Changes 4.2.2

From SEPsesam
Jump to: navigation, search

You will find complete GIT/CVS change reports here.

Major fixes and changes v4.2.2.40

  • BUGFIX: (#15978) If restore option for exchange database restore was set with wrong count of Quotes (") then the restore tried to overwrite the original database

Known Issues

  • Issue: (#16876) Windows UNC path backup does not exclude directories after update to 4.2.2.35. Now the exclude_type 'pattern' is the default but this exclude type does not skip the 'UNC' prefix to match with \\server\share\path. Workaround: Use exclude_type 'regexp' for Network Shares.
  • Issue: (#16913) Windows exclude patterns with regular expressions are modified, e.g. correct regular exclude pattern "\.ini$" is mapped to "\\.ini$".

ORACLE Java 1.7 Update 51

  • BUGFIX: (#16793) The tmpdir of java is now set to ^%TEMP^% instead of the SEP sesam work dir to prevent possible not existing write permissions. The desktop GUI link will now always be replaced when executing an update.

GUI

  • CHANGE: (#16790) accept 'Netiq Directory' as path element from 'sm_sho' and use 'Novell Directory' logic at browser
  • BUGFIX: (#16867) GUI must set Exclude List for exclude_type 'regexp' with ending '/$' in case of Windows

Kernel

  • BUGFIX: (#16809) Since version 4.2.2.35 the automatic insert of hardware tape drives during installation into the Sesam database is not executed. sm_auto_conf_hw failed with DB_ACCESS returns 0 msg=near 'Bandlaufwerke': syntax error. The first argument must not be enclosed within quotes.

SBC Windows

  • SBC_WIN32_VERSION=4.2R13: BUGFIX (#16845) After upgrade to 4.2.2.35 the UNC backup gives a warning for every directory - Skip ACL processing for files and directories if option '-o skip_acl' is given.
  • BUGFIX: (#16832) Since version 4.2.2.35 the backup of Source "all" on Server 2003 (x86) processes VssXmlBackupDocuments only. If no 'System Partition Volume' exists then no volume was processed.
  • BUGFIX: (#16838) SBC_EXCLUDE in sm.ini does not work for all entries if '-o excl=regexp'
  • BUGFIX: (#16844) Restore filter with PATTERN does not accept '/' path delimiter - Substitute '/' with '\' if exclude/include with file pattern.
    Example with filter */sbc.exe:
 sbc.exe -r -s @sesam_bin.save -o excl=pattern -R d:/temp/exclude_check */sbc.exe

SharePoint

  • BUGFIX: (#16831) SharePoint Restore with -a dbpath=...,logpath= failed with 'Cannot create item ...' - Path browsed via GUI must be converted to Windows (with '\') before passing to vss.dll

Exchange Server

  • BUGFIX: (#16855) Insert a warning/error message for all Exchange Server clients without required .Net 4 during update of Sesam package. Opening the .Net 4 homepage in case .Net 4 could not be found.

SQL Server

  • BUGFIX: (#16609) MS-SQL Server 2012 Inc Backup loops with Invalid cursor state. Avoid loop: Do not call DB_ODBC_Result() if failure was retrieved
  • CHANGE: Do not fail with state 'cancelled' if an INCR backup is requested but one of the databases has SIMPLE recovery model. Now state 'warning' is returned

SBC_SMDR (NetIQ)

  • BUGFIX: (#16790) eDIR Backup doesn't work on OES11 SP2 because of a name change - Allow backup and browsing with old 'Novell Directory' mapped to 'NetIQ Directory'

Major fixes and changes v4.2.2.35

Version 4.2.2.35 is a Pre-Release (available under testing) with the following known issues:

Known Issues

  • Issue: (#16809) The automatic insert of hardware tape drives during installation into the Sesam database is not executed in version 4.2.2.35. sm_auto_conf_hw failed with DB_ACCESS returns 0 msg=near 'Bandlaufwerke': syntax error.
  • Issue: (#16832) Backup of Source "all" on Server 2003 (x86) processes VssXmlBackupDocuments only. If no 'System Partition Volume' exists then no volume was processed.

ORACLE Java 1.7 Update 51

  • BUGFIX: (#16793) insert the -Djava.io.tmpdir=...work in the Desktop and start menu shortcuts. This is needed since the new java 1.751 does not have write permissions to open the sep.ico file in the sesam directory which is the default of this variable.
  • BUGFIX: (#16792) add read permission to java.io.FilePermission and listen permission to java.net.SocketPermission to the sm_java.policy

License Check

  • BUGFIX: (#16785) SBS License was not recognized on Linux Sesam version 4.2.2.18 - Use '</dev/null' as null device for UNIX

SBC Windows

  • BUGFIX: (#16445) Volume information are not shown in case of all backup: Use sbc-3163 and sbc-3164 to display volume information

Exchange Server

  • CHANGE: Migrate the sbc_es projects to the .NET Framework 4.0 by the creating new one based on Runtime 4.0 and making some adoption to get PS cmdlets runnable

Major fixes and changes v4.2.2.32

Version 4.2.2.32 was a SEP sesam Release Candidate - never released.

ORACLE Java 1.7 Update 51

  • CHANGE: Add 'listen' to permission 'java.net.SocketPermission' because default permissions have be changed by Oracle in Java 1.7.0_51. Without this permission Sesam RMI server could not be started any more using this Java version (except ALL permissions are set)

Notification

  • change: inserted a notification for the new Sharepoint (MOSS) task.

Windows Patch Delivery

Windows Patches are now delivered with SFX Archives.

Hint: If automatic copy of files failed then the patch archive files will be located in the selected directory and the files can be copied manually to the Sesam bin directories.

Installation and Update

  • CHANGE: Backup type 'SharePoint Sites' since version 4.2.2 not longer supported so set protection to '0'
  • BUGFIX: (#16375) file copy of sesam installation file to RDS produces corrupt executable - End with error 'scc-1132: Error: STDIN channel broken: [nnn Bytes read but only mmm Bytes transferred to remote process stdin]' if transfer is broken
  • BUGFIX: (#16375) iSendBuffer() retry sending up to 300 times
  • change: inserted an update statement for the protection type of the SharePoint Sites. Inserted the ignore '-' switch to the CBT, Email and MOSS notification
  • bugfix: Complete update is done by sm_check_version now and not by sm_startup. Was already changed for Redhat and SuSE
  • bugfix: Check if 'gv_daylbl' can be read and use system state if not set (first startup and RDS)
  • BUGFIX: (#16549) Since postgreSQL version 9.2 the postmaster switch '-c silent-mode=false' is no longer supported hence a version check is executed, and the switch is only used for versions smaller 9.2
  • BUGFIX: (#16454) Set correct package name for Sesam Debian packages < 4.2.2.18
MSI Installer
  • BUGFIX: (#16626) inserted the /norestart switch for remote update. The silent update will not enforce a reboot.
  • BUGFIX: (#16626) Reboot will be forced if update could not replace certain files.
  • BUGFIX: (#16646) Auto-Update shows FILE_CLOSE error: Bad file descriptor - sm_ctrlc_lib.c rev. 1.75-1.77 already fixed that problem
  • bugfix: (#16748) fixed the copy error of the new sesam.dll to the original sesam.dll and inserted a deletion of the new sesam.dll at the end.
  • BUGFIX: (#16748) move and copy the sesam.dll before the actual update starts to prevent a reboot in case of sesam.dll in use.
  • bugfix: change: removed the exit of the installation if the gui link could not be created. Inserted a /D switch to the java -version check call for Windows. (revision 1.667)
  • change: removed Windows Server 2003 and Vista from the server package, only client and gui packages will support those operating systems in version 4.3
  • BUGFIX: (#16605) If the modification time of a zip file belonging to a python executable is not equal to the creation time, the zip file has not been replaced during an update. Now the zip files are removed before in the update procedure to ensure that all zip files are replaced.
  • BUGFIX: (#16581) system tray asks to get administrative rights when trying to start.
  • CHANGE: Windows: To avoid VSS_E_SNAPSHOT_SET_IN_PROGRESS the VSS values must be set to at least: VSS_MAX_RETRY=6, VSS_DELAY=90 and VSS_WAIT_FOR_ASYNC_OP=200
VMware 5.5 - SET_VDDK.PS1
  • (#16637) set_vddk.ps1: Script could not be executed on systems with Powershell 1
  • (#16637) Added additional check of zip file
  • bugfix: (#16637,#16599) Added new function to support VDDK 5.5. Enhanced error handling, if files cannot be moved
  • change: skip the build of the vddk5.5.0 zip file for x86

License Check

  • CHANGE: (#16723) NetApp License should be visible in License output - Violations not yet counted
  • CHANGE: (#16723) NetApp License should be visible in License output - Now license counts NetApp clients with NetApp tasks
  • BUGFIX: (Bug 16693) SEP sesam Backup Client für VMware ESX/vSphere counts Datacenter instead of vCenter. Now scan for 'vCenter Server :'
  • BUGFIX: (#15574) License enforcement for MS-Exchange: Exchange Server are counted via EXCHANGE_NODES ==== MSEXCHANGE + EXCHANGE_DAG_MEMBER
  • CHANGE: (#16469) ORACLE Database Agents must be counted! - Customer needs ORACLE Agent license for every ORACLE Client with configured ORACLE tasks
  • BUGFIX: (#16498) 'iFolder backup requires BSR_Linux license' - Sequence of Backup Types enum did not match with gszBackupLicenses[].eBackupType. Now program 'sesam_make_lic -p' must be used to check Backup Type sequence

GUI

  • bugfix: (#16764) Restore Wizard target browser does not show SQL Server databases - display databases (next level with logical file names)
  • BUGFIX: (#15830) Browse RDBs with Exchange Access fetched via saveset -> tasks.password: 'sm_sho -o restore -t "Exchange Server" -s <saveset> <host> /ExchangeServer:/'
  • BUGFIX: (#15856) Show Pop-up message during the Exchange client source browsing if necessary feature (e.g. granularity level) is not installed
  • change: inserted the rmiHttpPort to the UI section of the sm.ini.
  • change: reinserted the calls for the sm_gui.jar. The sm_ui.jar is used when ui=1 in the sm.ini. Otherwise the sm_gui.jar is used.
  • change: set VM log 'remove deprecated VM...' to 'remove VM...'
  • new: Add update to Sesam 4.3 including switch of Sesam GUI to new UI server
  • BUGFIX: reopen existing exchange restore task with correct information; init combobox 'Import to original database', if use 'existing RDB' fixed (use original for restoretype '?o','?r')
  • CHANGE: (#16732) if restore 'exchange 2013', use tooltip 'single item restore is available' at saveset browser for browsertype 'mail' (show tooltip for type 'Mn')
  • BUGFIX: set correct root path at relocation browser (expert opts) ('/' instead of
  • BUGFIX: set correct root path at target file browser ('/' instead of '/img'), if restore VM with mode 'attach'
  • CHANGE: allow work with GUI, while restore wizard dialog is open; set restore wizard dialog modality type to 'modeless'
  • BUGFIX: (#16278) hide backup type with protection=0 at task type combobox (taskdialog), when open existing task
  • BUGFIX: (#16278) handle task with unsupported backup type, when open taskdialog; hide backup type with protection=0 at task type combobox
  • BUGFIX: (#16713) show checkbox for 'sharepoint server' row at task browser, which set path to '/'
  • BUGFIX: (#16632) select combobox 'import DB to original' for all 'Exchange' restoretypes at restorewizard
  • BUGFIX: run VM action 'detach' fails due to nullpointerexception; check, (get root virtual discs from VM instead of snapshot, if VM has no snapshot)
  • BUGFIX: (#16884) if change 'exchange server' restore to path, hide combobox 'Auto recover and online' ComboBox' at path panel.
  • BUGFIX: (#16683) hide restore type popup, if run exchange restore as path restore with path type:'path' at target browser; switch restore panel depends on selected task type 'path', 'exchange' fixed; set
  • BUGFIX: (#16680) Browsing with button "RDB Browser" must set 'Restore to existing RDB'
  • BUGFIX: show granularity options at taskdialog, if tasktype 'Exchange' is selected
  • CHANGE: (#5958) improve restore target browser; if select incompatible path with current restore type (path flag, dump flag, restore type, flag), open dialog with backup types (same as expert options tab) and force set compatible type or cancel browsing.
  • BUGFIX:(#16656) Closing the immediate start window caused an immediate start of a schedule
  • BUGFIX: (#16651) run full VM backup with each VMDK to own saveset fixed, do not backup each VMDK to own saveset for full backup after backup used blocks of VMDKs
  • BUGFIX: (#16648) create VM task cause nullpointerexception fixed (check, if browser will be opened by restore wizard)
  • CHANGE: (#16601) set label of combobox 'LVM VM modul' to 'Backup with snapshot' and set task.snapshot_flags to 'TM', if selected and task will
  • CHANGE: (#16632) show path+exchange elements at browser of panel 'RDB paths for automatically RDB' (Restore exchange)
  • BUGFIX: init existing restore task with correct values from DB and correct target relocation pathes from sel file; improve performance, if open expert opts dialog; run fill browser data as own thread; init target exchange path with "/" for full restore
  • CHANGE: improve restore exchange, add new checkbox 'import to original DB' and checkbox 'with recover'
  • BUGFIX: (#16621) sm_cmd start restore created to long r_task names
  • CHANGE: (#16617) hide checkbox for browser element, if type equals 'Mn' (restore wizard exchange browser)
  • BUGFIX: (#16545) run CLI migrate from sm_event, which add empty argument at the end of command; allow empty arguments without parameter;
  • BUGFIX:(#16603) do not set '/' before '.' at excluded 'NSS file system' path (clientbrowser); set correct exclude type mode depends on selected
  • BUGFIX: (#16568) Changing the schedule assignment within a migration job was not saved.
  • BUGFIX: (#16573) Avoid InvocationTargetException when deleting a migration task used in a time table
  • BUGFIX: (#16536) data mover combobox is empty, if restore migrated VM backup fixed; (datamover of migrate result was not set, get datamover
  • BUGFIX: (#16572) bugfix: If recovery database name is set within tab 'Start Restore' then the insert into restore_tasks does not set this
  • BUGFIX: (#16571) set correct panel with RDB config, if restart sharepoint restore task and task use mail view
  • BUGFIX: (#16400) set correct location for new client, if create client and selected node is inactive client or master server at
  • BUGFIX: (#16138) prevent throw ArrayIndexOutOfBoundsException, if do right click at empty table header (right side) of restore wizard
  • BUGFIX: (#16187) insert/update datastore with long comment cause sqlexception, limit textfield 'comment' to '255' chars
  • BUGFIX: (#16503) prevent migrate,start backup,edit task at gui results view, if run gui with 'restore permission' rights
  • BUGFIX: (#16557) update commandevent with special chars fixed; set field 'commandevent.command' to '-'; command object contains execution command
  • BUGFIX: (#16556) browse target directory for restore type 'NET APP' at restore wizard cause no data found fixed
  • BUGFIX: (#16548) show/update event.follow_up with empty string at scheduleevents dialog fixed (limit follow_up textfield to 1024 chars)
  • BUGFIX: (#16502) Start time may not be correctly selected when the end time is set in the schedule dialog involved file: ScheduleDialogItemListener.java
  • BUGFIX: (#16452) mark excluded file with '$' at the end to guarantee that excluded file matches exactly (browse client with os 'windows').
  • BUGFIX: (#16452) mark excluded folder with '/$' at the end of path to guarantee, that excluded folder match exactly
  • BUGFIX: (#16409) Avoid possible NullPointerException when viewing the properties of a backup via schedules

QM

  • bugfix: (#16582) If to_qm pipe is missing then recreate it

Kernel

  • BUGFIX: (#16663) Backup uses tape from other media pool without permission (single drive) Check if pool is the requested media pool before calling GET_PREPARED_MEDIA(); change: adjusted the slu calls to the new slu parameters
  • BUGFIX: (#16645) Count directories as not processed items during restore with '-o plain' option, so restore does not return an error, because no item has been processed
  • BUGFIX: (#16560) Now sql update and insert statements are checked for single quotes in values and the single quotes are duplicated to avoid syntax problems
  • BUGFIX: (#16758) sm_config_client called for Windows clients with version ==== 4.x must use gv_ro_bin:sbc.exe as proper backup target
  • BUGFIX: (#16757) sm_java.policy vCenter 'java.net.SocketPermission *:443 ...' disappeared after removing user policy
  • BUGFIX: (#16763) if a schedule had a duration of more than 99 hours only the first two digits of the hours were passed to the sepuler.
  • change: sm_filter -t {file} (tee): Open file delayed only if something read to avoid accidentally truncated file
  • BUGFIX: (#16615): Fix core dump during analyse of log in 'sm_recover'. ( RTASK is set to NULL, if called by 'sm_recover')
  • change: Loader message 'Label' -> 'Medium'. ('Medium is mounted' instead 'Label is mounted') (rev. 1.517)
  • CHANGE: REMOTE_COPY_CLIENT() now with extra argument 'char *cpMsg' to return specific error messages (rev. 1.516)
  • NEW: Write 'exclude_type' (regex,pattern) into 'clients.exclude_type' (rev. 1.515)
  • change: added a new functionality to delete ini key value pairs needed to change the debug ini key name of SM_SLU to SLU. (rev. 1.514)
  • new: 'GET_DRIVE_PROPERTIES': Check, if drive is the 1st drive of a Si3 datastore. This flag can be used to allow certain action only with this drive (rev. 1.513)
  • new: Add 'grp_name' and 'grp_id' to DRV_PROP struct and read it from DB in 'i_GET_DRIVE_PROPERTIES' (rev. 1.512)
  • CHANGE: (#16483) Set state in 'migration_results' to error, if checksums from original and copied saveset do not match.
  • (State in results table for copied saveset has already been set to error before!) (rev. 1.1074)
  • new: Write date of last access attempt into 'clients.access_time' (rev. 1.81)
  • change: Update client package: Check for 'STATUS=SUCCESS' and set state to '0' even if exit code from remote command is <>'0'
  • (sm_ctrlc reports error, if 'sm_ctrld_main' is restarted during update) (rev. 1.80)
  • CHANGE: If backup interface has 'ftp:' prefix, then TCP port will be used from interface name instead of 'clients' table (ftp://sesam:11001).
  • This avoids duplication of port in 'sbc -S' parameter (rev. 1.1085)
  • BUGFIX: (#16595) Restore selection of a single file from INC backup only failed if directory and file are the only 2 items in LIS file
  • change: (#5958) sm_sho now with extra option '-o restore' (dummy)
  • bugfix: (#16590) Now 'sm_main stop node' does stop stpd, sms and passd and set the sm.ini values to 'off' and 'sm_main start node' does set the ini values to 'on' and restart the processes
  • NEW: sm_event: Send mail including NOT/BCK log file
  • CHANGE: (#16578) Adopt syntax to start a backup to 'sm_event backup task <task name>'
  • change: Adapted error message for EHOSTUNREACH (113) or WSAEHOSTUNREACH (10065): No route to host. No valid routing table entry matches the destination address, host is down or blocked by its firewall.
  • BUGFIX: On Ubuntu the sort program must be called with LC_ALL=POSIX to get the traditional sort order that uses native byte values,
  • uppercase before lowercase
  • change: XX_EVENT_LOCK_TERM 'I009-SEPULER Blocking schedule for event type %s, task %s %s'
  • BUGFIX: (#16509) Daylight-Saving Time (DST): Some backup jobs have been dismissed - Do not check the start_time in time_range() function any more - check only the end_time
  • BUGFIX: (#14065) A cyclic scheduled task for last hour of Daylight-Saving Time (2:00 o'clock) is executed multiple times - Compute next starting time always and compare execution time with current time afterwards
  • CHANGE: If sm_robot is called with fast switch -f and a specific drive -d {dn} then only the given drive is occupied via Semaphore 'ctrl_drive{dn}'. This allows starting several sm_robot processes for different drives of same loader for a faster archive adjustment
  • BUGFIX: (#16494) In case of backup type BSR_Windows and if processed items are 0 with backup level DIFF/INCR then do not overwrite backup status with 'No segment file found'
  • CHANGE: (#16483) Set state in 'migration_results' to error, if checksums from original and copied saveset do not match. (State in results table for copied saveset has already been set to error before!)

STPD

  • Bugfix: STPC error. QUIT FTP command added to close connection gracefully
  • BUGFIX: (#16489) The MS SQL Server MTF block MQDA is now recognized as a valid MTF block
  • BUGFIX: (#16475) STPD calculation of necessary block amount used wrong Tape Block Size for tape drives with maximal block size > 64K. SBC error message: 'Error while unpacking data. Bad archive (MTF_ERROR_BAD_ARCHIV)'
  • BUGFIX: (#16482) Reset errno to zero to avoid multiple messages: 'WARNING: Got EINVAL writing data to saveset (written 46272 < to_write 65472): STDLIB error: 22 - Invalid argument.'
  • BUGFIX: (#16480) STPD restore now sends error '553 RETR Failed. EOF reached during read from Datastore or DeDup' if read from data file ends with EOF before all selected items are retrieved
  • Bugfix: #16358. Adding HTTP protocol checking

SMS

  • BUGFIX: SMS Version 4.2.2.6 with fix (#16719) SMS ReadTape returns ERROR_INVALID_FUNCTION - Some tape driver return error 'invalid function' if blocksize bigger 64K
  • BUGFIX: (#16719) SMS ReadTape returns ERROR_INVALID_FUNCTION - Some tape driver return error 'invalid function' if blocksize > 64K is used in ReadTape()
  • BUGFIX: (#16588) Skip ERROR_CRC block until next tapemark cos skipping one block may result in rewind of tape >> endless loop
  • BUGFIX: (#16588) Media readcheck should not stop with 'FAILURE' if a ERROR_CRC block was read. Now the CRC error is displayed and the readcheck will be continue
  • BUGFIX: (#8037) Recalculate readcheck state after tape introduce, remove and init
  • BUGFIX: (#11760) Readcheck must not end with error if a saveset with data size = 0 was not found on the medium
  • change: Case ERROR_IO_DEVICE added to handle this GetTapeStatus() return code
  • change: SMS_VERSION=4.2R5 Keep first log file, e.g. sms-1_165124.0001.log
  • change: Avoid 'ERROR' output if situation is expected, e.g. got 'Got state ERROR_FILEMARK_DETECTED'
  • BUGFIX: (#16575) Skipping blocks on tape failed - Check if device handle is opened to determine if first connection is already established
  • change: Avoid 'ERROR' in message 'changing SMS status to old value...'
  • BUGFIX: (#16527) SAN drive backup failed with '550 CWD Failed. :1007: Connecting to tape device failed. The handle is invalid. (6)' - Semaphore locking must be thread aware
  • BUGFIX: x_Semaphore() is used with sm_data_server threads so now semaphores are set with PID and ThreadId
  • BUGFIX: (#16474) On Microsoft Windows Server 2003 ReadFile with Variable BlockSize may fail with ERROR_INVALID_FUNCTION even if drive supports variable Block Size
  • change: SMS_VERSION=4.2R4 due to BUGFIX (#16475) STPD calculation of necessary block amount used wrong Tape Block Size for tape drives with maximal block size > 64K
  • change: Instead of message 'ERROR: TC_ReadBlockFromTape:: Got error {code} - ERROR: SYSTEM code: {msg}' write 'TC_ReadBlockFromTape:: Got state {macro} {msg}' if error is not a real one { ERROR_FILEMARK_DETECTED, ERROR_SETMARK_DETECTED, ERROR_NO_DATA_DETECTED, ERROR_INVALID_BLOCK_LENGTH }
  • BUGFIX: (#16464) fixed the restore problem for datastores. The NumberOfBytes variable was not initialized correctly. Now it is initialized with 64K in case it is 0. Also included in branch v_4_2_2_18_branch.
  • Bugfix: #16358. Early error reporting for sms stor/retr operations

BSR Backup and Restore

  • BUGFIX: (#16720) BSR Backup on qswinfix failed with: SBC logging without final state message - Check if cbm.ini file is empty
  • NEW: If 'gv_stpd_auth' is set to 'NONE', then no name resolution based authentication will be done. (Same as setting 'STPD_USER=_TMP_' in 'stpd.ini'). Can be used to switch of authentication in case of BSR restore without changing INI file
  • NEW: If client hostname of local ip address could not be resolved use IP address for Sesam authentication
  • NEW: Switch of client authentication if 'sm.ini:auth_file' is set to 'none'
  • change: inserted an additional filter for the disaster restore lis directory. Only file containing DISASTER, DESASTER, SESAM, disaster, desaster, or sesam in the lis directory will be restored when the disaster restore is selected.

SBC

  • NEW: Set saveset format to 'none' for verify of raw saveset data (only checksum will be checked). Use only 'sbc' for verify (no extension specific executable like sbc_vadp)

SBC Macintosh Snapshot Based on Timemachine

  • NEW: Backup of Mac OS X using Time Machine snapshots. if 'tasks.snapshot_flags=TM'

SBC Windows

  • BUGFIX: (#16445) Volume information are not shown in case of VSS: 'all' backup
  • BUGFIX:(#16094) - Hyper V 3.0 on Windows 2012 and CSV 2.0
  • BUGFIX: (#16649) No Backup of System Reserved Volume with VSS - Skip volumes with 'nosbc' file reimplemented
  • bugfix: (#16649) Added System Reserved Partition to backup using VSS when target is - "all".
  • change: If a SparseReadRange error occurred then return the specific file name
  • CHANGE: moss.dll, moss_util.exe and sbc_moss.dll obsolete
  • BUGFIX: (#15607) Windows backup now exits with error if backup source could not be found (ERROR_FILE_NOT_FOUND)
  • BUGFIX: (#16564) Avoid creating Sparse File if restore called with option 'verify'
  • BUGFIX: (#16555) sbc Windows backup ends with: WideCharToMultiByte() returned error 'ERROR_INSUFFICIENT_BUFFER' (122): Wide character could not be converted! - Add dump for the wide char which could not be converted and do not exit
  • change: Avoid misleading error message: 'sbc-4000: Trace: BackupItems: Fatal error iPackDataRet = -9'
  • CHANGE: Support option '-o skip_acl' in case of backup too
  • change: SBC_WIN32_VERSION=4.2R11 Change in processing of Exclude: No adaption from patterns to regular expressions

Exchange Server

  • BUGFIX: (#16544) - EX2013 - No data restored if a complete user or "Top of Information Store" will be selected
  • BUGFIX: (#16541) Failed Exchange restore in to original target path. Do not set '-o recovery_database,...' and '-R VSS:/' if restore selection was done in Restore Wizard via File_View (no mail items selected)
  • BUGFIX: (#16636) - Exchange 2013: sesam is unable to remove log folder after exchange 2013 restore
  • change: In case of Exchange Server backup get Exchange Server Version, e.g. 'Exchange 2013' from NOT file's first sbc-3113 output line and store it in results.sub_type
  • BUGFIX: (#16730) Warning message with wrong text in case of diff or inc backup: 'Zero-processed user mailboxes...' - now print out 'No items retrieved for user mailboxes of database ...'
  • Bug 16736 - Exchange 2013 backup - print out the version of the update (CU1, CU2 or CU3)
  • BUGFIX: (#16736) - Exchange 2013 backup - print out the version of the update (CU1, CU2 or CU3)
  • BUGFIX: (#16741) - Exchange 2013 backup - create user mailbox for sesam user in case of CU1, CU2 and higher versions
  • BUGFIX: (#16667) ME2013 Restore to generated RDBs and import to original DB DIFF or INC failed - Logic in sm_restore improved: If restore into given RDB with '/' then restore into generated RDB
  • CHANGE: SBC_WIN32_VERSION=4.2R12 Support for Exchange Server 2007/2010/2013 with relocated restore of mailbox folders
  • bugfix: do not wait on restore request operation is completed if one is failed
  • bugfix: [CExchangeServer::Restore - ExchangeManagementShell::RestoreImap: Failed for imap [Inbox]. Mailbox "ArnoB" doesn't exist on database "7a9a1999-161d-4ab9-a39a-34332d365c93"]
  • bugfix: [CVssServer::SetBackupSucceeded: IVssBackupComponents:SetBackupSucceeded] - VSS_E_OBJECT_NOT_FOUND;
  • bugfix: do not involve recovery database in case of Exchange 2013 backup
  • bugfix: CExchangeServer::Restore - ExchangeManagementShell::RestoreImap: Failed for imap [Inbox]. Cannot process command because of one or more missing mandatory parameters: RecoveryMailbox
  • BUGFIX: (#16410) - Exchange Mailbox Item Restore into Folder (Exchange 2007/2010)
  • NEW: (#16617) Exchange Server 2013 does no allow to restore single items - Print IMAP lines with type 'Mn' for 'not selectable' items in case of 'Exchange Server 2013' backup
  • BUGFIX: (#16410) - Exchange Mailbox Item Restore into Folder
  • change: In case of Exchange Server backup get Exchange Server Version, e.g. 'Exchange 2013' from NOT file's first sbc-3113 output line and store it in results.sub_type
  • BUGFIX: (#16530) Exchange Restore from DIFF failed with: ERROR: Cannot find matching LIS line for item ... - Position to container directory with saveset ID to avoid match with other identical offsets
  • BUGFIX: (#16521) - Show up the version of the Exchange 2013 during the backup operation
  • BUGFIX: (#16519) - Exchange Restore ends with 'Restore was not successful. Warning: DB Module: [ [ReadFileContents:CreateFileW] The system can not find the path specified.] - Use appropriate 'x_AllTrim(...)' instead of 'strlcpy( cpInstr,cpInstr + start)'. (revision 1.1077)
  • BUGFIX: (#16428) - Exchange 2007 CCR backup results in NULL-pointer error

Share Point

  • BUGFIX: (#16587) - SharePoint 2013 database restore with renaming/relocation into recovery one on the MS SQL 2012
  • BUGFIX: (#16567) - CVssServer::PrepareRestore(): IVssBackupComponents:PreRestore - NULL-pointer of the 'CRelocate' object] - Invalid pointer
  • BUGFIX: #(16566) - DB Module: [ [CVssServer::InitializeForRestore(): IVssBackupComponents::InitializeForRestore] - VSS_E_INVALID_XML_DOCUMENT]
  • BUGFIX: (#16515) - Selection of target disk/path for SharePoint (SQL) recovery database
  • BUGFIX: Restore command for SharePoint document needs '-o generate' switch
  • BUGFIX: Improved the mechanism of handling backup source and prevent the situation when source is finished by the comma (etc. "VSS:/SqlServerWriter/<server>/<db1>,SqlServerWriter/<server>/<db2>,"). It can leads to involving all writers/components into backup process.
  • BUGFIX: (#16450) SharePoint restore failed if SQL Server database is located in different volume than Sesam (xml backup documents)

MS SQLServer

  • BUGFIX: (#16499) Use wide char for database names with non ASCII characters
  • bugfix: (#16501) Restore of MS SQL Server ends with error: 'RESTORE cannot process database 'DbName' because it is in use by this session' - Add database name to connect string only in case of backup to allow 'CHECKPOINT' command for INCR backup

Lotus Notes

  • BUGFIX: (#16660) Windows: Lotus Notes Restore skips all selected items - SBC: In case of Lotus Notes with DB type DB_FILE_STREAM the include matching is done within SBCLN.DLL. SBCLN: Check if file name is able to keep temporary directory path with MAX_FULL_PATH
  • bugfix: Preset RecoverDateTime with empty string
  • change: Windows Lotus Notes x64: Use LIBGSK=gsk8iccs_64 or LIBGSK=gsk8iccs depending on architecture
  • BUGFIX: (#16465) Defines for build fetched from notes 9.0 sdk files linux64.mak and linux.mak
  • change: According to http://www-12.lotus.com/ldd/doc/domino_notes/8.5.3/api853ref.nsf/ 'After ND8.0, DHANDLE will be used instead of HANDLE'
  • BUGFIX: (#16465) Notes 9 backup stays suspended - Problem with different datatype declaration for datatype HANDLE in case of Linux

ORACLE

  • CHANGE: New restore option '-o validate' to create RMAN script with 'restore database validate; restore archivelog all validate;'
  • change: New revision 1.64 generated during checkin of -m sob -d su/src/sob sbc_oracle_rman.cmd
  • CHANGE: New restore option 'validate' to create RMAN script with 'restore database validate; restore archivelog all validate;'

Novell File System

  • bugfix: Check filter item during restore (sbc_smdr)
  • NEW: Support restore level '-l list' to display content of saveset (sbc_smdr)
  • BUGFIX: (#16562) TSA Backup cut always the last sign of the exclude string - DisplaySelectionList() Use strlcpy() with strlen(dest)+1 to avoid truncation of exclude pattern

Novell GroupWise 2012

  • BUGFIX: (#15859) GroupWise 2012 Backup based on TSAFS -> Size of GW folder was displayed with wrong size. Now extra 'du' command is used to get the size
  • BUGFIX: (#16487) GroupWise 20xx GW object type may be displayed as [TMP] instead of [PO] or [DOM] - Now check via server=.../{binary} either 'gwpoa' or 'gwmta'

Zarafa

  • change: Improve Zarafa restore of mailbox: Only mailbox folders must be put into msl file
  • BUGFIX: (#16529) As Zarafa restore with mailbox folder is supported by sbc_zarafa so extra blow_up (rev. 1.732) is not needed, only the folder id must be inserted into msl file.
  • BUGFIX: (#16529) Since rev. 1.441 Zarafa restore with selected Mailbox or Folder did not blow up the emails for MSL file. With empty MSL file the message 'No record keys given to restore' appears.

Major fixes and changes v4.2.2.18

Installation

  • BUGFIX: Update failed due to incorrect Sesam database export with values including single quote (') (PostgreSQL only)

Restore

  • BUGFIX: Restore from tape devices supporting block size > 256 K works again (Windows only)

Major fixes and changes v4.2.2.16

Installation

  • BUGFIX: (#16084) Installation failed due to incorrect side by side configuration (Windows system without .Net 3.5) - inserted the VC90 manifest file since some of the DLLs in the SMS directory are linked to the msvcr90.dll
  • BUGFIX: set_vddk.ps1 may copy VDDK libeay32.dll into SEP sesam bin directory even if DLL references are not satisfied.

Kernel

  • BUGFIX: (#15824) Exchange generation restore does not include INC backup, when choosing the whole mailbox - iREMOVE_DOUBLES(): In case of a mailbox line the saveset's datetime must be compared to keep the recent lines (sm_lis_items.exe)
  • BUGFIX: (#internal) Migration failed for 'Windows Server 2008 R2': select * from result_lbls where num=1 and saveset='SF20130711130814693@Wtj9-qXfqAm' and label like '2_____' - Use 'DB_OS_LEN' (sm_sms_copy.exe)
  • BUGFIX: (#16095) If drive requires cleaning tape then tape status check returns drive online even if there is no tape in drive: The customer's device driver returned ERROR_NOT_READY instead of ERROR_NO_MEDIA_IN_DRIVE (fetched from rev. 1.284) (sm_client.exe)
  • BUGFIX: (#2013061880000219) (Saytec) Migration failed because media was dismounted after 3 minutes shortly before migration started - now extended to 5 minutes.
  • BUGFIX: (#16107) SAYTEC - DISK_CHANGE: Avoid unloading the media due to reaching the sm_sms_watch idle timeout (sm_sms_watch.exe)
  • BUGFIX: Using 'drive_groups.grp_id' instead of 'grp_name' for check, if a medium can be loaded into a certain drive, because function 'i_CheckDriveLabel' compares over id
  • BUGFIX: (#15879) Remove ' from backup message, because it leads to a SQL error. (Example: If a CBT backup fails, then it stays active in Sesam DB).
  • BUGFIX: (#15916) single item restore of IMAP backups not possible - Avoid mode 'M' for higher level use 'U'ser and 'm'ailbox to allow browsing with GUI Restore Wizard

Kernel - SM_RESTORE

  • BUGFIX: (#16125) Exchange Mailbox Restore: sm_restore has skipped transactions logs - Now the matching algorithm was improved
  • BUGFIX: (#15978) If restore option for exchange database restore was set with wrong count of " then the restore tried to overwrite the original database - Now an enclosing " is added.
  • BUGFIX: (#15836) If there are several Exchange emails with identical Subject in same Mail folder then wrong email may be restored. Now check with create and access date time too
  • BUGFIX: (#15246) If during Exchange Server generation restore the given saveset is a migrated saveset then the check for needed savesets (FDI chain) must be done with original saveset
  • BUGFIX: CREATE_SLL_LINES(): Call i_UNCOMPRESS_FILE(cpLisFile, strLog); to uncompress lis files for Exchange Server restore if compressed meanwhile
  • BUGFIX: #16052 SM_RESTORE with log level 1 produces big log files - Produce "got(ld) {item}" only once per directory
  • BUGFIX: (#15936) Exchange restore failed with "Didn´t find matching LIS line for item ..." - If some email items are selected from same database then the database directory must not be processed twice
  • BUGFIX: (#16125) Exchange Mailbox Restore: sm_restore has skipped transactions logs - Now the matching algorithm was improved

GUI

  • BUGFIX: (#15497) Browsing Backup source or selecting Restore source with Chinese (Simplified, PRC) chars is displayed incorrectly
  • BUGFIX: (#16108) Folder with Umlaut not selectable in Restore Wizard - Now the search pattern is retried to work around a double UTF-8 encoded string
  • BUGFIX: (#15815) add new task to tree under client after create new client with task (component 'task by client')

VMware vSphere

  • BUGFIX: (#15220) restore VM with version '8' fixed (remove incompatible virtual devices from OVF file, this is used for VM creation)
  • BUGFIX: (#15220) restore VM with version '8' fixed (set virtual CD-ROM type from 'iso' to 'remoteatapi' or 'atapi')
  • BUGFIX: (#15912) exclude VMDKs from VM with parameter '-x vmdk<0-9>, when restore VM solved
  • BUGFIX: (#15912) Exclude several VMDKs fixed (-x vmdk1,vmdk3,... or -x vmdk1 -x vmdk3 -x ...)
  • BUGFIX: (#15912) Restore VM with exclude VMDK, if its backup uses one saveset fixed
  • bugfix: (#15935) CBT backup with Falconstor VM fixed (write changes to correct CBT file depends on VMDK)
  • BUGFIX: (#15937) restore Falconstor VM fixed (use correct VMDK paths depends on drivenum, when execute restore VMDK data command)
  • BUGFIX: (#15937) restore VM with count of VMDK greater than '7' fixed, set correct SCSI node, if append VMDK to SCSI controller (SCSI node position '7' is reserved by controller)
  • BUGFIX: (#15980) Restore VM to empty ESX server, that does not contain any VMs fixed
  • BUGFIX: (#16026) Restore VM from VM backup with ISO CD-ROM fixed (modify OVF file: if CD-ROM block contains element 'HostResource', change it to comment)
  • BUGFIX: (#16026) restore VM from VM backup with ISO CD-ROM fixed, set device type of CD-ROM to 'client tool' (modify CD/DVD block of ovf file, set ResourceSubType='vmware.cdrom.iso' to 'vmware.cdrom.remoteatapi')
  • BUGFIX: (#16048) full, differential, incremental backup with CBT of VM, which already has snapshots defined fixed
  • BUGFIX: backup offline VM with SAN access fixed (set missing SBC argument 'ssmoref')
  • change: (#15511) if run FULL VM backup and no used blocks of VMDK will be found (changedArea=null), write value '0 0' to CBT file instead of stopping the whole backup procedure
  • change: set correct error instead of nullpointerexception, if run FULL backup of VM and changeID config file fails to write due to missing VMDK change IDs
  • new: vSphere actions 'stop', 'restart' VM, requires only vSphere authentication data and action parameter -A <stop|restart> or -a action=<stop|restart>

Extensions

- SBC Windows, inclusive Exchange and MS SQL 2012, MySQL on Windows

  • BUGFIX: (#16132) Differential Savesets of UNC Paths are being reported SUCCESSFUL even if the specified share is unavailable!
  • BUGFIX: (Ticket#2013030780000021) Avoid duplicate '/' when restoring to another volume and only drive letter is given as target path - Error message: 'sbc-2044: Warning: Cannot create item [D:\\backup\:[5] WIN32 API error]: 5 - Permission denied.'
  • BUGFIX: #15947 - sbc scans excluded subdirectories for contents (if called with -T {since}): Now the return of iExcludeItem() is checked and if FLAG_EXCLUDE is returned for a directory the whole directory is skipped
  • BUGFIX: (#15876) Restore of exchange based on path with '-o plain' does not work because of duplicate slash added to restore target
  • BUGFIX: (#15944) Backup: If option -o skip_data then store only ACL
  • BUGFIX: (#15944) Restore - Avoid using \\?\UNC long UNC file prefix, use UNC format \\{server}\{share}\... instead
  • BUGFIX: (#15947) Windows: SBC scans excluded subdirectories for contents - If iExcludeItem(...) returns that directory is excluded by exclude patterns then skip that directory
  • BUGFIX: (#15947) Windows: SBC scans excluded subdirectories for contents - VSS.DLL must use iExcludeItem() instead of obsolete FilterITEM()
  • BUGFIX: (#16017) VM Restore with Exception: ... - Since revision 1.177 non file system VMDK items are counted as not processed
  • BUGFIX: (#16017) VM Restore with Exception: [Exit code from sbc: [1] - warning] - Do not print warning message 'sbc-2038: Warning: Setting of standard attributes for item [\\{server}\{host}\[...Store...]\] failed.' in case of VSPHERE_CBT restore
  • BUGFIX: (#16132) Differential savesets of UNC Paths are being reported SUCCESSFUL even if the specified share is unavailable - Now the return codes ERROR_BAD_NETPATH, ERROR_BAD_NET_NAME, ERROR_LOGON_FAILURE are aborting the retrieve items loop.
  • BUGFIX: [Ticket#2013031580000079] PCH Exchange Sicherung
  • BUGFIX: 15811 - Exchange restore deletes all RDB, which exist on the system
  • BUGFIX: 15968 - Restore of Exchange mailbox folder does not work.
  • BUGFIX: 15976 - Problem: Storage group [RSG] was removed after restore.
  • BUGFIX: 15977 - Failure of folder item restore with symbols "\ / ? : * " > < |'"
  • BUGFIX: 16024 - Info status - Restore with parameter "Restore to existing RDBs and import to original db" - missed extern function;
  • BUGFIX: 16111 - VSS_E_OBJECT_NOT_FOUND when backing up a specific volume and drives without drive letter exist
  • BUGFIX: Backported to use ntwdblib to support MS SQLServer 2000 which does not fully support ODBC - backing up databases with multiple data files becomes stalled
  • BUGFIX: Bug 15811 - Exchange restore deletes all RDB, which exist on the system;
  • BUGFIX: Bug 15823 - "Specified method is not supported" when disk of original database runs out of space; Bug 14956 - wenn eine Datenbank nicht gemountet werden kann, sollte es eine richtige Fehlmeldung geben
  • BUGFIX: Bug 15923 - "Object reference not set to an instance of an object." when adding mailbox permission via sbc_es.exe
  • BUGFIX: Bug 15945 - Failed Exchange restore in original path in Windows Event Log
  • BUGFIX: Bug 15979 - Windows Event Log with Error in case "Restore to exiting RDB" for DIFF and INCR type (Exchange 2007)
  • BUGFIX: Return message to caller in case of an error during extern restore connection (Oracle, vSphere)
  • BUGFIX: Fixing Bug #16015 with restoring MS SQL database.
  • BUGFIX: Bug 16010 - replacement for sp_dboption SQL script functionality (MS SQL 2012).