GUI Bugfixes: 0932e6b978 change: Fix the calculation of the "saveset_cnt" without losing the information about savesets continued on another media bugfix: The "Saveset" tab shows only saveset which starts on the selected media. Query misses savesets started on another media but continued on the selected media. BUGFIX: (#30706) - GUI doesn't show check marks for media that are encryption capable or contain encrypted blocks BUGFIX: (#30695) - Creating a datastore with name 'Server-Data' fails with error "primary key ('Server-Data') is not unique" BUGFIX: (#30679) - Property "Encryption Capable" of a drive shows no information bugfix: The backup source browser allows to select the virtualization service root nodes bugfix: (#30577) - Items from main selection can be opened multiple times bugfix: "Manage VM Tasks" action not available for Citrix XenServer clients bugfix: Resized run command dialog to fit all text fields bugfix: (#30648) - Modifying drive number of existing restore task is ignored new: (#30641) - Show number of SMS channels of drives as tooltip in migration task change: (#30639) - Disable 'Show Path' in backup source selection for clients with PROXY access bugfix: (#30627) - Since the new LOCK rule (via result_lbls) it is no longer possible to lock but also not to un_lock a single saveset change: (#30640) - Change data store BLOB tab to 'Storage Backend' bugfix: Refresh button in client file view not visible NEW: (#30546) - Add DB:task_group.options field NEW: (#30549) - Add DB:task_events.options field BUGFIX: (#30624) - vSphere host only visible in flat view NEW: (#30625) - Add medienpool colum to tab 'Info 1' in the client properties BUGFIX: (#30626) - In the properties of a run migration, a savesetID is displayed in the MEDIENPOOL column, not a media pool change: Fix handling of add/remove permission to/from ACL change: When building the string representation of the ACL value property, sort the list of users before printing in order to get a reproducible and comparable output. change: Adjust .gitignore for IDEA configuration files change: Rework addPermission/removePermission of AclsServiceImpl change: Rework creation of new ACL objects in AclsServiceImpl. Take default ACL into account. NEW: (#30523) - Add possibility to enter DB:*_events.user_comment bugfix: (#30610) - CI test for GUI authentication fails change: (#30602) - 30602 - Change option for cleanup data store action to create DB:media_results entry change: Add workaround for 29754 (Action 'cleanup' of a datastore does not work correctly when started via GUI on a SEP sesam server running on Windows) change: When the data store clean up action is called from the UI, no entry in DB:media_results is created. change: Fix SQL exception getting the next free ID of DB:notifications when accepting the EULA BUGFIX: (#30584) - sm_cmd send mail doesn't support -m gv_dayfile: parameter BUGFIX: (#30419) - ACL settings for GUI user must be adapted for each object change: Rework Ut8Utils to unify converting foreign character sets to UTF-8. BUGFIX: (#30545) - Checkbox 'Start ... in spite of ... error' is not correctly initialized if pre or post of restore task is enabled BUGFIX: (#30578) - Customer gets an error when he selects the lowest menu item in 'Manage VM orders' and confirms with ACCEPT new: Add "getItems" method to SepComboBoxNew change: Update project dependencies to rule out vulnerable libraries BUGFIX: (#30568) - vSphere restore ends with exception "Ungültiges Byte 1 von 1-Byte-UTF-8-Sequenz." if .conf file contains umlauts and Server is on Windows BUGFIX: (#30548) - Command events are executed with wrong user when filtering commands new: (#30593) - API notifications/find: filter for "module" should support multi selection (array) BUGFIX: (#30581) - Changing EOL via GUI at 'Job State > Backups' does not work anymore BUGFIX: (#30583) - Email attachments are sent twice bugfix - user table in permission dialog is now updated when group of user has changed bugfix: (#30585) - vSphere sandbox restore fails with NPE change: Switch more generic long DAO mapper to determine the next free ID by SQL script. BUGFIX: (30584) - sm_cmd send mail doesn't support -m gv_dayfile: parameter BUGFIX: (#30577) - Items from main selection can be opened multiple times CHANGE: (#30570) - Quiescing the file system of a VM causes multiple log entries for the same error CHANGE: (#30536) - Simpler creation of SiS datastore change: (#30565) - SESAM_BACKUP causes smtp mail error "Error executing 'send mail sesam -M" change: Enable reset client version info action by default in expert mode. change: Fix labeling of "SEP sesam Immutable Storage" BUGFIX: (#30531) - NPE when open calendar elements BUGFIX: (#30529) - Spring4shell vulnerability BUGFIX: (#30442) - An exception occurs in immediate start dialog if the used media pool was deleted shortly NEW: (#30415) - Medium initialisation event doesn't show a protocol BUGFIX: (#30412) - It is not possible to expire media _in_ the media-properties, but it is possible with a right-click to the media BUGFIX: (#30512) - Changing channels of a datastore drive shows the error '!DriveDialog.Message_driveNoAlreadyInUse_EnterDifferentNumber!' BUGFIX: (#30514) - A & in a password for a Sesam user makes it impossible to log in to the WebUI or RestoreWizard BUGFIX: (#30405) - NPE and wrong behaviour when changing email settings and try the test-send-button BUGFIX: (#30404) - NPE when using wrong password or username in the email settings NEW: (#30506) - UI should visualize the 'Community' edition NEW: Support certificate based authentication with LDAP/AD authentication source. BUGFIX: (#30452) - 30452 - Restore with Exchange Recovery Pro fails with "WARNING: EDB or LOG path not found in LIS file" bugfix: (#30485) - GUI reports 'The requested data is not available' for certain savesets change: Remote note on AD external authentication source settings panel. new: Enable to set the manager DN also for AD external authentication sources. The manager DN is needed to enabled advanced features like automatic user synchronization or certificate based authentication. bugfix: Fix NPE when drive number of selected backup or saveset cannot be determined for "remove_saveset" function. new: Add README.MD to project root to describe the setup of the development environment. change: Rework user authentication to search existing user by name, origin and user principal name. This will allow in future to support a "Same username, different domain" use case. NEW: (#30398) - GUI login via AD authentication with nested groups fails change: (#30457) - Add user@host to sm_break -u parameter change: Print the local system file encoding information in the "JAVA information" section when the VADP starts. new: Added a few Intellij SSH configurations change: Fix copyrights and spotbugs warnings new: Added Intellij data source to connect with local host Sesam PostgreSQL DB Bug: (#29675) - Rework the "sm_cmd migrate ..." command to use the REST API V2 bugfix: (#30396) - "backups/findTasks" returns the task in wrong order when the filter property "taskGroup" is set bugfix: (#30407) - Add a link to the backup task group to the backup properties dialog change: Move NestedLdapAuthoritiesPopulator to separate package to protect copyright header of file change: Adjust SEP copyright profile settings change: Define correct Intellij scope for applying the SEP copyright change: (#30394) - Add a link to the backup task group to the backup properties dialog change: Up-rev spring security to version 5.6.2 change: (#30188) - V2 API: Get rid of 'DataStoresDto' object new: (#30398) - GUI login via AD authentication with nested groups fails change: Create PROXY and VIRTUAL clients as "accessible" by default. change: Slightly change the size of the "new authentication source" dialog new: Added REST server localhost:5005 run configuration change: External backups not shown in Web UI "Backups" widget because of "skipChildren=true" filter setting. Also fix result paging when ACLs are enabled. change: Fix certificate based CLI authentication does not work on Windows. new: Add CLI "start task" run configuration change: Allow to call "sm_cmd send mail ..." without the "account" parameter. CHANGE: (#30391) - Change window header for 'Email Settings' change: Adjust query to determine backup types for clients. Now include also the backup types for virtual clients. change: Adjust XML code style settings change: When evaluating the ACLs of a task connected to a virtualization server client and a virtual client, take the ACL from the virtual client before taking it from the virtualization server. change: "sm_cmd check acl ..." prints wrong parent object when the user is allowed to see a virtual client only and not the corresponding vCenter (for VMware vSphere backup task). change: Adjust Intellij Markdown auto format settings change: "sm_cmd check acl ..." prints wrong parent object when the user is allowed to see a virtual client only and not the corresponding vCenter (for VMware vSphere backup task). change: Synchronize Intellij run configurations bugfix: error message when no ID is provided in cli new: (#30376) - For restores/find call restoreTask need to support multiselection change: Fix result filter property "lastBackupState" and "filterSavesets" are mutual exclusive change: (#30375) - Optimize info message if client can not be updated because Server is higher version new: (#30357) - Problems after VADP restore due to missing properties in .vmx file change: (#29677) - V2 API: Rework CLI *Params classes to use V2 API - TaskGroup, TaskGroupRelations BUGFIX: (#30172) - If the checkbox "Client is a VM" is ticked for a V-Center, no new VM Tasks can be added. change: Fix RSS feed "equal" operator change: Adjust label "Column.Stored" change: Add backup task name to migration log name for sub migration results. bugfix: (29677) - "sm_cmd start task" did not work change: (29677) - V2 API: Rework CLI *Params classes to use V2 API - TaskGroup, TaskGroupRelations change: Adjust data store type sorting in data store properties dialog change: (#30329) - sbc prints stack trace if datacenter of vSphere backup job does not exist change: Adjust the data store size check to distinguish between "capacity < min capacity" and "disk free < capacity". change: Fix exception when creating the data store failed. change: Reports can now configure a default date column to overwrite the global default "sesam_date". If not specified from outside via "-D unusable drives change: Adjust XML code style settings change: Use "try-with-resource" when reading the output of an executed command change: (#30189) - Permission/user management: do not allow creating new user "root" change: (#30142) - sm_cmd does not check syntax of datastore path new: (#30185) - Set and modify DB:data_store.uuid by CLI BUGFIX: (#30173) - VMDK's can't be browsed if client is a standalone ESX host new: Prepare to add vsphere 7 ESX host agent specific tests BUGFIX: (#30166) - vSphere VM's containing Umlauts can't be backed up in Jaglion change: (#30146) - Layout and labeling of S3 blob access dialog new: Add more SEP sesam client run configuraitons bugfix: Selected interface lost when changing the media pool in a task event. change: Fix handling of different BLOB store types in the BLOB store credentials panel. new: Add "Si Appliance" credential type change: (#30146) - Layout and labeling of S3 blob access dialog BUGFIX: (#30127) - Wrong media pool is displayed when I try a migration from a saveset what is deleted and only the migration exists BUGFIX: (#30129) - In Version 5.0.0.3 it is not possible to migrate an already copied saveset via context menu change: Fix build error after cherry-pick from release branch. change: Mark "getFirst" endpoint of "ParamsFixDao" as cyclic, as the UI is calling this endpoint every 5 seconds. bugfix: (#30135) - Protocol of parent migration job is sometimes hidden new: Add "notScheduled" task group filter property and improved SQL statement for "notScheduled" task filter property. new: Add "withoutTasks" client filter property. change: Allow "executeSql" to be invoked for "SELECT" statements by any user. Check for "SUPERUSER" permissions when a modification statement is detected. change: Fix some NPEs in license service when collecting L2/L3 license data. change: Resolve variables passed in via environment variables or system properties. Enable Intellij run configurations to autoconfigure themselves to the corresponding development branch. bugfix: (#30141) - sm_cmd does not prevent changing datastore type new: Allow using variables in UI and REST server command line parameters. change: Adapt project environment settings to allow upper/lowercase transformations. change: Adjust code style in PrintSystemProperties change: Adapt PrintSystemProperties example to look at StringSubstitutor new: Move code examples from 'gui-misc' project to 'gui' project in module 'sep-examples'. new: Added 'ProjectEnv' Intellij plug-in new: Add AclObjectUtil unit tests new: Add unit test to make sure that a AclUser cannot be added twice to an ACL. change: Update REST API documentation to refer to Direwolf bugfix: (#30111) - Error in API Documentation: 3.14. Data Store Service bugfix: (#30078) - The behavior when selecting VIEW (top right in the GUI) is wrong bugfix: (#30053) - When download of log file is not possible no error is shown by GUI bugfix: (#30109) - Exception occurs when setting encryption keys longer than 28 chars for a task new: Added new Sesam Client launch configuration (panoramix) bugfix: (#30114) - Wrong backup source selected for KVM backup: backup fails with "not enough values to unpack" bugfix: (#30113) - Archive Adjustment with virtual Loader does not allow to change the drive group change: (#29996) - The interface set in the drive group is not used (if no interface is set in the backup) change: Return correct type for "next event" API endpoint when event is a replication. change: Handle "Replication" type correctly new: (#30092) - Add DB:hw_drives.identifier to drive properties change: Upgrade H2 database to version 2.0.206 bugfix: (#30096) - adding taskgrouprelations without tasks is returning SUCCESS without adding anything change: Update sample file list for "PostgreSQL" backup type BUGFIX: Login with user 'administrator' fails on Windows platform and SQLite DB. SQLite DB is configured to work with case insensitive "LIKE", changed SQL statement to use '='. change: (#29989) - Parse of "client dir" output to JSON format change: Allow linking AZURE credentials with Si3 NG data stores via CLI bugfix: Swapped methods attached to view selection toolbar button action listeners after refactoring. bugfix: (#29758) - api/v2/statistics/find shows more results than possible change: Hide the "Clone State" column by default in the data store views. new: (#30025) - Add support for Si3-NG Azure as storage backend to GUI and kernel new: Add Intellij remote debug launch configuration template bugfix: (#30030) - When applying new license via GUI a strange behavior with file filter occurs new: Add flag to enable Hyper-V RCT mount change: (#30051) - Gui does not remember checkbox "use custom SEP sesam package source" BUGFIX: (#30041) - Immediate start at schedules always uses priority 0 bugfix: (#30044) - Wrong date is used in log messages during VADP backup bugfix: (#30007) - If 'sm_cmd modify' is given an invalid access mode, the access mode is set to NULL bugfix: (#29973) - sm_cmd start restore restoretask modifies DB:restore_tasks change: Avoid querying the DB:params_fix table again and again. Use the cache instead. change: Replace VMNetworkDto with newer object DefaultBufferNetworkSummaryObject change: Clean out dependencies of "sm_ui.jar" change: Fix German translation of "Button.ResetVersionInfo" change: Remove fragments of VADP from sm_ui.jar. Stop building not necessary "sm_server.jar" and "sm_gui.jar". change: Do not print stacktrace when exception message contain "Exit code from". change: Update copy scripts to place sbc_vadp.jar into new location (.../bin/sesam). new: Introduce possibility to choose between refresh modes FULL and UPDATE when triggering the update of a tree/table dockable center panel. This is needed to rebase the "Job Status" views to the tree/table framework eventually. change: Fix Proxmox-VE host system display in client browser. change: Always print VM exceptions with stack trace if an NPE is detected. new: (#29974) - Add snapshot information to "dir" command output change: Fine tune query of virtual machine current snapshot bugfix: For long running VADP backups/restores, the created Sesam snapshot is not removed anymore, because VADP reports "snapshot not found" NEW: (#29921) - Add the information from gv_server_not_available in the GUI CHANGE: (#29966) - An easier way to see the correct update history for the server at the about windows new: Copy NVRAM from/to data mover when NVRAM file exist. NEW: (#20629) - Enforce FULL backup, if latest FULL has failed change: Copyright Header in Modeltests and some mode.type tests bugfix: (#29950) - NPE during restore process if executables such as sbc_proxy can't be executed. change: After upgrading the H2 database to 2.0.202, several SQL syntax exceptions popped up when creating the test database, as the new version has much stricter syntax rules applied as the previous versions. Fix the syntax errors when generating the test database initialization scripts. change: Add SepDate sorting tests. change: Switch back to direct HTML reporting for Dependency Check plug-in change: Upgrade a set of project dependencies to newer versions. change: Upgrade H2 database dependency (for unit tests) to version 2.0.200. change: Switch check-dependency report to XML output format. change: (#29942) - Progress bar does not close if restore job is removed bugfix: Fix reference to "r.saveset_pool" in backup storage report bugfix: "Super user" group is listed in "Groups" combobox when creating a new user. new: Explicitly add required "maven-site-plugin" and "maven-project-info-reports-plugin" to Maven "reporting" section. change: Fix type in permissions declaration of "notifications/count" API new: Add "dependency-check-maven" plug-in to maven to generate reports for dependencies, versions and security issues. change: Avoid checking for applicable ACLs twice. Drop the check in "checkPermissions" to improve performance. change: When a user is forced for "sm_cmd check acl", clone the session context with the new user and pass in the cloned session context to the ACL manager. Do not use thread local storage of forced user/groups in the ACL manager. change: Pass in the session context to the ACL manager from the caller. Do not try to determine the session context from inside the ACL manager itself. All information needed to evaluate the access to a given item or the filtering of a list of items has to be passed in. change: Disable debug output of ACL handling and result filtering change: Fallback to original implementation of AbstractAclEnabledDao.isBypassAcl change: Fallback to default security context holder strategy. change: Fix handling of bypassing the ACLs for the local thread context. In some place, bypass had been set for current thread but not revoked properly afterwards. change: Rework ACL filtering to correctly evaluate applicable ACLs before processing large lists and reduce the number of database access operations during filtering. change: (#29938) - Interactive GUI update does not work anymore when having Windows GUI client change: Disable obsolete parsing the saveset ID from SBC output. change: Create output print stream with UTF-8 charset instead of CP850. bugfix: Temporary configuration directories not removed from server or data mover at the end of backup or restore operations. change: Improve condition to avoid potential NPE change: Improve matching of CD-ROM and/or DVD section to also allow to use a regular expression when matching the node name. change: Improve message when downloaded file cannot be renamed on UI update. change: Disable loading of output template for license service to make the build work again. BUGFIX: (#29868) - Report "Client Backup Summary" shows wrong / incomplete information change: (#29920) - Encryption/compression tab is not shown for existing Oracle tasks change: Treat the unit test H2 database like PostgreSQL. Turn on all PostgreSQL feature when running the unit tests. change: (#29923) - Fails to install on windows: SQL error or missing database (RIGHT and FULL OUTER JOINs are not currently supported) change: (#29911) - Revise usage of -S option for taskevent object bugfix: (#29909) - List of savesets for tape medium shows savesets starting on it only change: Fix layout of "Expire media" dialogs bugfix: User fails to authenticate if password and certificate is provided and password is valid but certificate not. change: (#29880) - Sandbox restore failed with ".. has no SCSI/AHCI controller configured .." bugfix: (#29823) - Backup-Source-Browser content is confused - not at the same place bugfix: (#29888) - After GUI server is reachable again an Operation Not Possible Exception may occurs change: (#29830) - GUI stuck for no reason during usage since updating to version 5001 bugfix: (#29876) - sm_cmd start migration does not accept -G 1 group flag parameter change: (#29877) - Browsing Informix shows the instance as child elements under the instance bugfix: (#29857) - OperationNotPossibleException with invalid state may occurs when attaching VM saveset bugfix: (#29849) - sm_cmd restore -t parameter does not work change: Allow "sm_cmd refresh vsphere" to be called without parameter to refresh all servers. NEW: (#29856) - Support of LIS v2 in restore wizard change: (#29789) - Enable filter functionality in every section of the 'Manage VM tasks' window bugfix: (#29851) - Enter key does not work on interface nor on note windows at clients bugfix: (#29790) - Filter 'VM name does not contain' in 'Manage VM tasks' window has no effect bugfix: (#29794) - Dialog 'Create VM tasks' does not adapt settings anymore for task change: Use StringUtils for string operations change: Adapt license params tests to use common infrastructure from TestUtils to load test data. bugfix: (#29766) - Migration event dialog must not query for all savesets when specific saveset is chosen new: (#29782) - Add 'cancel' to migration result context menu change: (#17922) - Include list/pattern for backup change: [Maven] SpotBugs still not working when called from Maven. change: Determine mount type from mount session, if not provided by DTO when calling "cancel". change: Add more information to "Unable to connect ..." errors change: (#29832) - Exchange Recovery Pro restore via WebUI is marked as canceled change: [Maven] Re-add Java SpotBugs plug-in. change: Rename SpotBugs configuration file change: (#29831) - No restore log is created when SFR is done via WebUI change: Pass on the RDS mount client to the REST API V1 mount service when mounting an Exchange backup. bugfix: (#29808) - Wrong preferred media pool is displayed, when reopening a saved restore task bugfix: (#29818) - "Source List Editor" is empty change: (#29819) - vSphere browse fails after freshly started RMI service change: VIJava version of the VMware vSphere web services service instance is not connected anymore and therefore the backup/restore is not possible. bugfix: (#29814) - Browsing Vsphere fails: java.lang.NoClassDefFoundError change: [Maven] Adapt the copy rule for JAR artefacts to sep-launcher/target/updates. new: (#29813) - Accept names too for "sm_cmd add acl -i" change: Improve remote buffer unit test performance. change: Fix handling of parent resource pool for remote buffer objects change: Try to get the detailed stacktraces for the exception happening during the tests change: Fix association of parent resource pool to remote buffer summary objects. change: Try to get the detailed stacktraces for the exception happening during the tests change: [Maven] Clean up overlapping dependencies change: Try to get the stacktrace for the exception happening during the tests change: Trigger the fork join common pool initialization explicitly but do not call "System.setSecurity" manager. Method is marked as to be removed in future release. change: Trigger the fork join common pool initialization explicitly with no security manager set. change: Add "childEntity" to list of folder properties to query change: Reset the security manager for sure when the applications starts change: Reset the system wide security manager set by the Rythm template engine change: Fix handling of "hostFolder" property for data stores. change: Fix a hidden NPE in VSphereAbstractService change: [Maven] Adjust XML SOAP framework dependencies. Avoid duplicated classes in JAR file. CHANGE: (#29780) - Restore users are unable to see any savesets in Jaglion, because the default ACL for all media pools is set to deny change: Handle NumberFormatException in ContextLogger more gracefully change: Fix StackOverflowError in AclsDaoImpl.filter when "formatCandidate" is called bugfix: (#29792) - Add maxResults, offset, orderBy and asc to notifications/find filter change: Make sure the full list of configured virtualization hosts is used in the remote buffer. change: Fix imports in REST server classes to use correct VMware vSphere web service dependencies. change: Source side deduplication checkbox is not enabled because the installed version is empty. maven: Fallback to jaxws-rt 2.3.1 to keep Java 8 compatibility CHANGE: Integrate new remote buffer implementation, based on official VMware vSphere web services API, into master branch. new: (#29734) - Allow to add a new user to an existing group right away when creating the user bugfix: (#29750) - When clicking on the "Test" button of the "Authentification Configuration" without entered text a NullPointerException occurs change: (#29773) - REST API: change type of notification.id from Long to String change: (#29742) - Prepare Java 17 support for all Java based components bugfix: (#29771) - Taking a file path filter from the list in expert options does not work change: Certificate path is set only when no user is specified at command line bugfix: Certificate based authentication does not work anymore for user 'sesam'. bugfix: (#29765) - Data Mover selection at migration event must involved systems only change: (#29742) - Prepare Java 17 support for all Java based components change: (#28851) - Change delete saveset dialog to use remove_saveset -s change: Enable 'Cancel' toolbar action also available when the task / task group state is "in queue". CHANGE: When a user is specified on CLI command line (sm_cmd -U ...), this should have priority over the certificate. Previous implementation ignored -U/-P when the certificate is found and readable. change: Make 'Cancel' context menu action also available when the task / task group state is "in queue". change: Fix sorting of external strings CHANGE: Improve the presentation of the standard groups in the user and groups settings panels and dialogs. bugfix: The special group "SUPERUSER" is visible in the permission management dialogs and panels but should not. new: Added add command to AclParams change: Improve data store tooltip to include consistency state also. bugfix: "sm_remove_saveset" called for a data store action other than "purge". maven: Remove deprecated password validator package 'nbvcxz'. new: Added base project for development of the next generation VMware SBC. bugfix: Changing a value in the "Multiple Drive Dialog" does not have any effect when trying to apply the change. change: To apply a value to multiple rows, the must be selected via the "Selected" checkbox in the first table row. bugfix: ID label in CLI help was printed in wrong place change: Re-connect of UI after connection lost to REST server broken. UI does not get out of the connection lost state ever again. bugfix: (#29732) - Removing Si3 datastore via sm_cmd fails with: CLI: remove: External process failed: executeSMDatastore change: Fix "sm_cmd list acls" returns with unclear error when default ACL is present in DB:ui_acls change: (#29691) - Use 'sm_remove_saveset' instead of 'sm_data_store' for purge process SRC Bugfixes: b05104f615 bugfix: (#30803) A Python exception appears in successful vSphere backup. Problem: Introduced with commit 37a7625. Solution: Close file stream which points to NOT file bugfix: (#30801) SBC error message not parsed correctly any more. Solution: Recent fix for '#30800 Exchange Backup finishes successfully despite missing log' introduced overwriting of ANALYSE_BACKUP_LOG() result error message with COPY_REFORMAT_LIS() message bugfix: (#30800) Exchange backup finishes successfully despite missing logs. Solution: COPY_REFORMAT_LIS() state must be checked even if processed items counter is not estimated, so add processed directories and files (from LIS) to condition bugfix: (#30783) vSphere mount fails with exception: local variable 'source_prefix' referenced before assignment. Problem: Recent commit introduced problem BUGFIX: (#26965) Use the FQFN for deletion of the patch subdirectories. CHANGE: (#30735) URL encode exclude list in saveset info (sbc -i ...) BUGFIX:(#30735) sm_stpd_http_conn may crash in migration task during read from SI3NG BUGFIX: (#30760) License output truncate. Increase buffers bugfix: (#30712) Sesam can't create VSS Snapshot (IVssBackupComponentsEx3::GetWriterStatusEx)Problem: Since commit 43cfd56 GetWriterStatusEx may throw exception if writer return with retry-able error. Solution: If GetWriterStatusEx return retry-able error for writer - log it and exit bugfix: (#30765) sm_restore cores during FDI Restore if diff backup is stored on a different media pool. Solution: Use correct restore task ID from main task bugfix: (#30768) Nutanix: sbc_proxy wrong parameter - mount not possible. Fix for '#30732 SIR mounts one VD only' commit 4817c0e introduced extra ':'. Solution: Prefix must always end with ':' change: (#30707) Check for existing VFS mount for PATH mounts. Fixed return status of check for existing VFS mounts for sm_vmdk_interface change: If VFS is already mounted, returne mounted drives at once without executing a failing VFS. Implementation for Linux for sm_vmdk_interface bugfix: Fixed mount of VHD types. Backup type set for guest_mount had an additional : bugfix: (#30710) MSSQL restore with relocation via WebUI fails with 'DB API call is failed with code ''DB_NUM_E_FAILED''. Solution: Turn delimiter in relocation MOVE option to '\' with szPath2Dos() bugfix: (#28389) sm_reformat_lis does not consider the exclude of a VHD. Problem: AutoRecovery avhd(x) not excluded. Solution: AutoRecovery avhd(x) excluded if its parent vhd(x) excluded BUGFIX: (#30733) Restore of XenServer CBT backup fails if migrated saveset is used BUGFIX: (#30685) SAP ASE restore fails: Failed to open session. Problem: pclose() returned -1 with errno ECHILD 'No child processes' after the original signal mask was restored BUGFIX: (#30733) Restore of XenServer CBT backup fails if migrated saveset is used BUGFIX: (#30737): Promox backup fails Error: Unable to get information about ZFS pools via API: nodes: check if API response has nodes value set bugfix: (#30732) Fixed mount of multiple virtual disks and mount of VHD with FDI chain BUGFIX: (#30732) Fixed mount of VMs with multiple virtual disks. loop in sm_vfs guest_mount added only the last disks to the list for mount command change: Remove 'beta' tag for release build bugfix: (#30710) MSSQL restore with relocation via WebUI fails with 'DB API call is failed with code ''DB_NUM_E_FAILED''. Solution: Turn delimiter in relocation MOVE option to '\' change: Use 'ONE' as short name for OpenNebula BUFIX: (#30719) Fixed Open Nebula mount. Session source contains slashes and VFS could not map source to VHD file. Use session src split by slash for mapper CHANGE: New version 5.0.0.8 for release build bugfix: (#30724) sm_dedup_interface get doesn't work for 'Path' data stores any more. Problem: In case of EOF data write skipped. Solution: Write data and only after check for EOF bugfix: Fixed mount of multiple VSphere VMDKs bugfix: (#30717) Hyper-V single item restore via mount fails with 'Could not find platform for client use loaded tape instead of retrieved change: Trace in iReadFromHandle2() revised. Macro WITH_TIMEOUT allows select with 5 seconds timeout CHANGE: New version 5.0.0.8 for release build Fix merge conflict BUGFIX: (#30675): Guest agent not detected on windows guests: detect guest agent on windows which has different application name bugfix: Error 'Subprocess sm_sms_backup undefined broken' was not detected since commit ab57b3e "Use 'i_cm_popen()' function to execute 'sm_sms_backup'" bugfix: initdb does not work correctly if var\db folder has been created separately on Windows. initdb will now create the folder itself. bugfix: (#29530) Exchange Server Backup - Avoid possible risk of accidental Log truncation. Problem: sm_ssh stdin thread was not finished after exit of process bugfix: (#29900) Cancelling a backup while it is performing the alivetest. Solution: Windows: iReadFromHandle2() now with WaitForMultipleObjects() with timeout to return into function PIPE_READ_CMD() to recognize signal bugfix: (#29530) Exchange Server Backup - Avoid possible risk of accidental Log truncation. Problem: Use binary mode to read output, line reader may not read full outout before sending data to stdin. Abort handling improved. bugfix: Do not truncate path in sbc_ndmp dir function bugfix: (#27565) Return starting status of Si3 for all sm_dedup_interface commands. Problem: dir command hangs. Solution: Check dir response for error NEW: (#23929) Enhance security for encrypted Si3 store. Remove password entry from INI file copied into data store root diretory bugfix: (#26965) Add more log lines to figure out why folders are not deleted. BUGFIX: (#26965) Use dir /B instead of FindAllDirs function to obtain all folders matching a wildcard. bugfix: Si3-NG correct purge handling after service block BUGFIX: (#26965) Use dir /B instead of FindAllDirs function to obtain all folders matching a wildcard. Revert "wip: (#30608) Option 'wait_for_stdin' just works when server is running on Windows and SMSSH is set. Workaround: Use new approach only if GLBV gv_conf_no_wait_for_stdin is set to 'N' for Linux" bugfix: (#29530) Exchange Server Backup - Avoid possible risk of accidental Log truncation. Problem: If sbc was called with 'wait_for_stdin' then even in error case the 'STATUS=2' must be sent via stdin bugfix: (#30656) SQL error when saving cryptographic hash of the password. 'szTable' variable has probably been overwritten by longer password NEW: (#30465) Add deprecation warning for MS-SQL 'all' backup source change: (#27786) Call insert for 'DB:performance' with 'no DB error' flag, so no alert message will be issued, if insert fails. It doesn't lead to problems in Sesam in this case. TODO: Problem during creation of unique ID has to be fixed. BUGFIX: (#26965) Remove the wildcard at the end of the patchbackup folder name, maybe the function FindAllDirs does not support wildcards. BUGFIX: (#30644) Skip method WriteExecuteSequenceInfo in case of uninstall. This method writes several information about installation, update, reinstall in a tmp file. In addition write in the log file if the folder SEPsesam\var\tmp has been deleted successfully. BUGFIX: (#26965) Added several log message to find out why the patch folders are not being deleted. BUGFIX: (#30642) Insert not only update message and state from sm_setup in sm.ini but also information about msi. This information can be helpful e.g. if a system restart is required. Update message was only set if the field update_message already existed in the sm.ini. Now the field is being inserted newly if it does not exist yet. new: (#30553) - Block certain actions on Si3 store bugfix: (#30608) Option 'wait_for_stdin' does not works when server is running on Linux. Solution: Read output logic moved from sm_sms_backup.c back to sm_lib.c to new function PIPE_READ_CMD() BUGFIX: (#27786) Concurrent processes created same unique ID. Solution: Windows: Logic in semaphor processing improved - avoid access in case of new process with same PID new: Si3-NG stop/start multiply services with single DI call bugfix: (#27786) Concurrent processes created same unique ID. Solution: Windows: Use function ftime_s() instead of ftime(). Additional: 'sm_info get_unique_id S {count}' checks if ID is unique new: Allow configuration of Sesam Immutable Store as default BUGFIX: (#30466) Fixed mount of Opennebula backup. Check if file exists before copy map file to mount client host bugfix: add time delays to fix DiskObjectsSpec test bugfix: fix requirements for JProcess BUGFIX: flush page-counter right after creation new: Si3-NG selective PCCK/OCCK mode added CHANGE: (#30072) Start full FSCK for Si3-NG every 3 days as default bugfix: (#30175) Cleaning cartridge property is not correctly named in sm_config_drives log bugfix: (#30591) Split DF, DU, getlabel, getpath, getuuid and STATUS APIs new: (#30027) - Add list file function on Si3 store NEW (#30447) Semaphore function with signal handler for Windows. Add msleep(500) to wait 1/2 second BUGFIX: (#30635) 'sm_disaster' is not executed during the migration of the Sesam backup. Use DB:result_lbls.backup_id to find correct DB:results entry change: sm_qm_main potential core dump fix due to missing table literal change: improve BDA logging BUGFIX: (#30619) sm_filter -t ends with Application Error since update to version 5.0.0. Visual Studio 2019 doe snot allow to close file descriptor obtained with function fileno() new: add source rename ability for BDA backups (sbc -b -s @nul BDA:C:/ProgramData/SEPsesam/var/ini/stpd.ini==::==myini) bugfix: (#30018) - read if only 1 byte is read and add error reporting for unexpected return code bugfix: fix DI warning new: add source rename ability for BDA backups (sbc -b -s @nul BDA:C:/ProgramData/SEPsesam/var/ini/stpd.ini==::==myini) change: Organize the key order in the DB:defaults initialization blocks a bit better to ease finding certain keys bugfix: removed blocking code when determining if server is ready new: Allow configuration of Sesam Immutable Store as default BUGFIX: (#30466) Fixed mount of Opennebula backup. Check if file exists before copy map file to mount client host bugfix: add time delays to fix DiskObjectsSpec test bugfix: fix requirements for JProcess BUGFIX: flush page-counter right after creation new: Si3-NG selective PCCK/OCCK mode added CHANGE: (#30072) Start full FSCK for Si3-NG every 3 days as default bugfix: (#30175) Cleaning cartridge property is not correctly named in sm_config_drives log bugfix: do not log credentials from config file; change pattern for PCCK logs NEW: (#30447) Semaphore function with signal handler for Windows. Add msleep(500) to wait 1/2 second BUGFIX: (#30635) 'sm_disaster' is not executed during the migration of the Sesam backup. Use DB:result_lbls.backup_id to find correct DB:results entry change: sm_qm_main potential core dump fix due to missing table literal change: improve BDA logging BUGFIX: (#30619) sm_filter -t ends with Application Error since update to version 5.0.0. Visual Studio 2019 doe snot allow to close file descriptor obtained with function fileno() new: add source rename ability for BDA backups (sbc -b -s @nul BDA:C:/ProgramData/SEPsesam/var/ini/stpd.ini==::==myini) BUGFIX: (#30155) - slu DEVICE -J short displays drive hardware encryption capabilities even if tape is not loaded bugfix: (#30632) Diff backup of master database fails with 'Mismatch database_backup_lsn'. Solution: Diff backup is executed as Full backup therefore check of LSN must be skipped bugfix: (#30609) slu cores while setting encryption key for tape drive bugfix: (#30018) - read if only 1 byte is read and add error reporting for unexpected return code new: add source rename ability for BDA backups (sbc -b -s @nul BDA:C:/ProgramData/SEPsesam/var/ini/stpd.ini==::==myini) change: Added volume uuid,name.description,size and device to Xen volumes dir print bugfix: fix local pages are not uploaded to s3 trash on gc; modify corruption logic to track missing pages using page-index bugfix: (#30621) Differencial all backup fails if system has empty volumes CHANGE: New version 5.0.0.8 for master build new: set physical and logical offsets for BDA backups new: (#27079) - Si3-NG snapshot function bugfix: (#29760) If an active task canceled, it should finish with state cancelled, not with error. Problem: Message 'sbc-1009: Error: XBSA Call BSACreateObject failed with message: Access to the requested object is not possible. STOR failed. NEGATIVE reply: 520 Operation aborted by Sesam Server. (0)' was treated as error BUGFIX (#29580) Avoid multiple update of backup status during 'break' operation. Solution: Avoid multiple update by adding clause: and state in ('q','a') BUGFIX: (#27717) uninstall of a client did not work since the password decryption failed. Now more information is logged and the password decryption works too. In addition the client name was not passed to the uninstall subprocess. bugfix: Do not increase 'counters' (like 'i++') in 'TR()' function, because execution depends on trace level BUGFIX: (#30614) EOL adjustment of saveset does not work when logging for 'sm_sesam' is disabled. Loop variable has been increased in 'TR()' function BUGFIX: sbc_netapp: catch attribute error if netapp api does not correctly report vserver ip address BUGFIX: sbc_netapp: catch Attribute error if ontap api does not correctly report size of value, default to size 0 new: (#30582) Abort of path restore leaves broken file. Problem: If a path restore is aborted the last file which is being restored will be broken. Solution: Remove item in progress, if restore has been interrupted and restore to tmp file and call move afterwards (atomic operation) BUGFIX: (#28207) sm_execute_update reports a successful update if 'rpm -U --test' was unsuccessful. Obtain the status of the zypper directly in the if statement.. BUGFIX: (#28537) Adjusted the info text as decribed in the call to clarify the procedure. CHANGE: (#30602) Create DB:media_results entry for every data store 'fsck', if not switched off by '-I none' bugfix: (#30608) Option 'wait_for_stdin' just works when server is running on Windows and SMSSH is set. Problem: Appending '2>&1' lead to error BUGFIX: (#28644) During a Linux update the field release has not been inserted as well as the field brand. The update of these fields is done differently for Windows and Linux. change: Si3NG di service start|stop gc|backup|restore|delete|purge bugfix: (#30608) Option 'wait_for_stdin' just works when server is running on Windows and SMSSH is set. Solution: In case of access mode CTRL use command sm_ctrlc.exe instead of library bugfix: (#30609) 'slu' cores while setting encryption key for tape drive. Replace 'strncpy()' by 'strlcpy()' change: (#27079) - Si3-NG snapshot function bugfix: (#30605) Generation restore fails with ''Generations restore couldn't find any savesets'' if gv_conf_fdi_chain_recursive_select is set to 'no'. Solution: Since commit ac725e9 the 'with recursive' was used in recursive calls due to default value 'true' change: Avoid compilation warning 'format-truncation' BUGFIX: (#30507) The servicepack information will now be deleted from sm.ini during the update process even if the version number is equal to the installed Sesam version. BUGFIX: (#30540) exit with 0 even though the SEP sesam BSR Pro service might not be started or stopped. If it is not installed if may not be stopped. BUGFIX: (#29754) Action 'cleanup' for a datastore does not work correctly when started via GUI on Sesam server running on Windows BUGFIX: (#29760) If an active task canceled, it should finish with state cancelled, not with error. Solution: Set state to cancelled if GET_MEDIUM was interrupted. Do not set final parent task state as long as some migrations are in state 'A'ctive (waiting for target drive) new: 'sm_prot -m ' option to set module name. Remove not used functions to translate log messages bugfix: return error in JSON format if query parameter is missing in Si3 API BUGFIX: (#30597) SAP/R3 backup fails because argument list for sbc is too long. Increase size of internal variable used to build sbc command new: (#30582) Abort of path restore leaves broken file. Problem: If a path restore is aborted the last file which is being restored will be broken. Solution: Remove item in progress, if restore has been interrupted and restore to tmp file and call move afterwards (atomic operation) BUGFIX: (#30597) SAP/R3 backup fails because argument list for sbc is too long. Increase size of internal variable used to build sbc command new: i_NOTIFICATION(): Log 2nd message argument new: Support -f option in 'sm_sesam notification' bugfix: Set correct interface name in 'sm_sesam notification' BUGFIX: (#30597) SAP/R3 backup fails because argument list for sbc is too long. Increase size of internal variable used to build sbc command new: filelists were moved to separate repository sesam_versions. BUGFIX: (#29760) If an active task canceled, it should finish with state cancelled, not with error. Solution: Migration task for target drive with state 'A' must alos be set to cancelled BUGFIX: (#30588) Verify of windows ALL generation backup fails with 'Error while unpacking data. Bad archive'. Problem: Backup with -T {since} tried to process file 'Boot\BCD' twice. Solution: Avoid processing similar file again, print message 'sbc-3254: Info: Current item [...]' bugfix: (#30589) Generational Restore is set to successful even if verify fails. Problem: SBC status do not changed to ERROR in case of MTF_ERROR_BAD_ARCHIVE. Solution: Set SBC status to ERROR if error detected bugfix: (#30590) Core dump of sm_dedup_interface at df command change: for SP use change report created by set_tag and adjust it. bugfix: (#30495) call sm_main start qm db correctly using new parameter -i. bugfix: (#30574) EOL of all vSphere child savesets must be adjusted correctly. Solution: Recursive result set must be ordered by session and child saveset. Note: Since commit ac725e9 the 'with recursive' select statement is used to get FDI_CHAIN. bugfix: (#30077) Re-write of 'sm_dedup_interface'. Add OCCK/PCCK command family new: (#30077) Re-write of 'sm_dedup_interface'. Add AzureBlob command family CHANGE: (#30072) Start FSCK for Si3-NG every 30 minutes, instead of every minute. Value can be set by 'gv_conf_dedup_fsck_loop[_]= bugfix: (#30580) Status of aborted restores changed. Solution: Add message code 'sbc-1028' to set 'aborted' flag in case of aborted SBC - until now only 'sbc-1145' for aborted STPD data transfer was checked BUGFIX: (#28939) Revert the pattern match adjustment. NEW: (#30575) Rewrite STPD ini and re-start Si3 service after encryption password has been changed new: Reset DB:hw:drives.previous_password, if all page files are not encrypted new: (#30572) Communication to HTTP(S) could take some seconds at the beginning due to DNS issues. Problem: Due to missing DNS or wrong configured DNS retrieve hostname and IP from Socket may take some time. Solution: Add new ini flag to allow skip hostname/IP retrieve from Socket change: (#27115) - Encryption of page files for Si3-NG BUGFIX: (#30518) sm_sms_backup may crash with Application Error during 'REMOTE_FILE_DELETE' if new-reader is in use. Solution: Move declaration of overlapped structure from function body to DB_PROC_PIP structure bugfix: (#30576) Remove temporary files created by 'sm_reformat_lis' for Exchange Server bugfix: (#30574) EOL of all child savesets must be adjusted correctly. Solution: Set results.based_on before calling ADJUST_EOLS() change: (#27115) - Encryption of page files for Si3-NG change: Change log level for some messages from 'sm_sms_watch' which where logged in lower level for debugging purposes BUGFIX: (#28939) The pattern matching for still active stpd did not match correctly. Also some trace messages were adjusted for a better readablity. bugfix: (#30571) Kopano backup fails with Python error in process.py. Correct check for 'os.popen.close()'. TODO: Use 'Process()' function bugfix: allow GC deleting not linked page if content is missing; fox disk page iterator to not skipping single page body file if content is missing BUGFIX: (#30550) Jobs may stuck in pending queue though drive ressource would be free. Solution: Reset allocation change: during SEP sesam uninstall change the log location to the temp folder to make error analysis possible. change: (#27115) - Encryption of page files for Si3-NG bugfix: (#30550) - qm_main should set allocated flag to true BUGFIX: (#29994) Correct check of 'DB:clients.access_state'. Do not compare string with integer bugfix: return PageMetaInfo on error in Azure change: using Netty for Azure client change: (#27115) - Encryption of page files for Si3-NG change: changed the max connection pool size to 3000 because 2048 concurrent requests can be made BUGFIX: (#30321) HCL Domino fails with 'backupItem [ITEM_NAME=\nlogctrl.lfh] packing failed'. Solution: Avoid to add nlogctrl.lfh if Circular log style is enabled and remember backup path if backup of logctrl.lfh is reasonable change: (#8554) Source code format change bugfix: (#30564) Possible data loss on Dedup_NG when timeout reached for PUT on S3. Problem: dedup_finish_backup error interpreted as non fatal. Solution: If dedup_finish_backup error happen - backup should be finished with error CHANGE: New version 5.0.0.6 for release build new: (#27115) - Encryption of page files for Si3-NG new: (#30336) Backup may fail with 'The data size of the backup data (SBC) is different from the size of the stored data (STPD)'. Adapt 'sm_sesam analyze_log -f [-I 0|1|X] [-s ]' for new logic in ANALYSE_BACKUP_LOG(): Fetch data_size of given saveset or use '-1' to indicate missing data_size new: (#30077) Re-write of 'sm_dedup_interface'. Add PCCK/OCCK command families bugfix: pcck/occk tasks must be forwarded to sm_dedup_cli new: (#27115) - Encryption of page files for Si3-NG new: (#30336) Backup may fail with 'The data size of the backup data (SBC) is different from the size of the stored data (STPD)'. Adapt 'sm_sesam analyze_log -f [-I 0|1|X] [-s ]' for new logic in ANALYSE_BACKUP_LOG(): Fetch data_size of given saveset or use '-1' to indicate missing data_size new: (#30077) Re-write of 'sm_dedup_interface'. Add PCCK/OCCK command families bugfix: (#30495) For Windows during postgreSQL installation a db check does not work. If the db check should not be executed use the newly introduced switch -i for sm_main. bugfix: (#30495) During postgreSQL installation a db check does not work. If the db check should not be executed use the newly introduced switch -i. bugfix: (#30495) For Windows during postgreSQL installation a db check does not work. If the db check should not be executed use the newly introduced switch -i for sm_main. bugfix: (#30495) During postgreSQL installation a db check does not work. If the db check should not be executed use the newly introduced switch -i. bugfix: pcck/occk tasks must be forwarded to sm_dedup_cli change: opennebula: strip newlines from reported errors, use common function new: add passHash and chunks size to mate info bugfix: (#30556) Powershell based pre and post scripts cannot be executed anymore on Windows. Solution: Call bCheckFileExtension() with bExtensionGiven to allow specific logic change: Adjust default Restore wizard defaults change: Set default to use legacy Java restore wizard in Jaglion V2. change: (#8554) Remove unused function change: opennebula browse: use latest history entry to get vm host NEW: opennebula: show host system during browse of virtual machines NEW: openebula: if more than 100 domains are found, use multithreaded approach to fetch domain information for sbc_dir NEW: (#24995) Delay start of saveset purge by 'gv_conf_wait_purge' minutes after last Sesam startup. Default is 10 minutes NEW: openebula: if more than 100 domains are found, use multithreaded approach to fetch domain information for sbc_dir change: Do not build release RPMS with patch level BUGFIX: (#30554): opennebula browse does not show all virtual machines: add --no-pager option to onevm list command on opennebula 6.2 CHANGE: New version 5.0.0.5 for master build BUGFIX: (#30323) SAP ASE backup fails with 'The specified module could not be found'. Problem: Since changes for multiple oracle stream the xbsa structures have been changed. Solution: Include 'sbc_def_common' after xbsa headers to set PATH_MAX to 260 new: (#30510) Check for NDMP host OS type is NetApp before setting snapshot management extension. Isioln for some reason reports this extension and then returns error when we try to set it to use. Fixed full volume restore, nlist destination path must be set to source filesystem, instead of / Revert "new: (#30510) Check for NDMP host OS type is NetApp before setting snapshot management extension. Isioln for some reason reports this extension and then returns error when we try to set it to use. Fixed full volume restore, nlist destination path must be set to source filesystem, instead of /" new: (#30510) Check for NDMP host OS type is NetApp before setting snapshot management extension. Isioln for some reason reports this extension and then returns error when we try to set it to use. Fixed full volume restore, nlist destination path must be set to source filesystem, instead of / bugfix: (#30530) Backup of a volume (specified by id) does not work like before change: opennebula: do not add vm objects during browse in for loop used to calculate disk size change: sbc_ndmp - prevent variable overwrite when decrypting password change: do not print 'No running job > keep allocation' that often to log file change: Use 'f' Python format string Revert "BUGFIX: (#30495) Check if PostgreSQL DB is running for 60 seconds before trying to export DB during update. If DB is still not running after 60 seconds abort the update procedure." Revert "BUGFIX: (#30495) Add a new sm_db switch called check to find out it the postgres db is up and running. It returns 0 if db check is ok and 2 otherwise. Use findstr on Windows to check if postgres db already exists." BUGFIX: (#30495) Also use checkdb for Windows to wait until the PostgreSQL DB is online. change: (#30077) Change 'sm_dedup_interface replicate start' command change: (#30077) Re-write of 'sm_dedup_interface'. Split global and task specific options change: (#30082) Revise message in case data store purge has nothing to delete new: Jenkins develop branch pipline file change: (#30077) Re-write of 'sm_dedup_interface'. Remove --boost option bugfix: (#21987) Simpler sm_db_update process. Problem: Error occured during execution of 'gv_ro_bin:sbc -b -s @sesam_db_$old_version_:gv_date:.sbc gv_rw_db: -x "/backup/" gv_rw_ini:'. Solution: Exclude pattern argument -x '...' must be given before backup source new: add fast files metadata check on PCCK bugfix: insert several log messages to find out why the var\tmp folder is not deleted during uninstall. BUGFIX: (#30500): add browse and backup functionality for opennebula 6.2 bugfix: (#30406) execute sm_set_windows_defender_preferences.ps1 -remove during sm_setup delete instead of executing it in the installshield project directly. BUGFIX: (#30495) Check if PostgreSQL DB is running for 60 seconds before trying to export DB during update. If DB is still not running after 60 seconds abort the update procedure. BUGFIX: (#30495) Add a new sm_db switch called check to find out it the postgres db is up and running. It returns 0 if db check is ok and 2 otherwise. Use findstr on Windows to check if postgres db already exists. bugfix: (#30527) When sm_sshd gets stopped it throws an exception 'ValueError: file descriptor cannot be a negative integer (-1)'. Problem: Listening socket was closed while new connection accepted. Solution: Check for specific exceptions while listening socket select or accept BUGFIX: (#30468) Add sm_remove_saveset.exe to the executables list that is killed during sm_os_startup stop. change: use sp2 for tagging instead of sp1 change: (#27115) - Encryption of page files for Si3-NG BUGFIX: (#30525) 'sm_pur_status' does not use configured log file retention time. Use correct 'DB:defaults.key' name 'eol_sesam_logs' bugfix: (#30490) - sm_qm_main prevent NULL for resources string bugfix: (#30456) Backup of UNC Path fails with source ends with * or *.* BUGFIX: (#30417) AbortBackup must be called during DoSnapshotSet when backup is canceled. Problem: Asynchronous operations like DoSnapshotSet could not be aborted by termination signal. Solution: Run asynchronous operations Wait in loop, which could be aborted at any time BUGFIX: (#30271) Copy of bigger files via sm_ssh exits with error code 255, no error shown. Problem: sm_sshd close connection before sm_ssh/ssh get all data. Add sm_ssh/sm_sshd to srv and cli SP new: (#29094) Add log about 'sm_dedup_interface' will be executed on source client for HPE Catalyst Replication NEW: (#29094) HPE Catalyst replication fails, if FC interface is used for backups. Use source side interface from DB:migration_tasks to access source Catalyst store change: HPE log level have to be adopted if DI started with higher log level new: (#29112) Start Catalyst replication on source node. Add -E to overwrite ini hostname for HPE replication CHANGE: (#27389) Do not create 'mtime' trigger for DB:performance new: (#29094) Add log about 'sm_dedup_interface' will be executed on source client for HPE Catalyst Replication CHANGE: sbc_one: report errors during command execution in one line, dont log complete exception string BUGFIX: (#30501) remove snapshot only if vm is in running mode BUGFIX: (#30517) MaxDB restore: Use saveset id returned by open job call, to allow restore using migrated savesets new: (#30515) - vfs supports line by line identification of LIS/LSLv2 file. This is achieved with option -o disable_catalog_type_detection=1 BUGFIX: (#30513) Unable to cancel opennebula backup. Solution: Search for sbc_proxy process NEW: (#30501) sbc_one: Create disk snapshot before cloning image new: (#30511) Create QM directories with group write permissions NEW: (#30508) Add Windows 8 and Windows Server 2012 to the list of unsupported systems for Sesam version 5.0. Revert "NEW: (#30501) sbc_one: Create disk snapshot before cloning image" new: (#30511) Create QM pipes with group write permissions BUGFIX: (#30472) Generational restore of MSSQL Database creates database with / in name. Solution: If only database name is given then avoid extra '/' BUGFIX: (#29450) Hana Restore waited for specific drive while other 'read' drive was available. Reimplement after fix for '#29659 Restore request from external caller gets wrong drive, if drive group spawns over multiple data stores'. Set retrieved drive if only one drive or multiple drive pathes are retrieved. bugfix: (#30492)(#30108) Encrypted and compressed backup with ADS (Alternate Data Streams) may fail with 'MTF descriptor block ID' error. Solution: ADS stream (ADAT) is always followed by at least 2 EH frame headers where remaining size of first header is less then data length NEW: (#30501) sbc_one: Create disk snapshot before cloning image change: HPE log level have to be adopted if DI started with higher log level new: New 'sm_common3.py' to migrate common Python functions to pure Python 3 change: HPE log level have to be adopted if DI started with higher log level bugfix: (#30490) - possible cause of qm crash due to too small allocation change: opennebula: gracefully exist backup/error operation if snapshot or restored image has state ERROR bugfix: (#30420) Adjust section Windows_migrate_sqlite_postgres to do the complete sqlite to postgres migration. A new section was added which tests if a migration from sqlite to postgres would work, called Windows_test_migrate_sqlite_postgres. All recommended adjustments were also made. Revert "wip: (#28447) Refactoring 'sm_trace' class'" BUGFIX: (#30500): pass object to all required functions BUGFIX: (#30500): add browse functinality for opennebula 6.2 and older change: (#27115) - Encryption of page files for Si3-NG BUGFIX: (#30500) use findall to find datastores from xml file generated on opennebula 6.3 change: (#27115) - Encryption of page files for Si3-NG BUGFIX: (#29450) Hana Restore waited for specific drive while other 'read' drive was available. Reimplement after fix for '#29659 Restore request from external caller gets wrong drive, if drive group spawns over multiple data stores'. Set retrieved drive if only one drive or multiple drive pathes are retrieved. NEW: (#30488) Linux file backup: Open files with 'O_NOATIME' bit, if '-o noatime' option is set, so access time isn't modified during backup. Super user rights are necessary NEW: (#27306) Option to reduce default log messages in sbc Linux. If '-o shortlog' option is set, information like Sesam kit name or default exclude list isn't logged. Useful for backup types like SAP Hana, where 'sbc' is called 1000 times an hour, to reduce redundant information bugfix: (#30458) - set server to empty if server does not exits on queue BUGFIX: (#30492) Encrypted backup with ADS (Alternate Data Streams) fails with: 'The number of saved objects according SBC and LIS file of backup does not match.'. Solution: ADAT stream may have remaining length 0 and remaining length may be less than stream data length change: (#27115) - Encryption of page files for Si3-NG new: (#30418) IVssAsync::Cancel should be called when backup operation is canceled after snapshot creation. Add log message 3456 about IVssAsync::Cancel called NEW: (#29094) HPE Catalyst replication fails, if FC interface is used for backups. Use source side interface from DB:migration_tasks to access source Catalyst store new: (#29112) Start Catalyst replication on source node. Add -E to overwrite ini hostname for HPE replication bugfix: (#30463) Incremental backup of MSSQL database with SIMPLE recovery model ends with 'Error: Saveset zero size'. Solution: If only database is given then CheckDbState() must not return with 'not saveable'(2) NEW: (#29094) HPE Catalyst replication fails, if FC interface is used for backups. Use source side interface from DB:migration_tasks to access source Catalyst store BUGFIX: (#30406) Only add the main folders to the Windows Defender exclusion list since all subdirectories are recursively included. Add datastore specific exclusions to the registry to be able to delete them during uninstall of sesam. BUGFIX: (#30487): If snapshot_timeout option is set,backup may fail with ">= not supported between instances of int and str" BUGFIX: (#29812) Add sm_vfs.exe and CbFS driver to client and gui kit. bugfix: (#30483) sm_rexec should return with 127, if command could not be found NEW: (#30364) Enforce UTF-8 file encoding in all Sesam Java components change: (#30429) - Command to stop running and pending Si3-NG replications, backups and restores bugfix: (#30483) sm_sshd should return with 127, if remote command could not be found new: (#29932) Nutanix on Windows - connected to ISCSI Nutanix target volume group with Windows MSiSCSI service. change: (#30429) change "replicate abort " call to Si3-NG Fix merge conflict BUGFIX: (#30478) put the sm.ini path in quotes to make sure that the servicepack information is inserted in the sm.ini during servicepack installation. CHANGE: (#30364) vSphere backup: Write backup log file with 'UTF-8' encoding bugfix: (#30271) Copy of bigger files via sm_ssh exits with error code 255, no error shown. Problem: sm_sshd close connection before sm_ssh/ssh get all data. Solution: Join STDIN thread with timeout change: (#30429) - Command to stop running and pending Si3-NG replications, backups and restores BUGFIX: (#30392) Windows backup could fail with 'Unknown MTF descriptor block ID. (52)' when backing up alternate data streams. Solution: Continue processing read_stream() even if the remaining size is equal to the MTF block size new: (#30429) add "backup|restore abort " call to Si3-NG change: (#30429) - Command to stop running and pending Si3-NG replications, backups and restores new: add "backup|restore abort " call to Si3-NG change: (#30429) - Command to stop running and pending Si3-NG replications, backups and restores bugfix: (#30108) Encrypted/Compressed backup of Windows client may fail with 'MTF descriptor block ID' error. Avoid 'WARNING:' message prefix BUGFIX: (#30108) Encrypted/Compressed backup of Windows client may fail with 'MTF descriptor block ID' error. Solution: Process frame headers bugfix: (#30434) Cancelling path backup logs 'The number of saved objects according SBC does not match' to day log. Commit 08d750a introduced wrong logic so backup state may be set to 'X' even if LIS items > SBC items (which is not a failure) BUGFIX: (#30468) Add sm_remove_saveset.exe to the executables list that is killed during sm_os_startup stop. bugfix: (#29481) Incorrect formatted 'dir' meta-data for MSSQL module BUGFIX: (#30463) Incremental backup of MSSQL database with SIMPLE recovery model ends with 'Error: Saveset zero size'. Solution: If only one database is given then backup level is turned into differential. Note: Such databases are still skipped if source covers multiple databases BUGFIX: (#30406) insert DataStore Path to exclusion list of Windows Defender. change: (#30429) - Command to stop running and pending Si3-NG replications, backups and restores change: (#30465) Remove MS-SQL 'all' backup support. Problem: its possible to select multiply databases as a backup source. Solution: Restrict number of databases as backup sources by adding "single_selection" change: (#30465) Remove MS-SQL 'all' backup support. Problem: its possible to select "MS SQL Server:" as a backup source. Solution: add "not_selectable" to "MS SQL Server:" so GUI will disable its select BUGFIX: (#30448) Fixed Dovecot backup. Added utf-8 encoding on opening LIS file and mail file in sm_reformat_lis BUGFIX: (#30462) MSSQL backup with source all logs 'The number of saved objects according SBC .. does not match'. Logic adapted new: #(30429) - Command to stop running and pending Si3-NG replications, backups and restores bugfix: (#30416) Method GatherWriterStatus() must be called before BackupComplete() to allow correct handling of VSS Writer status information. Problem: GatherWriterStatus() called after successful BackupComplete(). Solution: Change function call order GatherWriterStatus() before BackupComplete() bugfix: (#30434) Cancelling path backup logs 'The number of saved objects according SBC does not match' to day log. Problem: Solution introduced problem with exchange server backup BUGFIX (#29700) Cannot cancel active REAR backup. Solution: Search for sbc_proxy process change: (#30454) - Unify response code/JSON for all API calls NEW: (#30457) Add user, which has called 'sm_break' to Sesam log message CHANGE: (#8554) Do not call 'sm_kill_stpd' in 'sm_sms_backup' anymore, which was sometimes necessary to terminate hanging STPD processes in the past CHANGE: (#30154) Oracle restore on older systems fail in combination with newer Sesam server version. Do not return 'size=' tag for client version < '4.4.3.84'. Do check for backupt type Oracle only change: Use 'Could not' instead of 'Couldn´t' in Sesam log message to avoid problems with parsing BUGFIX: (#30446) Update of servicepack number in db did not work since version comparison was wrong. In addition check update should not be aborted in case access check using credentials fails. bugfix: (#8554) Termination of all Sesam actions: Do not try to block and start queues for drive groups without configured drive bugfix: (#8554) Termination of all Sesam actions: Do not try to reload SMS on clients with accessmode 'PROXY'. (Usually drives used for snapshot backups) CHANGE: (#29903) Change severity of message from ERROR to INFO for 'I015-DRIVES Encryption analysis in drive failed' new: add "backup list" call to Si3-NG BUGFIX: (#30364) 'sm_remove_saveset' process hangs during purge, if vSphere backups with special characters in source should be deleted new: add "backup abort" and "restore abort" calls to Si3-NG bugfix (#30451) Aborting queued restore prints 'Drive -3 not yet configured.' to day protocol. Solution: Avoid drive dismount if called with negative drive number. Additional: End with state '3' and warning message change: Added try catch to smctrace.trt() function to prevent unhandled exception on encoding or other errors bugfix: (#30167) Restore of a vSphere VM using migrated saveset with not possible, if backup and migration has been done with 4.4.3.X. Add sm_restore to file list (srv) bugfix: (#19572) - reenable multi submit queue and append fix: set resource type to arbitrary if no prefered resource given new: (#29404) VMware backup using hardware snapshots. Improve BDA module to work with RAW backups bugfix: (#29896) Final state message is misleading if task is cancelled while waiting: 'Timerange ... already expired'. Solution: End with appropriate message if called with special argument '-d -3' CHANGE: (#30161) Use exit code from 'sbc_restore_post' to decide, how final restore state will be set new: (#29404) VMware backup using hardware snapshots. Improve BDA module to work with RAW backups bugfix: (#30434) Cancelling path backup logs 'The number of saved objects according SBC does not match' to day log. Solution: Check processed items counter after others bugfix: (#30435) Cancelling backup results in 'Invalid usage of PROT:' log message in day protocol. Solution: Set default values if option '-C sesam_server' is not set BUGFIX: (#30431) sm_break -d does not cancel running backups. Solution: Add missing colum 'pid' to select statements bugfix: (#30427) VDDK missing during client installation. Recent commit did not take care about full path with '.ps1' powershell script extension bugfix: (#30428) script for applying the windows defender exclusions has been moved to bin\sesam directory. Adjust call in sm_setup. new: (#30181) Add STPD INI file template to RDS package new: Template Si3-NG 'stpd_conf' file BUGFIX: (#30423) Remove 'meta.vmx' meta file during purge of savesets change: (#27271) - Si3-NG doesn't log delete of object in dedup server log file change: (#30277) - Rotate log during Si3 startup CHANGE: (#30154) Oracle restore on older systems fail in combination with newer Sesam server version. Do not return 'size=' tag for client version < '4.4.3.84' bugfix: (#19572) - copy string to tmp before applying strtok_r on it bugfix: (#19572) - list of preferred resources propery initialised BUGFIX: (#30406, #29847) Add the scricpt sm_set_windows_defender_preferences.ps1 to the sesam folder to be able to add exclusions for Windows Defender using parameter -ExclusionPath. BUGFIX: (#30406, #29847) Add exclusions for all subdirectories of sesam even though they should already be included according to Microsoft since subfolders of an excluded folder are included automatically in the exclusion list. Insert a parameter -ExclusionPath which excludes a given path. This can be used to exclude certain folders e.g. the si3 folder. BUGFIX: (#30423) Some meta files are not removed during purge of backups BUGFIX: (#30422) The DataStore media filled does not count migrations and replications. Solution: Since Jaglion result_lbls.backup_id must be used to joind with results.saveset. Additional: Same logic for mount/umount saveset 'SEP Linux LVM Store' new: (#29404) VMware backup using hardware snapshots. Improve BDA module to work with RAW backups change: (#30077) Re-write of 'sm_dedup_interface'. Revise FSCK and GC commands BUGFIX: (#30399) Migration executes update statements for non existent backup on results table. Solution: Update statement with backup instead of saveset new: (#30277) - Rotate log during Si3 startup change: (#29932) Nutanix build on Windows. Few non used Linux specific library imports were removed. Added sm_iscsi_win.py for executing calls to MSiSCSI. It connects to the given portal and executes discovery. change: use different executables for testing the sp build. change: print out a more meaningful error message if no pdb files could be found. change: use folder master also for branch develop to build SPs since this is only for testing. bugfix: copy PDB files from the correct location which is different for VS2019 than for VS2010. Now they are copied from a subfolder in the local repository not the remote build location. bugfix: (#19572) - check if share is specified before printing it BUGFIX: (#30387) sbc_mysql on Windows systems ends with "Invalid backup source". Problem: All backup sources checked for existence in file system. Solution: Skip source existence check for backups which read data from pipe (-f @@) bugfix: (#29993) clean up stopping data when start full; save stopping point on server stop; change: (#25757) Set data size on RHEV CBT backup openjob BUGFIX: (#30387) sbc_mysql on Windows systems ends with "Invalid backup source". Problem: All backup sources checked for existence in file system. Solution: Skip source existence check for backups which read data from pipe (-f @@) BUGFIX: Remove SLES 11 from Sesam build version 5.0 CHANGE: (#30390) Print out help, if no parameter given CHANGE: (#30390) Use 'sm_cmd send mail' instead of old 'Mailer' function from UI server in 'sm_smtp' bugfix: (#30312) soblog contains massive amount of loglines in case of encrypted oracle backup and default loglevel change: (#30386) Reset status and message of Sesam server disk information, if enough free space is available again BUGFIX: (#30386) Correct values for total, used and free for Sesam server data file system Fix build error: sbt_helper.c(490): error C4700: uninitialized local variable 'szPadding' BUGFIX: (#29919) Oracle backups are not encrypted despite encryption option for task is enabled. Problem: Padding bytes read skipped. Solution: We have to read padding bytes except BSA_RC_NO_MORE_DATA already received by previous read (saveset padding already avoided for all DS except Tapes) BUGFIX: (#29919) Oracle backups are not encrypted despite encryption option for task is enabled change: remove obsolete code BUGFIX: sm_sms_interface getlabel during EOM failed with "Cannot Login to SMS 3009: Unable to open first free detected pipe". Solution: Increase SMS answer timeout to 2 x 20 seconds. Note: The key 'Answer_Timeout' in file sms.ini, section [SMS_Server] allows to set individual values. BUGFIX: (#29994) Update of some clients should not be tried but canceled right away. Added Windows Server 2008/R2 and Windows 7 to the list of unsupported systems again. Evaluate the access_state of the clients before starting the update. Only if the client is reachable/accessable the client update will be queued. Revert "BUGFIX: (#30108) Encrypted/Compressed backup of Windows client may fail with 'MTF descriptor block ID' error. Solution: Add stream parsing in case of encrypted or compressed data" (missing parsing of Compress/Encrypt Frame Headers) change: (#30379) sm_dedup_interface get enters endless loop if object to receive is corrupt. Revise log level bugfix: fix deep pcck new: (#30181) Add scripts to create Si3 appliance to RDS package change: Remove support for building Sesam kernel Python modules using Python 2 new: (#30181) Add cron job to delete Sesam logs new: (#30362) - Faster PCCK for Si3-NG with S3 backend bugfix: change api call attempt timeout BUGFIX: (#29919) Oracle backups are not encrypted despite encryption option for task is enabled. Problem: Padding bytes read skipped. Solution: We have to read padding bytes except BSA_RC_NO_MORE_DATA already received by previous read (saveset padding already avoided for all DS except Tapes) new: (#30362) - Faster PCCK for Si3-NG with S3 backend bugfix: increase read timeouts for s3 bugfix: avoid thread blocking on delete object request bugfix: (#29993) fix resumed FSCK when there is no info about last successful run bugfix: fix RoutesSpec test BUGFIX: (#30314) Show actual '.vmx' belonging to VM only bugfix: fix "sending early message" warning on server task start BUGFIX: (#30375, #30006) Write information about update status in the update_message of the clients table. Print new available version as well as possible update in case of equal version and if client available version is larger than sesam version in the update_message field. This will be displayed in the GUI from now on. If version is equal it will not be considered in the amount of possible update clients anymore. change: print update_message with a capital letter at the beginning to easen the reading of the message. change: Use '-o use_dump' option for DB export in case of postgreSQL during update new: (#30362) - Faster PCCK for Si3-NG with S3 backend BUGFIX: (#30108) Encrypted/Compressed backup of Windows client may fail with 'MTF descriptor block ID' error. Solution: Add stream parsing in case of encrypted or compressed data (#19572) - reenabled submit to several resources, fixed some memory corruption/leaks BUGFIX: (#29995) use a semaphore to prohibit update of one client several times. new: add sm_startup to file list for servicepack. BUGFIX: (#30384) After 20 retries return error if exit code of 'sm_main status' is equal to 1 or 2 not only to 1. BUGFIX: (#30384) Update of sesam server fails with exit -3 if sds does not start fast enough. Solution: Continue with startup even if sds is not done with the startup since if is a minor deamon. BUGFIX: (#30384) After 20 retries return error if exit code of 'sm_main status' is equal to 1 or 2 not only to 1. BUGFIX: (#30384) Update of sesam server fails with exit -3 if sds does not start fast enough. Solution: Continue with startup even if sds is not done with the startup since if is a minor deamon. CHANGE: (#30383) Purge of savesets: Restrict number of parallel delete operations to 4 for Si3-NG/S3 backend new: Add data store backend type (S3, Azure...) to class 'DataStore' BUGFIX: (#29995) Do not allow to update a client if the client update is already active. Use semaphore to prohibit update several times. BUGFIX: sm_sms_interface getlabel during EOM failed with "Cannot Login to SMS 3009: Unable to open first free detected pipe". Solution: Increase SMS answer timeout to 2 x 20 seconds. Note: The key 'Answer_Timeout' in file sms.ini, section [SMS_Server] allows to set individual values. BUGFIX: (#30381)i Remove saveset called by sm_sms_copy doesn't work. Do not check parent, if main migration task for saveset is still active new: (#30362) - Faster PCCK for Si3-NG with S3 backend bugfix: (#30309) vSphere CBT restore from using other DM then for backup runs with warning: "Cannot create directory [\\?\S:\SEPsesam]". Problem: CBT restore do not relocate temporary files (metafile). Solution: Temporary file should be relocated while CBT restore similar to DB restore change: (#30348) Build of Python executable without copy of shared libraries. Add PYCPY build flag to skip shared libs copy bugfix: -o use_dump option for DB export in case of postgreSQL. This saves a lot of time during update. bugfix: (#20993) Close 'tmp' file created by 'sm_reformat_lis' for 'Path' backups before removing it bugfix: (#20993) Remove 'tmp' file created by 'sm_reformat_lis' for 'Path' backups change: (#20993) Cleanup code in 'sm_reformat_lis' Revert "new: (#30362) - Faster PCCK for Si3-NG with S3 backend" change: (#30348) Build of Python executable without copy of shared libraries. new: (#30362) - Faster PCCK for Si3-NG with S3 backend change: (#30348) Build of Python executable without copy of shared libraries. SCC python binaries bugfix: (#30263) Show more information on .not log at Hyper-V backup. MAC address formatting. new: (#30362) - Faster PCCK for Si3-NG with S3 backend bugfix: fix logging BUGFIX: (#30347) KVM backup failed with core dump in sm_sbc_com. Solution: Avoid endless loop in ADJUST_EOLS bugfix: request import cause temporary file generation, and this fail not removed in case of sm_main status change: Set smctr to SetupExceptionHookof sm_reformat_lis and sm_vmdk_interface change: Added attribute smctr to SetupExceptionHook of sm_exception for writing starck trace of non Sesam exceptions to Sesam log. UseUs it in sm_data_store BUGFIX: (#30375) Optimize info message if client can not be updated because Server is higher version. Also optimize the info message if the package version is equal to the installed one and if the system is discontinued. BUGFIX: (#30016) Do not execute automatic client update to a higher version then of Sesam server itself except if -F switch is given. BUGFIX: (#30016) Do not execute automatic client update to a higher version then of Sesam server itself. Only SEP sesam Server can be updated to a higher version. Revert "change: Added smtrace written to log file on ehandled exception processed in sm_exception.py, adopted class smctrace, set logger in exception hook of sm_reformat_lis" change: sbc_rear: allow customer to disable SAN backup bugfix: (#30263) Show more information on .not log at Hyper-V backup. Problem: No MAC address info in log. Solution: Add MAC address to log. bugfix: (#28876) made PCCK parallel, speed up s3 pages operations change: (#27115) - Encryption of page files for Si3-NG CHANGE: New version 5.0.0.5 for master build change: (#30236) sbc does not timeout if sm_stpd_http_conn goes into uninterruptable sleep. Remove obsolete code in XBSA change: (#30363) - Faster download of objects during Si3 startup using S3 backend change: removed AlwaysRetryCondition from S3 change: setup exception hook with smtrace logger in sm_vmdk_interface and sm_data_store change: Added smtrace written to log file on ehandled exception processed in sm_exception.py, adopted class smctrace, set logger in exception hook of sm_reformat_lis bugfix: di replicate show for HPE shows to big trans_netto value change: move glbv delete specific parts to the subfunction :WAIT_FOR_GLBV_TO_BE_DELETED and use sm_client to create, read and delete GLBVs instead of sm_glbv. change: adjust the datestring in case the day and/or month consists only of 1 digit to insert a 0 in front. Insert a missing echo statement. change: "enable" pcck again change: injected the correct content loader for Azure disk page modifier bugfix: inject the correct S3 content loader NEW: (#18646) Write more information about replication into main log file BUGFIX: (#30355) HPE Catalyst replication to Cloud store uses 'sm_dedup_interface put/get' instead of replicate feature bugfix: attempt to make loading of page content lazy (for encryption) bugfix: values less then 100 MiB printed in wrong format change: (#30077) Re-write of 'sm_dedup_interface' (getuuid/getpath/getlabel) change: changed connectionMaxIdleTime to 5 seconds to always use fresh connection change: reintroduce custom retry and changed the number of concurrent requests NEW: (#18646) Write more information about replication into main log file bugfix: fix build bugfix: di replicate show for HPE shows to big trans_netto value new: filter HPE replication tasks by time BUGFIX: (#30355) HPE Catalyst replication to Cloud store uses 'sm_dedup_interface put/get' instead of replicate feature BUGFIX: (#30016) Do not execute automatic client update to a higher version then of Sesam server itself except if -F switch is given. bugfix: copy the log file from and to the correct location to easen error analysis during SP build. change: (#27115) - Encryption of page files for Si3-NG Revert "NEW: (#19572) - sm_qm_main allows submit to several resources" Revert "bugfix: (#19572) - sm_qm_main - windows build issue" Revert "bugfix: (#19572) - sm_qm_main clearing warning" Revert "bugfix: (#19572) - fixing sm_drive_config - using server" Revert "bugfix: (#19572) - fix for empty resource" bugfix: copy the SP file instead of moving it to the local workspace in case it is still in use and delete it after copy. Execute CLEAN a second time in case it fails the first time. Remove the second /Y parameter for the xcopy statement of the log file. change: (#30060) Remove filesystem access for Si3 drives bugfix: only remove the servicepack files of the current SP e.g. server or client before moving them to the SP build folder. This is needed for the Jenkins artefacts. bugfix: fix DI Windows build bugfix: (#30326) sm_dedup_interface doesn't read from pipe correctly change: move Si3-NG JVM calculations into DI new: add Transfer Timeout for Si3-NG (-W ) new: add Connection Timeout for Si3-NG (-W ) new: add index size for Si3-NG purpose config bugfix: sm_dedup_interface get|put have to show error in case of wrong INI bugfix: sm_dedup_interface core while memcpy call on stack char array bugfix: di status core while Si3-NG startup. Problem: division by zero. Solution: Validate division arguments bagfix: improve Si3-NG status bugfix: Improve DI Si3NG status summary bugfix: (#29834) - Start Si3 network access earlier during startup bugfix: fix DI SLES11 build bugfix: fix DI Linux build Revert "bugfix: fix DI Linux build" bugfix: fix DI Linux build change: (#30077) Re-write of 'sm_dedup_interface'. Revise "replicate start/show/abort" commands change: (#29834) - Start Si3 network access earlier during startup new: (#23341) Add a switch to sm_dedup_interface dir/status/etc -> show values like df or du with switch -H change: changed API of Si3NG to list available cloud backends bugfix: (#30122) sm_dedup_interface dir full for 'Path' datastores doesn't work anymore. Keyword "full" have to be ignored for all stores except Path new: print information about all cloud backends bugfix: (#30077) Re-write of 'sm_dedup_interface'. Problem: di list for disk store interpret "full" as filter. Solution: "full" should be interpret as keyword bugfix: (#30077) Re-write of 'sm_dedup_interface'. sles11fix build issue bugfix: (#30077) Re-write of 'sm_dedup_interface'. Use inheritance new: (#30077) Re-write of 'sm_dedup_interface'. Use inheritance change: (#30033) Use GiB for all size related values in sm_dedup_interface output as default bugfix: fix repair command family for Si3-NG di bugfix: (#29850) DI Si3-NG related changes according replication abort and replication show change: (#27115) - Encryption of page files for Si3-NG new: allow sm_client fastdir print full path (-n 1) BUGFIX: (#30350) Automatic installation using sm_update_client fails for certain installation types due to a missing whitespace between the SERVER name and the CLIENT parameter. Add vcxproj files. bugfix: (#30085) Hyper-V restore fails, if data store path from original VM does not exist on target server. Problem: Incorrect volume processing for remote target node. Solution: Change processing logic. BUGFIX: start SEP sesam BSR Pro during SP installation only if BSR Pro is installed on the system. new: adjust client and server file lists containing sbc, sbc_proxy, etc. for SP 5.0.0.4. BUGFIX: (#30349) Applying SP for Windows client/RDS must not use sm_shutdown anymore. Use sm_os_startup stop instead to stop services of Server/Client/RDS. BUGFIX: (#30055) In the xcopy statement the language specific echo f statement was replaced with the parameter /I for xcopy. In case a file is renamed using xcopy a * is added at the end of the destination file name. Inserted variable for recursive copy of folder and subfolders. NEW: (#26746) sm_event: Start command, execute command and start other event after task group has finished. Add example NEW: (#26746) sm_event: Start command, execute command and start other event after task group has finished Revert "BUGFIX: (#30055) In the xcopy statement the language specfic echo f statement was replaced with the parameter /I for xcopy. Inserted variable for recursive copy of folder and subfolders." bugfix: (#30326) sm_dedup_interface doesn't read from pipe correctly bugfix: (#30064) Backup of "Host Component" selection fails. Problem: Host component is not contain any VM. Solution: Hide Host Component. bugfix: (#30326) sm_dedup_interface doesn't read from pipe correctly change: (#27115) - Encryption of page files for Si3-NG BUGFIX: (#30016) Do not execute automatic client update to a higher version then of Sesam server itself. Only SEP sesam Server can be updated to a higher version. BUGFIX: (#29934) Replication for source pool spawning over multiple data stores. Set source side drive parameter correctly according to saveset location for SI3->SI3-NG replciation bugfix: is branch_name is equal to develop or master a master kit should be build. Replace the NEQ with EQU to check. bugfix: correct wrong variable branch_version with branch_name to ensure that the right build takes place. bugfix: (#19572) - fix for empty resource BUGFIX: (#29934) Replication for source pool spawning over multiple data stores. Set source side drive parameter correctly according to saveset location for SI3->SI3-NG replciation BUGFIX: (#30336) Backup may fail with 'The data size of the backup data (SBC) is different from the size of the stored data (STPD)': Do not compare sizes as strings but unsing 'long', because it's not an error if STPD size > sbc side bugfix: (#19572) - fixing sm_drive_config - using server bugfix: (#19572) - sm_qm_main clearing warning bugfix: (#19572) - sm_qm_main - windows build issue NEW: (#19572) - sm_qm_main allows submit to several resources NEW: (#26953) Increase default limit of allowed drive numbers to 799. Drive numbers >= 800 are reserved for Sesam internal use change: Added readme doc info for Nutanix -a deletafter new: (#27115) - Encryption of page files for Si3-NG change: (#30152) NDMP restore: Do not set '-a seldir=1' flag, if 'gv_conf_use_ndmp_only_dirs_or_files_' is set to '0' BUGFIX: (#30073) Backup from NFS Filesystem responds with warning "Warning: List of EA returned with error: errno = 95, error = Operation not supported." Problem: Warning printed if listxattr operation not supported. Solution: Print warning only if file exists and operation supported (not ENOENT and not ENOTSUP) BUGFIX: (#30258) Hana backup exits with an error after read retry, although everything was backed up. Solution: Use message level 'information' for message 'System call [...] timed out after [..] seconds' to avoid error exit code change: enebled Nutanix option to delete Sesam volume groups during cleanup operation (#29957) Fixed mount of Nutanix using migrated saveset. Original saveset was written to LSL instead of migrated. change: (#27115) - Encryption of page files for Si3-NG NEW: sbc_rhev: introduce common function that waits for disk status CHANGE: sbc_rhev: log error received from API if snapshot creation fails BUGFIX: (#30184) sbc_rhev: wait for imported RHEV virtual machine to be in status DOWN until further operation BUGFIX: (#30197) Use sm_popen for RHEV backup and restore to print SBC logs in real time BUGFIX: (#30184) Fixed RHEV Full CBT restore BUGFIX: (#30251) Better error handling if detaching disks from data-mover fails CHANGE: (#30253) Report used pycurl and imagio client version in RHEV module BUGFIX: (#30262) Fixed options for memory persistent snapshots in RHEV module CHANGE: (#30260) End backup with warning if virtual machine has no Guest Agent installed CHANGE: Do not output storage domains twice in RHEV module, split information about data-centers and available storage domains CHANGE: Improved snapshot logic in RHEV module CHANGE: (#30247) Implemented RHEV restore with overwrite bugfix: (#30009) Fixed VMware IMG umount on Linux server and Windows mount client. Linux names were set for Windows vmdk files because data store was on the Linux server and RDS client was used instead of mount client BUGFIX: (#30048) Restore from 'NetApp Snap Store' using snapshot mount doesn't work. Set correct drive number for queue submit BUGFIX: (#30167) Restore of a vSphere VM using migrated saveset with not possible, if backup and migration has been done with 4.4.3.X. Build correct segment file name for migrated savesets with deprecated DB:results entry BUGFIX: (#30139) Show task type 'NetApp' in addtion to 'NDMP' during backup source browsing on a NetApp filer BUGFIX: (#29949) Do not update 'DB:results.data_size' by 'sm_sms_watch' afer after backup has already finished. Fix for source side deduplication CHANGE: (#30240) Increase the calculated Si3-NG default index value bugfix: (#30147) Fixed sbc command in sbc_proxy NetApp module to prevent calling sbc_find. Added = sign before SSL file name that is read by sbc NEW: (#30031) Add Sesam server version to backup log file, if started by a task group bugfix: (#29998) Fixed preparation of VMDK meta files for IMG backup so that relative path is written instead of absolute path CHANGE: (#30022) Un-install of Sesam RPM on SLES15 fails because of dependency problems. Add dependency for 'libgnutls' package, so system provided RPM will be installed too bugfix: (#30105) Exchange Server backup must retry operation specific to method(). Problem: Backup simply aborted if retriable error caught. Solution: Retiable error should cause operation retry with some delay change: correct spelling bugfix: (#30105) Exchange Server backup must retry operation specific to method(). Problem: Backup simply aborted if retriable error caught. Solution: Retiable error should cause operation retry with some delay bugfix: (#30136) SharePoint backup fails because browsing via SharePoint VSS writer does not return any objects. Problem: "SharePoint Services Writer" missing in writers list. Solution: Add "SharePoint Services Writer" to writers list bugfix: (#30244) sbc cores during SESAM_BACKUP. Problem: Recursive directory iterate may core on low disk space systems. Solution: Use try-catch block to avoid future cores bugfix: Remove obsolete pre-compiler flag x64/X64. Only x64 platform is supported since Beefalo bugfix: sbc hangs for 5 minutes after DoSnapshotSet fails and throw exception. Problem: Exception handled on high level and do not set Event. Solution: Event should be set immediately, and exception re-thrown BUGFIX: (#30107) Exchange Server backup must call AbortBackup() when an error occurs after PrepareForBackup() Problem: AbortBackup called by Abort and None status also. Solution: Do not call AbortBackup for Abort and None status BUGFIX: (#30107) Exchange Server backup must call AbortBackup() when an error occurs after PrepareForBackup() Problem: If backup failed due to timeout AbortBackup not called. Solution: Call AbortBackup for any Async operation fails with Timeout BUGFIX: (#30104) Exchange server backup failed due to retry with 'Microsoft Exchange Replica Writer' Problem: If 'Microsoft Exchange Replica Writer' it rewrite origin one. Solution: do not change writer on retry BUGFIX: (#30107) Exchange Server backup must call AbortBackup() when an error occurs after PrepareForBackup() Problem: If backup finished in state stateDoSnapshotSet no AbortBackup called. Solution: Call AbortBackup for all states except stateBackupComplete, additional log message 3455 should notify about AbortBackup call CHANGE: New version 5.0.0.5 for release build CHANGE: New version 5.0.0.4 for release build BUGFIX: (#30256) Correct handling of Si3->Si3-NG replications NEW: (#30291) Saveset replication: Remove savesets from target store, if size is smaller then on source side change: (#27115) - Encryption of page files for Si3-NG BUGFIX: (#29901) Kill STPD process: Change search pattern to check for successful 'kill' operation. Error massage has to be empty Add vcxproj files. bugfix: (#30085) Hyper-V restore fails, if data store path from original VM does not exist on target server. Problem: Incorrect volume processing for remote target node. Solution: Change processing logic. bugfix: (#29952) modify s3 workers new: (#27271) add logs on object delete change: (#27115) - Encryption of page files for Si3-NG bugfix: (#29852) Restore 'nvram' file for vSphere virtual machine. Problem: File upload does not work on Linux due to wrong function call. Solution: Replace EmptyLL with correct IsEmptyLL call BUGFIX: (#30241) The original error message is overwritten by 'No segment file {...}.dat found'. Solution: Overwrite backup message only in case of 'success' or 'with warnings' bugfix: Recent commit lead to compilation error on Windows: error C2480: 'dfd': 'thread' is only valid for data items of static extent change: first checkout branch_version and then git pull to ensure that git pull works as expected. BUGFIX: (#30055) In the xcopy statement the language specfic echo f statement was replaced with the parameter /I for xcopy. Inserted variable for recursive copy of folder and subfolders. change: execute git pull before building the submodules. new: remove submodule build for VS2010 and update the submodule build VS2019 including all submodules in one folder and the sbc submodule separately. new: (#28939) Set 'DB:migration_results.pid/entry' to NULL, during exit of 'sm_copy' NEW: (#28939) Break for migrations: Loop over QM entries using main session id only one time new: (#28939) Set 'DB:migration_results.pid/entry' to NULL, during exit of 'sm_copy' change: only copy sesam folder to remote location for submodule sbc. bugfix: (#30066) Hyper-V RCT Backup fails to restore into new VM. Failed to read metadata. Problem: Metadata file name was based on job name. Solution: Base file name on saveset name. bugfix: (#30311) sm_ssh in 5.0.0.4 runs into a BrokenPipeError if output of lines is limited. Problem: "head" or other application close its pipe end as soon as got all needed data. Solution: If pipe is closed, but we still have date to write in - write into devnull NEW: (#30247) Implemented RHEV restore with overwrite. If VM with the same name as the restore target exists, delete it and execute restore if -o over is set. If -o over is not set, exit with log message 1939. bugfix: page was deleted during pcck check and wasn't marked as deleted change: (#30247) Implemented RHEV restore with overwrite. If VM with the same name as the restore target exists, delete it and execute restore if -o over is set. If -o over is not set, exit with log message 1939. BUGFIX: Optimize SQL statement to get savesets, which should be migrated NEW: (#28939) Break for migrations: Send signal to main task first to stop further execution. Re-read still active DB:migration_results entries, after pending migrations have been terminated change: (#30308) HPE StoreOnce should run at least with LogLevel Info (not quiet). Solution: Map STPD defailt log level 2 to OSCLT_LOG_LEVEL_INFO new: add Si3-NG .gitignore bugfix: (#30271) Copy of bigger files via sm_ssh exits with error code 255, no error shown. Problem: sm_sshd close connection before sm_ssh/ssh get all data. Solution: Join STDIN thread with timeout change: make one complete build and build only sbc separat. change: 'Use 'find' command in sm_main to set file and directory permissions to allow different modes Use getops to set EOL and drive number change: (#27115) - Encryption of page files for Si3-NG bugfix: change method to check s3 file existence; generate pege-counter if initial si3 connects to already existed store change: insert a switch for cleaning before building the submodules. change: enhance log messages. change: change last-run.conf file so it become valid JSON change: obtain the correct name of the current branch. change: move the define of the sesamdev further up. change: insert a missing routine name and insert the clean before starting to build the modules. change: build in local workspace\exe folder and copy to remote location after build is finished to ensure that artefacts are available for Jenkins locally. Also enhance error handling. new: remove submodule build for VS2010 and update the submodule build VS2019 including all submodules available. new: (#30121) - Add datastore name as prefix on Si3 S3/Azure store bugfix: (#30105) Exchange Server backup must retry operation specific to method(). Problem: Backup simply aborted if retriable error caught. Solution: Retiable error should cause operation retry with some delay new: (#30121) - Add datastore name as prefix on Si3 S3/Azure store NEW: (#30291) Copy Si3-NG log configuration file (sm_sdslog2.xml) from 'templates' to 'ini' folder during RDS update change: correct spelling new: Use faster compress algorithm to build RPM package new: (#30121) - Add datastore name as prefix on Si3 S3/Azure store BUGFIX: (#28939) When cancelling migration main task some sub entries stay in state 'active'. Solution: Check interrupt flag in the loop before inserting more migration results change: (#30274) - Allow configuration of multiple Si3 stores as default change: disable PCCK for S3 store BUGFIX: (#30297) Fixed VMWare CBT mount with Windows RDS and Linux mount client. Separator from Drive class was useed in preparation of metadata for filename in prepare_vmdk_for_cbt() merge: Merged with recent develop branch BUGFIX: (#30295) UI does not show TraceLog for backup tasks which have been pending for at least one day. Solution: Trace function must not add '_' when log file name is given with ending '.log' bugfix: (#30105) Exchange Server backup must retry operation specific to method(). Problem: Backup simply aborted if retriable error caught. Solution: Retiable error should cause operation retry with some delay new: change s3 method isObjectExist bugfix: change PCCK page read for s3; change s3Provider timeouts NEW: (#30291) Saveset replication: Check size of saveset on target store and compare it with backup data size, if no information could be found in replication log BUGFIX: (#30073) Backup from NFS Filesystem responds with warning "Warning: List of EA returned with error: errno = 95, error = Operation not supported." Problem: Warning printed if listxattr operation not supported. Solution: Print warning only if file exists and operation supported (not ENOENT and not ENOTSUP) Revert "new: (#30121) - Add datastore name as prefix on Si3 S3/Azure store" new: (#30121) - Add datastore name as prefix on Si3 S3/Azure store bugfix: change s3 provider backoff strategy bugfix: (#30282) catch queue.Empty exception in get_nowait() of sm_Popen3Full(). Add DONE messages to queue as byte and use decode() to prevent errors on reading byte values from the queue bugfix: (#30182) Exchange 2019 DAG backup fails "Failed to detect source. Name is "\Microsoft Exchange Replica Writer\". Problem: "Microsoft Exchange Replica Writer" is not available if on host both active and passive DB available. Solution: If DB not reachable over "Microsoft Exchange Replica Writer" try with "Microsoft Exchange Writer" change: Set Sanity State to OK while server loading bugfix: (#30157) correct the step number. bugfix: replace the slashes with backslashes to ensure that old sp files are deleted from the local workspace folder before new ones are copied. bugfix: (#30197) Replaced sm_Popen3Full() with sm_popen() for RHEV backup and restore. Fixed processing of output returned from sm_popen() so that sbc closejob works.Temporary solution until sm_Popen3Full() does not work well BUGFIX (#30281) 'sm_sesam remove backup' doesn't delete replicated saveset file. Call 'iREMOVE_SAVESET()' only for last saveset of backup CHANGE: (#30157) Reset servicepack related information in DB:clients after kit update change: move Si3-NG JVM calculations into DI CHANGE: (#30157) Remove deprecated 'servicepack' entry from 'sm.ini' during update BUGFIX (#30281) 'sm_sesam remove backup' doesn't delete replicated saveset file CHANGE: (#30157) Use new key 'servicepack_package' from 'sm.ini' to show currently installed SP change: move Si3-NG JVM calculations into DI new: Extra Si3-NG log file for basic backup information NEW: (#29719) Correc check of used space by Sesam for Si3-NG stores using S3 or Azure storage backend Fix build error in 'sm_lib.c' bugfix: change return of dedup_read_restore to -1 in case of exceptions BUGFIX: (#30206) MS-SQL backup fails while looking for a strange file '?? GLBV gv_savset_6 cannot be created: No such file or directory.dat'. Solution: Avoid writting to obsolete 'gv_savset_{drive}' - concurrent process has written and renamed same GLBV file Fix build error in 'sm_pur_status' new: copy pdb.zip file to remote location after successful build. Use dist_with_old for release build. change: insert a variable for master build even if on branch develop. BUGFIX: (#30211) execute sm_os_startup stop for Windows for all installation types during update. BUGFIX: (#30211) Execution of 'sm_os_startup stop' failed since output text file name was missing. BUGFIX: (#30211) Problem: During update of RDS and client not all processes were ended correctly this may lead to a reboot of the system after update or a failed update in case of remote update. Solution: Execute 'sm_os_startup stop' instead of only executing 'sm_main stop' for RDS, client and GUI update. In addition end O&O GUI and O&O Tray in case they are running during sm_os_startup stop. bugfix: change Path creation constructor bugfix: (#29958) detect changes of page body file new: 'sm_cross_check' Do not add files younger then EOL to file list new: (#29993) resume stopped fsck wip change: x_Snprintf(): Windows: 'Beginning with the UCRT in Visual Studio 2015 and Windows 10, vsnprintf is no longer identical to _vsnprintf. The vsnprintf function complies with the C99 standard' (SEP sesam uses Visual Studio 2019) bugfix: (#30197) Use sm_Popen3Full() function for Copy level RHEV backup and restore. NEW: (#30181) Si3 appliance: Read UUID from running server and update DB:data_stores.uuid with it change: sbc_rhev: rework snapshot logic BUGFIX: (#30171) Enforce full backup when error occurs at parsing of segment file. Solution: Use function b_ADJUST_TASK_STATUS() in sm_backup and sm_sms_backup change: (#30260) sbc_rhev: use more sophisticated approach to query if virtual machine guest agent is running and reachable change: sbc_rhev: do not output storage domains twice, split information about data-centers and available storage domains new: added store name as a prefix for Si3 blobs in Azure change: (#30262) adopt -a qui=0|1 option, be backward compatible with old option BUGFIX: (#30260) sbc_rhev: issue warning during backup if we cant detect any guest agent within the virtual machine change: (#30262) report information about persistent memory snapshot during snapshot creation BUGFIX (#30262): sbc_rhev: quiscence option does not have any effect, restore behavior and set persist_memorystate for snapshot to quiscence option change (#30253): sbc_rhev: report used pycurl and imagio client version in rhev module BUGFIX: (#30256) Correct check for finished replication for Si3->Si3-NG replication BUGFIX: (#30258) Hana backup exits with an error after read retry, although everything was backed up. Solution: Use message level 'information' for message 'System call [...] timed out after [..] seconds' to avoid error exit code bugfix: (#30136) SharePoint backup fails because browsing via SharePoint VSS writer does not return any objects. Problem: "SharePoint Services Writer" missing in writers list. Solution: Add "SharePoint Services Writer" to writers list change: improve sms/stpd doc bugfix: (#30244) sbc cores during SESAM_BACKUP. Problem: Recursive directory iterate may core on low disk space systems. Solution: Use try-catch block to avoid future cores Revert "bugfix: (#30136) SharePoint backup fails because browsing via SharePoint VSS writer does not return any objects. Problem: Since commit ceeaeef (bugfix (#28598) - Backup of multiple sources ends in endless loop if one source is not existent anymore) SharePoint browse failed. Solution: Detect Writer/Component the same way as it was before commit ceeaeef" bugfix: use build folder O:\sesam\build\repository instead of O:\z\prekits to ensure that kits are stored for future use. change (#30251): sbc_rhev: sleep some time before attempting to remove de-activated disks from data-mover: rhev sometimes responds with internal engine error if things are moving to quickly bugfix: (#30136) SharePoint backup fails because browsing via SharePoint VSS writer does not return any objects. Problem: Since commit ceeaeef (bugfix (#28598) - Backup of multiple sources ends in endless loop if one source is not existent anymore) SharePoint browse failed. Solution: Detect Writer/Component the same way as it was before commit ceeaeef new: add Transfer Timeout for Si3-NG (-W ) new: add Connection Timeout for Si3-NG (-W ) change: (#27115) - Encryption of page files for Si3-NG bugfix: (#30085) Hyper-V restore fails, if data store path from original VM does not exist on target server. ÐProblem: VSS trying to process original VM's path. Solution: Skip processing. BUGFIX: (#30184) use vm object returned during wait_for_vm function for further operation BUGFIX: (#30184) Added 5 seconds wait and retry loop before attching disks after RHEV VM has been created. Initial 5 seconds wait before going into loop prevented exception in tests, but loop with additional 10 seconds sleeps implemented for additionaly robustness Change trace BUGFIX: (#30228) Replication between Si3 and Si3-NG doesn't work if RDS in involved. Start read 'sm_dedup_interface' on source RDS now new: (#30210) Add cyclic performance information to STPD log in default log level new: add support of extended jvmconfig command to di new: add index size for Si3-NG purpose config new: add parameter index_size to jvm propose function bugfix: sm_dedup_interface get|put have to show error in case of wrong INI bugfix: fix server hanging when trying to delete not existing s3 object bugfix: (#20993) Close 'tmp' file created by 'sm_reformat_lis' for 'Path' backups before removing it new: New targets test_util\mtfana and test_util\sm_pytest bugfix: removed warning when parsing azure blob config bugfix: (#30066) Hyper-V RCT Backup fails to restore into new VM. Failed to read metadata. Problem: Metadata file name was based on job name. Solution: Base file name on saveset name. change: (#27115) - Encryption of page files for Si3-NG bugfix: print correct value bugfix: (#20993) Remove 'tmp' file created by 'sm_reformat_lis' for 'Path' backups bugfix: Use correct trace function during NFS tool chain check bugfix: Remove 'sm_postgres' from service pack file list BUGFIX: (#30226) Re-enable purge of whole Sesam day, if no valid backups exist anymore, with corrected SQL statement bugfix: (#30226) Do not execute 'Delete everything older then oldeset Sesam day' section in 'sm_pur_status'. After recent commit compilation on Windows complained about not used variables Fix merge conflict BUGFIX: (#30226) Do not execute 'Delete everything older then oldeset Sesam day' section in 'sm_pure_status', because valid backup could accedently be removed change: (#30213) Postgresql server could not be started using 'setpriv': Add 'sesam' user to '.profile' path change: Removed conditions to build with Python3 in sbc_proxy make bugfix: (#30197) Fixed processing after sm_popen call in RHEV backup and restore BUGFIX: (#30198) Purge does not delete meta files from LIS folder, if migrated saveset is purged before original one change: (#30197) Use sm_popen for RHEV backup and restore to print SBC logs in real time CHANGE: Add '.meta' and '.not.mail.gz' to list of file extension removed during saveset purge BUGFIX: (#30184) sbc_rhev: wait for imported RHEV virtual machine to be in status 'DOWN# until further operation change: (#27115) - Encryption of page files for Si3-NG BUGFIX: (#30171) Enforce full backup when error occurs at parsing of segment file. Solution: Use function b_ADJUST_TASK_STATUS() in sm_backup and sm_sms_backup change: normalize path in sm_remove_saveset BUGFIX: (#29764) If path backup without VSS of level diff/incr does not find a file to back up it ends with 'Warning: Cannot get item security data for []'. Solution: In this case an empty saveset with only TAPE, SSET and ESET must be created 29764 - If path backup without VSS of level diff/incr does not find a file to back up it ends with 'Warning: Cannot get item security data for []'. Solution: In this case an empty saveset with only TAPE, SSET and ESET must be created BUGFIX: (#30174) vSphere backup could fail with 'Call regcomp() exclude failed for' when specific disk exclude is set. Problem: VMWare config backup forward excludes to sbc_vmware. Solution: Ignore regex compile error for VSPHERE_CONFIG new: 'sm_cross_check': Parse LIS files created by Oracle and Informix backups new: 'sm_cross_check cleanup_lis': Print out full path name of LIS file, because of date based sub-directories change: sbc_rhev: log error received from API if snapshot creation fails BUGFIX: (#30207) Correct calculation of amount of data stored on a data store, if savesets are migrated to this store bugfix: (#30203) sbc bda restore fails on windows: Error while unpacking data. WIN32 API error: 87 - The parameter is incorrect. Solution: Fix for '#30177 Selective path restore causes sbc core dump' introduced check of OS ID and version but neither VSPHERE nor BDA are setting these values in case of MTF FILE or DIRB, so revert fix new: (#30202) Purge of meta files during transition to Sesam date based LIS directory structure bugfix: sm_dedup_interface core while memcpy call on stack char array new: 'sm_cross_check': Parse LIS files created by Oracle and Informix backups bugfix: (#30196) Use 'su - sm_postgres' approach to start Sesam DB. Remove wrong PATH setting new: (#29852) Restore nvram file for vSphere virtual machine. SBC upload file to VSPHERE VM folder BUGFIX: (#30168) Kopano backup: Set status to failed if mail subject index couldn't be created new: (#28547) 'sm_sms_watch': Call 'waitpid()' for running performance read jobs on RDS after drive becomes idle to avoid temporary 'zombie' processes new: 'sm_cross_check': Allow setting of protection time range by '-e ' parameter bugfix: (#30195) Restore of data backed up from subvolume fails with "Error while unpacking data. Error while reading file." Additional: Set OS_ID = MTF_OS_WINDOWS_NT in case of DIRB for a volume BUGFIX: (#30198) Purge does not delete meta files from LIS folder bugfix: (#30195) Restore of data backed up from subvolume fails with "Error while unpacking data. Error while reading file." Solution. Since commit 1425617 (#30177) 'Selective path restore causes sbc core dump' a directory descriptor block with vendor ID '0' was treated as invalid but SEP sesam uses such MTF DIRb blocks for root directories bugfix: Avoid string overflow in sUnicodePath2String() new: add mtfana test util build for MSVC bugfix: (#30147) Fixed sbc command in sbc_proxy NetApp module to prevent calling sbc_find. Added = sign before SSL file name that is read by sbc bugfix: (#30196) Use 'setpriv' command to start Sesam DB with Sesam user on newer Linux systems. On older systems (Debian 9, RHEL 7) use old 'su -' approach bugfix: (#30134) Selective restore from RDX backup does not work. Solution: Data request > 4GB could fail. Do not use same BCK_MESSAGE structure for move and read operation change: Log message 'serv_move_tape:: Rewinding disk device.' will now appear in level 4 bugfix: Remove obsolete pre-compiler flag x64/X64. Only x64 platform is supported since Beefalo bugfix: (#25824) - Microsoft Azure BLOB Store connectivity new: (#30094) - Forward correct error message from Azure BLOB API to user bugfix: sbc hangs for 5 minutes after DoSnapshotSet fails and throw exception. Problem: Exception handled on high level and do not set Event. Solution: Event should be set immediately, and exception re-thrown new: (#29852) Restore nvram file for vSphere virtual machine. SBC upload file to VSPHERE VM folder bugfix: fix test of ObjectStoreSpec bugfix: do not update dir trackers when deleting not loaded object bugfix: use general remove method for deleting corrupted objects bugfix: allow moving to trash not loaded objects new: (#28480) sbc_com_interface hangs, after reboot of Sesam server. Enable keep-alive for XBSA bugfix: di status core while Si3-NG startup. Problem: division by zero. Solution: Validate division arguments bagfix: improve Si3-NG status BUGFIX: (#30134) Selective restore from RDX backup does not work. Solution: RestoreFilter() must return segment number to match with line from .dat (segment) file change: Fix DB:clients.vm_server_type for "ESX-Server" clients. change: Obsolete comment '//TODO' removed BUGFIX: (#30134) Selective restore from RDX backup does not work. Solution: 'Total:' line in segment file was interpreted as segment change BUGFIX: (#30177) Selective path restore causes sbc core dump. Solution: Avoid interpretation of data if corrupt data is read new: (#29993) resume stopped fsck new: (#30034) - Use same key names in STPD ini files for similar information new: (#29834) - Start Si3 network access earlier during startup BUGFIX: (#30165) NSS backup stops at 'NWSMTSConnectToTargetServiceEx CL3PROG_SERVER (NWSM_AUTH_UTF8_DATA - encoded)'. Solution: Problem introduced with commit d316128 (#28092) 'Adopt make files in master to build with ... and SSL 111 by default'. NSS requires SSL 1.0.x change: (#27902) Use sm_Popen3Full in Nutanix restore BUGFIX: (#30167) Restore of a vSphere VM using migrated saveset with not possible, if backup and migration has been done with 4.4.3.X. Build correct segment file name for migrated savesets with deprecated DB:results entry BUGFIX: (#30166) - vSphere VM containing special characters can't be backed up. Add URI encoding of backup source for new backup sub type 'VSPHERE_CBT:' bugfix: Improve DI Si3NG status summary change: (#27306) Added SBC -o shortlog to Nutanix restore BUGFIX: (#30107) Exchange Server backup must call AbortBackup() when an error occurs after PrepareForBackup() Problem: AbortBackup called by Abort and None status also. Solution: Do not call AbortBackup for Abort and None status new: add Solution Filter files to allow separate build of SMS and SBC components bugfix: (#29834) - Start Si3 network access earlier during startup change: (#20993) Cleanup code in 'sm_reformat_lis' change: (#29834) - Start Si3 network access earlier during startup NEW: (#30160) Add Sesam 'skel/templates' folder to Sesam backup during installation change: (#30152) NDMP restore: Do not set '-a seldir=1' flag, if 'gv_conf_use_ndmp_only_dirs_or_files_' is set to '0' BUGFIX: (#30139) Show task type 'NetApp' in addtion to 'NDMP' during backup source browsing on a NetApp filer change: (#27306) Removed error from Windows SBC when -o option is not recognized, log a 3500 message instead. Added -o shortlog back to Xen copy backup for Windows change: (#30077) Re-write of 'sm_dedup_interface'. Revise "replicate start/show/abort" commands change: (#29834) - Start Si3 network access earlier during startup BUGFIX: (#30107) Exchange Server backup must call AbortBackup() when an error occurs after PrepareForBackup() Problem: If backup failed due to timeout AbortBackup not called. Solution: Call AbortBackup for any Async operation fails with Timeout change: (#29834) - Start Si3 network access earlier during startup BUGFIX: (#30104) Exchange server backup failed due to retry with 'Microsoft Exchange Replica Writer' Problem: If 'Microsoft Exchange Replica Writer' it rewrite origin one. Solution: do not change writer on retry change: (#27306) Removed -o short log from Xen copy backup on Windows. It is causing error "Option -o not available in this mode." Added error handling if SBC fails to download VM export data from Xen server change: (#27306) Added -o shortlog to Nutanix backup, Citrix full and copy backup BUGFIX: (#30156) A the start of Nutanix hotadd backup and restore, check if a proxy VM exists on the Nutanix server. If proxy VM does not exists, exit with error before start of any operations NEW: (#19587) Allow lazy resource allocation. NEW: Only submit_lazy/submit_sync_lazy treats {resource} as preferred. Now a submit with resource (drive number) will not change drive if drive is blocked. bugfix: (#30076) Hyper-V backup of a VM fails with error: The file or directory is not a reparse point. Problem: Cluster methods are no longer supported. Solution: Use non cluster methods. bugfix: (#30159) Update from Beefalo to Jaglion purges EOL free savesets which are locked. Set backup result.locked='-' if migrated saveset is locked BUGFIX: (#30159) Update from Beefalo to Jaglion purges EOL free savesets which are locked. Set backup result.locked if migrated saveset is locked BUGFIX: (#30158) NSS File System backup failed with 'E020-BACKUP The data size of the backup data (SBC) is different from the size of the stored data (STPD)'. Problem: sbc_smdr showed several lines of "sbc-3006: Info: Saveset size" in not log new: (#23341) Add a switch to sm_dedup_interface dir/status/etc -> show values like df or du with switch -H BUGFIX: (#28844) Command sm_ssh got exit code 255 at sbc_es.exe -browse databases/replicas. Solution: Sometimes GetExitCodeThread() may return STATUS_THREAD_IS_TERMINATING 0xC000004B (related to AV software e.g. McAfee) then process exit code must be used. change: changed API of Si3NG to list available cloud backends bugfix: (#30122) sm_dedup_interface dir full for 'Path' datastores doesn't work anymore. Keyword "full" have to be ignored for all stores except Path BUGFIX: (#30119) insert the zlib.dll from sesamdev\vmware\vddk_... subfolder to the build process of srv, cli and gui package. BUGFIX: (#30048) Restore from 'NetApp Snap Store' using snapshot mount doesn't work. Set correct drive number for queue submit change: (#30137) Fixed compiler error BUGFIX: (#30137) Fixed NDMP restore backward compatibility for restore of directories with option -a seldir. Moved parsing of SEL lines to C++ function addRestoreItem(). Elements of a SEL line are split into a std vector of strings and read inode value from the correct position in the vector bugfix: (#25824) - Microsoft Azure BLOB Store connectivity NEW: (#30126) Add information about Sesam server startup in progress to 'sm_main status' output NEW: (#29774) Start Si3 DeDuplication store with Java 17 BUGFIX: (#30107) Exchange Server backup must call AbortBackup() when an error occurs after PrepareForBackup() Problem: If backup finished in state stateDoSnapshotSet no AbortBackup called. Solution: Call AbortBackup for all states except stateBackupComplete, additional log message 3455 should notify about AbortBackup call bugfix: (#30123) sm_main reload sds-2 stops sds-28 too BUGFIX: (#30118) Encoding problem in 'sm_reformat_lis' processing VMware backups BUGFIX: (#29080) Exchange backup fails with "a bytes-like object is required, not 'str'". Problem: Since Python3 gzip open file in binary mode. Solution: Open gzip files in text mode BUGFIX: (#30110) Microsoft Sharepoint backup failed since update to 5.0.0.1 with 'Exception Caught sm_reformat_lis.py: 688 a bytes-like object is required, not 'str'. Problem: Since Python3 file read done in binary mode. Solution: Convert binary data to text BUGFIX: (#29080) Exchange backup fails with "a bytes-like object is required, not 'str'". Problem: Since Python3 gzip open file in binary mode. Solution: Open gzip files in text mode BUGFIX: (#30118) Encoding problem in 'sm_reformat_lis' processing VMware backups BUGFIX: (#30097) Fixed sm_reformat_lis for NDMP backup. Added utf-8 encoding to read from .lnd file and and write to temporary .lis file bugfix: (#30077) Re-write of 'sm_dedup_interface'. Problem: di list for disk store interpret "full" as filter. Solution: "full" should be interpret as keyword CHANGE: (#20404) Enforce FULL level backup, if backup was successful on client side (sbc), but something wents wrong during post processing on server side. This done for all backup types now. TODO: Check for which type we can avoid this BUGFIX: (#30117): decode virtual machine config before attempting to define new: (#29530) Exchange Server Backup - Avoid possible risk of accidental Log truncation. Additional: Add description of gv_conf_all_backup_wait_for_stdin, gv_conf_no_wait_for_stdin and gv_conf_wait_for_stdin_timeout NEW: (#29530) Exchange Server Backup - Avoid possible risk of accidental Log truncation. Solution: Now use option '-o wait_for_stdin[=timeout]' for Exchange Server [DAG] backup types. GLBV gv_conf_wait_for_stdin_timeout allows to set specific timeout values NEW: (#30089) Sesam server shutdown hangs, because DB isn't reachable any more. Start Sesam PostgreSQL server using 'setpriv' instead of 'su/runuser', because 'systemd' terminates 'su' sessions before stopping SEPsesam service change: (#30089) fix chmod on socket directory: use correct variable BUGFIX: (#30117): KVM restore fails with: Restore was not successful. UUID object has no attribute get_hex: port uid generation to python3 bugfix: (#30077) Re-write of 'sm_dedup_interface'. sles11fix build issue BUGFIX: (#30116): KVM Restore fails with: Error reading config file: 'str' object has no attribute 'decode' : open config file in binary mode BUGFIX: (#30008) Log Python stacktrace in case of an unexpected exception BUGFIX: (#30008) Open local Sesam package file using binrary mode BUGFIX: (#30008) Problem: Package download by sm_update_client using 'https://' fails. Solution: If download url is https use CA certificate file also for varification of download packages not only for index.txt. BUGFIX: (#30098) sm_data_server cores during 'getlabel' on RDX drive. Solution: Close file handle after reading from DISK_HARD with appropriate command bugfix: (#30077) Re-write of 'sm_dedup_interface'. Use inheritance NEW: sbc_rhev: introduce common function that waits for disk status BUGFIX: (#30059) Fixed parsing of Kopano .not file in sm_reformat_lis by adding decode with arguments utf-8 and replace BUGFIX: (#30106) RHEV: restore of VM may fail with: 'Disk in VM is already marked as boot' if first processed disk during restore is marked as bootable': initalize bootable flag for each disk BUGFIX: (#30097) Fixed reading of NDMP LIS file in sm_reformat_lis by decoding read lines with arguments utf-8 and replace bugfix: include the python2 folder explizitly to the self extracting zip file. Only tag during server build if called with parameter all. bugfix: change the method to obtain the current branch name to ensure that the branch name and not the tag name is used for checkout. change: Always create change report. bugfix: use switch -r for building the self extracting exe file for the SP since the python2 subfolder has to be included as separat folder in the SP. BUGFIX: (#30097) Fixed encoding of lis file in sm_reformat_lis for NDMP bugfix: (#30012) Browsing backup target core on specific environment: 'sbc -D /'. Solution: Since commit 91f047c (#26435) 'REMOTE_COPY_CLIENT does not close input data file handle' the DB_PROC_PIP structure must be initialized bugfix: remove outdated call for vcvars64.bat while sm_dedup build. Assume each build operation already started from CMD with vcvars64.bat set bugfix: check for value of set_tag not for existence since it is initialized with 1. bugfix: (#25824) - Microsoft Azure BLOB Store connectivity BUGFIX: (#30088) Re-activate '-p ' option in 'sesam_gui' script NEW: (#30089) Sesam server shutdown hangs, because DB isn't reachable any more. Start Sesam PostgreSQL server using 'runuser' instead of 'su', because 'systemd' terminates 'su' sessions before stopping SEPsesam service BUGFIX: (#30087) Backup to tape with EOM fails with 'Failed to detect lis file' when 'sm_reformat_lis' is called. Adopt SQL statement to build LIS file name to not returning multiple entries in subquery from DB:result_lbls bugfix: copy sp files and pdb files to patch Jenkins folder. This is needed for providing artefacts in jenkins. And only generate and show release notes files if setting the tag is enables. CHANGE: (#20404) Enforce FULL level backup, if backup was successful on client side (sbc), but something wents wrong during post processing on server side. This done for all backup types now. TODO: Check for which type we can avoid this changed logger for Si3-NG BUGFIX: (#30097) Default timeout to cancel running jobs for external backups must be at least 1 day. Revert CHANGE: (#17427) Set stop time for external to at least one day for restores only change: Increase maximum password size (I_MAXENCPASS) to 1024 characters bugfix: (#26093) fix race condition between backup and gc change: (#30074) add option no_ea to sbc help output: skips backup/restore of extended attributes change: (#30069) Moved NFS showmount check from sm_vmdk_interface to sbc_proxy bugfix: (#21470) New Python sm_popen() function. Problem: sm_Popen3Full cause high CPU load due to infinite poll loop. Solution: Replace poll loop with wait bugfix: (#30054) Hyper-V restore finishes successful, but VM is unusable. Problem: Wrong condition for successful completion. Solution: Change condition. NEW: (#28593) Implement correct start of full FSCK for Si3-NG data stores. Use correct option to start incr. FSCK change: (#28668) Make version reporting in STPD/SMS independent from the operating system change: (#1142) Backup of sparse files. Usage revised BUGFIX: (#30068) BSR Windows backup with more than 2 partitions does not work. Solution: Avoid check if LIS items match to SBC items in case of BSR Windows CHANGE: (#30022) Un-install of Sesam RPM on SLES15 fails because of dependency problems. Add dependency for 'libgnutls' package, so system provided RPM will be installed too BUGFIX: (#30052) Selective restore from tape fails with 'no valid MTF Stream ID was found', but complete restore works. Solution: Since commit 996785b the block size is not retrieved from segment file in case of selective restore change: (#30033) Use GiB for all size related values in sm_dedup_interface output as default BUGFIX: (#30058) External restore fails when backup has been done with old LIS file director structure, but new date based one set after backup. Call 'i_get_lisfile_name_task()' with 'get' option for restore operation NEW: (#29574) Support for Windows 11. Add oper_systems and syntaxes entries for 'Windows 11' BUGFIX: (#30047) Backup EOL is not set to maximal saveset EOL. Solution: SQL statement to get max(eol) must retrieve all rows with saveset or backup_id BUGFIX: (#30043) SGM and LIS file are searched on wrong host after migration, if backup has done with SSDD flag set new: 'exec_di_status()' function to call sm_dedup_interface and return 'STATUS and MSG' bugfix: (#30037) Enhance output of sm_rescan.ps1 script. Did some enhancements new: (#29853) Makefile to sign Linux packages and create Sesam repository after kit build change: (#28668) Make version reporting in STPD/SMS independent from the operating system change: (#1142) Backup of sparse files. Usage revised BUGFIX: (#30068) BSR Windows backup with more than 2 partitions does not work. Solution: Avoid check if LIS items match to SBC items in case of BSR Windows bugfix: remove clean if build executed using variable JENKINS_BUILD. bugfix: add extra check for missing during pcck pages; increase timeout for creating of shared page from 15 sec to 2 min CHANGE: (#30022) Un-install of Sesam RPM on SLES15 fails because of dependency problems. Add dependency for 'libgnutls' package, so system provided RPM will be installed too BUGFIX: (#30052) Selective restore from tape fails with 'no valid MTF Stream ID was found', but complete restore works. Solution: Since commit 996785b the block size is not retrieved from segment file in case of selective restore change: (#30033) Use GiB for all size related values in sm_dedup_interface output as default BUGFIX: (#30058) External restore fails when backup has been done with old LIS file director structure, but new date based one set after backup. Call 'i_get_lisfile_name_task()' with 'get' option for restore operation NEW: (#29574) Support for Windows 11. Add oper_systems and syntaxes entries for 'Windows 11' BUGFIX: (#30047) Backup EOL is not set to maximal saveset EOL. Solution: SQL statment to get max(eol) must retrieve all rows with saveset or backup_id BUGFIX: (#30043) SGM and LIS file are searched on wrong host after migration, if backup has done with SSDD flag set new: 'exec_di_status()' function to call sm_dedup_interface and return 'STATUS and MSG' bugfix: (#30037) Enhance output of sm_rescan.ps1 script. Did some enhancements new: (#29853) Makefile to sign Linux packages and create Sesam repository after kit build NEW: (#30023) Delete NetApp snapshots used as Sesam backup savesets by 'sm_remove_saveset purge' operation bugfix: (#29952) change default future timeout for s3 messages to inf bugfix: (#29952) decrease workers count; fixe error logging on custom retry BUGFIX: (#30032) Fixed restore of Citrix Xen server VM to selected target data store. Selected SR was not used by sbc_proxy to do restore, instead, default SR was searched for. When the default SR could not be found, restore failed NEW: (#30026) Add backup target media type (LTO, Si3 Deduplication Store...) to backup main log file BUGFIX: (#30017) Fixed false error message in mount of VMWare image backup change: Do not write error message into Sesam main log, if 'sm_replicate' could not find information for saveset in replication log. This is handled by 'sm_replicate' itself correctly change: Do not write error message into Sesam main log, if 'sm_rescan.ps1' execution returns an error. This has to be handled by caller correctly BUGFIX: (#30035) Exception in 'sm_event' during start of migration for a task group, if some tasks are still active NEW: (#30031) Add Sesam server version to backup log file, if started by a task group change: Changed warning and error messages for Nutanix backup crash consistent/application consistent backups by Sabina`s recomendations NEW: (#28593) Implement correct start of full FSCK for Si3-NG data stores. Use data store specific configuration change: Do not call sm_lis_items to get lsl files of session saveset of IMG backup during VMWare mount. LSL files are empty for session savesets of IMG backups and sm_lis_items for them is causing error logs in Sesam main log change: Updated readme doc for Nutanix module NEW: (#28593) Implement correct start of full FSCK for Si3-NG data stores bugfix: (#29697) Before cleanup of leftover LVM snapshot, check if mount points still exists and do not try to delete the snapshot if they do NEW: (#29574) Support for Windows 11. Solution detect Windows 11 ( CurrentBuild = 22000 ) change: (#30003) UNIX socket access for GO Sesam DB on Linux BUGFIX: (#29415) sm_vmdk_interface checks for an active NFS server on a Linux Sesam server or RDS client before starting attach action change: sm_vfs logging on lvl3 prints out some more info about SHA1 BUGFIX: (#30011) Systemd unit file calls 'sm_shutdown' instead of 'sm_main stop' to stop Sesam on client. Change unit file on Debian before update starts BUGFIX: (#30011) Systemd unit file calls 'sm_shutdown' instead of 'sm_main stop' to stop Sesam on client. Change unit file for RPM packages before update starts. bugfix: (#29541) Browsing VSS Writers not possible: "sbc" "-D" "VSS:" does not list system writers. Problem: Since commit ceeaeef browse target parsing fails. Solution: Empty VSS browsing target should be parsed as ALL bugfix: (#30012) Browsing backup target core on specific environment: "sbc -D /". Problem: Missing variable initialization. Solution: Initialize variable with zeros before reference it BUGFIX: (#30011) Systemd unit file calls 'sm_shutdown' instead of 'sm_main stop' to stop Sesam on client. Change unit file on Debian before update starts BUGFIX: (#30011) Systemd unit file calls 'sm_shutdown' instead of 'sm_main stop' to stop Sesam on client. Change unit file for RPM packages before update starts. bugfix: (#29541) Browsing VSS Writers not possible: "sbc" "-D" "VSS:" does not list system writers. Problem: Since commit ceeaeef browse target parsing fails. Solution: Empty VSS browsing target should be parsed as ALL bugfix: Return correct error message, if remote command in 'sm_config_client' return with internal error (e.g. invalid access mode) bugfix: (#30012) Browsing backup target core on specific environment: "sbc -D /". Problem: Missing variable initialization. Solution: Initialize variable with zeros before reference it CHANGE: New version 5.0.0.5 for master build BUGFIX: (#30011) Systemd unit file calls 'sm_shutdown' instead of 'sm_main stop' to stop Sesam on client new: (#26899) Semaphore function with signal handler bugfix: (#30004) Perf file for data stores drives includes all streams. Problem: Perf file is written according to STPD streams number, and without reference to Store Drives. Solution: Perf files should be unique and contain info reference to current Drive only, skip IDLE streams info in Perf file bugfix: (#30009) Fixed VMware IMG umount on Linux server and Windows mount client. Linux names were set for Windows vmdk files because data store was on the Linux server and RDS client was used instead of mount client Revert "change: sm_vfs prints SHA1 of data read when log lvl > 2" bugfix: (#30004) Perf file for data stores drives includes all streams. Problem: Perf file is written according to STPD streams number, and without reference to Store Drives. Solution: Perf files should be unique and contain info reference to current Drive only, skip IDLE streams info in Perf file bugfix: (#26757) Recursive deletion of files older then X days. Problem: misleading test in help. Solution: change help according to implementation bugfix: (#30004) Perf file for data stores drives includes all streams. Problem: Perf file is written according to STPD streams number, and without reference to Store Drives. Solution: Perf files should be unique and contain info reference to current Drive only, skip IDLE streams info in Perf file bugfix: (#26757) Recursive deletion of files older then X days. Problem: sm_client drop core if too long pattern passed. Solution: Only one path should be accepted, all other - ignored bugfix: (#26757) Recursive deletion of files older then X days. Problem: sm_client drop core if too long pattern passed. Solution: Cut off pattern to fit buffer size NEW: Write message into Sesam main log by 'sm_event', if program it has been called for, has terminated abnormally (core dump...) new: Write errors into Sesam main log by Python 'prot()' function too bugfix: (#29212) Changed sbc_proxy toolcheck so that it returns message to the calling module and more detailed message about missing executable is returned in case of error. Reorganized the code in sbc_proxy toolcheck function to be better for maintainance bugfix: (#20629) Enforce FULL backup if last FULL is older then N-days. Check doesn't work any more, if enforce FULL after error option has been set too bugfix: (#29978) STPD_Main cores for unknown reason at ConnectionHandlerCb. Problem: sockaddr_storage struct conversation may fail. Solution: Use original sockaddr structure bugfix: (#29998) Fixed preparation of VMDK meta files for IMG backup so that relative path is written instead of absolute path bugfix: (#29952) use custom retry policy bugfix: (#30001) Adopt data store 'fsck' for snapshot volumes for Python 3 bugfix: (#29227) Added additional log info for partitions returned from sh.virt_filesystems and if no parition is found in the virtual disk file, log a warning BUGFIX: (#29949) Do not update 'DB:results.data_size' by 'sm_sms_watch' afer after backup has already finished. Fix for source side deduplication bugfix: (#29990) Fixed Hyper-V network mount on Windows client CHANGE: New version 5.0.0.5 for master build BUGFIX: (#30011) Systemd unit file calls 'sm_shutdown' instead of 'sm_main stop' to stop Sesam on client new: (#26899) Semaphore function with signal handler bugfix: (#30004) Perf file for data stores drives includes all streams. Problem: Perf file is written according to STPD streams number, and without reference to Store Drives. Solution: Perf files should be unique and contain info reference to current Drive only, skip IDLE streams info in Perf file bugfix: (#30009) Fixed VMware IMG umount on Linux server and Windows mount client. Linux names were set for Windows vmdk files because data store was on the Linux server and RDS client was used instead of mount client change: sm_vfs prints SHA1 of data read when log lvl > 2 BUGFIX: (#27702) BSR Restore fails with: BSAGetData: Data receiving failed. Solution: Next read must reset EOF flags after previously hitting EOF. Note: Appeared only with FTP coz HTTP uses separate sessions. BUGFIX: (#28202) BSR restore on Windows doesn't work with FTP interface. Solution: BSR is using several threads but since Oracle multi stream implementation XBSA uses thread local socket for FTP Control socket. bugfix: (#29881) sbc core dump with -o rename option. Problem: WSAStartup call wasn't handled. Solution: handle function call. bugfix: sm_event: Allow new '-F' option in 'sm_backup' during parsing backup command change: Use 'sm_remove_saveset' to remove saveset after migration bugfix: (#29985) correclty remove sbc_vadp key if present in sm.ini by using section name and key name instead of key name and value. NEW: (#29970) Do not purge savesets, which are currently used by migration. This should usually not happen, because EOL has to be longer then migration will need NEW: (#24735) remove the key sbc_vadp from the sm.ini if it exists and $old_version < 5.0.0.4. BUGFIX: (#29881) Handling WSAStartup call. BUGFIX: (#27702) BSR Restore fails with: BSAGetData: Data receiving failed. Solution: Next read must reset EOF flags aftere previously hitting EOF. Note: Appeared only with FTP coz HTTP uses separate sessions. BUGFIX: (#28202) BSR restore on Windows doesn't work with FTP interface. Solution: BSR is using several threads but since Oracle multi stream implementation XBSA uses thread local socket for FTP Control socket. change: sbc_rhev: add prefix to log messages related to hotadd functionality for easier analysis of logfiles change (#29935): sbc_rhev: add new options to help output bugfix: Remove superfluous DB query with empty SQL statement NEW: (#29935) if -a use_virtioscsi is set, use ID_SCSI_SERIAL instead of ID_SERIAL to identify attached disks bugfix: use official sbt provider in build.sh bugfix: (#29983) Fixed execution of Hyper-V mount PC script. When copy to RDS client is not needed, then set script file on the local server to be used for mount NEW: (#24735) move sbc_vadp.jar from gui folder to sesam folder during installation and update. bugfix: Do not use 'sctRL' struct (used by select in outer loop) to update of 'DB:results.throughput' in case of source side deduplication. bugfix: Typo in 'I004-STARTUP Starting SEPsesam service' message BUGFIX: (#5884) Copy saveset information into target store after migration using correct interface. Former change has been overwritten change: (#29970) Preset DB:migration_results.status='X' to avoid empty status in case of DB problems bugfix: (#29978) STPD_Main cores for unknown reason at ConnectionHandlerCb. Problem: sockaddr_storage struct conversation may fail, as already seen on Windows. Solution: Manually copy sockaddr_storage structure new: 'I004-STARTUP Starting SEPsesam service' message new: DB:data_stores.options in data store class bugfix: add object corruption store to the pcck master arguments list bugfix: (#29886) create /tmp-pages directory to store pages during gc; clean up tmp files on server start; use Files.size as bounder in pageBodyReader instead of FileChannel.size new: (#29914) Check if data store is available before starting 'sm_remove_saveset purge' bugfix: (#29965) Core dump in sm_cmd, if '-f' option is not used to specify file. Problem: sm_cmd do not see difference between "-f file|mail|option". Solution: Open and read file only if "-f file" is used bugfix: sm_cmd should make difference between "-f file" and "-f option" bugfix: sm_cmd core then failed to open/read file passed with -f bugfix: use official sbt provider in build.sh bugfix: (#29983) Fixed execution of Hyper-V mount PC script. When copy to RDS client is not needed, then set script file on the local server to be used for mount NEW: (#24735) move sbc_vadp.jar from gui folder to sesam folder during installation and update. change: sbc_rhev: add prefix to log messages related to hotadd functionality for easier analysis of logfiles change (#29935): sbc_rhev: add new options to help output Add quotes for '-i' paramter for 'sbc_proxy' call bugfix: Remove superfluous DB query with empty SQL statement bugfix: Do not use 'sctRL' struct (used by select in outer loop) to update of 'DB:results.throughput' in case of source side deduplication. bugfix: Typo in 'I004-STARTUP Starting SEPsesam service' message BUGFIX: (29947) SQL syntax error caused by [='] in statement. Solution: Skip already processed string BUGFIX: (#5884) Copy saveset information into target store after migration using correct interface. Former change has been overwritten change: (#29970) Preset DB:migration_results.status='X' to avoid empty status in case of DB problems bugfix: (#29978) STPD_Main cores for unknown reason at ConnectionHandlerCb. Problem: sockaddr_storage struct conversation may fail, as already seen on Windows. Solution: Manually copy sockaddr_storage structure bugfix: use switch -e instead of -f to include a cmd from the skel folder. change: insert additional log messages including method names to easen error analysis. Rename some variables since they were also used in batch files that are called which leads to problems during build. Improve error handling. Errors are thrown and build is aborted for any error that might occur. change: if pwd is already set do not set it again. BUGFIX: (#29881) Handling WSAStartup call. new: 'I004-STARTUP Starting SEPsesam service' message new: DB:data_stores.options in data store class change: (#29724) Fixed preparation if guestmount command on Windows bugfix: add object corruption store to the pcck master arguments list bugfix: (#29886) create /tmp-pages directory to store pages during gc; clean up tmp files on server start; use Files.size as bounder in pageBodyReader instead of FileChannel.size new: (#29914) Check if data store is available before starting 'sm_remove_saveset purge' bugfix: (#29965) Core dump in sm_cmd, if '-f' option is not used to specify file. Problem: sm_cmd do not see difference between "-f file|mail|option". Solution: Open and read file only if "-f file" is used bugfix: sm_cmd should make difference between "-f file" and "-f option" bugfix: sm_cmd core then failed to open/read file passed with -f bugfix: (#29960) do not try to execute vmdk cleanup for Nutanix umount, everything is performed by the sbc_proxy umount BUGFIX: (#29963) Execute alive test on data mover defined on task instead of client level change: Use common sbc message for logging VM backup transport mode NEW: (#20629) Enforce FULL backup: Change message in NOT log (#29941) Variable snapshotType was not set when Nutanix backup was started with qui=0 (#29941) Nutanix backup default warning on crash consistent backup, qui=0 for no warning BUGFIX: (#29934) Replication for source pool spawning over multiple data stores. Set source side drive parameter correcltly according to saveset location (#29957) Fixed mount of Nutanix using migrated saveset. Original saveset was written to LSL instead of migrated. (#29961) Implemented -A delsnap function for Nutanix new: erase all x86 sections from SP creation since x86 is not supported anymore. Adjust all build sections to use Visual Studio 2019 specific build routine. Erase single file build section since the basis for the build is now the targets.txt file and all targets are treated equally. Inserted many examples in the usage. new: a file that contains all possible build targets für SP build. bugfix: fix repair command family for Si3-NG di bugfix: (#29960) do not try to execute vmdk cleanup for Nutanix umount, everything is performed by the sbc_proxy umount BUGFIX: (#29963) Execute alive test on data mover defined on task instead of client level change: Use common sbc message for logging VM backup transport mode NEW: (#20629) Enforce FULL backup: Change message in NOT log (#29941) Variable snapshotType was not set when Nutanix backup was started with qui=0 (#29941) Nutanix backup default warning on crash consistent backup, qui=0 for no warning BUGFIX: (#29934) Replication for source pool spawning over multiple data stores. Set source side drive parameter correcltly according to saveset location (#29957) Fixed mount of Nutanix using migrated saveset. Original saveset was written to LSL instead of migrated. NEW: (#29935) if -a use_virtioscsi is set, use ID_SCSI_SERIAL instead of ID_SERIAL to identify attached disks (#29961) Implemented -A delsnap function for Nutanix new: erase all x86 sections from SP creation since x86 is not supported anymore. Adjust all build sections to use Visual Studio 2019 specific build routine. Erase single file build section since the basis for the build is now the targets.txt file and all targets are treated equally. Inserted many examples in the usage. new: a file that contains all possible build targets für SP build. bugfix: fix repair command family for Si3-NG di NEW: (#29935) add option -a use_virtioscsi to allow disk attachment to data-mover via bus type VIRTIO_SCSI, default is VIRTIO bugfix: move tmp object file to trash if fails during backup; repair pages marked missing on occk bugfix: (#25155) - Support of Exchange 2019. Accept ExchangeServer 2019 as valid exchange server version bugfix: vm_folder name missing last character change: sm_qm_main refactoring check_expired_jobs function bugfix: (#29783) - introduced mechanism to flag log entries. Mechanism prevents flooding of messages to qm main log file bugfix (#29928) Remove Nutanix disks from VM config before creating restored VM also when protocol is NFS NEW: sbc_rhev: add option -a use_dev: searching for hot-add attached devices is then allways attempted via search in /dev/disk/by-id/ instead of using udev bugfix: (#29798) Exchange Server Backup - Log truncation after VSS_S_ASYNC_PENDING. Problem: Commit 8fe72f4 change Error processing workflow, which may cause issue. Solution: Revert recent changes bugfix: (#29941) Sesam defaults to CRASH_CONSISTENT Snapshots during Nutanix backup and finishes with warning if VM guest tools are not enabled bugfix: Call 'QUE_DEL_ENTRY_EXT()' function in 'FINISH()' of 'sm_restore' to release semaphore so 'sm_sbc_com_ext' will terminate correctly bugfix: 29946 - Do not fallback to Nutanix backup type NFS if iscsi data service cannot be connected. Exit with error instead bugfix: (#29943) Fixed restore of a VM with CDROM. In function removeDisksFromVMSpec, original disk list is removed from VM spec. Fixed so that If CDROM is found in the list, it is added back after disks are removed. But it is added as en empty CD, without the image reference. new: Add message tag for 'E001_DS_READ_CHECK_ERROR' message BUGFIX: (#29949) Do not update 'DB:results.data_size' by 'sm_sms_watch' afer after backup has already finished bugfix: (#29945) - si3ng function called after destructor change: (#29931) Changing global variables must be done atomic. Create 'tmp' file and use atomic 'rename()' function to create file bugfix: (#29936) Fixed preparation of meta files for VMDK mount of sub type IMG, on a Windows mount client and Linux Sesam server BUGFIX: (#29940) 'sm_sms_watch' resend notification for data store corruption every minute. Use correct variable for SQL statement BUGFIX: (#29937) Migration log file cannot be loaded by GUI when main job is started before newday, but child job after newday. Use Sesam date set by parent job BUGFIX: (#29939) Migration or replication job tries to copy savesets from former failed migrations. Check for DB:result_lbls.segment<>'-1' NEW: (#29935) add option -a use_virtioscsi to allow disk attachment to data-mover via bus type VIRTIO_SCSI, default is VIRTIO bugfix: move tmp object file to trash if fails during backup; repair pages marked missing on occk bugfix: (#25155) - Support of Exchange 2019. Accept ExchangeServer 2019 as valid exchange server version change: sm_qm_main refactoring check_expired_jobs function BUGFIX: (#29324) Additional data size check during saveset replication: Do update former failed 'DB:migration_results' if insert into 'DB:result_lbls' hasn't been done bugfix: (#29783) - introduced mechanism to flag log entries. Mechanism prevents flooding of messages to qm main log file bugfix (#29928) Remove Nutanix disks from VM config before creating restored VM also when protocol is NFS NEW: sbc_rhev: add option -a use_dev: searching for hot-add attached devices is then allways attempted via search in /dev/disk/by-id/ instead of using udev bugfix: (#29798) Exchange Server Backup - Log truncation after VSS_S_ASYNC_PENDING. Problem: Commit 8fe72f4 change Error processing workflow, which may cause issue. Solution: Revert recent changes NEW: (#29930) Adopt source side directory for restore via mount for dynamic backup saveset selection bugfix: Set task group flag correctly in 'sm_sepul_event' NEW: (#20629) Enforce FULL backup, if latest FULL has failed and DB:task_events.enforce_full='1' set change: Exit trace line in 'sm_sbc_com_ext' after 'PROT()' bugfix: Set task group flag correctly in 'sm_sepul_event' new: Template based restore using VD mount: Do not try to adopt SEL file bugfix: change s3 http client configuration bugfix: (#29925) Sandbox restore is successful, but analysis log function turns state to failed. Do not print sbc_proxy version log line, because 'ANALYZE_RESTORE_LOG()' starts reading after last one change: Do not print output of sbc_proxy to stderr when calling toolcheck nfs from sm_vmdk_interface attac. Changed out parameter for sm_rexec bugfix: change s3 http client configuration change: (#29114) VMDK mount: Set full CBT flag automatically, if backup sub type isn't set to '_IMG' for level copy backups too change: (#29114) VMDK mount: Set full CBT flag automatically, if backup sub type isn't set to '_IMG'. Correct mount of copy _IMG backups change: (#29114) VMDK mount: Set full CBT flag automatically, if backup sub type isn't set to '_IMG' change: (#29924) in case nutanix module is unable to correctly receive information of a container, object is not subscriptable. Handle Type Error and set usage/total to 0 in order to avoid exception BUGFIX: (#29919) Oracle backups are not encrypted despite encryption option for task is enabled bugfix: If sm_backup is started manually then set sesam_date and update results.state='a' to allow proper further processing, e.g. use of correct NOT file new: GLBV 'gv_conf_num_threads_remove_saveset' to set number of threads used by 'sm_remove_saveset'. Can be overwritten by '-N' option new: Add description for new 'GLBVs' Add new message 'E022_BASICS_STOP_SESAM_BELOW_FREE_SPACE' bugfix: (#29854) Fixed tool check for VDDK attach on Windows. We must check for WinNFSd instead of exportfs. Removed block of code in vmdk.py where exportfs is set to command to be executed on Windows. That code was never executed.' bugfix: (#29916) Oracle restore via HTTP becomes very slow. Problem: Increasing CURLOPT_BUFFERSIZE buffer cause longer wait in XBSA/cURL for more data until it will be passed for restore and this wait drastically impact restore performance. Solution: Do not change CURLOPT_BUFFERSIZE, leave it equal to default 16kb NEW: (#20629) Enforce FULL backup, if latest FULL has failed and DB:task_events.enforce_full='1' set change: cleanup binaries before build change: (#29835) copy binaries to subfolder so build job can add them to jenkins artifacts bugfix: If sm_backup is started manually then set sesam_date and update results.state='a' to allow proper further processing, e.g. use of correct NOT file bugfix: (#29912) core dump with sm_stpc list. Problem: Out of buffer cause core. Solution: Fix iteration over buffer change: (#29870) Canceled BSR backup is removed from DB:results. Do not call 'remove saveset' function in case of BSR Windows. Correct fix would be to remove this call on BSR side bugfix: (#29908) Finished HTTP/HTTPS connections remain in the 'syst' output of STPD. Problem: STPD_HTTP PID still remain. Solution: STPD_HTTP PID should become zero then process exit bugfix: (#26757) Recursive deletion of files older then X days. Problem: "sm_client drop help" core. Solution: Additional check for correct path set bugfix: (#29652) Exchange databases get backed up twice when at least two databases are specified as backup source. Problem: Its possible to add duplicate items into backup list. Solution: Check for unique before list insert bugfix: (#28876) increase api call attempt timeout change: (#29854) Added tool check for exportfs before executing VFS mount for VMDK attach operation bugfix: Cut long datastore messages to fit into DB:data_stores.msg bugfix: (#29850) DI Si3-NG related changes according replication abort and replication show bugfix: (#29850) add flag "release" to replicate show command to clean up sessions data bugfix: (#29783) - qm_del_entry - prevent using possibly obsolete pointer CHANGE: (#29892) - Database of users should be checked and cleaned before an update is carried out. change: Added tool check for exportfs before executing VFS mount for VMDK attach operation bugfix (#29451) Windows start/stop of sesam service may run into troubles. Problem: Commit c8f32d6 removed the semaphore which assures that there is only one sm_startup process running and which is checked in sm_setup.c. Note: Seamphore 'sepuler' is only blocked for a while to let sepuler wait for sm_startup new: (#29890) Avoid parallel run of purge processes for one data store in 'sm_remove_saveset' (has been implemented in sm_data_store already). Do not check semaphore, if one saveset (-s ) is given change: (#29870) Canceled BSR backup is removed from DB:results. Do not call 'remove saveset' function in case of BSR Windows. Correct fix would be to remove this call on BSR side bugfix: (#29858) - add size of EA ACL stream to relative offset. Complete write LIS line when data part arrives new: (#29890) Avoid parallel run of purge processes for one data store in 'sm_remove_saveset' (has been implemented in sm_data_store already). Do not check semaphore, if one saveset (-s ) is given bugfix: (#29850) fix abort fucntion; add support of replication abort by object id; add sid for pending replications bugfix: (#29844) SM_SSH process does not terminate the executing process when signal is caught. Problem: if sm_sshd child terminated executable keep running on Windows. Solution: Use JobObject to manage whole process tree from sm_sshd main NEW: (#29825) Do not call 'sm_main reload sms' in 'sm_break2' if called during Sesam server shutdown BUGFIX: (#29884) Execute 'sm_os_startup stop' in 'sm_shutdown' even if 'sm_main' is already down to stop Postgres server started during installation