GUI Bugfixes: BUGFIX:(#20467) The Activities -> Immediate Start functino was corrupt BUGFIX:(#20535) Error when clicking on Tasks, Task Groups, Migrations in Backup Plans bugfix: (#20537) "OK" button is enabled at schedule dialog for weekly schedule although no weekday is selected bugfix: (#20537) "OK" button is disabled in schedule with execution daily, weekly, monthly or yearly bugfix: (#20545) Remove taskgroup failed due to dependencies of schedule, which does not exist bugfix: mouse contextmenu show 'create new migration task' instead of 'create new replicaton task' under tree 'SI3 replications change:(#20539) Added trace code in the CacheUpdateHandlerClient to see the cacheGetDiff working change: hide button 'delete', when create location bugfix: (#20480) Error "Action invalid or not found null" occurs when setting "Client is a virtual machine" BUGFIX:(#20507) Avoid ClassCastException when breaking a media action change: (#20472) Adjusted the other tabs in the ResultsDialog and removed old panel2 change: (#20472) Fix resizing problems in ResultsDialogPanel2 bugfix: (#20524) Remove drive cause duplicate drive entries of all drives at tree change: (#15441) remove update results which will be done, when command 'adjust_eol' will be invoked, after change eol at datastore saveset table change: (#20472) Rebuild ResultsDialogPanel2 as replacement of ResultsDialogPanel2NB which was built with Netbeans bugfix: (#20486) empty remove device server combobox under drive, when open gui with only tab drive and create/edit drive bugfix: (#20505, #20492) when updating a drive, dont "delete" it completely from cache bugfix: (#20450) avoid creating an extra dbconnection on startup bugfix: (#15441) command 'adjust_eol' will not be invoked, when change eol at datastore saveset table bugfix: (#20486) empty remove device server combobox under drive, when open gui with only tab loader and edit drive under loader bugfix: (#15441) leave resultdialog with changed eol, but not confirmed with key 'return' does not apply changes; change: (#20472) Move the Locked checkbox to the 1st TAB of results dialog bugfix:(#20472) Fixed layout of 2nd TAB of results dialog with metal look and feel bugfix: (#20487) NPE when trying to delete a client from "components" - "clients" bugfix: (#15441) columns 'hidden_column_eol_changed', segment still visible; change eol cause classcastexception (Date <-> GregorianCalendar) bugfix: (#20481) add additional value for data size without migrated tasks (sum2) change: (#20472) Move EOL and locked to 1st TAB of results dialog bugfix: (#20465) File -> Import Database: file selection ends in Exception change: build failed, add missing renderer for column eol at Results Dialog, which render date+time change: (#20472) Move EOL and locked to 1st TAB of results dialog bugfix: cli 'add datastore' param -o does not work bugfix: (#20466) GUI: Activities -> immediate Start -> backup ends in exception bugfix: Added missing title in the "Loader action started" dialog bugfix: (#15441) execute 'sm_sesam adjust_eol with param '-E fix: invalid datatype in resultlbls fix: correctly update drives with drive groups BUGFIX: (#20451) Do not handle a value like 1,3,7 as invalid slot range when starting an archive adjustment bugfix: wrong data type VARCHAR for results_lbls.saveset_count (BIGINT) bugfix: (#15441) execute 'sm_sesam adjust_eol -s ', when results.eol changed; execute 'sm_sesam adjust_eol -s -m ', when results_lbls.eol changed bugfix: (#17470) New notifiation icon does not disappear, when no new notication exists bugfix: (#17470) New notification icon not triggered, when new notication are incoming; new notifiation icon does not appear, if no new notication exists BUGFIX: (# 20451) Avoid "Operation not possible" exception when starting archive adjustment bugfix: (#20454) cli add notification confirm/popup,... cause JSonMappingException - no suitable constructor found. .Solution: add default constructor for class 'NotificationInfo' change: register CLI notification action 'modify' with param '-O ' as PK change: set acknowledgement state via object (param -A -O ) BUGFIX: (#20447) After update license information and EULA not shown change: (#15441) hide column 'segment', use column name 'media sequence' for column 'num' SRC Bugfixes: new: changed version number from 4.4.3.16 to 4.4.3.17. bugfix: (#20406) Remember last (full) backup in DB:tasks. Problem: 4 fields required to set correct based_on (task_type specific). Solution: Add last_diff_backup, last_incr_backup to last_full_backup and last_fdi_backup. Set them after successful backup new: Windows: Function get_free_drive_letter [-m a] return next or all free drive letters change: (#20217) sm_vfs windows NFS support change: (#20406) Remember last (full) backup in DB:tasks. Problem: Do not set tasks.last_full_backup/.last_fdi_backup during update until further logic is implemented. bugfix: (#20440) BSR GPT/EFI restore fails with error: Operation failed! Problem: On log level 3 and higher (VSS Traces shown) restore failed - wrong drive letter read from registry. Solution: Use ANSI functions while reading drive letter from registry new: Use recent Python modules (pyudev, lxml) for RHEV backup bugfix: (#20499) SBC Windows: No LIS file generated if backup task runs with compression or encryption bugfix: Reduce global offset if 'block already fetched' otherwise traced position is wrong after that situation change: Trace line added with: 'mtfana: print_TAPE: Use logical block size ...' bugfix: (#20484) Selective path restore fails with 'Error while unpacking data. Bad archive'. Problem: Since commit 6ccb51b '(#16272) MTF analyse errors im STPD log' some offsets in the LIS are reduced by 1024. Solution: If buffer contains not enough bytes then remember size and increase global offset. NEW: Allow set of 'JAR' file used for vSphere backups by '-a jar=' option BUGFIX: (#20523) Fix core dump in 'GET_VOLUME->loader_handler()' because message from 'sm_loader' can become longer than message size used by caller new: Add build rule for Sesam SAP HANA install script new: Check for valid JSON output from API call change: ADJUST_EOLS(): Avoid 2nd DestLL( LL_SavesetList ) change: check MSExchangeIS and MSExchangeRepl services status before the backup operation is started. bugfix: (#20499) SBC Windows: No LIS file generated if backup task runs with compression or encryption. Problem: block_type folowed by attributes. Solution: Check only for strlen(...) < 4 bugfix: (#15441) Modifying save set EOL does not adjust the FDI save set chain. Problem: A lower EOL was not set. Solution: The argument cCFDI is checked for valid 'C','F','D' or 'I'. If not set the eol is enforced even if it is lower. Note: During backup and migration the CFDI is set new: Template for Sesam backup client for Fujitsu CD10k with Medusa backup API. bugfix: (#15441) Modifying save set EOL does not adjust the FDI save set chain. Problem: Missing argument in update result s statement bugfix: (#15441) Modifying save set EOL does not adjust the FDI save set chain. Problem: Missing argument in update results statement change: Trace call arguments bugfix: (#15441) Modifying save set EOL does not adjust the FDI save set chain. Problem: If results.eol is changedn then migrated saveset's eol must be updated too. bugfix: (#15441) Modifying save set EOL does not adjust the FDI save set chain. Problem: adjust_eol -m {mediapool} may be called with eol set to previous datetime -> set media.eol to max(eol) in this case too bugfix: (#15441) Modifying save set EOL does not adjust the FDI save set chain. Allow to use adjust_eol with -m {label} (convert to mediapool) BUGFIX: (#15441) Modifying save set EOL does not adjust the FDI save set chain. Problem: There is no common function to update complete saveset chain for FDI when either results.eol or result_lbls.eol is changed. Solution: New 'sm_sesam adjust_eol -s [-E ] [-m ] [-l ]' allows to set either all result_lbls.eol (if no media_pool is given) or adjust results.eol and saveset chain if media_pool is set. The media eol is adjusted to max(eol) after the eol change. bugfix: Call SRV_POST() even if CHECK_PERMIT(), GET_BACKUP_MEDIUM() or START_STOP_WATCH_DOG() failed bugfix: (#20455) sm_support --pathes throws exception. Problem: 'ValueError: too many values to unpack' when 'auth.ldap.userpattern=uid={0},ou=people' is scanned for '='. Solution: Use line.split(=,1) NEW: sm_install_saphana for automatic, parameter-based installation of SAP HANA extension bugfix: Preset last fsck full time to year 2000 bugfix: (#20440) BSR GPT/EFI restore fails with error: Operation failed! Problem: "Operating system architecture [x86] is not supported!" even if binaries and system is x86. Solution: Perform sbc_win32 clean from main makefile.win32 new: Start 'fsck' for Sesam deduplication store at least every 7 days as 'full fsck'. Intervall can be changed by 'gv_dedup_fsck_max_inc' Change log message bugfix: Last commit to write correct segment number info 'DB:result_lbls' doesn't work, because media pool name wasn't derived correcly from media label BUGFIX: Set 'DB:result_lbls.segment' for saveset spreading over several media (EOM) for 2nd and further tapes. Since 2010 'unknown segment' was set only change: Compiler warnings resolved BUGFIX: (#20207) Selective restore from replicated saveset fail, if original one has been purged. Obsolete line removed BUGFIX: (#20207) Selective restore from replicated saveset fail, if original one has been purged. Problem: The drive must be switched to the drive for the replicated saveset. Solution: If original saveset does not exist with original media pool in table result_lbls then use the replicated saveset with target_drive from migration_results bugfix: (#18465) Avoid wcs functions in packdata.c (performance issue) BUGFIX: (#20404) Enhance check if full backup has to be enforced. Problem: If FDI_CHAIN is corrupt then the tasks.task_sts must be set to 'F' to assure next run with level FULL. Solution: Check return from ADJUST_MEDIA_EOLS and set task_task_sts='F', add messages to NOT file and to PROT ('W010-BACKUP Adjusting the EOL for the saveset chain of ...'/'W011-BACKUP The data integrity check on the backup task ...' new: Si3 deduplication: Automatic start of GC and FSCK can be switched of by setting 'gv_dedup_no_fsck_start' and 'gv_dedup_no_fsck_start' global variable change: Add SM_SBC_COM_LIB to sm_sesam BUGFIX: (#20261) The EOL of a migrated saveset isn't increased in spite of further DIFFs/INCs. Problem: If adjusting the EOL for all necessary savesets in the saveset chain cannot find original/replicated/migrated savesets with valid eol >= now the try to increasing EOL of already outdated original/replicated saveset if still available. Update the results.eol for migrated saveset too. bugfix: (#18465) Avoid memset as far as possible without side effect change: Add UPDATE 0100 into demo license BUGFIX: (#20425) Delete all 'waiting' (state='w') entries from 'DB:cal_sheets' for current day and not the future ones only. 'sm_sepuler_cal' calulates entries using all 'DB:terms' entries, even from the past. This could happens for example, if 'sepuler' is blocked (newday) and leads to DB errors during insert BUGFIX: (#18984) The BSR Pro service should be restarted after a crash/abort. Problem: After abort of sbc_bsr backup a following backup is stalled due to service is not responding. Solution: Restart BSR Pro service always before backup (until version 11 BSR Pro with 'oodicmdc.exe /stop') change: Mistyped 'database', show traceback with loglevel '-v 2' change: 'i_RCMD' -> 'RemoteCmd' in Sesam main log switch sm_rexec log off change: Start 'sm_rexec' from Sesam Python 'sm_rexec()' function without Sesam log switch. Avoids many useless logs in Sesam day log. Can be switched on by 'prot=1' parameter din function call NEW: Add 'oracle-java8-jre' or 'openjfx' as dependency for Debian Sesam server/GUI package. 'openjfx' should depend on 'openjdk-8-jre' bugfix: (#20271) SMSSH sends exit code 3 occasionally, if system is on heavy load. Problem: GetExitCodeThread could return 259 - STIL_ALIVE which is not error code. Solution: In case of 259 ignore Thread exit code BUGFIX: Report error during import into 'DB:cal_sheets' in Sesam main log file. new: Use 'sm_db import' for import instead of executing DB specific coammds. new: Retry failed import without using a transaction. bugfix: 'sm_sepuler_cal': Report exit code from DB import and exit with this value itself Fix trace line (too many arguments) new: Return system error message, if 'dir' command fails BUGFIX (#17019) Do not 'purge' savesets, if saveset file list from data store coudn't be retreived BUGFIX (#17019) Do not 'purge' savesets, if saveset file could not be deleted. If '-o FORCE' option is set, DB entry will be deleted anyway add comment out code bugfix: (#20271) SMSSH sends exit code 3 occasionally, if system is on heavy load. Problem: GetExitCodeThread could return 259 - STIL_ALIVE which is not error code. Solution: Wait for real exit code in case of 259 Add log line BUGFIX: (#20420) Correct return value from 'i_FILE_DELETE()' function, if wild cards are used and at least one file couldn't be deleted. Function returns now '0=OK,1=Warning (at least one file was deleted and one not),2=Error (no file has been deleted) change: Avoid calling trap_db_err() with exit flag TRUE coz queues may remain in state 'blocked' when one SQL statemant cannot be executed, e.g. due to connection limit exceeded for PostgreSQL new: Add message to backup log, if a full backup has been enforced bugfix: (#20367) Exchange backup does not start due to new flags according DAG. Problem: The non-DAG backup was treated the same as DAG backup. Solution: If non-DAG Exchange backup, then backup database from the host ignoring active/passive flag bugfix: Comment out 'trace' function in 'sm_popen()', because it's not called from all modules correctly BUGFIX (#16206) - Validating Backup Integrity By Using CHKSGFILES. Log and result message rvised NEW: Purge data store at once after saveset migration and 'delete after migration' flag set change log message BUGFIX (#16206) - Validating Backup Integrity By Using CHKSGFILES. If sm_reformat_lis failed for DIFF/INCR then task_task.sts must be set to 'F' to assure that job runs next time with level FULL bugfix: (#20324) Very high system load can lead to Exception in sm_sshd "Error reading SSH protocol banner". Problem: Paramiko shows error in case of socket termination. Its common problem of paramiko library. Solution: Hide messages from paramiko except sm_ssh log level set to DEBUG BUGFIX (#16206) - Validating Backup Integrity By Using CHKSGFILES. Traceback is removed. bugfix: VMDK mount: Correct un-mount of all mounted VMDKs in case of basic full image backup