10 - Deprecated ``admin_up`` field. Instead, ``admin_state`` is introduced,
11 with 3 possible values -- ``up``, ``down`` and ``offline``.
12 - Replaced ``--disks`` option of ``gnt-instance replace-disks`` with a
13 more flexible ``--disk`` option. Now disk size and mode can be changed
15 - Removed deprecated ``QueryLocks`` LUXI request. Use
16 ``Query(what=QR_LOCK, ...)`` instead.
17 - The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
18 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES` and
19 :pyeval:`luxi.REQ_QUERY_GROUPS` are deprecated and will be removed in
20 a future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
26 *(Released Thu, 12 Apr 2012)*
28 Incompatible/important changes and bugfixes
29 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
31 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
32 resource's ``ip_check`` parameter changed from ``True`` to ``False``
33 to match the underlying LUXI interface.
34 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
35 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
36 not maintain backwards-compatibility as the underlying operation
37 changed in an incompatible way. The RAPI client can talk to old
38 servers, but it needs to be told so as the return value changed.
39 - When creating file-based instances via RAPI, the ``file_driver``
40 parameter no longer defaults to ``loop`` and must be specified.
41 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
43 - Support for the undocumented and deprecated RAPI instance creation
44 request format version 0 has been dropped. Use version 1, supported
45 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
46 - Pyparsing 1.4.6 or above is required, see :doc:`installation
47 documentation <install>`.
48 - The "cluster-verify" hooks are now executed per group by the
49 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
50 if you just run ``gnt-cluster verify``, which generates one opcode per
52 - The environment as passed to the OS scripts is cleared, and thus no
53 environment variables defined in the node daemon's environment will be
54 inherited by the scripts.
55 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
57 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
58 support operations involving multiple node groups.
59 - Offline nodes are ignored when failing over an instance.
60 - Support for KVM version 1.0, which changed the version reporting format
62 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
64 - ``Makefile`` is now compatible with Automake 1.11.2
65 - Includes all bugfixes made in the 2.4 series
70 - The ganeti-htools project has been merged into the ganeti-core source
71 tree and will be built as part of Ganeti (see :doc:`install-quick`).
72 - Implemented support for :doc:`shared storage <design-shared-storage>`.
73 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
74 GPT-style partition tables (requires `parted
75 <http://www.gnu.org/s/parted/>`_).
76 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
77 - Allowed adding tags on instance creation.
78 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
79 - Allow instances to be started in a paused state, enabling the user to
80 see the complete console output on boot using the console.
81 - Added new hypervisor flag to control default reboot behaviour
82 (``reboot_behavior``).
83 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
84 - Improved out-of-band management support:
86 - Added ``gnt-node health`` command reporting the health status of
88 - Added ``gnt-node power`` command to manage power status of nodes.
89 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
91 - Instance migration can fall back to failover if instance is not
93 - Filters can be used when listing nodes, instances, groups and locks;
94 see *ganeti(7)* manpage.
95 - Added post-execution status as variables to :doc:`hooks <hooks>`
97 - Instance tags are exported/imported together with the instance.
98 - When given an explicit job ID, ``gnt-job info`` will work for archived
100 - Jobs can define dependencies on other jobs (not yet supported via
101 RAPI or command line, but used by internal commands and usable via
104 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
107 - Instance failover is now available as a RAPI resource
108 (``/2/instances/[instance_name]/failover``).
109 - ``gnt-instance info`` defaults to static information if primary node
111 - Opcodes have a new ``comment`` attribute.
112 - Added basic SPICE support to KVM hypervisor.
113 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
115 Node group improvements
116 ~~~~~~~~~~~~~~~~~~~~~~~
118 - ``gnt-cluster verify`` has been modified to check groups separately,
119 thereby improving performance.
120 - Node group support has been added to ``gnt-cluster verify-disks``,
121 which now operates per node group.
122 - Watcher has been changed to work better with node groups.
124 - One process and state file per node group.
125 - Slow watcher in one group doesn't block other group's watcher.
127 - Added new command, ``gnt-group evacuate``, to move all instances in a
128 node group to other groups.
129 - Added ``gnt-instance change-group`` to move an instance to another
131 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
132 per-group operations.
133 - Node groups can be tagged.
134 - Some operations switch from an exclusive to a shared lock as soon as
136 - Instance's primary and secondary nodes' groups are now available as
137 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
138 and ``snodes.group.uuid``).
143 - Numerous updates to documentation and manpages.
145 - :doc:`RAPI <rapi>` documentation now has detailed parameter
147 - Some opcode/job results are now also documented, see :doc:`RAPI
150 - A lockset's internal lock is now also visible in lock monitor.
151 - Log messages from job queue workers now contain information about the
152 opcode they're processing.
153 - ``gnt-instance console`` no longer requires the instance lock.
154 - A short delay when waiting for job changes reduces the number of LUXI
155 requests significantly.
156 - DRBD metadata volumes are overwritten with zeros during disk creation.
157 - Out-of-band commands no longer acquire the cluster lock in exclusive
159 - ``devel/upload`` now uses correct permissions for directories.
165 *(Released Fri, 23 Mar 2012)*
167 This was the sixth release candidate of the 2.5 series.
173 *(Released Mon, 9 Jan 2012)*
175 This was the fifth release candidate of the 2.5 series.
181 *(Released Thu, 27 Oct 2011)*
183 This was the fourth release candidate of the 2.5 series.
189 *(Released Wed, 26 Oct 2011)*
191 This was the third release candidate of the 2.5 series.
197 *(Released Tue, 18 Oct 2011)*
199 This was the second release candidate of the 2.5 series.
205 *(Released Tue, 4 Oct 2011)*
207 This was the first release candidate of the 2.5 series.
213 *(Released Wed, 31 Aug 2011)*
215 This was the third beta release of the 2.5 series.
221 *(Released Mon, 22 Aug 2011)*
223 This was the second beta release of the 2.5 series.
229 *(Released Fri, 12 Aug 2011)*
231 This was the first beta release of the 2.5 series.
237 *(Released Thu, 27 Oct 2011)*
239 - Fixed bug when parsing command line parameter values ending in
241 - Fixed assertion error after unclean master shutdown
242 - Disable HTTP client pool for RPC, significantly reducing memory usage
244 - Fixed queue archive creation with wrong permissions
250 *(Released Tue, 23 Aug 2011)*
254 - Fixed documentation for importing with ``--src-dir`` option
255 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
256 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
262 *(Released Fri, 5 Aug 2011)*
264 Many bug-fixes and a few small features:
266 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
267 issues when you tried to add an instance with two MAC addresses in one
269 - KVM: fixed per-instance stored UID value
270 - KVM: configure bridged NICs at migration start
271 - KVM: Fix a bug where instance will not start with never KVM versions
273 - Added OS search path to ``gnt-cluster info``
274 - Fixed an issue with ``file_storage_dir`` where you were forced to
275 provide an absolute path, but the documentation states it is a
276 relative path, the documentation was right
277 - Added a new parameter to instance stop/start called ``--no-remember``
278 that will make the state change to not be remembered
279 - Implemented ``no_remember`` at RAPI level
280 - Improved the documentation
281 - Node evacuation: don't call IAllocator if node is already empty
282 - Fixed bug in DRBD8 replace disks on current nodes
283 - Fixed bug in recreate-disks for DRBD instances
284 - Moved assertion checking locks in ``gnt-instance replace-disks``
285 causing it to abort with not owning the right locks for some situation
286 - Job queue: Fixed potential race condition when cancelling queued jobs
287 - Fixed off-by-one bug in job serial generation
288 - ``gnt-node volumes``: Fix instance names
289 - Fixed aliases in bash completion
290 - Fixed a bug in reopening log files after being sent a SIGHUP
291 - Added a flag to burnin to allow specifying VCPU count
292 - Bugfixes to non-root Ganeti configuration
298 *(Released Thu, 12 May 2011)*
300 Many bug-fixes and a few new small features:
302 - Fixed a bug related to log opening failures
303 - Fixed a bug in instance listing with orphan instances
304 - Fixed a bug which prevented resetting the cluster-level node parameter
305 ``oob_program`` to the default
306 - Many fixes related to the ``cluster-merge`` tool
307 - Fixed a race condition in the lock monitor, which caused failures
308 during (at least) creation of many instances in parallel
309 - Improved output for gnt-job info
310 - Removed the quiet flag on some ssh calls which prevented debugging
312 - Improved the N+1 failure messages in cluster verify by actually
313 showing the memory values (needed and available)
314 - Increased lock attempt timeouts so that when executing long operations
315 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
316 too early and thus prevent the use of the 'fair' mechanism
317 - Changed instance query data (``gnt-instance info``) to not acquire
318 locks unless needed, thus allowing its use on locked instance if only
319 static information is asked for
320 - Improved behaviour with filesystems that do not support rename on an
322 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
323 kept locks on all nodes during the wipe, which is unneeded
324 - Fixed ``gnt-watcher`` handling of errors during hooks execution
325 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
326 10GiB) which caused the wipe to fail right at the end in some cases
327 - Fixed master IP activation when doing master failover with no-voting
328 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
329 the master node itself
330 - Fixed potential data-loss in under disk full conditions, where Ganeti
331 wouldn't check correctly the return code and would consider
332 partially-written files 'correct'
333 - Fixed bug related to multiple VGs and DRBD disk replacing
334 - Added new disk parameter ``metavg`` that allows placement of the meta
335 device for DRBD in a different volume group
336 - Fixed error handling in the node daemon when the system libc doesn't
337 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
338 - Fixed lock release during replace-disks, which kept cluster-wide locks
339 when doing disk replaces with an iallocator script
340 - Added check for missing bridges in cluster verify
341 - Handle EPIPE errors while writing to the terminal better, so that
342 piping the output to e.g. ``less`` doesn't cause a backtrace
343 - Fixed rare case where a ^C during Luxi calls could have been
344 interpreted as server errors, instead of simply terminating
345 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
347 - Added a few more parameters to the KVM hypervisor, allowing a second
348 CDROM, custom disk type for CDROMs and a floppy image
349 - Removed redundant message in instance rename when the name is given
351 - Added option to ``gnt-instance recreate-disks`` to allow creating the
352 disks on new nodes, allowing recreation when the original instance
353 nodes are completely gone
354 - Added option when converting disk templates to DRBD to skip waiting
355 for the resync, in order to make the instance available sooner
356 - Added two new variables to the OS scripts environment (containing the
358 - Made the root_path and optional parameter for the xen-pvm hypervisor,
359 to allow use of ``pvgrub`` as bootloader
360 - Changed the instance memory modifications to only check out-of-memory
361 conditions on memory increases, and turned the secondary node warnings
362 into errors (they can still be overridden via ``--force``)
363 - Fixed the handling of a corner case when the Python installation gets
364 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
365 to execute a command that doesn't exist
366 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
367 node of the instance returned failures during instance shutdown; this
368 adds the option ``--ignore-consistency`` to gnt-instance move
370 And as usual, various improvements to the error messages, documentation
377 *(Released Wed, 09 Mar 2011)*
379 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
380 the RAPI users file if run twice (even with ``--dry-run``).
382 The release fixes that bug (nothing else changed).
388 *(Released Mon, 07 Mar 2011)*
390 Final 2.4.0 release. Just a few small fixes:
392 - Fixed RAPI node evacuate
393 - Fixed the kvm-ifup script
394 - Fixed internal error handling for special job cases
395 - Updated man page to specify the escaping feature for options
401 *(Released Mon, 28 Feb 2011)*
403 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
404 that this feature wiped the disks of the wrong instance, leading to loss
409 - Fixed title of query field containing instance name
410 - Expanded the glossary in the documentation
411 - Fixed one unittest (internal issue)
417 *(Released Mon, 21 Feb 2011)*
419 A number of bug fixes plus just a couple functionality changes.
421 On the user-visible side, the ``gnt-* list`` command output has changed
422 with respect to "special" field states. The current rc1 style of display
423 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
424 the default output mode special fields states are displayed as follows:
426 - Offline resource: ``*``
427 - Unavailable/not applicable: ``-``
428 - Data missing (RPC failure): ``?``
429 - Unknown field: ``??``
431 Another user-visible change is the addition of ``--force-join`` to
436 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
437 - Fixed regression in RAPI/instance reinstall where all parameters were
438 required (instead of optional)
439 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
440 - Fixed iallocator usage (offline nodes were not considered offline)
441 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
442 - Fixed hypervisor and OS parameter validation with respect to
444 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
446 - Fixed ``tools/listrunner`` with respect to agent-based usage
452 *(Released Fri, 4 Feb 2011)*
454 Many changes and fixes since the beta1 release. While there were some
455 internal changes, the code has been mostly stabilised for the RC
458 Note: the dumb allocator was removed in this release, as it was not kept
459 up-to-date with the IAllocator protocol changes. It is recommended to
460 use the ``hail`` command from the ganeti-htools package.
462 Note: the 2.4 and up versions of Ganeti are not compatible with the
463 0.2.x branch of ganeti-htools. You need to upgrade to
464 ganeti-htools-0.3.0 (or later).
466 Regressions fixed from 2.3
467 ~~~~~~~~~~~~~~~~~~~~~~~~~~
469 - Fixed the ``gnt-cluster verify-disks`` command
470 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
472 - Fixed disk adoption breakage
473 - Fixed wrong headers in instance listing for field aliases
478 - Fixed corner case in KVM handling of NICs
479 - Fixed many cases of wrong handling of non-vm_capable nodes
480 - Fixed a bug where a missing instance symlink was not possible to
481 recreate with any ``gnt-*`` command (now ``gnt-instance
482 activate-disks`` does it)
483 - Fixed the volume group name as reported by ``gnt-cluster
485 - Increased timeouts for the import-export code, hopefully leading to
486 fewer aborts due network or instance timeouts
487 - Fixed bug in ``gnt-node list-storage``
488 - Fixed bug where not all daemons were started on cluster
489 initialisation, but only at the first watcher run
490 - Fixed many bugs in the OOB implementation
491 - Fixed watcher behaviour in presence of instances with offline
493 - Fixed instance list output for instances running on the wrong node
494 - a few fixes to the cluster-merge tool, but it still cannot merge
495 multi-node groups (currently it is not recommended to use this tool)
501 - Improved network configuration for the KVM hypervisor
502 - Added e1000 as a supported NIC for Xen-HVM
503 - Improved the lvmstrap tool to also be able to use partitions, as
504 opposed to full disks
505 - Improved speed of disk wiping (the cluster parameter
506 ``prealloc_wipe_disks``, so that it has a low impact on the total time
507 of instance creations
508 - Added documentation for the OS parameters
509 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
510 hypervisor is not responding
511 - Added display of blacklisted and hidden OS information in
513 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
514 NIC and node parameters, which might create problems with currently
515 invalid (but undetected) configuration files, but prevents validation
516 failures when unrelated parameters are modified
517 - Changed cluster initialisation to wait for the master daemon to become
519 - Expanded the RAPI interface:
521 - Added config redistribution resource
522 - Added activation/deactivation of instance disks
523 - Added export of console information
525 - Implemented log file reopening on SIGHUP, which allows using
526 logrotate(8) for the Ganeti log files
527 - Added a basic OOB helper script as an example
533 *(Released Fri, 14 Jan 2011)*
538 - Fixed timezone issues when formatting timestamps
539 - Added support for node groups, available via ``gnt-group`` and other
541 - Added out-of-band framework and management, see :doc:`design
542 document <design-oob>`
543 - Removed support for roman numbers from ``gnt-node list`` and
544 ``gnt-instance list``.
545 - Allowed modification of master network interface via ``gnt-cluster
546 modify --master-netdev``
547 - Accept offline secondaries while shutting down instance disks
548 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
549 - Added support for multiple LVM volume groups
550 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
552 - Added commands to list available fields:
554 - ``gnt-node list-fields``
555 - ``gnt-group list-fields``
556 - ``gnt-instance list-fields``
558 - Updated documentation and man pages
563 - Moved ``rapi_users`` file into separate directory, now named
564 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
566 - Added new tool for running commands on many machines,
567 ``tools/ganeti-listrunner``
568 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
569 improving the ``gnt-instance console`` output
570 - Allowed customisation of disk index separator at ``configure`` time
571 - Export node group allocation policy to :doc:`iallocator <iallocator>`
572 - Added support for non-partitioned md disks in ``lvmstrap``
573 - Added script to gracefully power off KVM instances
574 - Split ``utils`` module into smaller parts
575 - Changed query operations to return more detailed information, e.g.
576 whether an information is unavailable due to an offline node. To use
577 this new functionality, the LUXI call ``Query`` must be used. Field
578 information is now stored by the master daemon and can be retrieved
579 using ``QueryFields``. Instances, nodes and groups can also be queried
580 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
581 exposed via RAPI). The following commands make use of this
582 infrastructure change:
585 - ``gnt-group list-fields``
587 - ``gnt-node list-fields``
588 - ``gnt-instance list``
589 - ``gnt-instance list-fields``
590 - ``gnt-debug locks``
595 - New RAPI resources (see :doc:`rapi`):
599 - ``/2/groups/[group_name]``
600 - ``/2/groups/[group_name]/assign-nodes``
601 - ``/2/groups/[group_name]/modify``
602 - ``/2/groups/[group_name]/rename``
603 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
607 - Implemented ``no_install`` for instance creation
608 - Implemented OS parameters for instance reinstallation, allowing
609 use of special settings on reinstallation (e.g. for preserving data)
614 - Added IPv6 support in import/export
615 - Pause DRBD synchronization while wiping disks on instance creation
616 - Updated unittests and QA scripts
617 - Improved network parameters passed to KVM
618 - Converted man pages from docbook to reStructuredText
624 *(Released Mon, 20 Dec 2010)*
626 Released version 2.3.1~rc1 without any changes.
632 *(Released Wed, 1 Dec 2010)*
634 - impexpd: Disable OpenSSL compression in socat if possible (backport
635 from master, commit e90739d625b, see :doc:`installation guide
636 <install-quick>` for details)
637 - Changed unittest coverage report to exclude test scripts
638 - Added script to check version format
644 *(Released Wed, 1 Dec 2010)*
646 Released version 2.3.0~rc1 without any changes.
652 *(Released Fri, 19 Nov 2010)*
654 A number of bugfixes and documentation updates:
656 - Update ganeti-os-interface documentation
657 - Fixed a bug related to duplicate MACs or similar items which should be
659 - Fix breakage in OS state modify
660 - Reinstall instance: disallow offline secondaries (fixes bug related to
661 OS changing but reinstall failing)
662 - plus all the other fixes between 2.2.1 and 2.2.2
668 *(Released Tue, 2 Nov 2010)*
670 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
671 - Fixed master failover race with watcher
672 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
674 - Accept previously stopped instance for export with instance removal
675 - Simplify and extend the environment variables for instance OS scripts
676 - Added new node flags, ``master_capable`` and ``vm_capable``
677 - Added optional instance disk wiping prior during allocation. This is a
678 cluster-wide option and can be set/modified using
679 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
680 - Added IPv6 support, see :doc:`design document <design-2.3>` and
682 - Added a new watcher option (``--ignore-pause``)
683 - Added option to ignore offline node on instance start/stop
684 (``--ignore-offline``)
685 - Allow overriding OS parameters with ``gnt-instance reinstall``
686 - Added ability to change node's secondary IP address using ``gnt-node
688 - Implemented privilege separation for all daemons except
689 ``ganeti-noded``, see ``configure`` options
690 - Complain if an instance's disk is marked faulty in ``gnt-cluster
692 - Implemented job priorities (see ``ganeti(7)`` manpage)
693 - Ignore failures while shutting down instances during failover from
695 - Exit daemon's bootstrap process only once daemon is ready
696 - Export more information via ``LUInstanceQuery``/remote API
697 - Improved documentation, QA and unittests
698 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
699 restart if the file was created or changed
700 - Added LUXI protocol version sent with each request and response,
701 allowing detection of server/client mismatches
702 - Moved the Python scripts among gnt-* and ganeti-* into modules
703 - Moved all code related to setting up SSH to an external script,
705 - Infrastructure changes for node group support in future versions
711 *(Released Fri, 19 Nov 2010)*
713 A few small bugs fixed, and some improvements to the build system:
715 - Fix documentation regarding conversion to drbd
716 - Fix validation of parameters in cluster modify (``gnt-cluster modify
718 - Fix error handling in node modify with multiple changes
719 - Allow remote imports without checked names
725 *(Released Tue, 19 Oct 2010)*
727 - Disable SSL session ID cache in RPC client
733 *(Released Thu, 14 Oct 2010)*
735 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
736 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
738 - Fix problems with interaction between Curl and Python's HTTP server,
739 resulting in increased speed in many RPC calls
740 - Improve our release script to prevent breakage with older aclocal and
747 *(Released Thu, 7 Oct 2010)*
749 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
750 value retrieved from master
751 - Added support for blacklisted or hidden OS definitions
752 - Added simple lock monitor (accessible via (``gnt-debug locks``)
753 - Added support for -mem-path in KVM hypervisor abstraction layer
754 - Allow overriding instance parameters in tool for inter-cluster
755 instance moves (``tools/move-instance``)
756 - Improved opcode summaries (e.g. in ``gnt-job list``)
757 - Improve consistency of OS listing by sorting it
758 - Documentation updates
764 *(Released Fri, 8 Oct 2010)*
766 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
772 *(Released Mon, 4 Oct 2010)*
774 - Fixed regression in ``gnt-instance rename``
780 *(Released Wed, 22 Sep 2010)*
782 - Fixed OS_VARIANT variable for OS scripts
783 - Fixed cluster tag operations via RAPI
784 - Made ``setup-ssh`` exit with non-zero code if an error occurred
785 - Disabled RAPI CA checks in watcher
791 *(Released Mon, 23 Aug 2010)*
793 - Support DRBD versions of the format "a.b.c.d"
795 - Re-introduce support for usage from multiple threads in RAPI client
796 - Instance renames and modify via RAPI
797 - Work around race condition between processing and archival in job
799 - Mark opcodes following failed one as failed, too
800 - Job field ``lock_status`` was removed due to difficulties making it
801 work with the changed job queue in Ganeti 2.2; a better way to monitor
802 locks is expected for a later 2.2.x release
803 - Fixed dry-run behaviour with many commands
804 - Support ``ssh-agent`` again when adding nodes
805 - Many additional bugfixes
811 *(Released Fri, 30 Jul 2010)*
813 Important change: the internal RPC mechanism between Ganeti nodes has
814 changed from using a home-grown http library (based on the Python base
815 libraries) to use the PycURL library. This requires that PycURL is
816 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
817 against GnuTLS by default. cURL's support for GnuTLS had known issues
818 before cURL 7.21.0 and we recommend using the latest cURL release or
819 linking against OpenSSL. Most other distributions already link PycURL
820 and cURL against OpenSSL. The command::
822 python -c 'import pycurl; print pycurl.version'
824 can be used to determine the libraries PycURL and cURL are linked
827 Other significant changes:
829 - Rewrote much of the internals of the job queue, in order to achieve
830 better parallelism; this decouples job query operations from the job
831 processing, and it should allow much nicer behaviour of the master
832 daemon under load, and it also has uncovered some long-standing bugs
833 related to the job serialisation (now fixed)
834 - Added a default iallocator setting to the cluster parameters,
835 eliminating the need to always pass nodes or an iallocator for
836 operations that require selection of new node(s)
837 - Added experimental support for the LXC virtualization method
838 - Added support for OS parameters, which allows the installation of
839 instances to pass parameter to OS scripts in order to customise the
841 - Added a hypervisor parameter controlling the migration type (live or
842 non-live), since hypervisors have various levels of reliability; this
843 has renamed the 'live' parameter to 'mode'
844 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
845 logical volumes, meaning that cluster verify will ignore them and not
846 flag their presence as errors
847 - The watcher will now reset the error count for failed instances after
848 8 hours, thus allowing self-healing if the problem that caused the
849 instances to be down/fail to start has cleared in the meantime
850 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
851 check for, and warn, if the drbd module parameter ``usermode_helper``
852 is not consistent with the cluster-wide setting; this is needed to
853 make diagnose easier of failed drbd creations
854 - Started adding base IPv6 support, but this is not yet
855 enabled/available for use
856 - Rename operations (cluster, instance) will now return the new name,
857 which is especially useful if a short name was passed in
858 - Added support for instance migration in RAPI
859 - Added a tool to pre-configure nodes for the SSH setup, before joining
860 them to the cluster; this will allow in the future a simplified model
861 for node joining (but not yet fully enabled in 2.2); this needs the
862 paramiko python library
863 - Fixed handling of name-resolving errors
864 - Fixed consistency of job results on the error path
865 - Fixed master-failover race condition when executed multiple times in
867 - Fixed many bugs related to the job queue (mostly introduced during the
868 2.2 development cycle, so not all are impacting 2.1)
869 - Fixed instance migration with missing disk symlinks
870 - Fixed handling of unknown jobs in ``gnt-job archive``
871 - And many other small fixes/improvements
875 - Enhanced both the unittest and the QA coverage
876 - Switched the opcode validation to a generic model, and extended the
877 validation to all opcode parameters
878 - Changed more parts of the code that write shell scripts to use the
880 - Switched the master daemon to use the asyncore library for the Luxi
887 *(Released Thu, 17 Jun 2010)*
889 - Added tool (``move-instance``) and infrastructure to move instances
890 between separate clusters (see :doc:`separate documentation
891 <move-instance>` and :doc:`design document <design-2.2>`)
892 - Added per-request RPC timeout
893 - RAPI now requires a Content-Type header for requests with a body (e.g.
894 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
895 :rfc:`2616` (HTTP/1.1), section 7.2.1)
896 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
898 - Implemented initial support for running Ganeti daemons as separate
899 users, see configure-time flags ``--with-user-prefix`` and
900 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
902 - Instances can be removed after export (``gnt-backup export
904 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
905 key (instead of 1024 bit)
906 - Added new cluster configuration file for cluster domain secret
907 - Import/export now use SSL instead of SSH
908 - Added support for showing estimated time when exporting an instance,
909 see the ``ganeti-os-interface(7)`` manpage and look for
916 *(Released Tue, 16 Nov 2010)*
918 Some more bugfixes. Unless critical bugs occur, this will be the last
921 - Fix case of MAC special-values
922 - Fix mac checker regex
923 - backend: Fix typo causing "out of range" error
924 - Add missing --units in gnt-instance list man page
930 *(Released Tue, 24 Aug 2010)*
933 - Don't ignore secondary node silently on non-mirrored disk templates
935 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
936 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
937 - Properly document the usb_mouse parameter
938 - Fix path in ganeti-rapi(8) (issue 116)
939 - Adjust error message when the ganeti user's .ssh directory is
941 - Add same-node-check when changing the disk template to drbd
947 *(Released Fri, 16 Jul 2010)*
950 - Add an option to only select some reboot types during qa/burnin.
951 (on some hypervisors consequent reboots are not supported)
952 - Fix infrequent race condition in master failover. Sometimes the old
953 master ip address would be still detected as up for a short time
954 after it was removed, causing failover to fail.
955 - Decrease mlockall warnings when the ctypes module is missing. On
956 Python 2.4 we support running even if no ctypes module is installed,
957 but we were too verbose about this issue.
958 - Fix building on old distributions, on which man doesn't have a
960 - Fix RAPI not to ignore the MAC address on instance creation
961 - Implement the old instance creation format in the RAPI client.
967 *(Released Thu, 01 Jul 2010)*
969 A small bugfix release:
970 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
971 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
972 a non-existing option
973 - Fix instance create: the --force-variant option was ignored
974 - Improve pylint 0.21 compatibility and warnings with Python 2.6
975 - Fix modify node storage with non-FQDN arguments
976 - Fix RAPI client to authenticate under Python 2.6 when used
977 for more than 5 requests needing authentication
978 - Fix gnt-instance modify -t (storage) giving a wrong error message
979 when converting a non-shutdown drbd instance to plain
985 *(Released Fri, 18 Jun 2010)*
987 A small bugfix release:
989 - Fix live migration of KVM instances started with older Ganeti
990 versions which had fewer hypervisor parameters
991 - Fix gnt-instance grow-disk on down instances
992 - Fix an error-reporting bug during instance migration
993 - Better checking of the ``--net`` and ``--disk`` values, to avoid
994 silently ignoring broken ones
995 - Fix an RPC error reporting bug affecting, for example, RAPI client
997 - Fix bug triggered by different API version os-es on different nodes
998 - Fix a bug in instance startup with custom hvparams: OS level
999 parameters would fail to be applied.
1000 - Fix the RAPI client under Python 2.6 (but more work is needed to
1001 make it work completely well with OpenSSL)
1002 - Fix handling of errors when resolving names from DNS
1008 *(Released Thu, 3 Jun 2010)*
1010 A medium sized development cycle. Some new features, and some
1011 fixes/small improvements/cleanups.
1013 Significant features
1014 ~~~~~~~~~~~~~~~~~~~~
1016 The node deamon now tries to mlock itself into memory, unless the
1017 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1018 its logs, and falls back to console logging. This allows emergency
1019 features such as ``gnt-node powercycle`` to work even in the event of a
1020 broken node disk (tested offlining the disk hosting the node's
1021 filesystem and dropping its memory caches; don't try this at home)
1023 KVM: add vhost-net acceleration support. It can be tested with a new
1024 enough version of the kernel and of qemu-kvm.
1026 KVM: Add instance chrooting feature. If you use privilege dropping for
1027 your VMs you can also now force them to chroot to an empty directory,
1028 before starting the emulated guest.
1030 KVM: Add maximum migration bandwith and maximum downtime tweaking
1031 support (requires a new-enough version of qemu-kvm).
1033 Cluster verify will now warn if the master node doesn't have the master
1034 ip configured on it.
1036 Add a new (incompatible) instance creation request format to RAPI which
1037 supports all parameters (previously only a subset was supported, and it
1038 wasn't possible to extend the old format to accomodate all the new
1039 features. The old format is still supported, and a client can check for
1040 this feature, before using it, by checking for its presence in the
1041 ``features`` RAPI resource.
1043 Now with ancient latin support. Try it passing the ``--roman`` option to
1044 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1045 (requires the python-roman module to be installed, in order to work).
1050 As usual many internal code refactorings, documentation updates, and
1053 - Lots of improvements and cleanups to the experimental Remote API
1054 (RAPI) client library.
1055 - A new unit test suite for the core daemon libraries.
1056 - A fix to creating missing directories makes sure the umask is not
1057 applied anymore. This enforces the same directory permissions
1059 - Better handling terminating daemons with ctrl+c (used when running
1060 them in debugging mode).
1061 - Fix a race condition in live migrating a KVM instance, when stat()
1062 on the old proc status file returned EINVAL, which is an unexpected
1064 - Fixed manpage checking with newer man and utf-8 charachters. But now
1065 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1071 *(Released Fri, 7 May 2010)*
1073 Fix a bug which prevented untagged KVM instances from starting.
1079 *(Released Fri, 7 May 2010)*
1081 Another release with a long development cycle, during which many
1082 different features were added.
1084 Significant features
1085 ~~~~~~~~~~~~~~~~~~~~
1087 The KVM hypervisor now can run the individual instances as non-root, to
1088 reduce the impact of a VM being hijacked due to bugs in the
1089 hypervisor. It is possible to run all instances as a single (non-root)
1090 user, to manually specify a user for each instance, or to dynamically
1091 allocate a user out of a cluster-wide pool to each instance, with the
1092 guarantee that no two instances will run under the same user ID on any
1095 An experimental RAPI client library, that can be used standalone
1096 (without the other Ganeti libraries), is provided in the source tree as
1097 ``lib/rapi/client.py``. Note this client might change its interface in
1098 the future, as we iterate on its capabilities.
1100 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1101 replace the cluster's certificates and crypto keys. This might help in
1102 case they have been compromised, or have simply expired.
1104 A new disk option for instance creation has been added that allows one
1105 to "adopt" currently existing logical volumes, with data
1106 preservation. This should allow easier migration to Ganeti from
1107 unmanaged (or managed via other software) instances.
1109 Another disk improvement is the possibility to convert between redundant
1110 (DRBD) and plain (LVM) disk configuration for an instance. This should
1111 allow better scalability (starting with one node and growing the
1112 cluster, or shrinking a two-node cluster to one node).
1114 A new feature that could help with automated node failovers has been
1115 implemented: if a node sees itself as offline (by querying the master
1116 candidates), it will try to shutdown (hard) all instances and any active
1117 DRBD devices. This reduces the risk of duplicate instances if an
1118 external script automatically failovers the instances on such nodes. To
1119 enable this, the cluster parameter ``maintain_node_health`` should be
1120 enabled; in the future this option (per the name) will enable other
1121 automatic maintenance features.
1123 Instance export/import now will reuse the original instance
1124 specifications for all parameters; that means exporting an instance,
1125 deleting it and the importing it back should give an almost identical
1126 instance. Note that the default import behaviour has changed from
1127 before, where it created only one NIC; now it recreates the original
1130 Cluster verify has added a few new checks: SSL certificates validity,
1131 /etc/hosts consistency across the cluster, etc.
1136 As usual, many internal changes were done, documentation fixes,
1139 - Fixed cluster initialization with disabled cluster storage (regression
1140 introduced in 2.1.1)
1141 - File-based storage supports growing the disks
1142 - Fixed behaviour of node role changes
1143 - Fixed cluster verify for some corner cases, plus a general rewrite of
1144 cluster verify to allow future extension with more checks
1145 - Fixed log spamming by watcher and node daemon (regression introduced
1147 - Fixed possible validation issues when changing the list of enabled
1149 - Fixed cleanup of /etc/hosts during node removal
1150 - Fixed RAPI response for invalid methods
1151 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1152 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1153 from ide disks, etc.)
1154 - Allow OS changes without re-installation (to record a changed OS
1155 outside of Ganeti, or to allow OS renames)
1156 - Allow instance creation without OS installation (useful for example if
1157 the OS will be installed manually, or restored from a backup not in
1159 - Implemented option to make cluster ``copyfile`` use the replication
1161 - Added list of enabled hypervisors to ssconf (possibly useful for
1163 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1165 - A partial form of node re-IP is possible via node readd, which now
1166 allows changed node primary IP
1167 - Command line utilities now show an informational message if the job is
1169 - The logs of the master daemon now show the PID/UID/GID of the
1176 *(Released Fri, 12 Mar 2010)*
1178 During the 2.1.0 long release candidate cycle, a lot of improvements and
1179 changes have accumulated with were released later as 2.1.1.
1184 The node evacuate command (``gnt-node evacuate``) was significantly
1185 rewritten, and as such the IAllocator protocol was changed - a new
1186 request type has been added. This unfortunate change during a stable
1187 series is designed to improve performance of node evacuations; on
1188 clusters with more than about five nodes and which are well-balanced,
1189 evacuation should proceed in parallel for all instances of the node
1190 being evacuated. As such, any existing IAllocator scripts need to be
1191 updated, otherwise the above command will fail due to the unknown
1192 request. The provided "dumb" allocator has not been updated; but the
1193 ganeti-htools package supports the new protocol since version 0.2.4.
1195 Another important change is increased validation of node and instance
1196 names. This might create problems in special cases, if invalid host
1197 names are being used.
1199 Also, a new layer of hypervisor parameters has been added, that sits at
1200 OS level between the cluster defaults and the instance ones. This allows
1201 customisation of virtualization parameters depending on the installed
1202 OS. For example instances with OS 'X' may have a different KVM kernel
1203 (or any other parameter) than the cluster defaults. This is intended to
1204 help managing a multiple OSes on the same cluster, without manual
1205 modification of each instance's parameters.
1207 A tool for merging clusters, ``cluster-merge``, has been added in the
1208 tools sub-directory.
1213 - Improved the int/float conversions that should make the code more
1214 robust in face of errors from the node daemons
1215 - Fixed the remove node code in case of internal configuration errors
1216 - Fixed the node daemon behaviour in face of inconsistent queue
1217 directory (e.g. read-only file-system where we can't open the files
1219 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1220 now it either aborts cleanly or, if given the new "auto_promote"
1221 parameter, will automatically promote other nodes as needed
1222 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1223 completely prevent Ganeti from working
1224 - Fixed bug for instance export when not all disks were successfully
1226 - Fixed behaviour of node add when the new node is slow in starting up
1228 - Fixed handling of signals in the LUXI client, which should improve
1229 behaviour of command-line scripts
1230 - Added checks for invalid node/instance names in the configuration (now
1231 flagged during cluster verify)
1232 - Fixed watcher behaviour for disk activation errors
1233 - Fixed two potentially endless loops in http library, which led to the
1234 RAPI daemon hanging and consuming 100% CPU in some cases
1235 - Fixed bug in RAPI daemon related to hashed passwords
1236 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1238 - Enhanced compatibility with non-Debian OSes, but not using absolute
1239 path in some commands and allowing customisation of the ssh
1240 configuration directory
1241 - Fixed possible future issue with new Python versions by abiding to the
1242 proper use of ``__slots__`` attribute on classes
1243 - Added checks that should prevent directory traversal attacks
1244 - Many documentation fixes based on feedback from users
1249 - Added an "early_release" more for instance replace disks and node
1250 evacuate, where we release locks earlier and thus allow higher
1251 parallelism within the cluster
1252 - Added watcher hooks, intended to allow the watcher to restart other
1253 daemons (e.g. from the ganeti-nbma project), but they can be used of
1254 course for any other purpose
1255 - Added a compile-time disable for DRBD barriers, to increase
1256 performance if the administrator trusts the power supply or the
1257 storage system to not lose writes
1258 - Added the option of using syslog for logging instead of, or in
1259 addition to, Ganeti's own log files
1260 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1261 can indeed boot from a paravirtual NIC
1262 - Added a generic debug level for many operations; while this is not
1263 used widely yet, it allows one to pass the debug value all the way to
1265 - Enhanced the hooks environment for instance moves (failovers,
1266 migrations) where the primary/secondary nodes changed during the
1267 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1268 - Enhanced data validations for many user-supplied values; one important
1269 item is the restrictions imposed on instance and node names, which
1270 might reject some (invalid) host names
1271 - Add a configure-time option to disable file-based storage, if it's not
1272 needed; this allows greater security separation between the master
1273 node and the other nodes from the point of view of the inter-node RPC
1275 - Added user notification in interactive tools if job is waiting in the
1276 job queue or trying to acquire locks
1277 - Added log messages when a job is waiting for locks
1278 - Added filtering by node tags in instance operations which admit
1279 multiple instances (start, stop, reboot, reinstall)
1280 - Added a new tool for cluster mergers, ``cluster-merge``
1281 - Parameters from command line which are of the form ``a=b,c=d`` can now
1282 use backslash escapes to pass in values which contain commas,
1283 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1285 - For KVM, the instance name is the first parameter passed to KVM, so
1286 that it's more visible in the process list
1292 *(Released Tue, 2 Mar 2010)*
1294 Ganeti 2.1 brings many improvements with it. Major changes:
1296 - Added infrastructure to ease automated disk repairs
1297 - Added new daemon to export configuration data in a cheaper way than
1298 using the remote API
1299 - Instance NICs can now be routed instead of being associated with a
1301 - Improved job locking logic to reduce impact of jobs acquiring multiple
1302 locks waiting for other long-running jobs
1304 In-depth implementation details can be found in the Ganeti 2.1 design
1310 - Added chroot hypervisor
1311 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1313 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1314 ``bootloader_path`` and ``bootloader_args``)
1315 - Added the ``use_localtime`` option for the xen-hvm and kvm
1316 hypervisors, and the default value for this has changed to false (in
1317 2.0 xen-hvm always enabled it)
1318 - Added luxi call to submit multiple jobs in one go
1319 - Added cluster initialization option to not modify ``/etc/hosts``
1321 - Added network interface parameters
1322 - Added dry run mode to some LUs
1323 - Added RAPI resources:
1325 - ``/2/instances/[instance_name]/info``
1326 - ``/2/instances/[instance_name]/replace-disks``
1327 - ``/2/nodes/[node_name]/evacuate``
1328 - ``/2/nodes/[node_name]/migrate``
1329 - ``/2/nodes/[node_name]/role``
1330 - ``/2/nodes/[node_name]/storage``
1331 - ``/2/nodes/[node_name]/storage/modify``
1332 - ``/2/nodes/[node_name]/storage/repair``
1334 - Added OpCodes to evacuate or migrate all instances on a node
1335 - Added new command to list storage elements on nodes (``gnt-node
1336 list-storage``) and modify them (``gnt-node modify-storage``)
1337 - Added new ssconf files with master candidate IP address
1338 (``ssconf_master_candidates_ips``), node primary IP address
1339 (``ssconf_node_primary_ips``) and node secondary IP address
1340 (``ssconf_node_secondary_ips``)
1341 - Added ``ganeti-confd`` and a client library to query the Ganeti
1342 configuration via UDP
1343 - Added ability to run hooks after cluster initialization and before
1345 - Added automatic mode for disk replace (``gnt-instance replace-disks
1347 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1348 after catastrophic data-loss
1349 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1351 - Added ``gnt-instance move`` command to move instances
1352 - Added ``gnt-cluster watcher`` command to control watcher
1353 - Added ``gnt-node powercycle`` command to powercycle nodes
1354 - Added new job status field ``lock_status``
1355 - Added parseable error codes to cluster verification (``gnt-cluster
1356 verify --error-codes``) and made output less verbose (use
1357 ``--verbose`` to restore previous behaviour)
1358 - Added UUIDs to the main config entities (cluster, nodes, instances)
1359 - Added support for OS variants
1360 - Added support for hashed passwords in the Ganeti remote API users file
1362 - Added option to specify maximum timeout on instance shutdown
1363 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1364 - Added new helper script to start and stop Ganeti daemons
1365 (``daemon-util``), with the intent to reduce the work necessary to
1366 adjust Ganeti for non-Debian distributions and to start/stop daemons
1368 - Added more unittests
1369 - Fixed critical bug in ganeti-masterd startup
1370 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1371 now customisable at the cluster level for both the KVM and Xen
1372 hypervisors using the new ``migration_port`` parameter
1373 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1374 reinstalling an instance
1375 - Allowed ``@`` in tag names
1376 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1377 - Many documentation updates
1378 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1379 - ``gnt-instance reinstall`` can now reinstall multiple instances
1380 - Updated many command line parameters
1381 - Introduced new OS API version 15
1382 - No longer support a default hypervisor
1383 - Treat virtual LVs as inexistent
1384 - Improved job locking logic to reduce lock contention
1385 - Match instance and node names case insensitively
1386 - Reimplemented bash completion script to be more complete
1393 *(Released Thu, 4 Feb 2010)*
1395 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1396 - Fix a string formatting bug
1397 - Improve safety of the code in some error paths
1398 - Improve data validation in the master of values returned from nodes
1404 *(Released Thu, 17 Dec 2009)*
1406 - Fix security issue due to missing validation of iallocator names; this
1407 allows local and remote execution of arbitrary executables
1408 - Fix failure of gnt-node list during instance removal
1409 - Ship the RAPI documentation in the archive
1415 *(Released Wed, 30 Sep 2009)*
1417 - Fixed many wrong messages
1418 - Fixed a few bugs related to the locking library
1419 - Fixed MAC checking at instance creation time
1420 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1421 - Fixed a few issues related to signal handling in both daemons and
1423 - Fixed the example startup script provided
1424 - Fixed insserv dependencies in the example startup script (patch from
1426 - Fixed handling of drained nodes in the iallocator framework
1427 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1429 - Fixed error related to invalid job IDs in job polling
1430 - Fixed job/opcode persistence on unclean master shutdown
1431 - Fixed handling of partial job processing after unclean master
1433 - Fixed error reporting from LUs, previously all errors were converted
1434 into execution errors
1435 - Fixed error reporting from burnin
1436 - Decreased significantly the memory usage of the job queue
1437 - Optimised slightly multi-job submission
1438 - Optimised slightly opcode loading
1439 - Backported the multi-job submit framework from the development
1440 branch; multi-instance start and stop should be faster
1441 - Added script to clean archived jobs after 21 days; this will reduce
1442 the size of the queue directory
1443 - Added some extra checks in disk size tracking
1444 - Added an example ethers hook script
1445 - Added a cluster parameter that prevents Ganeti from modifying of
1447 - Added more node information to RAPI responses
1448 - Added a ``gnt-job watch`` command that allows following the ouput of a
1450 - Added a bind-address option to ganeti-rapi
1451 - Added more checks to the configuration verify
1452 - Enhanced the burnin script such that some operations can be retried
1454 - Converted instance reinstall to multi-instance model
1460 *(Released Fri, 7 Aug 2009)*
1462 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1463 to allow using the pre-2.0.2 behaviour in activation, if any existing
1464 instances have mismatched disk sizes in the configuration
1465 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
1466 any configuration mismatches for disk sizes
1467 - Added ``gnt-master cluste-failover --no-voting`` to allow master
1468 failover to work on two-node clusters
1469 - Fixed the ``--net`` option of ``gnt-backup import``, which was
1471 - Fixed detection of OS script errors in ``gnt-backup export``
1472 - Fixed exit code of ``gnt-backup export``
1478 *(Released Fri, 17 Jul 2009)*
1480 - Added experimental support for stripped logical volumes; this should
1481 enhance performance but comes with a higher complexity in the block
1482 device handling; stripping is only enabled when passing
1483 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
1484 affected even in the non-stripped mode
1485 - Improved resiliency against transient failures at the end of DRBD
1486 resyncs, and in general of DRBD resync checks
1487 - Fixed a couple of issues with exports and snapshot errors
1488 - Fixed a couple of issues in instance listing
1489 - Added display of the disk size in ``gnt-instance info``
1490 - Fixed checking for valid OSes in instance creation
1491 - Fixed handling of the "vcpus" parameter in instance listing and in
1492 general of invalid parameters
1493 - Fixed http server library, and thus RAPI, to handle invalid
1494 username/password combinations correctly; this means that now they
1495 report unauthorized for queries too, not only for modifications,
1496 allowing earlier detect of configuration problems
1497 - Added a new "role" node list field, equivalent to the master/master
1498 candidate/drained/offline flags combinations
1499 - Fixed cluster modify and changes of candidate pool size
1500 - Fixed cluster verify error messages for wrong files on regular nodes
1501 - Fixed a couple of issues with node demotion from master candidate role
1502 - Fixed node readd issues
1503 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
1504 - Added a new ``--no-voting`` option for masterfailover to fix failover
1505 on two-nodes clusters when the former master node is unreachable
1506 - Added instance reinstall over RAPI
1512 *(Released Tue, 16 Jun 2009)*
1514 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
1515 allow re-adding the start in single-user option (regression from 1.2)
1516 - the watcher writes the instance status to a file, to allow monitoring
1517 to report the instance status (from the master) based on cached
1518 results of the watcher's queries; while this can get stale if the
1519 watcher is being locked due to other work on the cluster, this is
1520 still an improvement
1521 - the watcher now also restarts the node daemon and the rapi daemon if
1523 - fixed the watcher to handle full and drained queue cases
1524 - hooks export more instance data in the environment, which helps if
1525 hook scripts need to take action based on the instance's properties
1526 (no longer need to query back into ganeti)
1527 - instance failovers when the instance is stopped do not check for free
1528 RAM, so that failing over a stopped instance is possible in low memory
1530 - rapi uses queries for tags instead of jobs (for less job traffic), and
1531 for cluster tags it won't talk to masterd at all but read them from
1533 - a couple of error handling fixes in RAPI
1534 - drbd handling: improved the error handling of inconsistent disks after
1535 resync to reduce the frequency of "there are some degraded disks for
1536 this instance" messages
1537 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
1538 error handling path called a wrong function name)
1544 *(Released Wed, 27 May 2009)*
1546 - no changes from rc5
1552 *(Released Wed, 20 May 2009)*
1554 - fix a couple of bugs (validation, argument checks)
1555 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
1556 - some small improvements to RAPI and IAllocator
1557 - make watcher automatically start the master daemon if down
1563 *(Released Mon, 27 Apr 2009)*
1565 - change the OS list to not require locks; this helps with big clusters
1566 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
1567 volume group is broken
1568 - ``gnt-instance info``, without any arguments, doesn't run for all
1569 instances anymore; either pass ``--all`` or pass the desired
1570 instances; this helps against mistakes on big clusters where listing
1571 the information for all instances takes a long time
1572 - miscellaneous doc and man pages fixes
1578 *(Released Wed, 8 Apr 2009)*
1580 - Change the internal locking model of some ``gnt-node`` commands, in
1581 order to reduce contention (and blocking of master daemon) when
1582 batching many creation/reinstall jobs
1583 - Fixes to Xen soft reboot
1584 - No longer build documentation at build time, instead distribute it in
1585 the archive, in order to reduce the need for the whole docbook/rst
1592 *(Released Fri, 27 Mar 2009)*
1594 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
1595 - Fix watcher startup sequence, improves the behaviour of busy clusters
1596 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
1597 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
1598 bind address and other places
1599 - Some documentation fixes and updates
1605 *(Released Mon, 2 Mar 2009)*
1607 - More documentation updates, now all docs should be more-or-less
1609 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
1611 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
1613 - Fix an issue related to $libdir/run/ganeti and cluster creation
1619 *(Released Thu, 19 Feb 2009)*
1621 - Xen PVM and KVM have switched the default value for the instance root
1622 disk to the first partition on the first drive, instead of the whole
1623 drive; this means that the OS installation scripts must be changed
1625 - Man pages have been updated
1626 - RAPI has been switched by default to HTTPS, and the exported functions
1627 should all work correctly
1628 - RAPI v1 has been removed
1629 - Many improvements to the KVM hypervisor
1630 - Block device errors are now better reported
1631 - Many other bugfixes and small improvements
1637 *(Released Mon, 26 Jan 2009)*
1639 - Version 2 is a general rewrite of the code and therefore the
1640 differences are too many to list, see the design document for 2.0 in
1641 the ``doc/`` subdirectory for more details
1642 - In this beta version there is not yet a migration path from 1.2 (there
1643 will be one in the final 2.0 release)
1644 - A few significant changes are:
1646 - all commands are executed by a daemon (``ganeti-masterd``) and the
1647 various ``gnt-*`` commands are just front-ends to it
1648 - all the commands are entered into, and executed from a job queue,
1649 see the ``gnt-job(8)`` manpage
1650 - the RAPI daemon supports read-write operations, secured by basic
1651 HTTP authentication on top of HTTPS
1652 - DRBD version 0.7 support has been removed, DRBD 8 is the only
1653 supported version (when migrating from Ganeti 1.2 to 2.0, you need
1654 to migrate to DRBD 8 first while still running Ganeti 1.2)
1655 - DRBD devices are using statically allocated minor numbers, which
1656 will be assigned to existing instances during the migration process
1657 - there is support for both Xen PVM and Xen HVM instances running on
1659 - KVM virtualization is supported too
1660 - file-based storage has been implemented, which means that it is
1661 possible to run the cluster without LVM and DRBD storage, for
1662 example using a shared filesystem exported from shared storage (and
1663 still have live migration)
1669 *(Released Tue, 13 Jan 2009)*
1671 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
1672 - Reuse the old instance mac address by default on instance import, if
1673 the instance name is the same.
1674 - Handle situations in which the node info rpc returns incomplete
1676 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
1677 - Improved version of batcher:
1679 - state file support
1680 - instance mac address support
1681 - support for HVM clusters/instances
1683 - Add an option to show the number of cpu sockets and nodes in
1685 - Support OSes that handle more than one version of the OS api (but do
1686 not change the current API in any other way)
1687 - Fix ``gnt-node migrate``
1688 - ``gnt-debug`` man page
1689 - Fixes various more typos and small issues
1690 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
1696 *(Released Wed, 24 Sep 2008)*
1698 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
1699 type of disk exported to fully virtualized instances.
1700 - provide access to the serial console of HVM instances
1701 - instance auto_balance flag, set by default. If turned off it will
1702 avoid warnings on cluster verify if there is not enough memory to fail
1703 over an instance. in the future it will prevent automatically failing
1704 it over when we will support that.
1705 - batcher tool for instance creation, see ``tools/README.batcher``
1706 - ``gnt-instance reinstall --select-os`` to interactively select a new
1707 operating system when reinstalling an instance.
1708 - when changing the memory amount on instance modify a check has been
1709 added that the instance will be able to start. also warnings are
1710 emitted if the instance will not be able to fail over, if auto_balance
1712 - documentation fixes
1713 - sync fields between ``gnt-instance list/modify/add/import``
1714 - fix a race condition in drbd when the sync speed was set after giving
1715 the device a remote peer.
1721 *(Released Tue, 22 Jul 2008)*
1723 - note: the allowed size and number of tags per object were reduced
1724 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1725 - fixed twisted 8.x compatibility
1726 - fixed ``gnt-instance replace-disks`` with iallocator
1727 - add TCP keepalives on twisted connections to detect restarted nodes
1728 - disk increase support, see ``gnt-instance grow-disk``
1729 - implement bulk node/instance query for RAPI
1730 - add tags in node/instance listing (optional)
1731 - experimental migration (and live migration) support, read the man page
1732 for ``gnt-instance migrate``
1733 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1734 has some small improvements in handling its state file
1740 *(Released Fri, 13 Jun 2008)*
1742 - Experimental readonly, REST-based remote API implementation;
1743 automatically started on master node, TCP port 5080, if enabled by
1744 ``--enable-rapi`` parameter to configure script.
1745 - Instance allocator support. Add and import instance accept a
1746 ``--iallocator`` parameter, and call that instance allocator to decide
1747 which node to use for the instance. The iallocator document describes
1748 what's expected from an allocator script.
1749 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1750 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1751 node is lost there is still enough memory to fail over the instances
1753 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1754 ``gnt-cluster verify``, to check for site-specific compliance. All the
1755 hooks will run, and their output, if any, will be displayed. Any
1756 failing hook will make the verification return an error value.
1757 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1758 primary and secondary interfaces
1759 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1760 that is still declared as part of the cluster and has failed.
1761 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1762 specifying output fields, that just adds the chosen fields to the
1764 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1765 export from the filesystem.
1766 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1767 have been added. Using them you can enable/disable acpi and pae
1768 support, and specify a path for a cd image to be exported to the
1769 instance. These parameters as the name suggest only work on HVM
1771 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1772 PAE support will be set to the previously hardcoded values, but the
1773 (previously hardcoded) path to the CDROM ISO image will be unset and
1774 if required, needs to be set manually with ``gnt-instance modify``
1776 - The address to which an instance's VNC console is bound is now
1777 selectable per-instance, rather than being cluster wide. Of course
1778 this only applies to instances controlled via VNC, so currently just
1779 applies to HVM clusters.
1785 *(Released Mon, 18 Feb 2008)*
1787 - more tweaks to the disk activation code (especially helpful for DRBD)
1788 - change the default ``gnt-instance list`` output format, now there is
1789 one combined status field (see the manpage for the exact values this
1791 - some more fixes for the mac export to hooks change
1792 - make Ganeti not break with DRBD 8.2.x (which changed the version
1793 format in ``/proc/drbd``) (issue 24)
1794 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1795 template, allowing migration from DRBD0.7+MD to DRBD8
1801 *(Released Wed, 30 Jan 2008)*
1803 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1805 - add command aliases infrastructure and a few aliases
1806 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1807 man pages and the ``--help`` option of ``gnt-node
1808 list``/``gnt-instance list``
1809 - fix ``gnt-backup list`` with down nodes (issue 21)
1810 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1811 - fix the dist archive and add a check for including svn/git files in
1813 - some developer-related changes: improve the burnin and the QA suite,
1814 add an upload script for testing during development
1820 *(Released Wed, 16 Jan 2008)*
1822 - experimental HVM support, read the install document, section
1823 "Initializing the cluster"
1824 - allow for the PVM hypervisor per-instance kernel and initrd paths
1825 - add a new command ``gnt-cluster verify-disks`` which uses a new
1826 algorithm to improve the reconnection of the DRBD pairs if the device
1827 on the secondary node has gone away
1828 - make logical volume code auto-activate LVs at disk activation time
1829 - slightly improve the speed of activating disks
1830 - allow specification of the MAC address at instance creation time, and
1831 changing it later via ``gnt-instance modify``
1832 - fix handling of external commands that generate lots of output on
1834 - update documentation with regard to minimum version of DRBD8 supported
1840 *(Released Tue, 4 Dec 2007)*
1842 - Log the ``xm create`` output to the node daemon log on failure (to
1843 help diagnosing the error)
1844 - In debug mode, log all external commands output if failed to the logs
1845 - Change parsing of lvm commands to ignore stderr
1851 *(Released Wed, 28 Nov 2007)*
1853 - Another round of updates to the DRBD 8 code to deal with more failures
1854 in the replace secondary node operation
1855 - Some more logging of failures in disk operations (lvm, drbd)
1856 - A few documentation updates
1863 *(Released Tue, 13 Nov 2007)*
1865 - Change configuration file format from Python's Pickle to JSON.
1866 Upgrading is possible using the cfgupgrade utility.
1867 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1868 faster replace disks and is more stable (DRBD 8 has many improvements
1869 compared to DRBD 0.7)
1870 - Added command line tags support (see man pages for ``gnt-instance``,
1871 ``gnt-node``, ``gnt-cluster``)
1872 - Added instance rename support
1873 - Added multi-instance startup/shutdown
1874 - Added cluster rename support
1875 - Added ``gnt-node evacuate`` to simplify some node operations
1876 - Added instance reboot operation that can speedup reboot as compared to
1878 - Soften the requirement that hostnames are in FQDN format
1879 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1881 - Removed dependency on debian's patched fping that uses the
1882 non-standard ``-S`` option
1883 - Now the OS definitions are searched for in multiple, configurable
1884 paths (easier for distros to package)
1885 - Some changes to the hooks infrastructure (especially the new
1886 post-configuration update hook)
1887 - Other small bugfixes
1889 .. vim: set textwidth=72 syntax=rst :