: Citrix Xen: Use 'sbc_proxy' as backup command and enable FULL/DIFF/INCR backups change: (#22504) Rename of log file in Sesam log function. Added date to moved log name NEW: Try to use same backup protocol (http/ftp) for Sesam server (sbc -b -C switch) as set for remote device server (-S). Use Sesam server name as 1st selection NEW: (#10197) Optional setting of Sesam server interface (sbc -b -C server) for 'sbc' calls via 'DB:task_events.srv_i_name'. Necessary, if client has to open sub-savesets (vSphere, Xen...) over 'sbc_com_interface', but couldn't reach Sesam server over 1st network interface bugfix: bugfix: (#18007) NetApp snapshots as Sesam backup saveset - dir set times to "----" for DataOntap8 change: (#22504) Rename of log file in Sesam log function. closing log before renaming (to all issues for now) change: (#22504) Rename of log file in Sesam log function. closing log before renaming change: (#22504) Rename of log file in Sesam log function. added path v_set_log_mover() bugfix: (#18007) NetApp snapshots as Sesam backup saveset - dir volumes print volume sizes" NEW: Try to use same backup protocol (http/ftp) for Sesam server (sbc -b -C switch) as set for remote device server (-S) change: (#22504) Rename of log file in Sesam log function. Added define for v_set_log_mover() change: Add Python version output bugfix: (#23125) Remote command fails during vSphere backup with 'RCMD Exit status = -1, msg [host TAUCHFIX not yet configured]'. Problem: Since last commit 3212cad the i_RCMD() checked for sesam server via name from table clients, but vSphere backup uses name from sm.ini [SERVER] which is written uppercase on tauchfix. Solution: Use name from table clients for command execution instead of sm.ini [SERVER] name=... change: (#22504) Rename of log file in Sesam log function. Small refactor of code BUGFIX: (#23124) Only uninstall and install the O&O package included in the installation package if the version number is higher than the one installed on the system. new: add stpd-https documentation BUGFIX: (#23118) Oracle backup: Return correct error message, if saveset couldn't be inserted into 'DB:results'. Before this fix, error message was set to 'Duplication of primary key' regardless of error returned from DB server BUGFIX: (#23100) sm_sms_watch cores because of too many open files (rsh_perf3.tmp). Problem: Windows: The thread function i_ASYNC_RCMD_THREAD() wrote into log file sm_client_{daylbl}.log but the log file handle was never closed. Solution: Close log file handle before exiting the thread CHANGE: sbc_rhev: better error handling CHANGE: sbc_rhev: fail better if config file not found during restore change: sm_vfs - no path conversion for unknown saveset type CHANGE: sbc_rhev: default values for variables to None NEW: sbc_rhev: fallback by searching devices in /dev/disk/by-id if not found using pyudev bugfix: Syntax error in Java 64 bit check for existing Java link CHANGE: 'DB:restore_tasks.r_task -> VARCHAR(128). In case of external started restore (Oracle) automatically created restore task can become longer then 49 characters new: Add a seperator line in backup log between EOL adoption messages and final backup state message change: sm_vfs saveset type 'raw' changed to 'none' change: sm_vfs supports RAW saveset (unknown saveset type - image new: Allow 'sbc' to fall back from full/diff/incr backup level to 'copy' (Citrix Xen < 7.3 for example) bugfix: sm_vfs mapper fix - missing else bugfix: (#23104) - Change automatically to Copy backup for older Citrix Xen versions : do not try to save config dir aftar full falled back to copy backup BUGFIX: (#23117) Incremental MS-SQL backup stucks during data transfer. Problem: The STATS thread is blocking further data transfer. Solution: Allow to skip BACKUP option STATS with setting '-a stats=-1' or '-a stats=0'. Note: '-a stats' will set option STATS without percentage (==: default 10%) Correct link rule for 'sm_arch' change: updated sm_vfs doc change: (#18812) - sm_vfs exposes saveset as a normal file Link slu and QM with 'cm_common.a' new: Do not log 'ARGUMENTS' twice remove commented line change: Follow up events: Check return code from 'sm_cmd' calls change: (#22504) Rename of log file in Sesam log function. Disabled double log new: Simpler replace of '' token in 'v_set_log_main() Revert "change: (#22504) Rename of log file in Sesam log function. Avoid log of parent task if log with PID is used" bugfix: (#23104) - Change automatically to Copy backup for older Citrix Xen versions : fixed double appending of SBC parametes list change: (#22504) Rename of log file in Sesam log function. Avoid log of parent task if log with PID is used bugfix: bugfix: (#18007) NetApp snapshots as Sesam backup saveset - delsnap command chenged to format "delsnap - s saveset" bugfix: (#23104) Change automatically to Copy backup for older Citrix Xen versions - adopted SBC command NEW: sbc_rhev: more debug output change: (#22504) Rename of log file in Sesam log function. Added strLog->bIsPidUsedInName flag bugfix: STPD HTTPS failed to start with new GnuTLS. Problem: Default priority string couldn't be parsed with new GnuTLS because some ciphers or protocols removed from "NORMAL". Solution: Build priority string according to GnuTLS version NEW: sbc_rhev: debug output bugfix: (#23083) - Hyper-V Host Component backup fails with "Backup finished. Status: ERROR Error: DB Module: [BackupProcessing: No virtual machine is found in the system]". Cause: something wrong with parsing time pameter fetched through WMI. Solution: back the CIMDateTime2FileTime() converter function. bugfix: (#21259) - Backup of VMs with Slash in name leads into exclude of Hyper-V VSS Writer. Solution: 1st step - make possible to do the VM backup by using its uuid. 2nd - would be making adaptation on the server side (sm_sms_backup.exe) change: (#21041): - sm_vfs maps specific file within saveset to a specified folder/filename change: (#22504) Rename of log file in Sesam log function. Added "" token parsing in v_set_log_main() change: (#22504) Rename of log file in Sesam log function. PID parsing from log file name bugfix (#23104) - Change automatically to Copy backup for older Citrix Xen versions change: (#20598) - Fallback to level COPY If backup level other then COPY is executed on Citrix version < 7.3 bugfix (#180079# - NetApp snapshots as Sesam backup saveset bugfix: Remove 'echo' from makefile, because it hides error return code from freeze process bugfix: (#23022) - A clustered VM is in state "not clustered" after restore with rename. Print the special log with id 3999 in case of backup. new: Si3/S3 clone: Create snapshot by 'sm_sms_watch' Add comment new: Function to read DB select result into Python dictionary. Attention: 'Dict' has no sort order Add trace messge new: Analyse restore log (used by sm_sms_copy too): Write 'errno' to trace file, if main log file couldn't be opened BUGFIX: (#21989) - ERROR: [CVssServer::CreateSnapshot(): IVssBackupComponents::DoSnapshotSet: WaitForAsyncOperation] - VSS_E_PROVIDER_VETO. Cause: Detected that sometimes it begins to occur if one volume has > 64 TB. Solution: WORKAROUND - would restart the operation if the VSS_PROVIDER_VETO is occurred. BUGFIX: (#22912) Improve security handling of sbc. Solution: Avoid to display sensitive data with higher log level bugfix: (#19092) Use a different temp name for the files that are moved to check if they are in use. This is needed to prevent a collision with the old installation on the system. BUGFIX: (#23085) EOL of a VM-backup (son) is set to 3 but the mediapool eol is 1. Problem: The EOL is set to the default 3 days for the child saveset (via SM_SBC_COM). Solution: Use common function i_INSERT_FIRST_LABEL() for SM_SBC_COM and SM_BACKUP bugfix: (#23082) Segmentation fault of sm_dedup_interface on SLES12. Problem: In case of error NUL returned but no check for it. Solution: Use variable only after NULL check NEW: (#16791) Use 'SMSSH' as default access mode for Sesam client access change: slu -z switch reverted change: block limits correctly displayed; slu supports -z to display block limits change: slu displays max_block_length and min_block_length for drives change: (#22504) Rename of log file in Sesam log function. Maked sm_sms_backup log unique bugfix: Hyper-V browse operation - wrong date/time is printed out, do not print date/time for "Host Component" BUGFIX: (#23006) Check for correct format of saveset stream info read by 'sm_sms_watch'. Avoid core dump in case of corrupted 'info' line new: (#22504) Rename of log file in Sesam log function. Also removed log rotating function from sbc bugfix: (#23083) - Hyper-V Host Component backup fails with "Backup finished. Status: ERROR Error: DB Module: [BackupProcessing: No virtual machine is found in the system]". Cause: re-factoring in the browse operation. Fix: re-implmented "Host Component" backup. change: (#22748) - sm_qm suports shocsv command; list of all jobs diplayed as csv kernel/unix_hal/sm_hal_unix.c bugfix: insert some log lines for clarification. NEW: (#23035) Compress LIS file on RDS before copy it to Sesam server. CHANGE: Search for file in journal files: If query string starts with '%' the use 'iregex' for compare NEW: (#23078) New BSR Pro 11.2. Add Uninstall keys to allow browsing and backup for new BSR Pro 2.2 (11.2) CHANGE: sbc_one: longer timeout during image creation wip: (#20598) Fixed snap=deletebefore so that task cannot delete snapshots made by other tasks and other sesam servers bugfix: (#20598) insert the .exe ending to the sbc exectable for windows to get rid of the access denied error. BUGFIX: (#23074) External backup could fail with 'Timeout reached, drive group qs-disk has no free streams anymore'. Problem: Submitted sm_sbc_com_ext did not yet start (5 seconds timeout). Solution: Wait up to 20 seconds for GLBV. Allow to override this timeout via GLBV 'gv_submit_ext_timeout' bugfix: (#17617] Delete Sesam snapshot on Citrix XenServer after backup is broken during sm_newday change: Calls to trap_par_err() revised new: (#22504) Rename of log file in Sesam log function. Added log rotating function new: (#22504) Rename of log file in Sesam log function. Added log rotating function (sbc Windows) bugfix: (#20598) missed a equal sign. bugfix: (#20598) accidently mixed up windows and linux cases. bugfix: (#23071) Use the new Installshield 2016 for building kits from now on. bugfix: (#20598) insert double quotes around the xen source for windows and change the BIN_DIR for sbc_proxy to sbc folder. new: (#23035) Optional compress of file before remote copy. Added -z option for sm_rexec to receive compressed file from sm_client, or for send compressed file bugfix: (#23035) declaration fix new: (#23035) Optional compress of file before remote copy. Added compression of file before sending via sm_rexec new: (#23035) Optional compress of file before remote copy. Added -z option for sm_rexec to receive compressed file from sm_client, that we decompressed locally after receiving wip: (#20598) - Added restore=true to VM meta-data import link for restore to original target BUGFIX: (#23067) Could not get volume list: Unable to find API: volume-get-iter. Problem: 'NetApp Release 8.2.3P6 7-Mode' does not provide method 'volume-get-iter'. Solution: Use method only for non '7-Mode' BUGFIX: (#23062) que_submit_sync could stuck sporadically. Problem: On SLES12 some sm_qm requests are lost. Solution: Avoid close and reopen of to_qm pipe new: (#23035) Optional compress of file before remote copy. Added decompression for sm_client bugfix: (23063) Custom setting of directory used for build CBT image during restore Re-enable log of VM uuid again new: Citrix XEN backup: Write VM uuid into 'DB:results.source_uuid', because VM name (= backup source) isn't unique new: (#23035) Optional compress of file before remote copy. Added -o comp[ress] to sm_client Revert "new: Set login backup options for all backup types, if set in DB. Revert because options are set from client options, if not set in task options. But this must not be correct for task" Revert "new: (#23035) Optional compress of file before remote copy. Added -o comp[ress] to sm_client" BUGFIX: (#20308) Restore by sbc_smdr: Write 'DB:restore_results.size' after restore has finished. Log line for saveset size has an older format, which hasn't been reconized by Sesam server bugfix: (#23035) fixed UNIX type conversion new: (#23035) Optional compress of file before remote copy. Added -o comp[ress] to sm_client bugfix: (#23022) - A clustered VM is in state "not clustered" after restore with rename. The new option is handled -a "clustered" for restore with clustelize. CHANGE: (#17975) Oracle backup: Move some re-oceuring logs to higher log level bugfix: (#15915) - [Bug 15915] Original Xen VM will be deleted, if VM is restored with new name BUGFIX: (#22836) Remote Sesam update: Modify Sesam systemd unit file, so sshd/ctrld will not be killed during update. Last commit hasn't work for older Sesam installations, where 'KillMode=' line wasn't present in unit file bugfix: Hyper-V browse opertion enchancement by the addtional information - ip addresses of the guest OS. Remove log of UUID until kernel has been adopted BUGFIX: (#19092) During remote update the zip file of the python sm_sshd.exe is being deleted even if the sm_sshd.exe is still in use. This could be one of the reasons for the 1601 error. new: Citrix Xen backup: Log VM uuid with special sbc message number bugfix: Hyper-V browse opertion enchancement by the addtional inmfomation fetched from guest OS. NEW: Allow to set network interface for MaxDB restore wip: (#20598) - Added VM disks virtual size to sbc_proxy -D wip: (#20598) - Added VM install and start time to sbc_proxy -D change: Add -EXPORT:strreplace for sesam.dll change: Improved sm_sms_tapeutil: Allow to read savest with Tape ID and Filemark into dump file bugfix: (#18501) sm_sms_interface hangs if SMS isn't online. Problem: Log show error lines with timeout. Solution: Avoid error logs due to timeout during connect to pipe NEW: (#22851) Extend Sesam search LIS function to filter results by task type: '-t ' is supported now NEW: (#22851) Extend Sesam search LIS function to filter results by task type (-t ) wip: (#20598) - Fixed direct restore for multiple disks wip: (#20598) - After VM.destroy, try to destroy VM disks, in case for disk data not removed from NFS storage change: Hyper-V browse imporved - print out the uuid of the VM. new: sbc Xen: Direct upload of full VDI saveset without temp. file BUGFIX: #23047 MaxDB: In case sbc_com_interface fails, no analysis is possible because loglevel cannot be raised. Solution: Use -v {loglevel} for 'sbc_com_interface ... open' too BUGFIX: #23047 MaxDB: In case sbc_com_interface fails, no analysis is possible because loglevel cannot be raised. Solution: Check for entry in utl file: SBC_COM_TRACE_LEVEL=... . If not set then set to 1. (#20598) Added argument direct for full restore to temporary file, http or pipe. CHANGE: (#23045) Update 'DB:results.fdi_type' to uppercase, as UI only shows results with upercase FDI type bugfix: Correct return value from sbc_pre operation from XEN module Revert "wip: (#20598) Fixed Copy backup - return values" Add log CHANGE: (#23045) Write 'DB:results.fdi_type' always in uppercase in DB for external backups. UIonly knows upper case states wip: (#20598) Fixed Copy backup - return values Add log bugfix: Correct return value from 'sbc_pre' in XEN module new: Call 'sm_dedup_interface dir' only once for a Si3 store, as there are noc directories for every media pool their wip: Adopt sbc_xen to run on Windows wip: Si3/S3 clone pool: Start 'sync' command in 'sm_sms_watch' loop wip: Si3/S3 clone pool: Set DB:media.last_init/sesam_date correctly change: Add 'DB:result_lbls.pool/id'. Currently there we have only 'DB:media.label' and pool/id has to be derived by splitting label. This does not allow to use simple joins and also the label format could not be changed to allow more then 99999 media in pool NEW: (#18492) Search files in Sesam LIS file by regular expressions. (-o strstr|regex|iregex) new: i_pcre_find()' with case insensitive search option bugfix: (#23027) vSphere backup fails with error: "Configuration files not found" and sbc cores. Solution: fixed array filling '-x' param filtering bugfix: (#23041) Browsing of "/Hyper-V:" resource is looping. Solution: Removed loop from -D param args parsing bugfix: fix to VSS (-D) cycle error fix syntax error new: 'DB:media.external_id' and 'DB:media_pools.pool_type new: Class 'Drive': add function to call 'sm_dedup_interface' by remote command bugfix: possible fix to VSS (-D) cycle error bugfix: Allow longer command line for 'java' call by 'sm_java' Remove RCSID log Correct checkif sbc.exe exists on Windows wip: (#20598) Fixed level copy backup and partly implemented optimization for level full restore BUGFIX: (#17770) Informix 12 on Windows. Problem: The inquire of an already stored saveset did not find any saveset due to wrong path delimiter: '\' ==> '/'. Sort order changed (now asc) to return last successful backed up saveset. change: stpc_recv() trace received bytes with log level 3 new: Set login backup options for all backup types, if set in DB change: (#23033) Write correct unit 'GiB' instead of 'GB' (what we've used as GiB in Sesam for over 10 years now) in 'sm_dedup_interface status' output new: Allow 'user=' and 'username=' for MySQL login bugfix: (#11419) Compress LIS files before copy from them RDS. Problem: SM_SSH compression doesn't work. Solution: SM_SSHD should allow compression BUGFIX: Sign dll, exe, pyd, ps1 files only once with SHA256 using the new certificate. change: (#23024) - Remember the salt used to encrypt the password bugfix: removed backslash for sbc_change_path, caused for skipping paths change: (#23024) - Remember the salt used to encrypt the password change: call_pass_api: Set request->szClientIP for Windows too change: PASSD: Authorized stpd connection now with FQDN, short name and alias check for given name (or alias). Former check of first 17 chars removed. bugfix: Fix XEN copy and meta-data backup change: removed double sbc info, cosmetic changes change: introduced slu -X media_and_drive_status wip: (#20598) Quote sbc command to work on Windows revert last change. Doesn't work change: Hyper-V browse operation - prints "single_selection" regardless of VM is clustered or not. Fix syntax error sm_sesam analyse_log: Call 'ANALYSE_RESTORE_LOG()' for '.mig' extension wip: (#20598) Add disks with device=xvda at the start of the list to be backed up first new (#20598): insert XEN module for Windows. Erase the chown import of the os module since it is not available for windows. BUGFIX: (#22837) The files in the rythm folder are now included in the installshield project. bugfix: (#23015) sbc -D command doesn't work anymore. Solution - merged nix and win dir logic change: 'DB:media.id' change INT to BIGINT, so save unique ids based on timestamp bugfix: change format of the output date for Hyper-V VMs, fixed "changed" date. wip: Read tape capacity by 'slu' and update 'DB:media' new: 'DB:media.capacity/free' to store this information read by 'slu' bugfix: (#22898) - Hyper-V backup with Resilient Change Tracking (equivalent to CBT). Wrapper around Virtual Disk API. wip: Read more information by slu command new: Read 'DB:backup_srv_options' into option struct in 'sm_sms_backup' Remove 'rcsid' from sm_sho' new: Add 'llTapeMaxCapacity' and 'llTapeRemainingCapacity' to struct SLU for media related valuse read by 'slu' change: adapt log levels in scsilib BUGFIX: (#21819) Backup ends with wrong error message '(0)'. Problem: Since this fix the message '1053: No further tape available' was not longer retrieved. Solution: Explicite check for message '1053: No further tape available' and for 'NEGATIVE reply: (0)' bugfix: (#22922) Create STPD "perf" files during startup. Problem: sm_sms_watch complains about missing perf file. Solution: Generate "perf" files on STPD_MAIN start change: (#22998) Remove old non SMS_INTERFACE code parts. Problem: We still maintain obsolete code. Solution: Remove obsolete code bugfix: (#18744) sm_sms_main must not write logfiles under C:/ if sms.ini is defunct. Problem: sesam creates millions of logfiles on the C: partition if Log_Directory not set appropriate. Solution: Notify about missing log file destination entry and abort application startup (SM_SSHD) BUGFIX: (#22837) fill the rythm folder during windows build. change: (#22348) Allow backup type 'NetApp' to be used for NDMP clients too BUGFIX: (#22511) Do not allow multiple VMs or 'All' to be saved in one task. Problem: GUI Hyper-V source browser allows to select 'all'. Solution: Display source with '/Hyper-V: ...,not_selectable' CHANGE: sbc_one: wait until snapshot has been removed change: Trace message improved change: SBC_WIN32_VERSION 4.4R4 now supports '-a password=...' instead of old '-a passwd=...' new: Unify setting of login and data mover parameter for backup NEW: insert the rythm folder in the windows kit build of the server kit. It is now present only in the server kit but it is still empty. wip: (#20598) Do not delete disks before deleting VM -> error VDI_IN_USE change: remove useless log "TP_OpenSavesetStart:: No Tape Trailer!" BUGFIX: (#18501) sm_sms_interface hangs if SMS isn't online. Problem: Process hangs during opening of pipe. Solution: Use pipe timeouts for command pipe (10s) and for answer pipe (1800s). Timeouts may be overwritten via sms.ini [SMS_Server] Connect_Timeout or Answer_Timeout bugfix: (#22694) moved magic's for encrypt to separate header, fixed duplicated and missed. Problem: Use string SZ_CPIO_ENCR_STUB instead of value SZ_CPIO_ENCR_MAGIC wip: (#20598) Fixed name of base_snap file - use task name instead of VM name. bugfix: (#22934) Hyper-V backup fails because of a leading space character in hostname. Problem: Browsing of cluster node now with 'Cluster Node Name: {hyperv_node}'. Solution: Remove leading blanc from node and from FQDN bugfix: (#22782) changed sbc_vmware minimum log to 4 instead of 3 bugfix: (#22782) changed vmware lib trace level to 3 (minimum needed for catching errors) bugfix: (#18744) sm_sms_main must not write logfiles under C:/ if sms.ini is defunct. Problem: sesam creates millions of logfiles on the C: partition if Log_Directory not set appropriate. Solution: Notify about missing log file destination entry and abort application startup (sm_ctrld) wip: (#20598) Switch to FULL backup if snap_file is corrupted or not found. Temporary disable logout xen session. bugfix: (#18744) sm_sms_main must not write logfiles under C:/ if sms.ini is defunct. Problem: sesam creates millions of logfiles on the C: partition if Log_Directory not set appropriate. Solution: Notify about missing log file destination entry and abort application startup bugfix: (#22007) Stpd cores during BSR Pro restore from tape. Problem: mistype in variable name. Solution: fix mistype change: Function generate_performance_data() moved to admin_mode.c. Write ouptut to perf...file only if output was changed change: Source code for login_to_SMS() revised change: Function description added wip: (#20598) DIFF backup implemented, delete old snapshots on every FULL backup, fixed snapshot cleanup not to delete base snapshots change: Trace levels in i_SEMA() and x_Semaphore() for Linux/UNIX revised to reduce log output change: trace() function output without level specification must only appear in log level 3 new: Log some important SMS log messages in all log levels change: Compiler warning resolved: missing prototype change: Indentation revised Correct message text wip: (#22915) STPD connection timeout, because EOL adaption and meta file deletion needs too long during EOM. Problem: Pipe name mismatch on windows. Solution: Avoid setting macro PIPENAME if already set change: (#17310) - LISv1 is generated from LISv2 upon restore CHANGE: (#18370) Set default segment size for SMS tapes to 20GB. This value is set here only during installation. TODO: Set it by 'sm_config_drives' bugfix: (#22782) VMware backup protocol does not include all VDDK logs. Solution: increased minimum VDDK log level Change TABs into SPACE in source code new: Add SMS state change logging BUGFIX: (#22915) STPD connection timeout, because EOL adoption and meta file deletion needs too long during EOM. Problem: Tape thread does not wait for communication pipe. Solution: Retry open for client's read only pipe Define 'iRet' before use BUGFIX: (#22915) STPD connection timeout, because EOL adoption and meta file deletion needs too long during EOM. Problem: Backup failed after last commit. Solution: Need while loop to connect to pipe to avoid going on without pending data NEW: Re-start data store FSCK operation, after removing orphand saveset sets or DB entries new: Citrix Xen server backup: Use tape server set for main saveset Change trace messages wip: Use active main job for Citrix Xen VHD restore, instead of starting an extra 'sm_sbc_com_ext' job BUGFIX: (#22915) STPD connection timeout, because EOL adoption and meta file deletion needs too long during EOM. Problem: Data thread is aborted 'ERROR: read_from_pipe:: client breaks.' but tape thread still waits on named pipe. Solution: Use function connect_pipe() with timeout to return after 10 seconds change: control StpdHttpChildWaitTimeout using stpd.ini. Problem: Default timeout 17 min (1000 sec) is too short. Solution: Increase default timeout to 30 min (1800 sec) bugfix: (#22988) Set SMS working state to CHANGING_MEDIA at once EOM area is hit. Problem: Working state changed to CHANGING_MEDIA too late, which allows to start new backups while EOM. Solution: Set CHANGING_MEDIA working state as soon as EOM area reached change: control StpdHttpChildWaitTimeout using stpd.ini. Default: 1000 sec as before wip: (#20598) Logout xenapi session after backup/restore/dir operations BUGFIX: SAP Hana backup: Check for task termination signal from Hana server send, just before returning answer to Hana client BUGFIX: SAP Hana backup: Check for task termination sinnal from Hana server send during fetching of backup medium new: Add a signal handler for 'sm_sbc_com_ext', so it could handle Sesam 'break' operation correclty change: Call trace(...) in qm_lib functions error_func(...) and usage_func() Revert "Revert "BUGFIX: (#22799) sm_sms_interface force_eom doesn't work any more. Problem: The communication to the EOM pipe does block other communication to the data server. Solution: Release the semaphore before calling call_tape_api(...)"" bugfix: Do not return before calling FindClose(hFind) in IsMountPoint() change: Use certificate with pfx file extension BUGFIX: (#22977) Segment number is set to '-1', when backup starts before and sm_sms_backup starts after newday. Instead of reading 'Sesam date' in 'sm_sms_backup' again (which is changed by newday), it is read from 'DB:results' bugfix: fixed crash in acl_to_info() when relocation path is used new: "di snapshot get " command added wip: (#17241) - Introduce Access Control Lists Revert "BUGFIX: (#22799) sm_sms_interface force_eom doesn't work any more. Problem: The communication to the EOM pipe does block other communication to the data server. Solution: Release the semaphore before calling call_tape_api(...)" change: STPD abort connection if tape in "CHANGING_MEDIA" state. Problem: STPD should check tape state and wait for correct one if possible. Solution: Run login_SMS in loop while tape state is different to correct one or timeout appears (default: retry 120 times every 30 sec) change: checking for backup source duplication bugfix: Hyper-V browse is failed if one VM has missed "InstallDate" property or "GetSystemFileSize" method is failed. wip: (#20598) Cleanup of forgotten snapshots before backup when -a snap=deletebefore bugfix: (#22898) - Hyper-V backup with Resilient Change Tracking (equivalent to CBT). Prepare enviriment. NEW: (#22966) License for OpenNebula. Soluton: Add License Keys OPENNEBULA_SERVER and OPENNEBULA_SERVER_SEATS to license the amount of OpenNebula servers and VMs BUGFIX: (#22970) crash fixed while vmware ini listing (extended buffer for data) BUGFIX (#19092): Get rid of check of variable Installed. This variable is not initialized correctly hence the check fails. change: Add 'DROP INDEX IF EXISTS results_original_saveset;' BUGFIX (#19092): Replace all occurences of ISMAJORUPGRADE and UPGRADINGPRODUCTCODE with IS_MAJOR_UPGRADE="Yes" BUGFIX: (#22969) - NPE comes up, while editing the properties of a immediate started task (with e.g. 1h offset) wip: (#20598) - Moved level copy restore sbc command call to sbc_proxy BUGFIX: (#22836) Remote Sesam update: Modify Sesam systemd unit file, so sshd/ctrld will not be killed during update. (#22197) Remote Sesam update: Retry update by 'zypper' to install missing dependencies, if update by 'rpm' fails BUGFIX: (#22836) Remote Sesam update doesn't work, because 'systemd' kills complete process group. In service file 'KillMode=Process' has been set, but must be 'KillMode=process' new: (#22963) Add 'DB:terms.over_next_exec' new: (#4676) User authentication and authorization. Solution: Add sqlite triggers for new table 'ui_acls' and remove obsolete triggers for 'ui_*locations*' tables BUGFIX: (#19092) Sesam services will now be started after the old sesam has been removed. This is needed for the major update. Otherwise the client is offline after major update. Revert "BUGFIX: (#22949) sbc_unix: fixed absent SBC_EA_ entry in saveset because of timeout issue" BUGFIX: (#22949) sbc_unix: fixed absent SBC_EA_ entry in saveset because of timeout issue NEW: Use 'sbc_proxy' to browse Citrix Xen Server new: Add 'name description' during browse of Citrix Xen CHANGE: sbc_one: work on restores of non persistent vms bugfix: fixed loop in XBSA if trying to close connection when no info when written wip: (#20598) - Check if multiple VMs with the same name as backup/restore source/target exist and skip action if so. Fixes for SR functionality NEW: sbc_one: allow backup of non-persistent images CHANGE: sbc_one: mark non persistent vms as not selectable BUGFIX: (#19092) Partially revert the commit 63b5adab207b since the sesaminstaller.ism was committed accidentally. change: minor changes. NEW: sbc_rhev: show vm comment during browse Revert "bugfix: fixed loop in XBSA if trying to close connection when no info when written" bugfix: Impoved the mechasnism for the aborting of backup operation. new: Dynamic detection of makefile path, so includes will work regardless of path used on command line CHANGE: (#22960) Sesam SAP Hana configuration: Correct usage example to reflect new option (-j) for backup task bugfix: insert a hint in the build routine of the installer. If a new GUID is introduced the user should add several files in the Setup.rul file of the installer. The automatic insertion would be too much work. BUGFIX: (#22810) Delete left over registry keys from 4.4.1, 4.4.2, 4.4.3, 4.4.4 and the UpgradeCode as well as the SEP Elektronik GmbH if no partial uninstall has been selected, during uninstall. CHANGE: sbc_one: skip disks with type SWAP bugfix (#12525) Adopted SR size to output format new: (#22920) sm_client delete file with multiple argument list. Problem: sm_client can delete only one file, or using wildcart. Solution: make sm_client call to delete several files including wildcards up to max. command line length bugfix (#12525) Adopted SR size to output format new: (#22920) sm_client delete file with multiple argument list. Problem: sm_client can delete only one file, or using wildcart. Solution: make sm_client call to delete several files including wildcards up to max. command line length NEW: (#17241) - Introduce Access Control Lists Use login parameter already parsed by 'sbc_opts' function, instead of parsing them from '-a ' switch again bugfix: fixed loop in XBSA if trying to close connection when no info when written bugfix (#12525) Adopted correct output format new: Add 'DB:oper_systems.login_type' to trigger setting of credentials in UI during client configuration (like for 'tasks') bugfix: (#22931) - Various VSS API errors during hyper-v csv backup. Cause: restored VMs into CSV got snapshot folder resided on the local volume instead CSV. Solution: Fix it and some minor bugs. bug (#12525) Fixed call to scan_vm BUGFIX: (#22950) Sesam uses sqlite3.exe in C:\Program Files (x86)\Intel\ASC, not the executable in \bin\db. Solution: chdir into gv_ro_db before command execution bug (#12525) Target storage selection for Citrix XEN restore bugfix: Correct check of return code from drive configuration (SM_SUCCESS/SM_OK) new: 'DB:task_types.targetstore' to trigger selection of target store (vSphere data store, Citrix Xen storage repository...) during restore bug: (#12525) Target storage selection for Citrix XEN restore - remove reduntant default SR message -l copy bug: (#12525) Target storage selection for Citrix XEN restore - changes to work also for -l copy CHANGE: sbc_one: set default encoding to UTF8 to work around problems with browsing VMs that contain special characters like umlauts wip: (#14840) Citrix Xen Server configuration data restore: Write data inot a file on data mover instead of trying to create a VM name 'meta-data' bug: (#12525) Target storage selection for Citrix XEN restore NEW: (#17060) Backup over proxy client: Change error message, if proxy isn't reachable, so user can see, which client access caused the problem new: Set 'DB:defaults.protection='r'(eadonly) for Sesam license entries new: Set 'DB:defaults.protection' to 'w'(rite) for entries, which can be changed in GUI bugfix: Keep the bin_gui_htmlComponent in the gui packages. bugfix: removed the build error 'ISDEV : error -5008: Intel64 or AMD64 must be specified in the template of the Summary Stream'. The new bin_gui_splashComponent was an AMD64 component and was not replaced correctly during build to be an x86 component. bugfix: changed the position of the sm_wol.exe and sm_scrdll.dll to the correct place in the file lists. change: Replace old restore log messages in sm_sms_copy and use specific migration related ones NEW: (#22913) Do not set encryption flags, if saveset is dumped into a file change: (#21260) Log the EOL adaption in Backup/Migration/Replication output. Solution: Add output to migration log new: Citrix Xen target storage repository by same command line option as for vSphere new: (#22919) 'DB:tasks.backup_srv_options' and 'DB:restore_tasks.srv_options' to set options, which are only used on server side and not set in 'sbc' call change: Code review FILL_SMS_INI_LIST(): Check if tape device supports encryption only in case of real tape device new: Citrix XEN restore: Check for active main restore task to avoid start of sub restore task new: Citrix XEN restore: Set session id correctly for external sub restore requests new: (#22858) Add Sesam server version to begin of backup log NEW: Added table 'ui_acls' and disabled all 'ui_*locations*' tables NEW: (#22837) Add some generic 'rythm' templates to Sesam 'skel' directory, to demonstrate use of new Sesam template engine wip: (#20598) Fixed meta-data backup and set session id to -j value on restore BUGFIX: (#22924) Migration based on monthly schedule with 3 days offset started with one hour delay after reset of Daylight Saving Time (DST). Problem: The 3 days offset added to Friday 2017-10-27 results in execution on Monday 2017-10-30 after the day of reset of DST (Sunday 2017-10-29) but calculation used Friday and added 1 hour (DST additional hour). Solution: Calculate DST changes based on execution date BUGFIX: (#21439) Message about failed EOL adoption leads to a SQL error. Problem: SQL with extra quotes. Solution: Avoid extra quotes in messages change: (#22135) -x [pattern] - now works in other OSes too, also some code formatting wip: (#20598) Use -j as session id on restore bugfix: (#22935) allow to switch LDAP/AD authentication on/off via sm_setup switch -l 0/1 and -d 0/1 using activate_uac. deactivate_uac turns LDAP/AD off. bugfix: (#22892) - sm_vfs parses complete date time and does not reconstruct it bugfix: (#22934) Hyper-V backup fails because of a leading space character in hostname. Problem: Browsing of cluster node now with 'Cluster Node Name: {hyperv_node}'. Solution: Remove leading blanc from node name BUGFIX: (#22901) Do not create a complete Si3 ini-file (sm_main tries to start a 2nd Si3 server), if number of streams for 2nd drive is changed add log in GET_DRIVE_PROPERTIES BUGFIX: (#22933) Copy data store INI file into SI3 data directory again, so it can be used for inital seed. Copy hasn't been made any more, because of wrong presetting of a variable NEW: (#22921) Add dependencies for REAR on SLES12. (bc binutils genisoimage) new: Set -j for Citrix Xen restore to track sessions opened by sbc_com bugfix: (#22135) -x [pattern] - replacing "," by "|" before calling regex new: Do not call any drive specific action, if migration is started with drive number < 0. Fix compilation error BUGFIX: (#22839) The subversion check only worked if the subversion the same length. It should work for all version length now. BUGFIX: (#22839) The subversion was not detected correctly. Now it should work. bugfix: the build of the installshield did not exit with error in case of an error. bugfix: insert the sm_vfs.exe in the exclude list of the client and gui kit. bugfix: change the position of the sm_wol.exe in the file list. change: (#21260) Log the EOL adaption in Backup/Migration/Replication output. Solution: Add EOL adaption to backup NOT file bugfix: (#22918) BSR Pro restore from multi-tape fails. Problem: Restore expect to have first saveset tape inserted into slot, which is not always true. Solution: While calculate restore detect expected tape and insert it if its not in slot already change: Use '#ifdef WIN32' instead of 'IPC_VERSION' as this shows better which part is used for Linux or Windows NEW: (#22917) Enable log rotate for sm_qm_exit log file. change: Remove '#ifdef UNIX/WindowsNT' from IPC/no IPC part, because Windows uses IPC and UNIX not Sort options in 'case' statement alphbetically NEW: (#21841) - Adjust protection mode for default entries in sesam_insert_start.sql for entries known to be written by the UI wip: (#20598) Fixed sbc arguments for backup/restore level copy BUGFIX: (#22839) Insert log lines to easen error detection BUGFIX: (#22801) insert the minor version of Java in the Installer Text to clarify why the update can not continue. CHANGE: sbc_one: sys.exit(2) in case of error CHANGE: sbc_rhev: sys.exit(2) in case of error CHANGE: sbc_kvm: sys.exit(2) in case of error wip: (#20598) copy backup: prevent additional call to sbc from sbc_proxy, do not write bas snap files on copy backup bugfix: (#22135) -x ALL now worked with all cases of vSphere backup NEW: (#22909) Modify 'DB:all_events' VIEW not to show 'next execution time', if schedule is deactivated sbc XEN: Exit with exit code=2 in case of error bugfix: (#22889) data_size and stored_size in the db are not equal to the values of sm_dedup_interface -d LW list of the same saveset. Problem: Last line in LIS file could contain wrong number of Total Dedup bytes (Windows). Solution: Use size_t in output format Citrix XEN cbt restore: Avoid restore of same saveset twice wip: (#20598) Remove restored config dirs after restore. change: HYPERV browse VM with full information. Integration into vss module with refactoring - remove sources change: HYPERV browse VM with full information. Integration into vss module with refactoring of exclusion logic. wip: Citrix Xen CBT restore: Do not create sub-restores, because FDI chain is restored by sbc XEN module by 'sbc_com' wip: (#20598) changed workdir name from XEN to xen change: Change 'XEN' into 'xen' in config directory path wip: (20598) Change work dir to saveset/XEN, delay closing last job on restore change: (#22370) Add 'direct forensic sector image' as imaging method for encrypted drives in BSR Pro. Change: Usage changed: '-a backup_method=used_sector|forensic|direct_forensic' wip: vSphere URL encoded source for 'VSPHERE_CONFIG' wip: Citrix Xen CBT restore: Do not create sub-restores, because FDI chain is restored by sbc XEN module by 'sbc_com' BUGFIX: (#22907) Client configuration: Fix core dump during access check of old Sesam clients. Check length of returned error line correctly and cut it off, it it's to long NEW: (#20811) Reduce number of entries in 'DB:media_results' for action 'check_space' wip: (#20598) - Restore work dirs for all INC savesets, exception check of enable_cbt call, re-enable delete of work dirs on backup new: Exlucde 'dbrecover/dbbackup' build for Domino client from dependency chech in RPM bugfix: (#22135) removed warning when NO config downloaded BUGFIX: (#19487) Lotus Notes DIFF skips newly added mailbox databases. Problem: Notes API NSFIsNewBackupNeeded(cpFile, dwComfortSpan, &dwBackupNeeded); does not return dwBackupNeeded for newly added NSF files. Solution: Check for return code ERR_RM_RECOVER_NONBACKUP - The database has not been backed up on this server/instance of the log bugfix: (#22135) removed displaying of NOT downloaded config fix merge conflict wip: Citrix XEN CBT restore: Do not search for VMDK saveset, if no backup source is set in 'sbc_com' call, but return parameters for requested configuration file backup change: (#22135) sbc -b .... -x ALL "VSPHERE_CONFIG:" now didnt download anything new: Add sm_wol.exe. Solution: Add copy command for cli and gui package NEW: Add sm_wol.exe NEW: (#22867) OpenNebula VM restore: Build restore target in the same way as for KVM. ('source VM==::==target VM') NEW: Windows: Add sm_wol.exe to windows kits BUGFIX: (#22839) A check for the java subversion greater than 1.8.0_111 was added. The Java Check was completely refactored. Java version 9 has been removed from the supported list since the GUI does not support it yet. BUGFIX: (#22839) To ensure that a java version 1.8.0_111 is larger than 1.8.0_71 a strlen comparison was added. Java version 9 has been removed from the supported list since the GUI does not support it yet. change: Add spadmin.dll to x64 packages gui and cli new: Add backup level to Citrix Xen restore command wip: (#20598) Fixed full backup/restore when VM has multiple disks. BUGFIX: (#22893) Hyper-V 2016 CSV VM restore fails with 'Unable to locate input assembly 'hyperv_import.dll' or one of its dependencies.'. Solution: Add missing binaries hyperv_import.dll and spadmin.dll to cli and gui packages. bugfix: (#22897) - Hyper-V restore creates useless directory names like "__RESTORED_VHD__{1a945756-2081-487a-bc61-80ec22f19747}". Changed to "restored_". BUGFIX: (#22893) Hyper-V 2016 CSV VM restore fails with 'Unable to locate input assembly 'hyperv_import.dll' or one of its dependencies.'. Solution: Add missing binaries bugfix: Client and GUI packages must include hyperv_import.dll, spadmin.dll and sm_vfs.dll change: HYPERV browse VM with full information. Integration into vss module with refactoring - added creation and modification date/time. change: HYPERV browse VM with full information. Integration into vss module with refactoring - added size. bugfix: (#22887) - slu introduces additional check to determine if device is capable of fast inventory bugfix: (#22694) moved magic's for encrypt to separate header, fixed duplicated and missed, removed obsolete traces change: HYPERV browse VM with full information. Integration into vss module with refactoring - test browse operation. change: (#21260) Log the EOL adaption in Backup/Migration/Replication output. Draft: Add information to NOT file (1st draft) change: Improved file_append( src, trgt ) change: Do not overwrite content of NOT file -> use file_append() change: Comment added to struct CLIENT.szLocation: '// location.id' coz this field does not store the location's name change: Trace message in file_append() adapted change: Now file_append( src, trgt ) returns system exit code similar to Windows file_append(...) BUGFIX: (#22874) Reduce the frequency of messages in the day log for sm_client df. Problem: The execution of the 'sm_client df' must not appear in SEP daily log. Solution: Avoid the logging of that command. Improve logic: Call sm_alarm with W007_MEDIA_SDS_MIN_FREE_REACHED or W008_MEDIA_MIN_FREE_REACHED if threshold is reached the first time wip: (#20598) Save dummy saveset for disks with zero changes on incremental backup change: HYPERV browse VM with full information. Integration into vss module with refactoring. change: Most log messages in qm_lib_get_answer_buff() moved to upper level to get proper output of command change: Typo in comment change: sm_vfs display more information on bac_lexical cast CHANGE: (#18152) sm_qm library. Problem: The QUE_SHO() functions did not check the state returned by qm_lib_close_answer() and therefore every call created a wrong error message 'E008-BASICS Error in command execution: ....' in SEP sesam daily log change: Remove DllExport from local function pointer gfp_sbc_log() change: (#22515) Extended client information function call. Problem: sbc windows did not print application info. Solution: Skip dummy function ktrace() change: (#22515) Extended client information function call. Solution: Use function sbc_log() for sbc_unix too CHANGE: (#22515) Extended client information function call. Problem: sbc windows did not print application info. Solution: Use function sbc_log() instead of ktrace() change: HYPERV browse VM with full information. Integration into vss module. wip: (#20598) INCR backup fix- incr backup saves changes after prev. incr or from full backup when it is first incr backup after full backup bugfix: (#22135) changed ALL INCLUDE var name to prevent to it to be free-ed by "INCLUDE*" mask BUGFIX: (#19092) the check of the old sesam version was accidentally done with the new sesam version. Now the old sesam version is used instead. change: di analyze output changed change: (#17909) Backup library for Sybase ASE. Problem: If saveset does not longer exist then COM returns message: 'sm_sbc_com-W002-SBC_COM CONNECT_RESTORE Saveset SF20171127172533829@UDGj1cD_cBS does not exist' change: (#22135) remove spaces at start of -x subparam (if exist) NEW: (#22135) sbc -b .... -x ALL "VSPHERE_CONFIG:" - to avoid download unneeded files BUGFIX: (#20657) The function iCheckValidJava did not return an ERROR in case the JAVA_HOME variable does not exist. Now an error is returned and an alternative java version is searched for. BUGFIX: (#22857) - use mutex to prevent concurrent operations on all types of savesets bugfix: (#22722) vSphere: fixed crash on vix disk close BUGFIX: (#19092) The setup type will now be obtained correctly. BUGFIX: (#22799) sm_sms_interface force_eom doesn't work any more. Problem: The communication to the EOM pipe does block other communication to the data server. Solution: Release the semaphore before calling call_tape_api(...) CHANGE: sbc_one: show vm name as comment during browse CHANGE: sbc_one: output only VM ID during browse bugfix: (#21109) Backup front end client for SAP ASE. Problem: If SAP ASE is not installed on Windows then command ends with ''UnboundLocalError: local variable 'OCS_dir' referenced before assignment''. Solution: Initialize 'OCS_dir' before use. bugfix: (#21523) Master ticket for Authentication ( UI and sm_cmd ) against AD or LDAP. Problem: After the update to 4.4.3.48 the table 'ui_group_external_group_relations' was empty. Solution: Export tables with name > 31 chars too (now up to 127 chars) change: Added 'enum LANGUAGES eLang = LANG_EN' again change: Added ktrace(..) dummy function again change: Comment line removed bugfix: Run restore session with SBC_COM session change: Startup log improved: Now if the startup log file is not set via environment var "SSA_LOGFILE" then log will be appended to file '/tmp/ssa_logfile.log' (Linux) or 'C:/ssa_logfile.log' (Windows) if existing. change: Logfile with header line "# SESAM SAP ASE LIBRARY [git ID]" change: Return syb error message together with last sbc_log(LOG_ERROR, ...) message - so SAP ASE error is more detailed. change: Build backup source with STOR command assembled from SBC_COM args. change: Do not call sbc_com_interface with empty error message -M "" - coz popen() will fail on Windows change: HYPERV browse VM with full information. Add new sources. BUGFIX: (#19092) To enable the new baviour of the major update a function had to be inserted which temporarily changes the product version in the sm.ini to a fake version. This enables the major update to work properly. change: Logging improved: SAP ASE may not allow to set SSA_LOGFILE in environment - so check for default log file /tmp/ssb_logfile.log or %TEMP%/ssb_logfile.log new: sbc_com inqure for Citrix Xen CBT restore bugfix: (#22845) sm_sms_watch uses qm_lib. Usage is controlled by gv variable: gv_use_qm_lib bugfix: qm_lib - size of buffer correctly returned change: Function qm_lib_get_answer_buff() revised change: Function logic for qm_lib_get_answer_buff() revised change: qm_lib windows fix bugfix: (#22722) vSphere: (relocation) added \Q \E non-escaping sequence bugfix: (#22722) vSphere: (relocation) avoid escaping already escaped symbols CHANGE: (#22265) EOL adjustment for failed Backup or Migration must be activated via default entries 'eol_adjust_failed_backup/migration', Default: YES. Problem: Entries must be added to table defaults. Solution: Add default entries if not yet existing change: qm_lib enhancements in get_buffer and close functions - return number of bytes bugfix: (#22722) vSphere: temporarily crash fix bugfix: (#22722) vSphere: added encoding of target for cbt restore, restored counting of saved/not save elements (added overflow check), fixed crash on VixDiskLib.Close (incorrect disk handle) error text NEW: (#22849) Search pattern in LIS files belonging to a certain client BUGFIX: (#22848) VMware attach does not work when server is Windows and RDS is Linux. Problem: The sm_rexec on windows passes enclosing quotes (') to the client: exportfs '*:/..'. Solution: Skip quotes from exporfs command BUGFIX: (#22848) VMware attach does not work when server is Windows and RDS is Linux. Problem: The gv_rw_tmp is set on server. Solution: Use path returned from command 'sm_client ... Dircreate ...' bugfix: (#22843) Restore via LOAD command failed for SAP ASE: 'Backup Server: 4.10.2.1: Label validation error: first label not VOL1.'. Problem: The tape={label} does not return the tape ID. Solution: If tape ID is missing then generate from tape label change: Improved log file handling bugfix: (#22843) Restore via LOAD command failed for SAP ASE: 'Backup Server: 4.10.2.1: Label validation error: first label not VOL1.'. Problem: The restore requires 'drive=SMS:{sms_device}'. Solution: Return SMS device instead of 'drive={drive_num}' bugfix: (#22843) Restore via LOAD command failed for SAP ASE: 'Backup Server: 4.10.2.1: Label validation error: first label not VOL1.'. Problem: The restore for a given saveset must inquire the saveset information. Solution: COM Inquire added change: Correct command 'LOAD DATABASE ... FROM' bugfix: insert the component for the gui\\splash directory into the list of unneeded files for the client package. change: (#17909) Backup library for Sybase ASE. Problem: Restore data with syb_read() does not stop when BSAGetData() returns BSA_RC_ABORT_SYSTEM. Solution: Logic improved change: (#17909) Backup library for Sybase ASE. Problem: Startup logs into 'C:\Windows\system32\saes_...log'. Solution: Startup now logs into file set with environment 'SSA_LOGFILE' and trace level is set with 'LOGLEVEL', default: 0 - (log errors) CHANGE: sbc_rhev: allow setting of cluster name via option (#22844) BUGFIX: (#22848) VMware attach does not work when server is Windows and RDS is Linux. Problem: Windows CRLF is not fully stripped off. Solution: Use macro REMOVE_LF() change: Do not build Domino server backup library into 'bin/sesam' directory, as SLES11 and RHEL6 don't support RPM dependency exclusion wip: (#20598) Implementation of incremental restore change: Refactoring of 'ANALYSE_RESTORE_LOG()' function, to make it easier to change log analyse change: 'UPDATE_LOADER_CONTENS()': Assume internal loader drive number, if set to <= 0, as internal number can be 0, whereas Sesam drive numbers starting with 1 NEW: (#18106) Build Sesam Domino Server backup library during default build and add it to RPM files. BUGFIX: (#19092) Change the behavior of the msi during major upgrade from 'Complete uninstall old setup before installing new setup' to 'Install setup then remove unneeded files'. This may solve the problem of remote major upgrade failing with error 1601. Some behavior of the sesam specific update procedure had to be adjusted. BUGFIX: (#22840) vSphere restore of incremental backups: CBT data has been written into wrong VMDK leading to unusable restored VM. Revert "bugfix: (#19632) cant restore backup on "offline" vSphere disk in "san" mode. Solution: enabled testing of write protect for CBT mode before restore" BUGFIX: (#22743) Fix crash during dowload of vSphere configuration files. Extend buffer to avoid overflow BUGFIX: (#21646) Update 'DB:loader_contents' for all drives in loader during archive adjustments new: di ddl analyser - log level 3 added, which allows to view referenced chunks by Refs chunk BUGFIX: (#22793) Si3 deduplication store: Do not start FSCK, if a GC request is pending. If start of GC fails, retry it during next Si3 specific status check loop bugfix: ANALYSE_RESTORE_LOG(): Do not strip last character from error message change: LISv2 controlled by variable gv_use_lis_v2 in stpd http BUGFIX: (#22827) Drive configuration for one drive: Correct SQL statement to get drive properties. This has not affected normal drive configuration on RDS or Sesam server base BUGFIX: (#22828) Saveset replication: Fix exception during 'DB:migration_results' update, if some values couldn't be fetched from replication log file new: Si3 dedup: Return message from 'gc start' in Sesam 'STATUS=... MSG=...' format, to parse it by 'RCMD()' function BUGFIX: (#22742) LTO tape encryption: Clear encryption correctly before new tape label is written to tape. Else tape header could be encrypted bugfix: (#22813) Exchange 2016 backup fails with 'The backup source [''VSS:/Microsoft Exchange Replica Writer...'' is not available'. Reviewed: 2 messages changed. BUGFIX: (#22813) Exchange 2016 backup fails with 'The backup source [''VSS:/Microsoft Exchange Replica Writer...'' is not available'. Problem: The client name did not match with the Exchange server name. Solution: 1. Use lowercase to check backup source against browsed output. 2. Use browsed server name - do not overwrite it with retrieved Sesam client's name change: (#22722) vSphere: disabled double decoding of target disk name wip: (#20918) Restart of failed migrations Revert "BUGFIX: (#19092) Change the behavior of the msi during major upgrade from 'Complete uninstall old setup before installing new setup' to 'Install setup then remove unneeded files'. This may solve the problem of remote major upgrade failing with error 1601." change: lisv2 - offset not written correctly in some cases BUGFIX: (#22768) - Do not start Exchange DAG backup on host, which is configured, but not in same location. Empty sm_backup.exe protokol issue. BUGFIX: (#22796) Exception during call of 'sm_event notify result sesam' to sent backup information by mail after backup has been finished. Variable 'notfile' hasn't be set correctly change: Use new array length SM_MAX_CMD_LEN+SL1 for LOG(), WriteTrace() and sbc_log message change: sm_lic.ini for UCS free edition with update allowed until 20471212 change: Trace PIPE_OPEN with 'strlen(cmd) %d,' wip: (#20598) Moved NBD client logs to loglevel 3 wip: (#20598) Full backup stores base_snapshots file to times dir and incr backup reads from it. BUGFIX: (#19092) Change the behavior of the msi during major upgrade from 'Complete uninstall old setup before installing new setup' to 'Install setup then remove unneeded files'. This may solve the problem of remote major upgrade failing with error 1601. bugfix: (#22722) vSphere: swapped decoding source names on restore remove 2 empty source code lines NEW: Compare Sesam server name from Sesam DB, with name (gv_server) set in 'sm.ini'. If values are not equal set name in 'sm.ini' to value read from Sesam DB. This avoids problems caused by Sesam server sees itself as client bugfix: (#22792) sbc produces a core dump if a file has '%20d' in it's name. Problem: sbc_log to a trace file uses arbitrary string as format string BUGFIX: (#22792) sbc produces a core dump if a file has '%20d' in it's name. Problem: trace function was called with file name insteda of format specifier. Solution: Call trace function with format specifier '%s' as first argument. bugfix: check stack on empty before doing call pop() function. bugfix: includes modified due to cross-platform compilation new: sm_rexec: Restrict max. command line length to 'SM_MAX_CMD_LEN=8191', which is supported on Unix and Windows (>=XP) new: Define 'SM_MAX_CMD_LEN=8192' for smallest 'maximal command line length + 1' supported by 'all' Sesam OS >= Windows XP change: (#17464) Backup to Sesam Datastore uses padding, backup to work directory does not. Problem: By default savesets a written with padding int DS. Solution: Change default value to skip padding for DataStores NEW: Remote command: Return with error, if command line becomes to long, instead of starting trimmed command BUGFIX: (#22784) Saveset .info file does not contain complete information on Windows SEP sesam server, on Linux it does. Problem: Command was shortened in PIPE_OPEN(). Solution: Increase local buffer to MAX_CMD_LEN bugfix: Problem: Unix sm_stpd_http_conn init always fails with timeout. Solution: Show real cause of sm_stpd_http_conn start fail change: (#22722) vSphere: changed original name saving method for encoded sources bugfix: (#22778) Core dump from sm_dedup_interface fsck. Problem: Unset optional Thrift variable cause core while assign. Solution: Remove variable as it doen't used sbc_rhev: CHANGE: do not add data-stores with type iso to selection as restore allways fails if selected bugfix: (#22698) Changed the error message if the user has no admin privileges the following message will now be shown: You do not have sufficient privileges to complete this installation for all users of the machine. Log on as administrator and then retry this installation. bugfix: (#22761) the powershell scripts add_vm_ware_machines.ps1 and sm_installnfs.ps1 are no longer needed hence they are removed from the sesam build. BUGFIX: (#22768) - Do not start Exchange DAG backup on host, which is configured, but not in same location. Compilation errors/warnings are fixed, error-handling is improved. CHANGE: (#22265) EOL adjustment for failed Backup or Migration must be activated via default entries 'eol_adjust_failed_backup/migration', Default: YES Problem: Customer expects that recent savesets are EOL adjusted to prevent data loss NEW: SAP Hana backup: New key 'COM_TIMEOUT=' to set wait timeout for 'sbc_com' calls BUGFIX: (#22768) - Do not start Exchange DAG backup on host, which is configured, but not in same location. Improved source code by using same location in Exchange node detection for DAG backup. change: (#22722) Fixed curl '&' escaping when getting config file names from server BUGFIX: (#22626) Exchange backup does not start due to new flags according DAG. Problem: If mode flag is not set then BuildExSource() searches for active and passive node - but if exchange server node name does not match then backup is aborted. Solution: Start backup without checking state of mailbox database sbc_rhev: BUGFIX:(#22773) Return error in case of missing disk during backup new: (#22664) SAP Hana: Kill 'sbc_com_interface' child process, if backint receives a signal BUGFIX: (#21336) Blocking dates with user defined calender is not working. Problem: If there are a few days in the calendar_events day range, e.g. 24.12.2017 - 26.12.2017 then the second day was skipped. Solution: Improve logic, do not add 2 days in case of a date range sbc_rhev: CHANGE: exit gracefully if disk saveset not found in sesam database sbc_rhev: CHANGE: trace only loglevel >=2 change: (#22722) vsphere_config backup file access error. Modified target filenames to URI encoded names. Lowercased target filename hex codes change: (#22722) vsphere_config backup file access error. Modified target filenames to URI encoded names bugfix: (#21134) Adjusted the text in the java info dialog so the customer knows that a 64-Bit java is required and not present. BUGFIX: (#22758) Restore of a single db from ms-sql 'all' backup fails with 'SQL Server does not exist or access denied.*. Problem: The default instance 'MSSQLSERVER' must not be set in the restore target, e.g. target '/sql_srv-2012/restore_db' but logic treated the target database as instance name therefore the connect failed. Solution: Improved parsing of the restore target. Now the instance is checked and if it is not existing on the Microsoft SQL Server then it is treated as database. Note: If the restore target does not specify the database name then the original database name is retrieved from the MTF stream. bugfix: (#21134) sm_setup as well as sm_set_java are now checking if the installed java is 64-bit. If this is not the case they exit with error. sbc_rhev: NEW: allow restore of virtual machines with IDE disks sbc_rhev: CHANGE: fix regressions caused by last commit during backup BUGFIX:(#22770) allow restore of virtual machines with IDE disk, fix detection of hot plugged disksin newer rhev versions BUGFIX (#20332) - If the DAG node with all the DB tasks is not reachable, Exchange DAG cluster backup fails with 'Network communication problem: SOCKET error: 10061 - Connection refused'. Adapt sms_sms_backup tool for using another node. new: Documentation for sbc external backup bugfix: (#22617) it is now possible to insert the switch version or -version for sm_java.exe to check the java version which is used for gui and sds. bugfix: Build 'sm_ctrl' modules again, which are not build at all any more after: change: (#22618) QM and SCC executables built when calling from kernel build. Solution: corrected makefiles for building only libs needed by kernel bugfix: Link 'sm_backup' with 'sm_exchange.o' Revert "bugfix: compilation error under UNIX system "sm_backup.c:760: undefined reference to bGetExchangePCS"" bugfix: compilation error under UNIX system "sm_backup.c:760: undefined reference to bGetExchangePCS" BUGFIX: (#21084) - If one DAG node is not existing as client in the sesam DB, Exchange DAG cluster backup fails with 'Network communication problem: SOCKET error: 10061 - Connection refused'. Solution: Choose available Exchange node based on "location" approach. change: lis v2 -> #LISVER replaced with attribute at the end of first line: 'Tape ID: 1, Tape Label: lisv200001 ... LIS: 2.0' BUGFIX: (#22751) Mailbox backup for user 'mstrauß' fails and finishes with state 'info' since 4.4.3.45. Problem: Since commit b536a44 'change: Windows: Use CodePage ISO-8859-1 (28591) for wcstombsl() and mbstowcsl()' a call to mbstowcsl() without locale used UTF-8. Solution: If no locale is set then use current process locale again new: (#20598) Implemented creation of virtual disks on restore. Moved restore to sbc_post change: qm_lib - select ignores signals - signals are processed in signal handler new: (#22739) Si3/S3 sync on snapshot base. Problem: S3 snapshot operation using DI. Solution: Add CLI for S3 snapshot operations new: Python 'sbccom' module: Preset 'ssid' variable at the beginning, if set already by caller NEW: (#22452) SAP Hana: Signal handler in Sesam backup module, so operation will be terminated correctly change: (#22722) Added VSPHERE_IMG opt new: Add 'dosfstools' as dependencies for REAR new: 'Strip' output lines from SAP ASE isql before log them wip: (#22664) If 'sbc_com closejob' is called for a pending backup in queue, set this job to expired, so 'openjob' will fail with 'sm_sbc_com-W007-SBC_COM Timeout reached, drive group dg_SLES11 has no free streams anymore' wip: 'vsphere://' as backup prefix wip: XEN restore for NBD backups: Restore config files before use them to create new VM change: If GetTapeParameters(GET_TAPE_MEDIA_INFORMATION/GET_TAPE_DRIVE_INFORMATION) returned with ERROR_INVALID_FUNCTION (1) then ignore this and go on with 64KB block BUGFIX: (#22717) Media init: Sesam initializes empty media but refuses to use it. Problem: In case of drive without loader the function DOIT_MANUALLY() returns new label together label_change set to 'd' and state ST_NO_LABEL. Solution: Check for this situation and accept new label change: (#22679) Connection to HTTP interface could fail with "Initialization timeout" at high system load. Problem: From logs its not clear which stpd_http_conn and when has finished. Solution: Improve sm_stpd_main (HTTP) logs BUGFIX (#22655) - Do not truncate logs of the MSSQL backup source until the INC backup has been completed successfully. Fix compilation error. bugfix: (#22722) Fixed segfault in cm_url_encode() when input param is NULL BUGFIX (#22655) - Do not truncate logs of the MSSQL backup source until the INC backup has been completed successfully change: (#22722) vSphere: ignoring "vmx=" param if config backup only and path is given change: (#22722) Correct free memory after URI encode/decode chenge: add total amount of bytes for each s3 list operation BUGFIX: (#21816) Restore of a migrated save set split on 2 tapes failed because selected data only on next tape. Problem: Failed in case of a Zarafa restore. Solution: In case of a mail mode restore the next label check must be done in 2nd run of BLOW_UP_SLL() BUGFIX: (#22655) - Do not truncate logs of the MSSQL backup source until the INC backup has been completed successfully. Use CoInitializeSecurity() function just in the one place. BUGFIX: (#22655) - Do not truncate logs of the MSSQL backup source until the INC backup has been completed successfully. Bug in the browse operation caused by the recent changes. change: Add libsepsybase.dll to SEP sesam cli and gui kit NEW: (#22735) Add 'sm_dedup_interface status' output to 'DB:media_results.long_message' during execution of 'sm_data_store check_space' command with data base entry flag set change: DB:media_results.long_msg->long_message new: SAP ASE client: Print content of INI file by sbc log function change: (#22722) preventing double-encoded vm folder name bugfix: Since commit 292e320 'BUGFIX: (#22655) Do not truncate logs of the MSSQL backup source until the INC backup ...' the state passed to i_ADJUST_FDI_MEDIA_EOLS() is not set proper for COPY or FULL backups change: (#22722) Changed vSphere temp config directory to as-is, as it getted from cli parser change: (#22722) URL encode backup source for vSphere. Parsing VMX path from source if "vmx=" param not given new: 'DB:media_results.long_message VARCHAR(4096)' for longer messages (e.g. sm_dedup_interface status) new: Si3 deduplication store: Reduce S3 related debug logs in default log level NEW: Add some basic information about Si3/S3 connection into 'DB:data_stores' change: (#22722) URL encode backup source for vSphere. Added URI-encoded VMX path support change: (#22722) URL encode backup source for vSphere. Added "vsphere+config://" support new: (#20598) save vmdk metadata on backup, import VM metadata on restore change: Do not process [RemoveObsoleteTables] and [RemoveUiPermissions] if old_version >= 4.4.3.42 change: (#22722) URL encode backup source for vSphere. Added "vsphere://" support for windows change: (#22655) Do not truncate logs of the MSSQL backup source until the INC backup has been completed successfully. Problem: Problem with EOL chain must lead to run next DIFF/INCR as FULL. Solution: Improved logic BUGFIX: (#22732) Update failed due to import error with wrong fdi_type in results entry of a replication task. Problem: A '^Z' byte lead to premature end of file during read. Solution: Use the sqlite2 -init {filename} to avoid end of file after '^Z' byte. BUGFIX: (#22732) Update failed due to import error with wrong fdi_type in results entry of a replication task. Problem: FDI_TYPE not set for non backup task events. Solution: Set default FDI_TYPE 'C' wip: URL encode backup source for vSphere BUGFIX: (#22655) Do not truncate logs of the MSSQL backup source until the INC backup has been completed successfully. Problem: For some backup types a failed DIFF or INCR must not enforce FULL backup instead of DIFF/INCR. Solution: Function GET_BACKUP_TYPE_ENFORCE_TASK_STS() will return required task_sts 'F' for DIFF/INCR, 'D' for INCR and ' ' if not required new: Enforce DIFF instead of INCR with tasks.task_sts set to 'D' wip: 'vsphere://' as backup prefix change: (#22722) URL encode backup source for vSphere. Added support safe character set to cm_url_encode() BUGFIX: (#22655) - Do not truncate logs of the MSSQL backup source until the INC backup has been completed successfully. Handle errors in the util function correctly and some improvements in the "Complete" and "Close" operations. change: Add 'ERROR'/'WARNING' message to the backup log (not file) if sm_reformat_lis returned with error or warning BUGFIX: (#22655) Do not truncate logs of the MSSQL backup source until the INC backup has been completed successfully. Problem: Backup does not enforce FULL if backup failed but did already receive some data. Solution: When DIFF/INCR backup was aborted or failed and if backup type is not a 'Path' backup and some data was transferred then enforce FULL for next DIFF/INCR change: (#22722) URL encode backup source for vSphere. Problem: when encoded URI path given for vSphere backup, we got an error. Solution: Fixed conditions, added URI decoder call when "vsphere://" used change: (#15899) URI/URL encode backup source and restore target. Got and error when trying to backup vmware source. Solution: fixed isCbt parser BUGFIX: (#22720) Follow up event fails if command is to long. Problem: The buffer was restricted to 512 characters. Solution: Increase to 2048 characters BUGFIX: (#22706) Migration fails, if tape server host name contains part of backup server name. Instead of string search, check over DB SQL statement CHANGE: Use 'http' interface as default for new Sesam installations change: (#22677) added regexp filter to sbc (-k) kill. Usage: sbc -k -l 0 -s CHANGE: (#19872) MSSQL: Full backup now issues warning if database in simple recovery model. Problem: The system database 'master' cannot be set into 'full revocery model' and for the database 'model' Microsoft recommends only full back ups. Solution: Skip warnings for the system databases 'master', 'model' and 'msdb' new: (#20598) Added/changed logs and renamed base snapshots file to *.l.cbt new: Citrix Xen backup: Set backup level for VMDK backups to level of main task change: Set first character of saveset id to 'V' for XEN VMDK backups change: (#22618) QM and SCC executables built when calling from kernel build. Solution: corrected makefiles for building only libs needed by kernel wip: Browsing of SAP ASE on Linux new: (#20598) Functionality for incremental backups change: Improved error message handling - get error line instead of 'Msg ..., Level ..., State ...' change: qm_lib be more robust when setting timeout for submit_sync bugfix: If backup failed then there is no load command output file BUGFIX: (#22655) - Do not truncate logs of the MSSQL backup source until the INC backup has been completed successfully. "Complete pending" class implemented with some refactoring of the current code. new: (#20598) Fixed SSL certification on connecting to NBD NEW: (#22695) vSphere single item recovery: Loop all available savesets to find one on a data store, which can be used to mount/attach it directly to VM BUGFIX: (#22707) 'sm_java' was missing in Sesam GUI kit, so GUI couldn't be started by 'sesam_gui' script any more. Fix: Copy 'sm_java_cli' as 'sm_java' into Sesam 'bin' directory BUGFIX: (#22655) - Do not truncate logs of the MSSQL backup source until the INC backup has been completed successfully. Refactoring sbcmlsq module. new: (#20598) NBD client changes, Incremental backup, code refactory and error handling bugfix: Windows sbc_com_interface core dumps during vtsbc_log() since commit b7d02f7f. Problem: Different sizes for I_MAXPATH 260 <> 4096. Solution: In case of Windows use #define I_MAXPATH MAXPATHLEN (=4096) BUGFIX: (#22704) Follow up event for saveset replication returns with 'I011-COPY All savesets have already been copied'. Problem: The check if repliation for that saveset is already running did fail due to 'active' state entry in table migration_results. Solution: Ignore own migration_id for replication during check (similar to migration) BUGFIX: (#22701) If external backup job copies the segment file before newday but calls GET_BLOCKS_FROM_SGM after newday then the segment is set to -1. Problem: The gv_daylbl is used after newday but segment file was written the day before. Solution: Do not use gv_daylbl to generate the segment file name -> use the sResults.szLisName instead new: bugfix: (#21109) Backup front end client for SAP ASE. Problem: The restore command should be printed into a temporary file. Solution: Write 'load database' command into file gv_rw_tmp:isql_load_sql-{task}-{level}.txt change: Windows: Delete sm_prof.ps1.?.?.?.* older than 14 days change: bugfix: In 4.4.3.45 sm_prof.bat and all it's backups are deleted 14 days after sesam update. Problem: Windows: Some pathes are given with delimier '/' instead of '\'. Solution: Use szPath2Unix() to work only with '/' delimiter bugfix: (#22590) sm_shutdown -f doesn't stop sm_main.exe and sm_qm.exe. Problem: The sm_qm_main process is stopped but a concurrent sm_qm.exe process is connected to qm_server pipe. Solution: Check if sm_qm_main is running before connecting to qm_server pipe (fetched from commit ca43773) bugfix: (#22697) sbc cores since installation of V4.4.3.46 during VMware backup. Problem: Since commit 02b41f7 'new: (#22515, #22677) extended -D output. Partial windows functionality added, sbc_kill kill_showlist saveset name added for process list ..' the vixdisklib_load.c LoadLibrary as buggy. Solution: Do not declare '(void *)VDDKLibraryFullPathName' bugfix: (#22696) Kopano backup fails due to sbc not able to backup ACL of files with large path name (length>1024). Problem: I_MAXPATH was set to 1024. Solution: Define I_MAXPATH with PATH_MAX or MAXPATHLEN (4096) and increase define I_MAXLINE to I_MAXPATH+128 change: sbc-2101/-2106 with extra argument 'Cannot get/set %sACL for '%s': %s.' to display strerror(errno) bugfix: Since commit 4d3d1c9 (#20072) STPD communication ends with misleading error '425 Can t open data connection. WINSOCK: Address family not supported by protocol family ...' the stpd_main log shows [WINSOCK: Successful WSAStartup not yet performed. (0x276d,10093)]. Solution: Call WSAStartup() before socket() BUGFIX: (#22655) - Do not truncate logs of the MSSQL backup source until the INC backup has been completed successfully. First draft change: (#22693) Backup log encrypted via blowfish includes thousands of "sbc_encrypt: md5 BF: (32)". Solution: changed level of traced strings new: Citrix Xen backup: Enable NBD on all network devices NEW: (#22677) added regexp filter to sbc (-k) kill. Usage: sbc -k -l 0 -s new: (#20598) Save VM configuration file on full backup. CHANGE: Do not protect 'checkspace' call for data stores by a semaphore, so long running 'checkspace' will not block other backups new: Si3/S3 connector: Try to get time of last S3 sync, only if S3 connector is configured change: Si3/S3 connector: Do try to get time of last S3 sync, because this causes an exception, if no S3 connection has been configured. TODO: Implement in correct way: Check, if S3 is available before calling function new: (#20598) - Implemented sbc_com backup for list of disks for VM change: Moved 'import sbc_xen' under try: ... except ImportError: BUGFIX: In 4.4.3.45 sm_prof.bat and all it's backups are deleted 14 days after sesam update. Problem: Microsoft findfirst() function returns sm_prof.bat when called with 'sm_prof.bat.?.?.?.*'. Solution: Add fnmatch to check if file is matching to given file pattern change: Replace 'Zarafa' by 'Kopano' in log messages for reformat of 'Kopano' backups NEW: Entry in 'DB:media_results' for Si3 dedup server purge action