Bacula Enterprise Edition 8.2.8

Ukazała się nowa wersja Bacula Enterprise Edition oznaczona numerem 8.2.8. Inteos oraz Bacula Systems zalecają swoim klientom upgrade do tej wersji. Aktualizacja do tej wersji z BEE 8.2.x nie wymaga aktualizacji katalogu. Aktualizacja z poprzednich wersji będzie tego wymagała, należy ją wcześniej zaplanować.

Aktualizacja dostępna jest dla wszystkich klientów posiadających aktualną subskrypcję oprogramowania Bacula Enterprise Edition dowolnego poziomu oraz dla wszystkich koientów usługi Inteos Backup.

Informacje szczegółowe:

             Release Notes for Bacula Enterprise 8.2.x

============== !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ===================
New Catalog format in version 8.2.0 and greater
-----------------------------------------------
This release of Bacula uses a new catalog format.  We provide a set of scripts
that convert 6.6.x and earlier catalog versions to the new 8.2.0 format (1017).
Please see the white paper "Migration-to-8.2" for detailed instructions on the
migration process.

The database upgrade from 1016 to 1017 is fast and simple.

The Bacula Enterprise RPM spec files have been modified to detect the database
change and do the upgrade for you (from 6.6 and 8.0).  We strongly recommend
that you save a copy of your existing database before upgrading using RPMs.  If
you want to have more control over the catalog upgrade step when upgrading using
RPMs, you may want to contact the Bacula Systems support team to get the
update_bacula_tables script and run it before the RPM installation step.

If you are starting from scratch, you don't need to run the update_bacula_tables
script because the create_bacula_tables script automatically creates the tables
using the new format.

New Solaris Packages
--------------------
This release of Bacula uses new Solaris packages, only the 64bit programs are
now available and important changes were made to the package layout . Binaries
are now installed under /opt/bacula/bin and Plugins are installed under
/opt/bacula/plugins. When upgrading from a 8.0.x or 6.6.x versions, it is
advised to backed up /opt/bacula, remove installed packages, and install new
ones. It might be necessary to update the PluginDirectory directive in
bacula-fd.conf and bacula-sd.conf to "/opt/bacula/plugins" to start daemons.

============== !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ===================

Release 8.2.8 31 August 2015

8.2.8 is a bug fix release and there are a few new features.

New Features:
 - Add %I option for runscripts to get the Copy/Migration JobId
 - dedup: dde_query.py: compare 2 DDE + make LZ4 optional
 - mysql: Add "innobackupex_tmpdir" option in plugin command line
 - Allow to enable new BSR optimization code at bacula-sd startup for beta testing
   (see with the Bacula Systems support team for more information.)

Fixes:
 - Fix bvfs_lsdir pattern parameter setting
 - Fix #898 truncate volumes larger than 200 bytes

 - dedup: include low level info in "Failed adding #%08x to Index" message
 - dedup: improve error message for a missing chunk

 - kvm: Use fusermount instead of umount when cleaning up the working directory
 - mysql: Fix #1241 about error on global-grants.sql generation
 - delta: Fix #1231 cancel the job if a directory is specified as source

 - bweb: Add new pool action does not cause back to BWeb admin part
 - bweb: Enable using redirection '>' character in where= restore parameter
 - bweb: Fix current language change
 - bweb: Disable prune action for volumes with volstatus different than Used and Full
 - bweb: Avoid warning when using CGI::param in array context.
 - bweb: Add possibility to use translated texts with javascripts 
 - bweb: Update translation files Makefile for Japanese language templates
 - vsphere: Fix #1191 Strip out floppy disk information during OVF import
 - vsphere: Try to use defaultOvf when getting simple Exception in importOvf()
 - vsphere: Add retry code after the exception com.vmware.vim25.ManagedObjectNotFound
 - vsphere: Allow to call vsphere-ctl from the PATH

Packages:
 - rpm: Fix btraceback permission in redhat spec file
 - Add Solaris 5.12 support
 - Add support for Solaris /etc/xxx.d configuration files
 - Add Suse 11 SP4 platform to bacula spec file
 - Add Fedora 22 platform

Bugs fixed/closed since last release:
1231 1241

Release 8.2.7 16 July 2015

8.2.7 is a bug fix release and there are a few new features.

Fixes:
------
 - Fix #1209 about bat segfault when clicking on Media
 - Fix #1099 about director crash with rescheduled jobs
 - Fix #1131 about Job::Next Pool resource precedence over the Pool::Next pool directive
 - Fix #1181 "Enable Client Rehydration" on Windows FD ERR=The data is invalid
 - Fix compilation issue on AIX 7.1
 - dedup: check FD<->SD message #1190
 - dedup: improve thread termination
 - snapshot: Fix #1210 Do not silently skip LVM volumes that are not suitable for snapshot
 - bweb: Fix restore option tab with MySQL catalog backend

New Features:
-------------
 - snapshot: Add fail_job_on_error parameter in bsnapshot.conf
 - bweb: Add folding arrows/triangles when details fields are showing/hidding 
 - bweb: Button "Add Client" in clients list now is referred to Add new Client wizard
 - mysql: Add support for multiple instances binary backup in the same fileset

Changes:
--------
 - remove all duped_bsock->destroy() and use bsock->cancel()
 - Qmsg(M_FATAL) set jcr->JobStatus to JS_FatalError immediately
 - snapshot: Abort the job by default if a snapshot creation fails
 - Remove recursion in match_findex()
 - Document the -T (use trace file for debug) option in man pages
 - snapshot: Try to detect LVM when the filesystem is ext3 or XFS
 - mysql: Avoid warning with abort_on_job plugin option

Bugs fixed/closed since last release:
1099 1131 1173 1209 1210 1190


==================================================================

Release 8.2.6 16 June 2015

8.2.6 is a major bug fix release.  We strongly advise all 8.2.x users to
upgrade your Director and all Storage daemons to version 8.2.6.

VirtualFull, Copy and Migration jobs performed with versions 8.2.x prior to
8.2.6 could potentially contain incorrect backup data if you were multiplexing
(running multiple simultaneous Jobs to the same Volume). In that case, we
recommend that after upgrading to 8.2.6, you do a new Full backup to replace
any old VirtualFull, Copy, or Migration jobs.

Fixes:
------
 - Fix #1150 about restore errors from VirtualFull/Copy/Migration jobs
 - Fix #536 about Copy/Migration/VF that should not use Client "Maximum Concurrent Jobs"
 - Fix potential segfault with unused ConfigFile objects
 - vsphere: Fix warning message when using abort_on_error option
 - oracle: Fix procedure to synchronize the local catalog with Bacula's catalog

Changes:
--------
 - #1108 Enhance setdebug help command and console completion
 - Do not print message about retention when using "prune volume yes" command
 - Add more JCR variables in lockdump procedure

Bugs fixed/closed since last release:
1150 1108 536

==================================================================


Release 8.2.5 12 June 2015

8.2.5 is a bug fix release that corrects an issue with the Global Endpoint
Deduplication engine and an issue with the catalog upgrade procedure for
PostgreSQL.

Fixes:
------
 - dedup: fix deadlock in grow()
 - Fix error in update_postgresql_tables.in

==================================================================

Release 8.2.4 08 June 2015

8.2.4 is a bug fix and performance enhancement release. Performance
is improved for restore, copy/migration and virtual full jobs (in a
larger scope than with 8.2.3).

New Features:
-------------
 - mysql: Add extra_file plugin command line option
 - oracle-sbt: Add restoreclient option in configuration file and RMAN SEND option
 - bweb: Implement "lastday" schedule keyword (#1106)
 - Add "-T" option in bacula-sd to write to a trace file

Fixes:
------
 - Fix #1127 about the repositioning enhancement during restore
 - Fix batch mode detection for SQLite3
 - Correct try_reposition() return code after a seek()
 - bweb: Fix free/total space labels in Storage Overview (bug #1121)
 - dedup: count orphan addresses in vacuum
 - Avoid problem with db_get_job_record() when SchedTime or RealEndTime is 
  NULL
 - Send console command output to Jmsg(INFO) if the UA_sock is not initialized 
 - vsphere: Avoid NullPointerException when vSphere returns null entities

Bugs fixed/closed since last release:
1106 1121 1127

==================================================================

Release 8.2.3 26 May 2015

8.2.3 is a bug fix and performance enhancement release. Performance
is improved for restore, copy/migration and virtual full jobs.

Changes:
--------
 - Fix #876 about SD reads too far with complex bootstrap
 - Fix compilation for Solaris9
 - Fix segfault on dot commands used in RunScript::Console directive
 - Fix #1092 about bad permissions on shstore plugin rpm package
 - Do not trash existing record during label of new volume
 - bweb: Enlarge clients list width in restore screen (fix #1067)
 - bweb: Replace nrs_table with bsTable library
 - vss: Try to be more flexible with Microsoft Partial files

Bugs fixed/closed since last release:
1067 1092 876

==================================================================

Release 8.2.2 13 May 2015

8.2.2 is a bug fix release that corrects issues with oracle-sbt, mysql and
postgresql plugins.  This release added the support for vSphere 6.0,
and there are a few new features.


New Features:
-------------
 - update vixdisk lib for vsphere 6
 - Add new %E job code to use non fatal job errors in scripts
 - lock the pid file using fcntl(F_SETLK) 

Changes:
--------
 - Fix #1080 about RMAN restore issue when debug turned on
 - Fix RMAN crosscheck issue introduced in 8.2.0
 - Fix #1075 The replace=never flag was not properly handled when combined with 
  database= option in mysql/postgresql plugin
 - Fix bug #1083 about Pmsg() function that do not display timestamps
 - Fix #1008 about status storage that displays "Writing" and "Reading" information 
  for the same DCR

 - dedup: avoid a race condition in vacuum when resizing the bitarray
 - dedup: Fix MA1022 starting a second vacuum was canceling the 1st one
 - dedup: fix a small bug in the error msg returned when starting vacuum twice
 - dedup: fix recovery not working when FSM index is corrupted
 - dedup: replace "*(int*)hash" typecasting by hash2int(hash) (part2)
 - dedup: delete bee_dde.tch.old before a new optimize_index()
 
 - configure.in: new HAVE_FCNTL_LOCK detect fcntl() locking capability
 - snap: Fix small initialization problem with LVM backend
 - bat: Fix #1047 about segfaults in Client, Media and Pool view


Bugs fixed/closed since last release:
1008 1047 1075 1080 1083 1022

==================================================================

Release 8.2.1 21 April 2015

8.2.1 is a bug fix release that corrects various minor issues and improves
the performance of the deduplication vacuum process.  This release also
contains a few new features:

New Features:
-------------
 - Allow to list restore jobs in the llist jobid= command
 - bweb: Add search field to configurable tables tools
 - bweb: Implement assign TLS certs/key to components while component cert is generated
 - bweb: Enable certificate generation for Console <=> Director communication
 - bweb: Add Director certificate type in cert properties
 - bweb: Add TLS certs/keys configuration for Console component in its Director resource
 - bweb: TLS encryption state in running jobs status (lock-pad icon)

Changes:
--------
 - dedup: Improve vacuum performance
 - Work bsnapshot for SLES12 and fix issue with ZFS
 - Fix #1021 about bad RPATH variable on tune-dde tool
 - Fix #858 invalid .mod command in BAT during restore
 - vss: Fix #1016 about MSSQL VSS warning when the range string points to a file
 - Fix #987 cannot copy/migrate jobs with a Level=VF in the job resource
 - Fix #940 about segfault in bat when doing an "update slots"
 - Fix #983 about segfault on win32 filedaemon when using bat to monitor the status
 - Fix #969 about a segfault while doing a cancel of a copy job
 - Fix #891 about bad "Requires" on demo rpm packages
 - Fix #972 about segfault in show command used with multiple resources
 - Fix #965 about an empty error message after a problem when sending accurate 
  file list
 - Fill errmsg after an error with FETCH query in db_big_sql_query()
 - Fix small memory leak in cancel command with ujobid and job parameters
 - Ensure that client resource is not freed during setbandwidth command
 - Fix errors in the use of a Mmsg() causing a segfault with bacula-sd -t command
 - Do not try to re-hydrate compressed stream on the client side
 - Fix Client Side Re-hydration that was disabled after the first error

 - kvm: Use libguestfs-tools-c instead of guestfish in kvm rpm package
 - kvm: Skip cdrom device in virsh domblklist
 - kvm: Fix #984 about missing option in guestfish command
 - kvm: Use umount instead of guestunmount command
 - kvm: Add missing working argument in HV_VM_MOUNT_FS procedure
 - kvm: Display drive list properly at the end of the vmmount.pl session
 - kvm: use vmmount script in exepath
 - kvm: Fix #984 about error with kvmmount
 - kvm: make install-kvm.sh script more user friendly

 - bweb: Fix zero value in CGI params (bugfix #1030)
 - bweb: Fix edit pool button in admin part Media => Pool (fix #995)
 - bweb: Fix double slash in ssl_dir path
 - bweb: Fix warning in log about uninitialized string
 - bweb: Fix set current component name if the name is not given in dir=name 
  param but as ComponentName GET param

Bugs fixed/closed since last release:
1016 1021 1030 858 891 940 965 969 972 983 984 987 858 995

==================================================================

Release 8.2.0 18 March 2015


The Director (Dir) and all Storage Daemons (SDs) must be version 8.2 because
they are not compatible with prior versions.


See above for the new catalog format.



New Features
------------
 - Snapshot Management
   Bacula Enterprise 8.2 is now able to handle filesystem snapshots on Linux/Unix
   systems using a similar approach that Bacula uses with Windows systems.
   Snapshots can be automatically created and used to backup files. It is now
   possible to manage snapshots from the Bacula bconsole tool through a unique
   interface. It is also possible to keep snapshosts after a backup and to keep
   track of them using Bacula's catalog.

   The following snapshot backends are supported with Bacula Enterprise 8.2:
    - ZFS
    - BTRFS
    - LVM


 - Global Endpoint Deduplication(TM)
   Storage to Storage Copy/Migration using the Global Endpoint Deduplication(TM)
   is now supported.

   New automatic deduplication index optimization added to the vacuum
   procedure.

   Part of the deduplication index can be locked into memory for performance.

   Users can configure parameters related to the size of the deduplication
   index and the amount of memory that can be used to cache the index. See
   the Global Endpoint Deduplication user's guide for more information.

 - New Hyper-V Plugin - Backing up and restoring Hyper-V virtual machines at
   block level is supported with Full level backups.

 - New KVM Plugin - Backing up and restoring KVM virtual machines at file level
   is now supported. The plugin uses the libvirt virtual machine management
   interface.

 - Windows Encrypted File System (EFS) support.

 - Copy/Migration/VirtualFull performance enhancements to handle very large jobs.

 - Storage Daemon reports disk usage in "status storage" command.

 - BWeb Management Suite new features:
   - FileSet Wizard
   - Storage Overview
   - TLS and certificate management
   - Job/Media table columns configuration

 - New vSphere plugin "quiesce_host=" option added to control
   Snapshot creation.

 - New Solaris packages are now available for Solaris 10, 11 Intel and Sparc
   64bit.


Changes with Directives
-----------------------
The FileDaemon "Dedup Index Directory" in bacula-fd.conf directive is
deprecated and replaced by "Enable Client Rehydration = " directive.

The "Enable Client Rehydration" FileDaemon directive is optional and allows Bacula to
try to do rehydration using existing local data on the Client at restore time.
In some cases, the use of this directive permits Bacula to transfer less data
over the network during a restore. The default value is "no".


-----------------------------------------------------------------------------


Changes:
 - Update extras package in debs/RPM package
 - Upgrade catalog version to 1017
 - Add VSS Hyper-V plugin
 - Add EFS in the client status flag list
 - Implement Windows EFS Support

 - bweb: Display VolBytes and total VolSize
 - bweb: Implement new FileSet wizard
 - bweb: Show local components versions with "Check Bweb Configuration" call
 - bweb: Implement Storage Overview subpage
 - bweb: Enable support for deploying Windows clients with encrypted connection
 - bweb: Implement security center for encrypted connection configuration
 - bweb: Implement OperatorCommand, MailCommand, MailOnSuccess and MailOnError
         on Messages
 - bweb: Add BaseJob option to Job and FileSet editing pages

 - dedup: SD-SD avoid rehydration
 - dedup: Add tune-dde utility
 - dedup: Enhance the vacuum process
 - dedup: Add support for VirtualFull between dedup and non-dedup device
 - dedup: Add a first cut of check_free_space() in dedupengine class
 - dedup: Add -t option to bextract to test the integrity of a backup
 - dedup: Allows "btools" to open the dedupengine readonly addendum
 - dedup: Remove OpenSSL dependency
 - dedup: Allows "btools" to open the dedupengine readonly
 - dedup: Update APIv2 to current "dedup usage" command
 - dedup: Display DDE information with variable units
 - dedup: New "opportunistic dedup restore" to replace "dedup client cache".
          The "opportunistic restore" principle is a replacement for the
          "dedup client cache" used in version 8.0.1. The goal is to avoid
          the transfer of chunks from the SD at restore time by reusing
          identical chunks that can be found on the target itself. The new
          design is much simpler and should perform just as well.
 - dedup: Calculate and maintain dedup ratio

 - Add KVM plugin
 - Add ReadBytes in the status client output
 - Add fs_get_free_space() function
 - Remove recursion from free_bsr() and free_bsr_item() to handle very large
   BSR
 - Add Accurate FileSet option "M" to compare ctime/mtime with the save_time
   like with normal Incrementals
 - Add index on Job(JobTDate) to MySQL catalog
 - Add read_only and autoselect to the device status ouput when using API v2
 - Allow the use of "device=" option in release/mount/unmount commands
 - Report device number, maximum concurrent jobs and total and free space for
   File devices
 - Add @encode and sp_decode functions for plugins
 - vsphere: Add new quiesce_host=Yes/no/try Plugin command line option
 - vsphere: Use getSnapshotInTree instead of getCurrentSnapshot to locate the
   new snapshot
 - vsphere: Use try as default value for quiesce_host parameter
 - mysql: Avoid warning with dump_opt option. Fix #798
 - Add abort_on_error parameter in postgresql and mysql plugins
 - postgresql: Display more debug information when starting PITR backup
 - postgresql: Avoid the use of "pg_restore -c -C" on old PostgreSQL databases
 - postgresql: Allow the restoration of databases with "." in the name, do not
   use "-e" switch in pg_restore

Fixes:
 - Fix #951 about SDCallsClient not used during restore jobs
 - Fix #431 upon upgrade, RPM resets group membership
 - Fix jobmedia creation bug
 - Fix #875 about bvfs repeats the same output many times
 - Fix segfault on Windows FD when stopping the service
 - Fix Windows VSS plugin memory leak
 - Fix #843 "show storage" option missing in the help command output
 - Fix use bzip2 for SLES dependency
 - Fix empty password in Windows 2012 R2 installer
 - Fix #737 Windows plugins not uninstalled properly
 - Fix #805 nextpool command line argument not properly used
 - Fix avoid segfault in connect_to_file_daemon() when jcr->client is NULL
 - Fix #776 volume created in the catalog but not on disk and #464 SD can't
       read an existing volume
 - Add schedule to show command tab completion
 - dedup: Fix #763 HASH mismatch, ASSERT replacement
 - Set file volumes that are not accessible to "Error" in mount_next_vol
 - Fix #769 by adding a dependency to bacula-enterprise-libs on dedup, shstore
       and aligned SD plugins RPMs
 - Fix #743 about bat permission conflict on /opt/bacula/etc
 - Fix #769 about dependency problem with dedup RPM package
 - Fix #747 about restore problem failing on "Unexpected cryptographic session
       data stream"
 - Fix bad check on bopen_rsrc return status. bug #2106
 - Fix checksum error in btape fill command
 - Fix double free in btape
 - Fix failed mount request in btape fill test
 - Avoid ASSERT() when using btape with vtape driver
 - Possible fix for NULL client bug #2105
 - Fix compilation of Nagios check_bacula
 - dedup: Fix #787 status displays funny chars
 - dedup: Remove the "dedup usage reset" from the "dedup" menu
 - dedup: lacc_calc_hash not incremented when dedup=bothsides
 - dedup: dedup=storage was calculating the hash twice for every chunk
 - Fix "chunk" was written as "chunck" in some output and for some variable names
 - dedup: Fix dedup usage command mismatch between used and unused diskspace
 - Fix #699 about duplicated job name when starting two jobs at the same time
 - Fix #701 about status schedule missing from tab completion and correct job filter
 - Fix #346 Add ipv6 support for Solaris
 - Fix #588 Improve SD BSR read performance
 - Fix #660 about Dedup Index on FD
 - Fix #692 about compatibility issue with community FD
 - Fix QT Windows build for 32bit
 - Fix tls-duplicate-job seg fault + harden pthread_kill() code
 - vsphere: Fix #13779 about error message during clean operation
 - vsphere: Fix #13736 about pre-freeze and post-thaw scripts not executed correctly
 - bweb: Fix #938 about a bug in "Backup Policy" wizard on MaxRunSchedTime
 - bweb: Fix for empty OS Type value in new client wizard
 - bweb: Fix #663 about bad generated FSType output when editing a FileSet
 - bweb: Fix #689 about a problem with back button in the "New Backup" wizard
 - bweb: Fix a problem in editing Messages resources caused by fixing #641
 - bweb: Replace "Base" option with "BaseJob" in FileSet edit dialog box
 - bweb: Fix #641 about not showing rename Name message while cancel button is
         clicked
 - bweb: Fix cleanup session variables for Integrate New Client wizard (fix #955)

Cleanup:
 - Simplify concatenations (cats) by eliminating the B_DB_PRIV class
 - Convert db_lock/unlock to be called via class
 - Remove more vestiges of Ingres
 - Add license to files without any
 - Make global "my_name" hold 127 chars
 - Add copyright to Makefiles
 - Fix btape fill command by removing some debug code in empty_block()
 - Call terminate_stored() properly in SD when pressing ctrl-C 

Bugs fixed/closed since last release:
1934 2105 2106 346 431 588 641 660 663 689 692 699 701 737 743 747 763 769 776
787 798 805 843 875 931 938 951 955 13779 13736