=== GUI === 

* bugfix: (#17671) fill target relocation path combobox at restore wizard expert opts with correct values depends on backup type
* BUGFIX: (#17671) Restore 'Exchange Server' with file view causes switching target node after activate relocation at expert option dialog. Relocation target pathes combobox at expert options does not offer path of target directory. Now the selected target node is memorized, if the expert option dialog is opened.
* BUGFIX: (#17709) If the EOL was changed from the Data Store View Dialog the GUI did not set the EOL time. Now the EOL time is set.
* BUGFIX: (#17669) When creating a new Data Store the browed path was not applied correctly. Now the path is set as expected.
* BUGFIX: (#17643) sm_client cannot access unc path with backslash in the end. Now the backslashes in the path for the Data Store are replaced with slashes due to an added backslash filter for the Data Store pathes.
* BUGFIX: (#17660) select preferred mediapool because restore failed. Solution: hide combobox preferred mediapool
* BUGFIX: (#17601, #17456) Update mail watcher thread to be the only one taking care of timeout (set smtp session timeout to infinity) and retry if the first time mail sending fails. The second try also enables full smtp session debugging to verify what exactly hangs. The sending timeout is based on the "mailSendTimeout" (-M) parameter and will be extended based on the attachment size (adds 1s per 5k) to prevent slow upload speeds from causing the mail sending to fail.
* change: adapt send check time to check mroe often, send mail again if timeout is reached (once) allow to force debug when sending mail
* BUGFIX: (#17627) Default backup interface is set to 'http', if host name is behind 'http' in lexical order at run backup dialog. Now the server name is the default backup interface unless a different default interface has been selected.
* change: (#17608) Remove Si3 data store files from disk if DS is removed. Add command 'sm_data_stores remove -L <dsname> -o FORCE' between commands 'sm_arch', 'sm_config_drives', when remove datastore with parameter -F YES
* NEW: (#17613) Add media pool option for "restore" to sm_cmd. For a restore event "-m media pool" can already be set, but not for "sm_cmd restore". Cli parameter '-M' for action 'restore' was added to sm_cmd, which sets the mediapool of restoreevent
* BUGFIX: (#17625) Using cli to add a datastore with type "-t 'SEP Si3 Store'" fails, since the wrong datastore type was insert into the DB. Now the cli help message was adjusted with the correct and valid datastore type and the Si3 datastore type was set to "SEP Si3 Deduplication Store", if type starts with 'SEP Si3'.
* BUGFIX: (#17576) Ignore version-check if guid.location is empty and remove timestamp from duplicate notification check. Only title and message are checked if they already exist in the db and are only added if not found. This way any rss feed document is only shown exactly once
* BUGFIX: extend notification fetcher to only run once every minute instead of every 10s to prevent notifications from popping up too often.
* BUGFIX: (#17607) Closing a notification of a given object also triggers that all previous notifications are resolved as well (updateNotificaitonWithActionCloseToResolved). This prevents the notification dialog from popping up multiple times for the same notification.
* BUGFIX: (#17562) Start restore task with type 'Exchange Server 2003' uses wrong source. Content of addition source textfield with init 'MAIL:user=, target=' will be used instead of primary source textfield. Hide additional textfield and set correct source. Further bugfix: Reopen existing restore task 'Exchange Server 2003' show other restorewizard layout instead of create new task.
* bugfix: refresh log button is visible, if open result dialog without active log tab
* BUGFIX: (#17607) filter notifications to retrieve only one per object type.

=== Installation and update ===

* BUGFIX: (#17695) Since SEP sesam version 4.4 Windows Server 2003 is not supported anymore. The sm_config_client and the sm_update_client will not try to update Windows Server 2003 systems. They will rather return an info message saying that the OS is not supported.
* bugfix: (#17678) set java path did not work correctly, neither in the server nor in the GUI package. The sm.ini path was not initialized correctly. Furthermore the gv_ variables have to be obtained differently due to the lack of registry entries in case of a GUI installation.
* BUGFIX: (#17642) On Windows, if java version 1.6 is installed on a system the check for the correct java version 1.7 or 1.8 failed. Since SEP sesam version 4.4.1.22 the check for the correct java version during the update failed. Hence the java process can not be started correctly.
* BUGFIX: (#17636,17498) Sesam services have been stopped during update but not started again if the update of a windows 32 bit system failed. Sesam services are not stopped anymore if the sesam installation kit is execute on a windows system which is not supported anymore. E.g. Windows Server 2003 x86 is not supported and after the update the services are running.
* BUGFIX: (#17636) To ensure that all sesam components use the same java executable (sm_setup) during installation and later on a java wrapper was introduced called sm_java.exe and sm_javaw.exe. This executable is now called from the GUI link to start the rmi client. The java wrapper class is responsible for obtaining the correct java version.
* BUGFIX: (#17642) The update set the correct java version but did not return the correct value. This lead to an error during sesam update saying no valid java version could be found.

=== Restore === 

* bugfix: (#17703) ME07,ME10 - Restore to original target path - W007-RESTORE Restore was successful with warnings. The changes for '(#17680) Restore must immediately rename files which are in use' cannot be applied for locked files. So old algorithm with restoring data into '{file}.${pid}_new' and delayed 'rename on reboot' must be reimplemented for such cases.
* BUGFIX: Do not set preferred media pool for sub-restore, because this version of 'i_FIND_MEDIUM' doesn't handle it correctly -> Error during restore: 'Could not find label'
* BUGFIX: (#17667) Cristie BSR restore fails with: "Failed to extract configuration from backup" if last backup has failed. Do not return BSR savesets, where one saveset of complete session has failed.
* BUGFIX: (#17680) Restore must immediately rename files which are in use. Use messages sbc-3183 and sbc-3217 to log the file in use and the file which will be deleted on reboot. Rename {file} into '{file}.delete_on_reboot' and only if this fails then rename into '{file}.${pid}.delete_on_reboot' to avoid creation of multiple '...delete_on_reboot' files.
* BUGFIX: (#17586) File restore without tree structure (plain) ends with 'Warnings'. Avoid restore of parent directories
* bugfix: (#17599) The warning 'Target file [...] exists but will be overwritten' occurs during restore to dump file. The restore to an existing dump file with option 'overwrite' ends always with state 'warning'. Avoid the warning message, now the info message sbc-3216 'Target file [...] exists and will be overwritten.' is displayed.
* NEW: vSphere Single-Item-Restore: Try to mount Novell NSS volumes
* BUGFIX: (#16980) Correct handling of files backed up from non-utf8 compatible clients. Restore may be aborted due to buffer overflow when calling 'sm_lis_items -i {sel_line}' to get original encoding. In case of a generation restore the ADAPT_SEL_FILE() must not be called for sub rtask
* NEW: Do not occupy restore semaphore in case of restore from Sesam data stores, so multiple restore can run at the same time
* bugfix: (#17529) Scheduled restore task does not find savesets, if restore task was defined with gen mode and task is COPY. If sm_lis_items is called for gen LIS but with COPY saveset then generate GEN LIS from COPY LIS file

=== Backup ===

* BUGFIX: (#17700) Lotus Notes backup terminates with 'Error: Operation with incomplete or broken data stream!'. If a file has extension '.nsf' but is not a database then this file must be backed up as a simple file. If NSFBackupStart() returns with error 'File is not a database' then process .nsf file as simple file too.
* BUGFIX: (#17604) Backups do not stop in case of an error in 'End-of-Media' action. (Tech.: Compare only the first word of value set from 'sm_sms_watch' like 'BREAK No further tape available')
* BUGFIX: (#17640) Netapp module expects volumes to be in /vol/, so backup fails if volume isn't mounted here. sbc_netapp: interpretation of option volume_path= for customers with special junction paths
* BUGFIX: (#17645) RMAN backup failed with (can't connect with Media Manager, Error 0) after using PORT 0,0,0,0,107,73 in stpd communication. Since support for IPv6 the retry with IP address from control socket may fail due to command 'setsockopt(SData, IPPROTO_IPV6, IPV6_V6ONLY,...)' which must not be executed for IPv4 sockets.  LINUX: The retry failed cos the memcpy command to set the FTP data port was invalid.
* BUGFIX: (#16980) Correct handling of files backed up from non-utf8 compatible clients. If special chars are encoded with 'iso8859', e.g. Umlaute, then store sub_type 'iso8859' for backup_type 'Path' and adapt the sel file during restore. If LIS contains 'no_utf-8_encoded' chars then set immediate sub_type to 'no_utf-8_encoded'
* BUGFIX: (#17596) Set correct client name of RDS, which has been overwritten by Sesam server name during external backups, if no RDS (-S) parameter has been set

=== Migration ===

* BUGFIX: (#17653) Migration task stalled during tape change - sm_sms_watch with 'Medium Medium ...'. Avoid string overflow in sm_sms_watch CHECK_AND_LOAD_NEXT_LABEL()

=== License ===

* BUGFIX: (#17191) License output too long for Sesam 'PROT' function. Now use buffer length 4*MSG_LEN (4*34500) instead of 10K
* CHANGE: (#17658) One license key for all SAP license keys: SAP, SAP_HANA and MaxDB. New license key 'SAP' is used for MaxDB, SAPR3 and SAP_HANA. If old license keys 'MaxDB' or 'SAP3' are added to 'SAP' license value (License key 'SAP_HANA' not longer supported)
* CHANGE: (#17412) Implement license with 'Volume Front Side' counter. New license Key 'Volume_FS_TB' for SEP price list 2014
* CHANGE: (#17659) License keys ESX_RAM_R001/R064, HYPR_... and CXS_... are obsolete. All implemented logic for license keys ESXR001,ESXR064,CXSR001,CXSR064,HYPR001,HYPR064,ESXServer_RAM,Citrix_RAM and Hyper-V_RAM removed

=== Media === 

* BUGFIX: (#17708) ERROR: Unload media File2_FULL00008 failed. If tape device is busy then retry up to 300 times (~15 minutes).
* bugfix: (#17685) sm_config_drives cores after FILL_SMS_INI_LIST was called. The stack allocation for large buffer is now avoided by allocating the output string statically

=== Kernel === 

* NEW: Wait longer time between start of Sesam PostgreSQL server and 'createuser' command so DB server has more time for startup
* NEW: Script to install Sesam SAP HANA extension
* BUGFIX: (#17656) Windows: Getting parent process pid via sm_list.exe is CPU consuming. Now use NtQueryInformationProcess() to get the parent process pid.
* BUGFIX: (#17593) ctrld loggings are located under ...\var. If part of the CTRLD log path is not a directory or a directory junction (e.g. a SYMLINK) then the MakeDirStructure() function will skip the rest of the log directory path. Now if that part exists then go on with directory creation.
* BUGFIX: (#17598) sm_client dir causes an error if a new empty partition without drive letter is browsed. Ignore ending '\*' pattern if called for a volume.
* BUGFIX: (#17591) localtime function causes deadlock in sbc_smdr. On Linux calling localtime() in signal_handler() may lead to deadlock. Avoid calling sbc_log() in signal_handler() if log level < 4.

=== Si3 Dedup ===

* NEW: Workaround: For data store on dedup volumes: Set balance weight for all queues to same value until we have a better calculation
* BUGFIX: (#17624) After deletion of the last si3 datastore the sds is still running. Now the deletion of the last si3 datastore stops the sds via sm_main stop sds.
* BUGFIX: Init of Si3 dedup store on Windows. Strip off '\' from the end of 'gv_rw_stpd' path, because java fails to start

=== STPD ===

* BUGFIX: Restict max. number of STPD connections to 128, because 'sm_stpd_utl_main' cores. If more streams are configured in 'stpd.ini' then reduce it to max. allowed value. Change 'C' to 'C++' comments