bugfix: (#25314) - With SP2 the Exchange backup works but due to missing eseutil output it is marked as failed. Cause: side effect of the small refactoring due to commit 955721182a NEW: (#22512) Allow DIFF backups for task type 'Kopano' bugfix: Correct check for DIFF backup during restore NEW: (#22512) Differential backup for Kopano mail server wip: (#22512) Diff backup for Kopano. Problem: Selective restore of several user mailboxes or folders with DIFF backup may fail due to repeated CREATE_SLL_LINES(). Solution: Remember already processed users BUGFIX: (#25017) Migration of a DIFF or INCR increases the EOL of the original FULL. Problem: On linux checking GLBV gv_adjust_eol_migration_increases_eol_on_other_pool returns true. Solution: Improve GLBV check >CHANGE: (#21832) Add possibility to increase EOL of another media pool for migrations. Solution: If default key 'eol_adjust_migration_on_other_pool' is set then allow to increase saveset's EOL with maximum eol on other pool in case of migration BUGFIX: (#21527) EOL Adjustment for Migration Tasks failed if previous FDI chained save set is concurrently migrated. Problem: Since commit 961522d the logic which allows a backup to increase the EOL of a saveset on another pool was skipped CHANGE: (#22320) EOL adjustment: Increase backup EOL must only extend the EOL of save sets which have already longest EOL. Solution: Change result_lbls.eol to given backup EOL according to given EOL > backup EOL then increase only the saveset EOL with the longest EOL else decrease all saveset EOL with saveset EOL > given EOL BUGFIX: (#24618) Adjust EOL returns with error, if only a migrated saveset exists NEW: Append EOL changes made by user to backup log BUGFIX: (#23969) After replication the original media EOL is updated too. Solution: Add media pool in SQL query to estimate the labels BUGFIX: (#24448) Failed XEN backup is recognized as successful by ANALYSE_BACKUP_LOG() function. Check, if there is SBC success message from main task BUGFIX: (#21527) EOL Adjustment for Migration Tasks failed if previous FDI chained save set is concurrently migrated. Solution: Do not update EOL of concurrently migrated savesets BUGFIX: (#21527) EOL Adjustment for Migration Tasks failed if previous FDI chained save set is concurrently migrated. Solution: Check if missing saveset is concurrently migrated BUGFIX: (#25235) - Sharepoint backup processes the content databases 20 times after installation of servicepack 2019-1. Cause: file duplication bug in case of backup few VSS components specified by the backup source. bugfix: compilation error: error C2552: 'dataInit' : non-aggregates cannot be initialized with initializer list 'TStaticString' : Types with user defined constructors are not aggregate BUGFIX: (#24804) - Hyper-V INCR restore with custom snapshot fails with VSS_E_BAD_STATE. Caused by the multi-xml files in the saveset because of bugged "RepeatOnError" function. change: improve trace mechanism by the DEBUG_TRACE macros allowing to trace minor logs change: Saveset migration: Do not use drives with acces mode 'write' as source drive CHANGE: (#24956) Specified source drive for saveset migration is overwritten by drive with access_mode 'read' BUGFIX: (#24345) Migration main log is written twice. Problem: Since commit d92fe07 the file_append() function is used to copy content from log file to NOT file. Solution: In case of migration the NOT file is already written NEW: (#25168) Data store FSCK: Add number of savesets with corrupted data files to overall summary message Comment out 'rcsid' BUGFIX: (#25253) backint for Oracle: Corret parsing of device, pool and task values returned by 'sbc_com', if values contain option parameter ('-d,'-m','-j') bugfix: Compile failed due to compile rule with new (in master) sm_kill_lib.c change: New macro define STARTSI(x,y) !strncasecmp(x,y,strlen(y)) change: sbc_ndmp - fix of multi session restore BUGFIX: (#25218) Migration crash during EOL calculation of a very long FDI chain. Solution: Incresae the stack size for sm_sms_copy.exe and sm_sbc_com.exe change: Enclose x_DumpEx(...) in #ifndef SESAM_VFS ... #endif to avoid use in sm_vfs BUGFIX: (#25058) Selective NDMP restore from tape takes an unreasonable amount of time. Problem: The first sbc_ndmp data request is started without following tape flag to read the saveset header however further data requests are started with the following tape flag. Solution: Do not implicite set the FTF flag bugfix: sbc_ndmp restore failed with error 'sbc-1500: Error: sbc_parse_item: Error: Maximum number of items in array is reached (10)'. Solution: sbc_update_dataarg() the arry char *cpField[10] must be initialized with NULL values change: Extra trace in ta_set_position(), e.g. 'ta_set_position:: command MTFSF, count 1' bugfix: (#25058) NDMP Restore takes a long time (4 GB isn't ready after many hours). Avoid extra getlabel for each data request bugfix: Initialize variables by empty string to avoid broken SQL statements bugfix: (#25034) - additional prefix check for NDMP. Used when making browse for NDMP NEW: (#25035) Support operating system Windows Server 2019 wip: (#25058) Selective NDMP restore from tape takes an unreasonable amount of time. Problem: Reading from tape with buffer requires to set the size for the buffer in the RETR command (here: NDMXBS_BUFF_SIZE 2MB) - otherwise the EOF check in BSAEndData() will read one extra byte change: sbc_ndmp - moved defines to header file change: sbc_ndmp - moved defines to header file change: sbc_ndmp fixed log message change: sbc_ndmp during restore to Netapp disable ENHANCED_DAR_ENABLED for netapp version < 8 change: sbc_ndmp - log NDMP appliance OS type and version change: sbc_ndmp - temporary set record_size back to 10K - it is causing problems during backup bugfix: (#21353) NDMP restore is much slower than the backup (Solution: Increased NDMP data mover record size from 10K to 128K) BUGFIX: (#25058) - Fixed positioning of the saveset read after DAR requests received from NDMP server change: sbc_ndmp - fixed selective restore from local data store change: Fixed logs in sbc_ndmp change: NDMP dir command: Return items with prefix 'NDMP:' instead of '/NDMP:' change: 'sbc_ndmp - selective restore from local saveset' change: sbc_ndmp adds RETR flag n for subsequent backup reads after the first read. change: Do not report 'mtime' and 'ctime' during NDMP dir operation, as these values aren't set bugfix: (#25061) sbc_ndmp core dump during browse. Solution: Check if spSMS->cpSaveset == NULL bugfix: Option -o next was not parsed. Solution: Call sbc_checkargs( spGlobal ) to parse arguments for RETR command BUGFIX: (#24860) All daemons except the sds have been set to on after update no matter what the original value was. This section is obsolete hence it is erased. All original values will be kept from now on during update. bugfix: if -x ALL is set for CONFIG backup no download of config files from VC should be done at all bugfix: if -x ALL is set for CONFIG backup no download of config files from VC should be done at all BUGFIX: (#25058) Selective NDMP restore from tape takes an unreasonable amount of time long time. Problem: sm_restore must not start with 2nd tape in case of NDMP bugfix: (#25108) Core dump in 'sm_restore', if 'LL_SavesetInfo' list hasn't been allocated BUGFIX: (#24903) NDMP restore with 4.4.3.64 does not work in case of DIFF/INC backup. Problem: Since bugfix '(#23416) Selective Restore from migrated NDMP saveset failed ...' the original saveset is required. Solution: Set to given saveset if original_saveset is not set change: Check if Generation LIS file is uncompressed BUGFIX: (#24846) Adjust Generation lis file for Hyper-V generation restore. Problem: In case of non-overwrite the restore must skip the older avhdx files. Solution: Create generation lis without older avhdx lines. Compare lines with create and modification time in sm_restore (BLOW_UP) BUGFIX: (#24720) Selective restore failed when data is selected from the second tape and a parent directory from first tape is added. Solution: In this case the parent directory label must be checked. BUGFIX: (#24551) SEP sesam service shuts down if DataStore directory runs out of free space. Solution: Writing to a file with sm_client must not use the kernel macro cm_fclose() which shuts down SEP sesam server if file system is full BUGFIX: (#25058) Selective NDMP restore from tape takes an unreasonable amount of time long time. Solution: In case of NDMP the segment file was not used to allow fast tape positioning NEW: (#25064) New BSR registry key for DiskImage 14.0 is missing. Add registry key detection for version 14.0.xx to executable sbc_bsr.exe and sm_client.exe bugfix: (#24258) Merge to v4_4_3_4 branch CHANGE: (#25095) Increase wait time to get backup stream for migration from 1 to 10 hours wip: (#24109) VSS path backup fails with 'Format logical block size is too small to store descriptor block data.'. Problem: Sparse file processing does only restore up to 18 ranges. Solution: Use SegmentHeader buffer to read data buffer with ReadSH() to store partially buffer and to return MTF_MORE_BLOCKS. wip: (#24109) VSS path backup fails with 'Format logical block size is too small to store descriptor block data.'. Problem: Sparse file processing does only backup 18 sparse ranges. Wip: If buffer was not sufficient then the offset in the Sparse Frame Header was not stored bugfix: Typo in trace line wip: (#24109) VSS path backup fails with 'Format logical block size is too small to store descriptor block data.'. Problem: Sparse file processing does only backup 18 sparse ranges. Wip: Check return from CopySH() and return MTF_MORE_BLOCKS if buffer size is not sufficient bugfix: (#24225) - Hyper-V full backup during pre backup merge must set the backup state to warning if user snapshots exist. Just print out warning without touching user snapshots. bugfix: (#24225) - Hyper-V full backup during pre backup merge must set the backup state to warning if user snapshots exist. BUGFIX: (#24654) - sbc sporadically produces an application error during VSS file backup on Windows. Cause: some CVssComponent variables are not initialized in the correct way, QueryStatus() should throw error in case of non-sufficient returned code. change: wrap the IVssAsync into class BUGFIX: (#24654) - sbc sporadically produces an application error during VSS file backup on Windows. Handle "timeout" errors for "PrepareForBackup" and "DoSnapshotSet" functions. BUGXIX: (#24855): sm_copy: Set string for Sesam main log message in case of warning, to fix core dump. TODO: Use real message, returned by sub copy task bugfix: Build for Windows x86 failed since commit 3f71c17 (#23448) Change the library build switch from NEWLIB to LIBVER and set it to 1 by default. Solution: Undefine LIBVER in case of x86 in sms\sms_interface\makefile.win32 BUGFIX: (#24982) Linux 'syslog' complains about 'unknown facility/priority', if Sesam disaster message becoms longer then 255 bytes CHANGE: (#24982) Add a space between '-R /' and newline in sbc disaster restore command, to avoid parsing errors from 'syslog' CHANGE: Java verion == 1.8 as requirement instead of >=1.8 BUGFIX: (#24970) SM_SMS_WATCH does not dismount tape when copying LIS from RDS needs more than 15 seconds. Solution: Increase summary and unload counters in this case CHANGE: (#23448) Change the library build switch from NEWLIB to LIBVER and set it to 1 by default. Using NEWLIB=0 builds using the old libraries. BUGFIX: Correct check of Sesam type (client, RDS..) used to detect, if source side deduplication is allowed. Check has returned 'false' for Linux every time Set build target from command line change: sbc_ndmp prints xbsa messges too often on windows bugfix: (#24374) - sbc_ndmp ignores printing message Could not delete ord restore meta file ... change: safer way to log 3108 messages bugfix: (#23730) - sbc_ndmp - do not call bsa init on iterative restores wq!bugfix: (#23730) - sbc_ndmp does not terminate BSA transaction every time. Only at the end of the session bugfix: (#23730) sbc_ndmp cores during selective NDMP restore bugfix (#21246) NDMP Oracle ZFS restore failure (Windows build) bugfix (#21246) NDMP Oracle ZFS restore failure BUGFIX: (#21291) Missing files in backup saveset of Oracle ZFS bugfix: (#24656) Store last mover seek position and use it for further comparison. Some variable renaming and reinitialization bugfix: (#24271) NDMP restore does not trace the restored items in the common way change: Welcome message for sbc_ndmp set to 'SESAM BACKUP CLIENT FOR NDMP ON Windows/UNIX' wip: (#14372) Add unique ID to RETR command to allow break of saveset operation without aborting concurrent restore or migration. Solution: Add '-i {saveset_information}' to the RETR command bugfix: (#23491) - sbc_ndmp does not always process buffers correctly. Buffer processing was rewritten for restore session. Additional logging was added (checksum). sbc_ndmp will report warnings/errors in case that some missing or corrupted block was found change: warnings cleanup change: improved sbc_ndmp build - no need to clean on changes in sesam_job.h bugfix: (#23735) - NDMP backup of NetApp SnapMirror volumes bugfix: add Nuitka support for sm_ssh utils bugfix: scc/python clean does not executed from main makefile new: add Nuitka support for sm_ssh utils bugfix: (#24849) SM_SSH connection sporadically fails with "Error reading SSH protocol banner". Problem: sm_sshd sub-thread keep running if all tasks finished. Solution: Improve logging thread termination BUGFIX: (#24849) SM_SSH connection sporadically fails with "Error reading SSH protocol banner". Problem: Logging from multi-process may cause dead lock and application termination by timeout which handled by Paramiko as missing banner error. Solution: Revise SM_SSHD logging, allow logging only from main process to avoid race conditions BUGFIX: (#24849) SM_SSH connection sporadically fails with "Error reading SSH protocol banner". Problem: Logging from multi-process may cause dead lock and application termination by timeout which handled by Paramiko as missing banner error. Solution: Revise SM_SSHD logging, allow logging only from main process to avoid race conditions BUGFIX: (#24851) - Backup of system_state may fail occasionally. Missed commit f733b04a328aa5a2b052fb93756b7c5ada3571f8. bugfix: (#24653) - add support for LTO-8 and LTO-M8 change: Citrix Xen - After VM is imported on COPY based restore, search for full template including savesetID and also take care for old backups compatibility BUGFIX: (#24600) Queue manager may not work anymore after newday. Problem: Pipe to_qm_pipe is removed by sm_qm when answer pipe from_qm_pipe... cannot be opened. Solution: Do not remove pipe to_qm_pipe Script to build Sesam release by Jenkins BUGFIX: (#24834) Do not remove files from 'smslis' directory on RDS older than 1 day, because there can be files from backups running longer. Now default EOL of 7 days is used, which should be sufficiant change: Citrix Xen - when overwriting existing VM, do not try to delete existing VM for the second time in COPY level restore new: Target 'build_static' to build all executables into subdirectory 'patch' with static linkage change: Citrix Xen - fixed processing of sbc arguments in sbc_proxy change: Citrix Xen - after existing VM is removed for restore with overwrite, do not try to import disks from original VM bugfix: (#24739) Restore of citrix copy backup fails with error: Curl: operation aborted by callback. Problem: Restore marked as failed. Solution: Revise log level bugfix: (#24739) Restore of citrix copy backup fails with error: Curl: operation aborted by callback. Problem: Citrix restore always finished using CURL_READFUNC_ABORT. Solution: Improve Citrix restore logic bugfix: (#24739) Restore of citrix copy backup fails with error: Curl: operation aborted by callback. Problem: Citrix restore always finished using CURL_READFUNC_ABORT. Solution: Improve Citrix restore logic bugfix: (#24772) - Citrix Xen restore tried to remove non existing (null reference) VM when overwrite was set but there was no matching VM to be overwritten. BUGFIX: (#24170) Oracle: Control file restore from AUTOBACK doesn't work BUGFIX: (#24257) xen module does not handle master/slave connection the right way change: fixed merge issues in XEN module BUGFIX: (#24371) XEN backup fails after NBD client cannot connect to server BUGFIX: (#24782) Problem: Files in the skel\rythm folder are not replaced with the new ones during update. Solution: Add a rule to remove existing files and then insert new ones in the skel\ryhtm folder during update. This rule is needed since the files in these folders are socalled unversioned files and those files are not replaced if they already exist. BUGFIX: (#24770) manually specified restore drive is overwritten by drive with access_mode 'read'. Solution: Implement one common function for migration and restore and avoid overwritting specified drive bugfix: (#19889) Backup of Citrix XEN can fail, but is reported as successful. Problem: "False" errors could be detected which cause backup abort. Solution: Check HTTP result code for last chunk only bugfix: (#19889) Backup of Citrix XEN can fail, but is reported as successful. Problem: "False" errors could be detected which cause backup abort. Solution: Check HTTP result code for last chunk only new: changed version number from 4.4.3.64 to 4.4.3.65. BUGFIX: (#15188) - Rescan Citrix Xen Server Storage before start of restore * BUGFIX: (#19267) - XEN: restore does not work if original VM exists and is renamed BUGFIX: (#17611) - restore: support FORCE option to workaorund possible problems with checksums in xva files bugfix (#24214) - FULL backup of XenServer 7.1 CU1 throws an 'Unhandled exception' while Enabling NBD on network interfaces CHANGE: (#24587) Enlarge 'DB:accounts.password' to CHAR(512). All other password columns are changed to same lenght too. NEW: Speed up SQL statement to get orphand migration 'DB:results' entries BUGFIX: (#24698) Restore shows message ''invalid state 'X' in table migration_results'' if state X is in table. Solution: Add constraint and state in ('0','1') to get only valid replicated savesets NEW: Speed up SQL statement to get orphand migration 'DB:results' entries CHANGE: Do not reset operating system 'OES-Linux' to 'LINUX'. 'OES-Linux' will enable 'Micro Focus (Netware)' credential tab in GUI, but it will not be visible for LINUX any more bugfix: make possible to compile with unique list items to be backed up. Missed some system writers during the SYSTEMSTATE, code optimization. new: improve DI to print all alive objects existing in S3 snapshots (di s3 object list alive) Sesam GUI... wip: Fix failing unit test wip: Fix failing unit test change: "Main Log" of migration result is editable and writes from right to left BUGFIX: (#25130) - For host-backup and meta-data the backup type and source are wrong BUGFIX: (25241) - After changing the media pool in a backup event read only drives are also selectable BUGFIX: (#25162) - Drive selection is disabled for immediate start of a saveset migration from results wip: (#24687) - GUI with parameter -M1 only the master is shown in the table view wip: (#24687) - GUI with parameter -M1 only the master is shown in the table view wip: (#24687) - GUI with parameter -M1 only the master is shown in the table view BUGFIX: (#24687) - GUI with parameter -M1 only the master is shown in the table view wip: Fix yearly failing term calculator test wip: Fix yearly failing term calculator test BUGFIX: (#24791) - Wrong BackupLevel is displayed - FULL instead of INC bugfix: (#25040) - ClassCastException open "edit multiple drives" dialog BUGFIX: (#25020) - Trying to view a cancelled readability check results in GUI NullPointerException BUGFIX: (#24896) - Manually changing task type causes an exception BUGFIX: (#24760) - Allow change of RDS, if a new drive for a PATH data store is added wip: (#24606) - extend AD authentication to use free selectable OUs for AD user and groups wip: (#24606) - extend AD authentication to use free selectable OUs for AD user and groups wip: (#24606) - extend AD authentication to use free selectable OUs for AD user and groups BUGFIX: (#24606) - extend AD authentication to use free selectable OUs for AD user and groups BUGFIX: (#24830) - It is not possible to change the 'Path for data store' in DATA STORES - DRIVE properties, of an existing drive wip: Fix failing unit tests wip: Fix failing unit tests change: Synchronize report style sheet between master and v4_4_3_4 change: Add extra image for active backups to the report BUGFIX: (#24752) - GUI option "Manage VM tasks" allows to mark multiple VM tasks without existing VM for deletion, but it just deletes one wip: Fix failing unit tests. change: (#24734) - Filling of vSphere VM buffer doesn't work reliable change: (#24734) - Filling of vSphere VM buffer doesn't work reliable change: (#24734) - Filling of vSphere VM buffer doesn't work reliable change: (#24734) - Generation of vSphere tasks: Section "VM tasks without existing VM" includes entries for exiting VMs change: (#24734) - Generation of vSphere tasks: Section "VM tasks without existing VM" includes entries for exiting VMs change: (#24734) - Generation of vSphere tasks: Section "VM tasks without existing VM" includes entries for exiting VMs change: (#24734) - Generation of vSphere tasks: Section "VM tasks without existing VM" includes entries for exiting VMs wip: Fix potential NPE in VMBuffer change: (#24734) - Generation of vSphere tasks: Section "VM tasks without existing VM" includes entries for exiting VMs change: (#24734) - Generation of vSphere tasks: Section "VM tasks without existing VM" includes entries for exiting VMs BUGFIX: (#24740) Set source drive and source interface in migration events dialog read-only (in case of an Immediate start) BUGFIX: (#24724) - Suppress flag for migration events isn't written to DB wip: (#24733) - The interfaces to the drives is no longer correct BUGFIX: (#24733) - The interfaces to the drives is no longer correct BUGFIX (#22885) - Adding to much text to a replication-task-comment gives an exception when the replication is started BUGFIX: (#24693) - Exchange path restore may overwrite original data when a new "New restore target" is selected BUGFIX: (#24706) - Task type selection "Hyper-V" ends in "NullPointerException" BUGFIX: (#24704) - NPE for unknown reason when changing task type wip: Fix failing unit test CHANGE: (#24587) Enlarge 'DB:accounts.password' to CHAR(512). All other password columns are changed to same length too. BUGFIX: (#24628) - NPE when expanding properties of a rds single tape drive BUGFIX: (#24652/#24613) - Migration using same drive group and multiple drives in group isn't possible any more BUGFIX: (#24158) - RHEV task type isn't shown by client source browser any more bugfix: IllegalArgumentException: New row height less than 1 bugfix: (#24190) Browing 'Citrix XEN' does not work correctly anymore. Problem: Since commit a4ac75d ''(#24158) RHEV task type isn't shown by client source browser any more'' the CITRIX XEN is mapped to RHEV. Solution: Correct logic