8 *(Released Tue, 24 Jul 2012)*
10 A small bugfix release, with no new features:
12 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
13 different default shell (e.g. Debian, Ubuntu)
14 - fixed KVM startup and live migration with a custom keymap (fixes Issue
15 243 and Debian bug #650664)
16 - fixed compatibility with KVM versions that don't support multiple boot
17 devices (fixes Issue 230 and Debian bug #624256)
19 Additionally, a few fixes were done to the build system (fixed parallel
20 build failures) and to the unittests (fixed race condition in test for
21 *FileID functions, and the default enable/disable mode for QA test is
28 *(Released Fri, 11 May 2012)*
30 A small bugfix release.
32 The main issues solved are on the topic of compatibility with newer LVM
35 - fixed parsing of ``lv_attr`` field
36 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
37 the ``--force`` flag is needed
39 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
40 to accept kernel 3.x too (was hardcoded to 2.6.*).
42 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
43 of hook results and that also made display of other errors suboptimal
44 was fixed; the code behaves now like 2.4 and earlier.
46 Another change in 2.5, the cleanup of the OS scripts environment, is too
47 aggressive: it removed even the ``PATH`` variable, which requires the OS
48 scripts to *always* need to export it. Since this is a bit too strict,
49 we now export a minimal PATH, the same that we export for hooks.
51 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
52 integrated into this release.
54 Finally, a few other miscellaneous changes were done (no new features,
55 just small improvements):
57 - Fix ``gnt-group --help`` display
58 - Fix hardcoded Xen kernel path
59 - Fix grow-disk handling of invalid units
60 - Update synopsis for ``gnt-cluster repair-disk-sizes``
61 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
67 *(Released Thu, 12 Apr 2012)*
69 Incompatible/important changes and bugfixes
70 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
72 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
73 resource's ``ip_check`` parameter changed from ``True`` to ``False``
74 to match the underlying LUXI interface.
75 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
76 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
77 not maintain backwards-compatibility as the underlying operation
78 changed in an incompatible way. The RAPI client can talk to old
79 servers, but it needs to be told so as the return value changed.
80 - When creating file-based instances via RAPI, the ``file_driver``
81 parameter no longer defaults to ``loop`` and must be specified.
82 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
84 - Support for the undocumented and deprecated RAPI instance creation
85 request format version 0 has been dropped. Use version 1, supported
86 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
87 - Pyparsing 1.4.6 or above is required, see :doc:`installation
88 documentation <install>`.
89 - The "cluster-verify" hooks are now executed per group by the
90 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
91 if you just run ``gnt-cluster verify``, which generates one opcode per
93 - The environment as passed to the OS scripts is cleared, and thus no
94 environment variables defined in the node daemon's environment will be
95 inherited by the scripts.
96 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
98 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
99 support operations involving multiple node groups.
100 - Offline nodes are ignored when failing over an instance.
101 - Support for KVM version 1.0, which changed the version reporting format
103 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
105 - ``Makefile`` is now compatible with Automake 1.11.2
106 - Includes all bugfixes made in the 2.4 series
111 - The ganeti-htools project has been merged into the ganeti-core source
112 tree and will be built as part of Ganeti (see :doc:`install-quick`).
113 - Implemented support for :doc:`shared storage <design-shared-storage>`.
114 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
115 GPT-style partition tables (requires `parted
116 <http://www.gnu.org/s/parted/>`_).
117 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
118 - Allowed adding tags on instance creation.
119 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
120 - Allow instances to be started in a paused state, enabling the user to
121 see the complete console output on boot using the console.
122 - Added new hypervisor flag to control default reboot behaviour
123 (``reboot_behavior``).
124 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
125 - Improved out-of-band management support:
127 - Added ``gnt-node health`` command reporting the health status of
129 - Added ``gnt-node power`` command to manage power status of nodes.
130 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
132 - Instance migration can fall back to failover if instance is not
134 - Filters can be used when listing nodes, instances, groups and locks;
135 see *ganeti(7)* manpage.
136 - Added post-execution status as variables to :doc:`hooks <hooks>`
138 - Instance tags are exported/imported together with the instance.
139 - When given an explicit job ID, ``gnt-job info`` will work for archived
141 - Jobs can define dependencies on other jobs (not yet supported via
142 RAPI or command line, but used by internal commands and usable via
145 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
148 - Instance failover is now available as a RAPI resource
149 (``/2/instances/[instance_name]/failover``).
150 - ``gnt-instance info`` defaults to static information if primary node
152 - Opcodes have a new ``comment`` attribute.
153 - Added basic SPICE support to KVM hypervisor.
154 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
156 Node group improvements
157 ~~~~~~~~~~~~~~~~~~~~~~~
159 - ``gnt-cluster verify`` has been modified to check groups separately,
160 thereby improving performance.
161 - Node group support has been added to ``gnt-cluster verify-disks``,
162 which now operates per node group.
163 - Watcher has been changed to work better with node groups.
165 - One process and state file per node group.
166 - Slow watcher in one group doesn't block other group's watcher.
168 - Added new command, ``gnt-group evacuate``, to move all instances in a
169 node group to other groups.
170 - Added ``gnt-instance change-group`` to move an instance to another
172 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
173 per-group operations.
174 - Node groups can be tagged.
175 - Some operations switch from an exclusive to a shared lock as soon as
177 - Instance's primary and secondary nodes' groups are now available as
178 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
179 and ``snodes.group.uuid``).
184 - Numerous updates to documentation and manpages.
186 - :doc:`RAPI <rapi>` documentation now has detailed parameter
188 - Some opcode/job results are now also documented, see :doc:`RAPI
191 - A lockset's internal lock is now also visible in lock monitor.
192 - Log messages from job queue workers now contain information about the
193 opcode they're processing.
194 - ``gnt-instance console`` no longer requires the instance lock.
195 - A short delay when waiting for job changes reduces the number of LUXI
196 requests significantly.
197 - DRBD metadata volumes are overwritten with zeros during disk creation.
198 - Out-of-band commands no longer acquire the cluster lock in exclusive
200 - ``devel/upload`` now uses correct permissions for directories.
206 *(Released Fri, 23 Mar 2012)*
208 This was the sixth release candidate of the 2.5 series.
214 *(Released Mon, 9 Jan 2012)*
216 This was the fifth release candidate of the 2.5 series.
222 *(Released Thu, 27 Oct 2011)*
224 This was the fourth release candidate of the 2.5 series.
230 *(Released Wed, 26 Oct 2011)*
232 This was the third release candidate of the 2.5 series.
238 *(Released Tue, 18 Oct 2011)*
240 This was the second release candidate of the 2.5 series.
246 *(Released Tue, 4 Oct 2011)*
248 This was the first release candidate of the 2.5 series.
254 *(Released Wed, 31 Aug 2011)*
256 This was the third beta release of the 2.5 series.
262 *(Released Mon, 22 Aug 2011)*
264 This was the second beta release of the 2.5 series.
270 *(Released Fri, 12 Aug 2011)*
272 This was the first beta release of the 2.5 series.
278 *(Released Thu, 27 Oct 2011)*
280 - Fixed bug when parsing command line parameter values ending in
282 - Fixed assertion error after unclean master shutdown
283 - Disable HTTP client pool for RPC, significantly reducing memory usage
285 - Fixed queue archive creation with wrong permissions
291 *(Released Tue, 23 Aug 2011)*
295 - Fixed documentation for importing with ``--src-dir`` option
296 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
297 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
303 *(Released Fri, 5 Aug 2011)*
305 Many bug-fixes and a few small features:
307 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
308 issues when you tried to add an instance with two MAC addresses in one
310 - KVM: fixed per-instance stored UID value
311 - KVM: configure bridged NICs at migration start
312 - KVM: Fix a bug where instance will not start with never KVM versions
314 - Added OS search path to ``gnt-cluster info``
315 - Fixed an issue with ``file_storage_dir`` where you were forced to
316 provide an absolute path, but the documentation states it is a
317 relative path, the documentation was right
318 - Added a new parameter to instance stop/start called ``--no-remember``
319 that will make the state change to not be remembered
320 - Implemented ``no_remember`` at RAPI level
321 - Improved the documentation
322 - Node evacuation: don't call IAllocator if node is already empty
323 - Fixed bug in DRBD8 replace disks on current nodes
324 - Fixed bug in recreate-disks for DRBD instances
325 - Moved assertion checking locks in ``gnt-instance replace-disks``
326 causing it to abort with not owning the right locks for some situation
327 - Job queue: Fixed potential race condition when cancelling queued jobs
328 - Fixed off-by-one bug in job serial generation
329 - ``gnt-node volumes``: Fix instance names
330 - Fixed aliases in bash completion
331 - Fixed a bug in reopening log files after being sent a SIGHUP
332 - Added a flag to burnin to allow specifying VCPU count
333 - Bugfixes to non-root Ganeti configuration
339 *(Released Thu, 12 May 2011)*
341 Many bug-fixes and a few new small features:
343 - Fixed a bug related to log opening failures
344 - Fixed a bug in instance listing with orphan instances
345 - Fixed a bug which prevented resetting the cluster-level node parameter
346 ``oob_program`` to the default
347 - Many fixes related to the ``cluster-merge`` tool
348 - Fixed a race condition in the lock monitor, which caused failures
349 during (at least) creation of many instances in parallel
350 - Improved output for gnt-job info
351 - Removed the quiet flag on some ssh calls which prevented debugging
353 - Improved the N+1 failure messages in cluster verify by actually
354 showing the memory values (needed and available)
355 - Increased lock attempt timeouts so that when executing long operations
356 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
357 too early and thus prevent the use of the 'fair' mechanism
358 - Changed instance query data (``gnt-instance info``) to not acquire
359 locks unless needed, thus allowing its use on locked instance if only
360 static information is asked for
361 - Improved behaviour with filesystems that do not support rename on an
363 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
364 kept locks on all nodes during the wipe, which is unneeded
365 - Fixed ``gnt-watcher`` handling of errors during hooks execution
366 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
367 10GiB) which caused the wipe to fail right at the end in some cases
368 - Fixed master IP activation when doing master failover with no-voting
369 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
370 the master node itself
371 - Fixed potential data-loss in under disk full conditions, where Ganeti
372 wouldn't check correctly the return code and would consider
373 partially-written files 'correct'
374 - Fixed bug related to multiple VGs and DRBD disk replacing
375 - Added new disk parameter ``metavg`` that allows placement of the meta
376 device for DRBD in a different volume group
377 - Fixed error handling in the node daemon when the system libc doesn't
378 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
379 - Fixed lock release during replace-disks, which kept cluster-wide locks
380 when doing disk replaces with an iallocator script
381 - Added check for missing bridges in cluster verify
382 - Handle EPIPE errors while writing to the terminal better, so that
383 piping the output to e.g. ``less`` doesn't cause a backtrace
384 - Fixed rare case where a ^C during Luxi calls could have been
385 interpreted as server errors, instead of simply terminating
386 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
388 - Added a few more parameters to the KVM hypervisor, allowing a second
389 CDROM, custom disk type for CDROMs and a floppy image
390 - Removed redundant message in instance rename when the name is given
392 - Added option to ``gnt-instance recreate-disks`` to allow creating the
393 disks on new nodes, allowing recreation when the original instance
394 nodes are completely gone
395 - Added option when converting disk templates to DRBD to skip waiting
396 for the resync, in order to make the instance available sooner
397 - Added two new variables to the OS scripts environment (containing the
399 - Made the root_path and optional parameter for the xen-pvm hypervisor,
400 to allow use of ``pvgrub`` as bootloader
401 - Changed the instance memory modifications to only check out-of-memory
402 conditions on memory increases, and turned the secondary node warnings
403 into errors (they can still be overridden via ``--force``)
404 - Fixed the handling of a corner case when the Python installation gets
405 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
406 to execute a command that doesn't exist
407 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
408 node of the instance returned failures during instance shutdown; this
409 adds the option ``--ignore-consistency`` to gnt-instance move
411 And as usual, various improvements to the error messages, documentation
417 *(Released Wed, 09 Mar 2011)*
419 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
420 the RAPI users file if run twice (even with ``--dry-run``).
422 The release fixes that bug (nothing else changed).
428 *(Released Mon, 07 Mar 2011)*
430 Final 2.4.0 release. Just a few small fixes:
432 - Fixed RAPI node evacuate
433 - Fixed the kvm-ifup script
434 - Fixed internal error handling for special job cases
435 - Updated man page to specify the escaping feature for options
441 *(Released Mon, 28 Feb 2011)*
443 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
444 that this feature wiped the disks of the wrong instance, leading to loss
449 - Fixed title of query field containing instance name
450 - Expanded the glossary in the documentation
451 - Fixed one unittest (internal issue)
457 *(Released Mon, 21 Feb 2011)*
459 A number of bug fixes plus just a couple functionality changes.
461 On the user-visible side, the ``gnt-* list`` command output has changed
462 with respect to "special" field states. The current rc1 style of display
463 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
464 the default output mode special fields states are displayed as follows:
466 - Offline resource: ``*``
467 - Unavailable/not applicable: ``-``
468 - Data missing (RPC failure): ``?``
469 - Unknown field: ``??``
471 Another user-visible change is the addition of ``--force-join`` to
476 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
477 - Fixed regression in RAPI/instance reinstall where all parameters were
478 required (instead of optional)
479 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
480 - Fixed iallocator usage (offline nodes were not considered offline)
481 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
482 - Fixed hypervisor and OS parameter validation with respect to
484 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
486 - Fixed ``tools/listrunner`` with respect to agent-based usage
492 *(Released Fri, 4 Feb 2011)*
494 Many changes and fixes since the beta1 release. While there were some
495 internal changes, the code has been mostly stabilised for the RC
498 Note: the dumb allocator was removed in this release, as it was not kept
499 up-to-date with the IAllocator protocol changes. It is recommended to
500 use the ``hail`` command from the ganeti-htools package.
502 Note: the 2.4 and up versions of Ganeti are not compatible with the
503 0.2.x branch of ganeti-htools. You need to upgrade to
504 ganeti-htools-0.3.0 (or later).
506 Regressions fixed from 2.3
507 ~~~~~~~~~~~~~~~~~~~~~~~~~~
509 - Fixed the ``gnt-cluster verify-disks`` command
510 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
512 - Fixed disk adoption breakage
513 - Fixed wrong headers in instance listing for field aliases
518 - Fixed corner case in KVM handling of NICs
519 - Fixed many cases of wrong handling of non-vm_capable nodes
520 - Fixed a bug where a missing instance symlink was not possible to
521 recreate with any ``gnt-*`` command (now ``gnt-instance
522 activate-disks`` does it)
523 - Fixed the volume group name as reported by ``gnt-cluster
525 - Increased timeouts for the import-export code, hopefully leading to
526 fewer aborts due network or instance timeouts
527 - Fixed bug in ``gnt-node list-storage``
528 - Fixed bug where not all daemons were started on cluster
529 initialisation, but only at the first watcher run
530 - Fixed many bugs in the OOB implementation
531 - Fixed watcher behaviour in presence of instances with offline
533 - Fixed instance list output for instances running on the wrong node
534 - a few fixes to the cluster-merge tool, but it still cannot merge
535 multi-node groups (currently it is not recommended to use this tool)
541 - Improved network configuration for the KVM hypervisor
542 - Added e1000 as a supported NIC for Xen-HVM
543 - Improved the lvmstrap tool to also be able to use partitions, as
544 opposed to full disks
545 - Improved speed of disk wiping (the cluster parameter
546 ``prealloc_wipe_disks``, so that it has a low impact on the total time
547 of instance creations
548 - Added documentation for the OS parameters
549 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
550 hypervisor is not responding
551 - Added display of blacklisted and hidden OS information in
553 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
554 NIC and node parameters, which might create problems with currently
555 invalid (but undetected) configuration files, but prevents validation
556 failures when unrelated parameters are modified
557 - Changed cluster initialisation to wait for the master daemon to become
559 - Expanded the RAPI interface:
561 - Added config redistribution resource
562 - Added activation/deactivation of instance disks
563 - Added export of console information
565 - Implemented log file reopening on SIGHUP, which allows using
566 logrotate(8) for the Ganeti log files
567 - Added a basic OOB helper script as an example
573 *(Released Fri, 14 Jan 2011)*
578 - Fixed timezone issues when formatting timestamps
579 - Added support for node groups, available via ``gnt-group`` and other
581 - Added out-of-band framework and management, see :doc:`design
582 document <design-oob>`
583 - Removed support for roman numbers from ``gnt-node list`` and
584 ``gnt-instance list``.
585 - Allowed modification of master network interface via ``gnt-cluster
586 modify --master-netdev``
587 - Accept offline secondaries while shutting down instance disks
588 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
589 - Added support for multiple LVM volume groups
590 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
592 - Added commands to list available fields:
594 - ``gnt-node list-fields``
595 - ``gnt-group list-fields``
596 - ``gnt-instance list-fields``
598 - Updated documentation and man pages
603 - Moved ``rapi_users`` file into separate directory, now named
604 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
606 - Added new tool for running commands on many machines,
607 ``tools/ganeti-listrunner``
608 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
609 improving the ``gnt-instance console`` output
610 - Allowed customisation of disk index separator at ``configure`` time
611 - Export node group allocation policy to :doc:`iallocator <iallocator>`
612 - Added support for non-partitioned md disks in ``lvmstrap``
613 - Added script to gracefully power off KVM instances
614 - Split ``utils`` module into smaller parts
615 - Changed query operations to return more detailed information, e.g.
616 whether an information is unavailable due to an offline node. To use
617 this new functionality, the LUXI call ``Query`` must be used. Field
618 information is now stored by the master daemon and can be retrieved
619 using ``QueryFields``. Instances, nodes and groups can also be queried
620 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
621 exposed via RAPI). The following commands make use of this
622 infrastructure change:
625 - ``gnt-group list-fields``
627 - ``gnt-node list-fields``
628 - ``gnt-instance list``
629 - ``gnt-instance list-fields``
630 - ``gnt-debug locks``
635 - New RAPI resources (see :doc:`rapi`):
639 - ``/2/groups/[group_name]``
640 - ``/2/groups/[group_name]/assign-nodes``
641 - ``/2/groups/[group_name]/modify``
642 - ``/2/groups/[group_name]/rename``
643 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
647 - Implemented ``no_install`` for instance creation
648 - Implemented OS parameters for instance reinstallation, allowing
649 use of special settings on reinstallation (e.g. for preserving data)
654 - Added IPv6 support in import/export
655 - Pause DRBD synchronization while wiping disks on instance creation
656 - Updated unittests and QA scripts
657 - Improved network parameters passed to KVM
658 - Converted man pages from docbook to reStructuredText
664 *(Released Mon, 20 Dec 2010)*
666 Released version 2.3.1~rc1 without any changes.
672 *(Released Wed, 1 Dec 2010)*
674 - impexpd: Disable OpenSSL compression in socat if possible (backport
675 from master, commit e90739d625b, see :doc:`installation guide
676 <install-quick>` for details)
677 - Changed unittest coverage report to exclude test scripts
678 - Added script to check version format
684 *(Released Wed, 1 Dec 2010)*
686 Released version 2.3.0~rc1 without any changes.
692 *(Released Fri, 19 Nov 2010)*
694 A number of bugfixes and documentation updates:
696 - Update ganeti-os-interface documentation
697 - Fixed a bug related to duplicate MACs or similar items which should be
699 - Fix breakage in OS state modify
700 - Reinstall instance: disallow offline secondaries (fixes bug related to
701 OS changing but reinstall failing)
702 - plus all the other fixes between 2.2.1 and 2.2.2
708 *(Released Tue, 2 Nov 2010)*
710 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
711 - Fixed master failover race with watcher
712 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
714 - Accept previously stopped instance for export with instance removal
715 - Simplify and extend the environment variables for instance OS scripts
716 - Added new node flags, ``master_capable`` and ``vm_capable``
717 - Added optional instance disk wiping prior during allocation. This is a
718 cluster-wide option and can be set/modified using
719 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
720 - Added IPv6 support, see :doc:`design document <design-2.3>` and
722 - Added a new watcher option (``--ignore-pause``)
723 - Added option to ignore offline node on instance start/stop
724 (``--ignore-offline``)
725 - Allow overriding OS parameters with ``gnt-instance reinstall``
726 - Added ability to change node's secondary IP address using ``gnt-node
728 - Implemented privilege separation for all daemons except
729 ``ganeti-noded``, see ``configure`` options
730 - Complain if an instance's disk is marked faulty in ``gnt-cluster
732 - Implemented job priorities (see ``ganeti(7)`` manpage)
733 - Ignore failures while shutting down instances during failover from
735 - Exit daemon's bootstrap process only once daemon is ready
736 - Export more information via ``LUInstanceQuery``/remote API
737 - Improved documentation, QA and unittests
738 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
739 restart if the file was created or changed
740 - Added LUXI protocol version sent with each request and response,
741 allowing detection of server/client mismatches
742 - Moved the Python scripts among gnt-* and ganeti-* into modules
743 - Moved all code related to setting up SSH to an external script,
745 - Infrastructure changes for node group support in future versions
751 *(Released Fri, 19 Nov 2010)*
753 A few small bugs fixed, and some improvements to the build system:
755 - Fix documentation regarding conversion to drbd
756 - Fix validation of parameters in cluster modify (``gnt-cluster modify
758 - Fix error handling in node modify with multiple changes
759 - Allow remote imports without checked names
765 *(Released Tue, 19 Oct 2010)*
767 - Disable SSL session ID cache in RPC client
773 *(Released Thu, 14 Oct 2010)*
775 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
776 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
778 - Fix problems with interaction between Curl and Python's HTTP server,
779 resulting in increased speed in many RPC calls
780 - Improve our release script to prevent breakage with older aclocal and
787 *(Released Thu, 7 Oct 2010)*
789 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
790 value retrieved from master
791 - Added support for blacklisted or hidden OS definitions
792 - Added simple lock monitor (accessible via (``gnt-debug locks``)
793 - Added support for -mem-path in KVM hypervisor abstraction layer
794 - Allow overriding instance parameters in tool for inter-cluster
795 instance moves (``tools/move-instance``)
796 - Improved opcode summaries (e.g. in ``gnt-job list``)
797 - Improve consistency of OS listing by sorting it
798 - Documentation updates
804 *(Released Fri, 8 Oct 2010)*
806 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
812 *(Released Mon, 4 Oct 2010)*
814 - Fixed regression in ``gnt-instance rename``
820 *(Released Wed, 22 Sep 2010)*
822 - Fixed OS_VARIANT variable for OS scripts
823 - Fixed cluster tag operations via RAPI
824 - Made ``setup-ssh`` exit with non-zero code if an error occurred
825 - Disabled RAPI CA checks in watcher
831 *(Released Mon, 23 Aug 2010)*
833 - Support DRBD versions of the format "a.b.c.d"
835 - Re-introduce support for usage from multiple threads in RAPI client
836 - Instance renames and modify via RAPI
837 - Work around race condition between processing and archival in job
839 - Mark opcodes following failed one as failed, too
840 - Job field ``lock_status`` was removed due to difficulties making it
841 work with the changed job queue in Ganeti 2.2; a better way to monitor
842 locks is expected for a later 2.2.x release
843 - Fixed dry-run behaviour with many commands
844 - Support ``ssh-agent`` again when adding nodes
845 - Many additional bugfixes
851 *(Released Fri, 30 Jul 2010)*
853 Important change: the internal RPC mechanism between Ganeti nodes has
854 changed from using a home-grown http library (based on the Python base
855 libraries) to use the PycURL library. This requires that PycURL is
856 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
857 against GnuTLS by default. cURL's support for GnuTLS had known issues
858 before cURL 7.21.0 and we recommend using the latest cURL release or
859 linking against OpenSSL. Most other distributions already link PycURL
860 and cURL against OpenSSL. The command::
862 python -c 'import pycurl; print pycurl.version'
864 can be used to determine the libraries PycURL and cURL are linked
867 Other significant changes:
869 - Rewrote much of the internals of the job queue, in order to achieve
870 better parallelism; this decouples job query operations from the job
871 processing, and it should allow much nicer behaviour of the master
872 daemon under load, and it also has uncovered some long-standing bugs
873 related to the job serialisation (now fixed)
874 - Added a default iallocator setting to the cluster parameters,
875 eliminating the need to always pass nodes or an iallocator for
876 operations that require selection of new node(s)
877 - Added experimental support for the LXC virtualization method
878 - Added support for OS parameters, which allows the installation of
879 instances to pass parameter to OS scripts in order to customise the
881 - Added a hypervisor parameter controlling the migration type (live or
882 non-live), since hypervisors have various levels of reliability; this
883 has renamed the 'live' parameter to 'mode'
884 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
885 logical volumes, meaning that cluster verify will ignore them and not
886 flag their presence as errors
887 - The watcher will now reset the error count for failed instances after
888 8 hours, thus allowing self-healing if the problem that caused the
889 instances to be down/fail to start has cleared in the meantime
890 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
891 check for, and warn, if the drbd module parameter ``usermode_helper``
892 is not consistent with the cluster-wide setting; this is needed to
893 make diagnose easier of failed drbd creations
894 - Started adding base IPv6 support, but this is not yet
895 enabled/available for use
896 - Rename operations (cluster, instance) will now return the new name,
897 which is especially useful if a short name was passed in
898 - Added support for instance migration in RAPI
899 - Added a tool to pre-configure nodes for the SSH setup, before joining
900 them to the cluster; this will allow in the future a simplified model
901 for node joining (but not yet fully enabled in 2.2); this needs the
902 paramiko python library
903 - Fixed handling of name-resolving errors
904 - Fixed consistency of job results on the error path
905 - Fixed master-failover race condition when executed multiple times in
907 - Fixed many bugs related to the job queue (mostly introduced during the
908 2.2 development cycle, so not all are impacting 2.1)
909 - Fixed instance migration with missing disk symlinks
910 - Fixed handling of unknown jobs in ``gnt-job archive``
911 - And many other small fixes/improvements
915 - Enhanced both the unittest and the QA coverage
916 - Switched the opcode validation to a generic model, and extended the
917 validation to all opcode parameters
918 - Changed more parts of the code that write shell scripts to use the
920 - Switched the master daemon to use the asyncore library for the Luxi
927 *(Released Thu, 17 Jun 2010)*
929 - Added tool (``move-instance``) and infrastructure to move instances
930 between separate clusters (see :doc:`separate documentation
931 <move-instance>` and :doc:`design document <design-2.2>`)
932 - Added per-request RPC timeout
933 - RAPI now requires a Content-Type header for requests with a body (e.g.
934 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
935 :rfc:`2616` (HTTP/1.1), section 7.2.1)
936 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
938 - Implemented initial support for running Ganeti daemons as separate
939 users, see configure-time flags ``--with-user-prefix`` and
940 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
942 - Instances can be removed after export (``gnt-backup export
944 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
945 key (instead of 1024 bit)
946 - Added new cluster configuration file for cluster domain secret
947 - Import/export now use SSL instead of SSH
948 - Added support for showing estimated time when exporting an instance,
949 see the ``ganeti-os-interface(7)`` manpage and look for
956 *(Released Tue, 16 Nov 2010)*
958 Some more bugfixes. Unless critical bugs occur, this will be the last
961 - Fix case of MAC special-values
962 - Fix mac checker regex
963 - backend: Fix typo causing "out of range" error
964 - Add missing --units in gnt-instance list man page
970 *(Released Tue, 24 Aug 2010)*
973 - Don't ignore secondary node silently on non-mirrored disk templates
975 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
976 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
977 - Properly document the usb_mouse parameter
978 - Fix path in ganeti-rapi(8) (issue 116)
979 - Adjust error message when the ganeti user's .ssh directory is
981 - Add same-node-check when changing the disk template to drbd
987 *(Released Fri, 16 Jul 2010)*
990 - Add an option to only select some reboot types during qa/burnin.
991 (on some hypervisors consequent reboots are not supported)
992 - Fix infrequent race condition in master failover. Sometimes the old
993 master ip address would be still detected as up for a short time
994 after it was removed, causing failover to fail.
995 - Decrease mlockall warnings when the ctypes module is missing. On
996 Python 2.4 we support running even if no ctypes module is installed,
997 but we were too verbose about this issue.
998 - Fix building on old distributions, on which man doesn't have a
1000 - Fix RAPI not to ignore the MAC address on instance creation
1001 - Implement the old instance creation format in the RAPI client.
1007 *(Released Thu, 01 Jul 2010)*
1009 A small bugfix release:
1010 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1011 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1012 a non-existing option
1013 - Fix instance create: the --force-variant option was ignored
1014 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1015 - Fix modify node storage with non-FQDN arguments
1016 - Fix RAPI client to authenticate under Python 2.6 when used
1017 for more than 5 requests needing authentication
1018 - Fix gnt-instance modify -t (storage) giving a wrong error message
1019 when converting a non-shutdown drbd instance to plain
1025 *(Released Fri, 18 Jun 2010)*
1027 A small bugfix release:
1029 - Fix live migration of KVM instances started with older Ganeti
1030 versions which had fewer hypervisor parameters
1031 - Fix gnt-instance grow-disk on down instances
1032 - Fix an error-reporting bug during instance migration
1033 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1034 silently ignoring broken ones
1035 - Fix an RPC error reporting bug affecting, for example, RAPI client
1037 - Fix bug triggered by different API version os-es on different nodes
1038 - Fix a bug in instance startup with custom hvparams: OS level
1039 parameters would fail to be applied.
1040 - Fix the RAPI client under Python 2.6 (but more work is needed to
1041 make it work completely well with OpenSSL)
1042 - Fix handling of errors when resolving names from DNS
1048 *(Released Thu, 3 Jun 2010)*
1050 A medium sized development cycle. Some new features, and some
1051 fixes/small improvements/cleanups.
1053 Significant features
1054 ~~~~~~~~~~~~~~~~~~~~
1056 The node deamon now tries to mlock itself into memory, unless the
1057 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1058 its logs, and falls back to console logging. This allows emergency
1059 features such as ``gnt-node powercycle`` to work even in the event of a
1060 broken node disk (tested offlining the disk hosting the node's
1061 filesystem and dropping its memory caches; don't try this at home)
1063 KVM: add vhost-net acceleration support. It can be tested with a new
1064 enough version of the kernel and of qemu-kvm.
1066 KVM: Add instance chrooting feature. If you use privilege dropping for
1067 your VMs you can also now force them to chroot to an empty directory,
1068 before starting the emulated guest.
1070 KVM: Add maximum migration bandwith and maximum downtime tweaking
1071 support (requires a new-enough version of qemu-kvm).
1073 Cluster verify will now warn if the master node doesn't have the master
1074 ip configured on it.
1076 Add a new (incompatible) instance creation request format to RAPI which
1077 supports all parameters (previously only a subset was supported, and it
1078 wasn't possible to extend the old format to accomodate all the new
1079 features. The old format is still supported, and a client can check for
1080 this feature, before using it, by checking for its presence in the
1081 ``features`` RAPI resource.
1083 Now with ancient latin support. Try it passing the ``--roman`` option to
1084 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1085 (requires the python-roman module to be installed, in order to work).
1090 As usual many internal code refactorings, documentation updates, and
1093 - Lots of improvements and cleanups to the experimental Remote API
1094 (RAPI) client library.
1095 - A new unit test suite for the core daemon libraries.
1096 - A fix to creating missing directories makes sure the umask is not
1097 applied anymore. This enforces the same directory permissions
1099 - Better handling terminating daemons with ctrl+c (used when running
1100 them in debugging mode).
1101 - Fix a race condition in live migrating a KVM instance, when stat()
1102 on the old proc status file returned EINVAL, which is an unexpected
1104 - Fixed manpage checking with newer man and utf-8 charachters. But now
1105 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1111 *(Released Fri, 7 May 2010)*
1113 Fix a bug which prevented untagged KVM instances from starting.
1119 *(Released Fri, 7 May 2010)*
1121 Another release with a long development cycle, during which many
1122 different features were added.
1124 Significant features
1125 ~~~~~~~~~~~~~~~~~~~~
1127 The KVM hypervisor now can run the individual instances as non-root, to
1128 reduce the impact of a VM being hijacked due to bugs in the
1129 hypervisor. It is possible to run all instances as a single (non-root)
1130 user, to manually specify a user for each instance, or to dynamically
1131 allocate a user out of a cluster-wide pool to each instance, with the
1132 guarantee that no two instances will run under the same user ID on any
1135 An experimental RAPI client library, that can be used standalone
1136 (without the other Ganeti libraries), is provided in the source tree as
1137 ``lib/rapi/client.py``. Note this client might change its interface in
1138 the future, as we iterate on its capabilities.
1140 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1141 replace the cluster's certificates and crypto keys. This might help in
1142 case they have been compromised, or have simply expired.
1144 A new disk option for instance creation has been added that allows one
1145 to "adopt" currently existing logical volumes, with data
1146 preservation. This should allow easier migration to Ganeti from
1147 unmanaged (or managed via other software) instances.
1149 Another disk improvement is the possibility to convert between redundant
1150 (DRBD) and plain (LVM) disk configuration for an instance. This should
1151 allow better scalability (starting with one node and growing the
1152 cluster, or shrinking a two-node cluster to one node).
1154 A new feature that could help with automated node failovers has been
1155 implemented: if a node sees itself as offline (by querying the master
1156 candidates), it will try to shutdown (hard) all instances and any active
1157 DRBD devices. This reduces the risk of duplicate instances if an
1158 external script automatically failovers the instances on such nodes. To
1159 enable this, the cluster parameter ``maintain_node_health`` should be
1160 enabled; in the future this option (per the name) will enable other
1161 automatic maintenance features.
1163 Instance export/import now will reuse the original instance
1164 specifications for all parameters; that means exporting an instance,
1165 deleting it and the importing it back should give an almost identical
1166 instance. Note that the default import behaviour has changed from
1167 before, where it created only one NIC; now it recreates the original
1170 Cluster verify has added a few new checks: SSL certificates validity,
1171 /etc/hosts consistency across the cluster, etc.
1176 As usual, many internal changes were done, documentation fixes,
1179 - Fixed cluster initialization with disabled cluster storage (regression
1180 introduced in 2.1.1)
1181 - File-based storage supports growing the disks
1182 - Fixed behaviour of node role changes
1183 - Fixed cluster verify for some corner cases, plus a general rewrite of
1184 cluster verify to allow future extension with more checks
1185 - Fixed log spamming by watcher and node daemon (regression introduced
1187 - Fixed possible validation issues when changing the list of enabled
1189 - Fixed cleanup of /etc/hosts during node removal
1190 - Fixed RAPI response for invalid methods
1191 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1192 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1193 from ide disks, etc.)
1194 - Allow OS changes without re-installation (to record a changed OS
1195 outside of Ganeti, or to allow OS renames)
1196 - Allow instance creation without OS installation (useful for example if
1197 the OS will be installed manually, or restored from a backup not in
1199 - Implemented option to make cluster ``copyfile`` use the replication
1201 - Added list of enabled hypervisors to ssconf (possibly useful for
1203 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1205 - A partial form of node re-IP is possible via node readd, which now
1206 allows changed node primary IP
1207 - Command line utilities now show an informational message if the job is
1209 - The logs of the master daemon now show the PID/UID/GID of the
1216 *(Released Fri, 12 Mar 2010)*
1218 During the 2.1.0 long release candidate cycle, a lot of improvements and
1219 changes have accumulated with were released later as 2.1.1.
1224 The node evacuate command (``gnt-node evacuate``) was significantly
1225 rewritten, and as such the IAllocator protocol was changed - a new
1226 request type has been added. This unfortunate change during a stable
1227 series is designed to improve performance of node evacuations; on
1228 clusters with more than about five nodes and which are well-balanced,
1229 evacuation should proceed in parallel for all instances of the node
1230 being evacuated. As such, any existing IAllocator scripts need to be
1231 updated, otherwise the above command will fail due to the unknown
1232 request. The provided "dumb" allocator has not been updated; but the
1233 ganeti-htools package supports the new protocol since version 0.2.4.
1235 Another important change is increased validation of node and instance
1236 names. This might create problems in special cases, if invalid host
1237 names are being used.
1239 Also, a new layer of hypervisor parameters has been added, that sits at
1240 OS level between the cluster defaults and the instance ones. This allows
1241 customisation of virtualization parameters depending on the installed
1242 OS. For example instances with OS 'X' may have a different KVM kernel
1243 (or any other parameter) than the cluster defaults. This is intended to
1244 help managing a multiple OSes on the same cluster, without manual
1245 modification of each instance's parameters.
1247 A tool for merging clusters, ``cluster-merge``, has been added in the
1248 tools sub-directory.
1253 - Improved the int/float conversions that should make the code more
1254 robust in face of errors from the node daemons
1255 - Fixed the remove node code in case of internal configuration errors
1256 - Fixed the node daemon behaviour in face of inconsistent queue
1257 directory (e.g. read-only file-system where we can't open the files
1259 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1260 now it either aborts cleanly or, if given the new "auto_promote"
1261 parameter, will automatically promote other nodes as needed
1262 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1263 completely prevent Ganeti from working
1264 - Fixed bug for instance export when not all disks were successfully
1266 - Fixed behaviour of node add when the new node is slow in starting up
1268 - Fixed handling of signals in the LUXI client, which should improve
1269 behaviour of command-line scripts
1270 - Added checks for invalid node/instance names in the configuration (now
1271 flagged during cluster verify)
1272 - Fixed watcher behaviour for disk activation errors
1273 - Fixed two potentially endless loops in http library, which led to the
1274 RAPI daemon hanging and consuming 100% CPU in some cases
1275 - Fixed bug in RAPI daemon related to hashed passwords
1276 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1278 - Enhanced compatibility with non-Debian OSes, but not using absolute
1279 path in some commands and allowing customisation of the ssh
1280 configuration directory
1281 - Fixed possible future issue with new Python versions by abiding to the
1282 proper use of ``__slots__`` attribute on classes
1283 - Added checks that should prevent directory traversal attacks
1284 - Many documentation fixes based on feedback from users
1289 - Added an "early_release" more for instance replace disks and node
1290 evacuate, where we release locks earlier and thus allow higher
1291 parallelism within the cluster
1292 - Added watcher hooks, intended to allow the watcher to restart other
1293 daemons (e.g. from the ganeti-nbma project), but they can be used of
1294 course for any other purpose
1295 - Added a compile-time disable for DRBD barriers, to increase
1296 performance if the administrator trusts the power supply or the
1297 storage system to not lose writes
1298 - Added the option of using syslog for logging instead of, or in
1299 addition to, Ganeti's own log files
1300 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1301 can indeed boot from a paravirtual NIC
1302 - Added a generic debug level for many operations; while this is not
1303 used widely yet, it allows one to pass the debug value all the way to
1305 - Enhanced the hooks environment for instance moves (failovers,
1306 migrations) where the primary/secondary nodes changed during the
1307 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1308 - Enhanced data validations for many user-supplied values; one important
1309 item is the restrictions imposed on instance and node names, which
1310 might reject some (invalid) host names
1311 - Add a configure-time option to disable file-based storage, if it's not
1312 needed; this allows greater security separation between the master
1313 node and the other nodes from the point of view of the inter-node RPC
1315 - Added user notification in interactive tools if job is waiting in the
1316 job queue or trying to acquire locks
1317 - Added log messages when a job is waiting for locks
1318 - Added filtering by node tags in instance operations which admit
1319 multiple instances (start, stop, reboot, reinstall)
1320 - Added a new tool for cluster mergers, ``cluster-merge``
1321 - Parameters from command line which are of the form ``a=b,c=d`` can now
1322 use backslash escapes to pass in values which contain commas,
1323 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1325 - For KVM, the instance name is the first parameter passed to KVM, so
1326 that it's more visible in the process list
1332 *(Released Tue, 2 Mar 2010)*
1334 Ganeti 2.1 brings many improvements with it. Major changes:
1336 - Added infrastructure to ease automated disk repairs
1337 - Added new daemon to export configuration data in a cheaper way than
1338 using the remote API
1339 - Instance NICs can now be routed instead of being associated with a
1341 - Improved job locking logic to reduce impact of jobs acquiring multiple
1342 locks waiting for other long-running jobs
1344 In-depth implementation details can be found in the Ganeti 2.1 design
1350 - Added chroot hypervisor
1351 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1353 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1354 ``bootloader_path`` and ``bootloader_args``)
1355 - Added the ``use_localtime`` option for the xen-hvm and kvm
1356 hypervisors, and the default value for this has changed to false (in
1357 2.0 xen-hvm always enabled it)
1358 - Added luxi call to submit multiple jobs in one go
1359 - Added cluster initialization option to not modify ``/etc/hosts``
1361 - Added network interface parameters
1362 - Added dry run mode to some LUs
1363 - Added RAPI resources:
1365 - ``/2/instances/[instance_name]/info``
1366 - ``/2/instances/[instance_name]/replace-disks``
1367 - ``/2/nodes/[node_name]/evacuate``
1368 - ``/2/nodes/[node_name]/migrate``
1369 - ``/2/nodes/[node_name]/role``
1370 - ``/2/nodes/[node_name]/storage``
1371 - ``/2/nodes/[node_name]/storage/modify``
1372 - ``/2/nodes/[node_name]/storage/repair``
1374 - Added OpCodes to evacuate or migrate all instances on a node
1375 - Added new command to list storage elements on nodes (``gnt-node
1376 list-storage``) and modify them (``gnt-node modify-storage``)
1377 - Added new ssconf files with master candidate IP address
1378 (``ssconf_master_candidates_ips``), node primary IP address
1379 (``ssconf_node_primary_ips``) and node secondary IP address
1380 (``ssconf_node_secondary_ips``)
1381 - Added ``ganeti-confd`` and a client library to query the Ganeti
1382 configuration via UDP
1383 - Added ability to run hooks after cluster initialization and before
1385 - Added automatic mode for disk replace (``gnt-instance replace-disks
1387 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1388 after catastrophic data-loss
1389 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1391 - Added ``gnt-instance move`` command to move instances
1392 - Added ``gnt-cluster watcher`` command to control watcher
1393 - Added ``gnt-node powercycle`` command to powercycle nodes
1394 - Added new job status field ``lock_status``
1395 - Added parseable error codes to cluster verification (``gnt-cluster
1396 verify --error-codes``) and made output less verbose (use
1397 ``--verbose`` to restore previous behaviour)
1398 - Added UUIDs to the main config entities (cluster, nodes, instances)
1399 - Added support for OS variants
1400 - Added support for hashed passwords in the Ganeti remote API users file
1402 - Added option to specify maximum timeout on instance shutdown
1403 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1404 - Added new helper script to start and stop Ganeti daemons
1405 (``daemon-util``), with the intent to reduce the work necessary to
1406 adjust Ganeti for non-Debian distributions and to start/stop daemons
1408 - Added more unittests
1409 - Fixed critical bug in ganeti-masterd startup
1410 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1411 now customisable at the cluster level for both the KVM and Xen
1412 hypervisors using the new ``migration_port`` parameter
1413 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1414 reinstalling an instance
1415 - Allowed ``@`` in tag names
1416 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1417 - Many documentation updates
1418 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1419 - ``gnt-instance reinstall`` can now reinstall multiple instances
1420 - Updated many command line parameters
1421 - Introduced new OS API version 15
1422 - No longer support a default hypervisor
1423 - Treat virtual LVs as inexistent
1424 - Improved job locking logic to reduce lock contention
1425 - Match instance and node names case insensitively
1426 - Reimplemented bash completion script to be more complete
1433 *(Released Thu, 4 Feb 2010)*
1435 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1436 - Fix a string formatting bug
1437 - Improve safety of the code in some error paths
1438 - Improve data validation in the master of values returned from nodes
1444 *(Released Thu, 17 Dec 2009)*
1446 - Fix security issue due to missing validation of iallocator names; this
1447 allows local and remote execution of arbitrary executables
1448 - Fix failure of gnt-node list during instance removal
1449 - Ship the RAPI documentation in the archive
1455 *(Released Wed, 30 Sep 2009)*
1457 - Fixed many wrong messages
1458 - Fixed a few bugs related to the locking library
1459 - Fixed MAC checking at instance creation time
1460 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1461 - Fixed a few issues related to signal handling in both daemons and
1463 - Fixed the example startup script provided
1464 - Fixed insserv dependencies in the example startup script (patch from
1466 - Fixed handling of drained nodes in the iallocator framework
1467 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1469 - Fixed error related to invalid job IDs in job polling
1470 - Fixed job/opcode persistence on unclean master shutdown
1471 - Fixed handling of partial job processing after unclean master
1473 - Fixed error reporting from LUs, previously all errors were converted
1474 into execution errors
1475 - Fixed error reporting from burnin
1476 - Decreased significantly the memory usage of the job queue
1477 - Optimised slightly multi-job submission
1478 - Optimised slightly opcode loading
1479 - Backported the multi-job submit framework from the development
1480 branch; multi-instance start and stop should be faster
1481 - Added script to clean archived jobs after 21 days; this will reduce
1482 the size of the queue directory
1483 - Added some extra checks in disk size tracking
1484 - Added an example ethers hook script
1485 - Added a cluster parameter that prevents Ganeti from modifying of
1487 - Added more node information to RAPI responses
1488 - Added a ``gnt-job watch`` command that allows following the ouput of a
1490 - Added a bind-address option to ganeti-rapi
1491 - Added more checks to the configuration verify
1492 - Enhanced the burnin script such that some operations can be retried
1494 - Converted instance reinstall to multi-instance model
1500 *(Released Fri, 7 Aug 2009)*
1502 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1503 to allow using the pre-2.0.2 behaviour in activation, if any existing
1504 instances have mismatched disk sizes in the configuration
1505 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
1506 any configuration mismatches for disk sizes
1507 - Added ``gnt-master cluste-failover --no-voting`` to allow master
1508 failover to work on two-node clusters
1509 - Fixed the ``--net`` option of ``gnt-backup import``, which was
1511 - Fixed detection of OS script errors in ``gnt-backup export``
1512 - Fixed exit code of ``gnt-backup export``
1518 *(Released Fri, 17 Jul 2009)*
1520 - Added experimental support for stripped logical volumes; this should
1521 enhance performance but comes with a higher complexity in the block
1522 device handling; stripping is only enabled when passing
1523 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
1524 affected even in the non-stripped mode
1525 - Improved resiliency against transient failures at the end of DRBD
1526 resyncs, and in general of DRBD resync checks
1527 - Fixed a couple of issues with exports and snapshot errors
1528 - Fixed a couple of issues in instance listing
1529 - Added display of the disk size in ``gnt-instance info``
1530 - Fixed checking for valid OSes in instance creation
1531 - Fixed handling of the "vcpus" parameter in instance listing and in
1532 general of invalid parameters
1533 - Fixed http server library, and thus RAPI, to handle invalid
1534 username/password combinations correctly; this means that now they
1535 report unauthorized for queries too, not only for modifications,
1536 allowing earlier detect of configuration problems
1537 - Added a new "role" node list field, equivalent to the master/master
1538 candidate/drained/offline flags combinations
1539 - Fixed cluster modify and changes of candidate pool size
1540 - Fixed cluster verify error messages for wrong files on regular nodes
1541 - Fixed a couple of issues with node demotion from master candidate role
1542 - Fixed node readd issues
1543 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
1544 - Added a new ``--no-voting`` option for masterfailover to fix failover
1545 on two-nodes clusters when the former master node is unreachable
1546 - Added instance reinstall over RAPI
1552 *(Released Tue, 16 Jun 2009)*
1554 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
1555 allow re-adding the start in single-user option (regression from 1.2)
1556 - the watcher writes the instance status to a file, to allow monitoring
1557 to report the instance status (from the master) based on cached
1558 results of the watcher's queries; while this can get stale if the
1559 watcher is being locked due to other work on the cluster, this is
1560 still an improvement
1561 - the watcher now also restarts the node daemon and the rapi daemon if
1563 - fixed the watcher to handle full and drained queue cases
1564 - hooks export more instance data in the environment, which helps if
1565 hook scripts need to take action based on the instance's properties
1566 (no longer need to query back into ganeti)
1567 - instance failovers when the instance is stopped do not check for free
1568 RAM, so that failing over a stopped instance is possible in low memory
1570 - rapi uses queries for tags instead of jobs (for less job traffic), and
1571 for cluster tags it won't talk to masterd at all but read them from
1573 - a couple of error handling fixes in RAPI
1574 - drbd handling: improved the error handling of inconsistent disks after
1575 resync to reduce the frequency of "there are some degraded disks for
1576 this instance" messages
1577 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
1578 error handling path called a wrong function name)
1584 *(Released Wed, 27 May 2009)*
1586 - no changes from rc5
1589 Version 2.0 release candidate 5
1590 -------------------------------
1592 *(Released Wed, 20 May 2009)*
1594 - fix a couple of bugs (validation, argument checks)
1595 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
1596 - some small improvements to RAPI and IAllocator
1597 - make watcher automatically start the master daemon if down
1600 Version 2.0 release candidate 4
1601 -------------------------------
1603 *(Released Mon, 27 Apr 2009)*
1605 - change the OS list to not require locks; this helps with big clusters
1606 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
1607 volume group is broken
1608 - ``gnt-instance info``, without any arguments, doesn't run for all
1609 instances anymore; either pass ``--all`` or pass the desired
1610 instances; this helps against mistakes on big clusters where listing
1611 the information for all instances takes a long time
1612 - miscellaneous doc and man pages fixes
1615 Version 2.0 release candidate 3
1616 -------------------------------
1618 *(Released Wed, 8 Apr 2009)*
1620 - Change the internal locking model of some ``gnt-node`` commands, in
1621 order to reduce contention (and blocking of master daemon) when
1622 batching many creation/reinstall jobs
1623 - Fixes to Xen soft reboot
1624 - No longer build documentation at build time, instead distribute it in
1625 the archive, in order to reduce the need for the whole docbook/rst
1629 Version 2.0 release candidate 2
1630 -------------------------------
1632 *(Released Fri, 27 Mar 2009)*
1634 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
1635 - Fix watcher startup sequence, improves the behaviour of busy clusters
1636 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
1637 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
1638 bind address and other places
1639 - Some documentation fixes and updates
1642 Version 2.0 release candidate 1
1643 -------------------------------
1645 *(Released Mon, 2 Mar 2009)*
1647 - More documentation updates, now all docs should be more-or-less
1649 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
1651 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
1653 - Fix an issue related to $libdir/run/ganeti and cluster creation
1659 *(Released Thu, 19 Feb 2009)*
1661 - Xen PVM and KVM have switched the default value for the instance root
1662 disk to the first partition on the first drive, instead of the whole
1663 drive; this means that the OS installation scripts must be changed
1665 - Man pages have been updated
1666 - RAPI has been switched by default to HTTPS, and the exported functions
1667 should all work correctly
1668 - RAPI v1 has been removed
1669 - Many improvements to the KVM hypervisor
1670 - Block device errors are now better reported
1671 - Many other bugfixes and small improvements
1677 *(Released Mon, 26 Jan 2009)*
1679 - Version 2 is a general rewrite of the code and therefore the
1680 differences are too many to list, see the design document for 2.0 in
1681 the ``doc/`` subdirectory for more details
1682 - In this beta version there is not yet a migration path from 1.2 (there
1683 will be one in the final 2.0 release)
1684 - A few significant changes are:
1686 - all commands are executed by a daemon (``ganeti-masterd``) and the
1687 various ``gnt-*`` commands are just front-ends to it
1688 - all the commands are entered into, and executed from a job queue,
1689 see the ``gnt-job(8)`` manpage
1690 - the RAPI daemon supports read-write operations, secured by basic
1691 HTTP authentication on top of HTTPS
1692 - DRBD version 0.7 support has been removed, DRBD 8 is the only
1693 supported version (when migrating from Ganeti 1.2 to 2.0, you need
1694 to migrate to DRBD 8 first while still running Ganeti 1.2)
1695 - DRBD devices are using statically allocated minor numbers, which
1696 will be assigned to existing instances during the migration process
1697 - there is support for both Xen PVM and Xen HVM instances running on
1699 - KVM virtualization is supported too
1700 - file-based storage has been implemented, which means that it is
1701 possible to run the cluster without LVM and DRBD storage, for
1702 example using a shared filesystem exported from shared storage (and
1703 still have live migration)
1709 *(Released Tue, 13 Jan 2009)*
1711 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
1712 - Reuse the old instance mac address by default on instance import, if
1713 the instance name is the same.
1714 - Handle situations in which the node info rpc returns incomplete
1716 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
1717 - Improved version of batcher:
1719 - state file support
1720 - instance mac address support
1721 - support for HVM clusters/instances
1723 - Add an option to show the number of cpu sockets and nodes in
1725 - Support OSes that handle more than one version of the OS api (but do
1726 not change the current API in any other way)
1727 - Fix ``gnt-node migrate``
1728 - ``gnt-debug`` man page
1729 - Fixes various more typos and small issues
1730 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
1736 *(Released Wed, 24 Sep 2008)*
1738 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
1739 type of disk exported to fully virtualized instances.
1740 - provide access to the serial console of HVM instances
1741 - instance auto_balance flag, set by default. If turned off it will
1742 avoid warnings on cluster verify if there is not enough memory to fail
1743 over an instance. in the future it will prevent automatically failing
1744 it over when we will support that.
1745 - batcher tool for instance creation, see ``tools/README.batcher``
1746 - ``gnt-instance reinstall --select-os`` to interactively select a new
1747 operating system when reinstalling an instance.
1748 - when changing the memory amount on instance modify a check has been
1749 added that the instance will be able to start. also warnings are
1750 emitted if the instance will not be able to fail over, if auto_balance
1752 - documentation fixes
1753 - sync fields between ``gnt-instance list/modify/add/import``
1754 - fix a race condition in drbd when the sync speed was set after giving
1755 the device a remote peer.
1761 *(Released Tue, 22 Jul 2008)*
1763 - note: the allowed size and number of tags per object were reduced
1764 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1765 - fixed twisted 8.x compatibility
1766 - fixed ``gnt-instance replace-disks`` with iallocator
1767 - add TCP keepalives on twisted connections to detect restarted nodes
1768 - disk increase support, see ``gnt-instance grow-disk``
1769 - implement bulk node/instance query for RAPI
1770 - add tags in node/instance listing (optional)
1771 - experimental migration (and live migration) support, read the man page
1772 for ``gnt-instance migrate``
1773 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1774 has some small improvements in handling its state file
1780 *(Released Fri, 13 Jun 2008)*
1782 - Experimental readonly, REST-based remote API implementation;
1783 automatically started on master node, TCP port 5080, if enabled by
1784 ``--enable-rapi`` parameter to configure script.
1785 - Instance allocator support. Add and import instance accept a
1786 ``--iallocator`` parameter, and call that instance allocator to decide
1787 which node to use for the instance. The iallocator document describes
1788 what's expected from an allocator script.
1789 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1790 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1791 node is lost there is still enough memory to fail over the instances
1793 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1794 ``gnt-cluster verify``, to check for site-specific compliance. All the
1795 hooks will run, and their output, if any, will be displayed. Any
1796 failing hook will make the verification return an error value.
1797 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1798 primary and secondary interfaces
1799 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1800 that is still declared as part of the cluster and has failed.
1801 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1802 specifying output fields, that just adds the chosen fields to the
1804 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1805 export from the filesystem.
1806 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1807 have been added. Using them you can enable/disable acpi and pae
1808 support, and specify a path for a cd image to be exported to the
1809 instance. These parameters as the name suggest only work on HVM
1811 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1812 PAE support will be set to the previously hardcoded values, but the
1813 (previously hardcoded) path to the CDROM ISO image will be unset and
1814 if required, needs to be set manually with ``gnt-instance modify``
1816 - The address to which an instance's VNC console is bound is now
1817 selectable per-instance, rather than being cluster wide. Of course
1818 this only applies to instances controlled via VNC, so currently just
1819 applies to HVM clusters.
1825 *(Released Mon, 18 Feb 2008)*
1827 - more tweaks to the disk activation code (especially helpful for DRBD)
1828 - change the default ``gnt-instance list`` output format, now there is
1829 one combined status field (see the manpage for the exact values this
1831 - some more fixes for the mac export to hooks change
1832 - make Ganeti not break with DRBD 8.2.x (which changed the version
1833 format in ``/proc/drbd``) (issue 24)
1834 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1835 template, allowing migration from DRBD0.7+MD to DRBD8
1841 *(Released Wed, 30 Jan 2008)*
1843 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1845 - add command aliases infrastructure and a few aliases
1846 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1847 man pages and the ``--help`` option of ``gnt-node
1848 list``/``gnt-instance list``
1849 - fix ``gnt-backup list`` with down nodes (issue 21)
1850 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1851 - fix the dist archive and add a check for including svn/git files in
1853 - some developer-related changes: improve the burnin and the QA suite,
1854 add an upload script for testing during development
1860 *(Released Wed, 16 Jan 2008)*
1862 - experimental HVM support, read the install document, section
1863 "Initializing the cluster"
1864 - allow for the PVM hypervisor per-instance kernel and initrd paths
1865 - add a new command ``gnt-cluster verify-disks`` which uses a new
1866 algorithm to improve the reconnection of the DRBD pairs if the device
1867 on the secondary node has gone away
1868 - make logical volume code auto-activate LVs at disk activation time
1869 - slightly improve the speed of activating disks
1870 - allow specification of the MAC address at instance creation time, and
1871 changing it later via ``gnt-instance modify``
1872 - fix handling of external commands that generate lots of output on
1874 - update documentation with regard to minimum version of DRBD8 supported
1880 *(Released Tue, 4 Dec 2007)*
1882 - Log the ``xm create`` output to the node daemon log on failure (to
1883 help diagnosing the error)
1884 - In debug mode, log all external commands output if failed to the logs
1885 - Change parsing of lvm commands to ignore stderr
1891 *(Released Wed, 28 Nov 2007)*
1893 - Another round of updates to the DRBD 8 code to deal with more failures
1894 in the replace secondary node operation
1895 - Some more logging of failures in disk operations (lvm, drbd)
1896 - A few documentation updates
1903 *(Released Tue, 13 Nov 2007)*
1905 - Change configuration file format from Python's Pickle to JSON.
1906 Upgrading is possible using the cfgupgrade utility.
1907 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1908 faster replace disks and is more stable (DRBD 8 has many improvements
1909 compared to DRBD 0.7)
1910 - Added command line tags support (see man pages for ``gnt-instance``,
1911 ``gnt-node``, ``gnt-cluster``)
1912 - Added instance rename support
1913 - Added multi-instance startup/shutdown
1914 - Added cluster rename support
1915 - Added ``gnt-node evacuate`` to simplify some node operations
1916 - Added instance reboot operation that can speedup reboot as compared to
1918 - Soften the requirement that hostnames are in FQDN format
1919 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1921 - Removed dependency on debian's patched fping that uses the
1922 non-standard ``-S`` option
1923 - Now the OS definitions are searched for in multiple, configurable
1924 paths (easier for distros to package)
1925 - Some changes to the hooks infrastructure (especially the new
1926 post-configuration update hook)
1927 - Other small bugfixes
1929 .. vim: set textwidth=72 syntax=rst :