10 Improvements and bugfixes
11 ~~~~~~~~~~~~~~~~~~~~~~~~~
13 - Support for kvm version 1.0, that changed the version reporting format
20 *(Released Thu, 27 Oct 2011)*
22 Incompatible/important changes and bugfixes
23 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
25 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
26 resource's ``ip_check`` parameter changed from ``True`` to ``False``
27 to match the underlying LUXI interface.
28 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
29 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
30 not maintain backwards-compatibility as the underlying operation
31 changed in an incompatible way. The RAPI client can talk to old
32 servers, but it needs to be told so as the return value changed.
33 - When creating file-based instances via RAPI, the ``file_driver``
34 parameter no longer defaults to ``loop`` and must be specified.
35 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
37 - Support for the undocumented and deprecated RAPI instance creation
38 request format version 0 has been dropped. Use version 1, supported
39 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
40 - Pyparsing 1.4.6 or above is required, see :doc:`installation
41 documentation <install>`.
42 - The "cluster-verify" hooks are now executed per group by the
43 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
44 if you just run ``gnt-cluster verify``, which generates one opcode per
46 - The environment as passed to the OS scripts is cleared, and thus no
47 environment variables defined in the node daemon's environment will be
48 inherited by the scripts.
49 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
51 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
52 support operations involving multiple node groups.
53 - Offline nodes are ignored when failing over an instance.
58 - The ganeti-htools project has been merged into the ganeti-core source
59 tree and will be built as part of Ganeti (see :doc:`install-quick`).
60 - Implemented support for :doc:`shared storage <design-shared-storage>`.
61 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
62 GPT-style partition tables (requires `parted
63 <http://www.gnu.org/s/parted/>`_).
64 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
65 - Allowed adding tags on instance creation.
66 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
67 - Allow instances to be started in a paused state, enabling the user to
68 see the complete console output on boot using the console.
69 - Added new hypervisor flag to control default reboot behaviour
70 (``reboot_behavior``).
71 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
72 - Improved out-of-band management support:
74 - Added ``gnt-node health`` command reporting the health status of
76 - Added ``gnt-node power`` command to manage power status of nodes.
77 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
79 - Instance migration can fall back to failover if instance is not
81 - Filters can be used when listing nodes, instances, groups and locks;
82 see *ganeti(7)* manpage.
83 - Added post-execution status as variables to :doc:`hooks <hooks>`
85 - Instance tags are exported/imported together with the instance.
86 - When given an explicit job ID, ``gnt-job info`` will work for archived
88 - Jobs can define dependencies on other jobs (not yet supported via
89 RAPI or command line, but used by internal commands and usable via
92 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
95 - Instance failover is now available as a RAPI resource
96 (``/2/instances/[instance_name]/failover``).
97 - ``gnt-instance info`` defaults to static information if primary node
99 - Opcodes have a new ``comment`` attribute.
100 - Added basic SPICE support to KVM hypervisor.
101 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
103 Node group improvements
104 ~~~~~~~~~~~~~~~~~~~~~~~
106 - ``gnt-cluster verify`` has been modified to check groups separately,
107 thereby improving performance.
108 - Node group support has been added to ``gnt-cluster verify-disks``,
109 which now operates per node group.
110 - Watcher has been changed to work better with node groups.
112 - One process and state file per node group.
113 - Slow watcher in one group doesn't block other group's watcher.
115 - Added new command, ``gnt-group evacuate``, to move all instances in a
116 node group to other groups.
117 - Added ``gnt-instance change-group`` to move an instance to another
119 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
120 per-group operations.
121 - Node groups can be tagged.
122 - Some operations switch from an exclusive to a shared lock as soon as
124 - Instance's primary and secondary nodes' groups are now available as
125 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
126 and ``snodes.group.uuid``).
131 - Numerous updates to documentation and manpages.
133 - :doc:`RAPI <rapi>` documentation now has detailed parameter
135 - Some opcode/job results are now also documented, see :doc:`RAPI
138 - A lockset's internal lock is now also visible in lock monitor.
139 - Log messages from job queue workers now contain information about the
140 opcode they're processing.
141 - ``gnt-instance console`` no longer requires the instance lock.
142 - A short delay when waiting for job changes reduces the number of LUXI
143 requests significantly.
144 - DRBD metadata volumes are overwritten with zeros during disk creation.
150 *(Released Wed, 26 Oct 2011)*
152 This was the third release candidate of the 2.5 series.
158 *(Released Tue, 18 Oct 2011)*
160 This was the second release candidate of the 2.5 series.
166 *(Released Tue, 4 Oct 2011)*
168 This was the first release candidate of the 2.5 series.
174 *(Released Wed, 31 Aug 2011)*
176 This was the third beta release of the 2.5 series.
182 *(Released Mon, 22 Aug 2011)*
184 This was the second beta release of the 2.5 series.
190 *(Released Fri, 12 Aug 2011)*
192 This was the first beta release of the 2.5 series.
198 *(Released Thu, 27 Oct 2011)*
200 - Fixed bug when parsing command line parameter values ending in
202 - Fixed assertion error after unclean master shutdown
203 - Disable HTTP client pool for RPC, significantly reducing memory usage
205 - Fixed queue archive creation with wrong permissions
211 *(Released Tue, 23 Aug 2011)*
215 - Fixed documentation for importing with ``--src-dir`` option
216 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
217 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
223 *(Released Fri, 5 Aug 2011)*
225 Many bug-fixes and a few small features:
227 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
228 issues when you tried to add an instance with two MAC addresses in one
230 - KVM: fixed per-instance stored UID value
231 - KVM: configure bridged NICs at migration start
232 - KVM: Fix a bug where instance will not start with never KVM versions
234 - Added OS search path to ``gnt-cluster info``
235 - Fixed an issue with ``file_storage_dir`` where you were forced to
236 provide an absolute path, but the documentation states it is a
237 relative path, the documentation was right
238 - Added a new parameter to instance stop/start called ``--no-remember``
239 that will make the state change to not be remembered
240 - Implemented ``no_remember`` at RAPI level
241 - Improved the documentation
242 - Node evacuation: don't call IAllocator if node is already empty
243 - Fixed bug in DRBD8 replace disks on current nodes
244 - Fixed bug in recreate-disks for DRBD instances
245 - Moved assertion checking locks in ``gnt-instance replace-disks``
246 causing it to abort with not owning the right locks for some situation
247 - Job queue: Fixed potential race condition when cancelling queued jobs
248 - Fixed off-by-one bug in job serial generation
249 - ``gnt-node volumes``: Fix instance names
250 - Fixed aliases in bash completion
251 - Fixed a bug in reopening log files after being sent a SIGHUP
252 - Added a flag to burnin to allow specifying VCPU count
253 - Bugfixes to non-root Ganeti configuration
259 *(Released Thu, 12 May 2011)*
261 Many bug-fixes and a few new small features:
263 - Fixed a bug related to log opening failures
264 - Fixed a bug in instance listing with orphan instances
265 - Fixed a bug which prevented resetting the cluster-level node parameter
266 ``oob_program`` to the default
267 - Many fixes related to the ``cluster-merge`` tool
268 - Fixed a race condition in the lock monitor, which caused failures
269 during (at least) creation of many instances in parallel
270 - Improved output for gnt-job info
271 - Removed the quiet flag on some ssh calls which prevented debugging
273 - Improved the N+1 failure messages in cluster verify by actually
274 showing the memory values (needed and available)
275 - Increased lock attempt timeouts so that when executing long operations
276 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
277 too early and thus prevent the use of the 'fair' mechanism
278 - Changed instance query data (``gnt-instance info``) to not acquire
279 locks unless needed, thus allowing its use on locked instance if only
280 static information is asked for
281 - Improved behaviour with filesystems that do not support rename on an
283 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
284 kept locks on all nodes during the wipe, which is unneeded
285 - Fixed ``gnt-watcher`` handling of errors during hooks execution
286 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
287 10GiB) which caused the wipe to fail right at the end in some cases
288 - Fixed master IP activation when doing master failover with no-voting
289 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
290 the master node itself
291 - Fixed potential data-loss in under disk full conditions, where Ganeti
292 wouldn't check correctly the return code and would consider
293 partially-written files 'correct'
294 - Fixed bug related to multiple VGs and DRBD disk replacing
295 - Added new disk parameter ``metavg`` that allows placement of the meta
296 device for DRBD in a different volume group
297 - Fixed error handling in the node daemon when the system libc doesn't
298 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
299 - Fixed lock release during replace-disks, which kept cluster-wide locks
300 when doing disk replaces with an iallocator script
301 - Added check for missing bridges in cluster verify
302 - Handle EPIPE errors while writing to the terminal better, so that
303 piping the output to e.g. ``less`` doesn't cause a backtrace
304 - Fixed rare case where a ^C during Luxi calls could have been
305 interpreted as server errors, instead of simply terminating
306 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
308 - Added a few more parameters to the KVM hypervisor, allowing a second
309 CDROM, custom disk type for CDROMs and a floppy image
310 - Removed redundant message in instance rename when the name is given
312 - Added option to ``gnt-instance recreate-disks`` to allow creating the
313 disks on new nodes, allowing recreation when the original instance
314 nodes are completely gone
315 - Added option when converting disk templates to DRBD to skip waiting
316 for the resync, in order to make the instance available sooner
317 - Added two new variables to the OS scripts environment (containing the
319 - Made the root_path and optional parameter for the xen-pvm hypervisor,
320 to allow use of ``pvgrub`` as bootloader
321 - Changed the instance memory modifications to only check out-of-memory
322 conditions on memory increases, and turned the secondary node warnings
323 into errors (they can still be overridden via ``--force``)
324 - Fixed the handling of a corner case when the Python installation gets
325 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
326 to execute a command that doesn't exist
327 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
328 node of the instance returned failures during instance shutdown; this
329 adds the option ``--ignore-consistency`` to gnt-instance move
331 And as usual, various improvements to the error messages, documentation
337 *(Released Wed, 09 Mar 2011)*
339 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
340 the RAPI users file if run twice (even with ``--dry-run``).
342 The release fixes that bug (nothing else changed).
348 *(Released Mon, 07 Mar 2011)*
350 Final 2.4.0 release. Just a few small fixes:
352 - Fixed RAPI node evacuate
353 - Fixed the kvm-ifup script
354 - Fixed internal error handling for special job cases
355 - Updated man page to specify the escaping feature for options
361 *(Released Mon, 28 Feb 2011)*
363 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
364 that this feature wiped the disks of the wrong instance, leading to loss
369 - Fixed title of query field containing instance name
370 - Expanded the glossary in the documentation
371 - Fixed one unittest (internal issue)
377 *(Released Mon, 21 Feb 2011)*
379 A number of bug fixes plus just a couple functionality changes.
381 On the user-visible side, the ``gnt-* list`` command output has changed
382 with respect to "special" field states. The current rc1 style of display
383 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
384 the default output mode special fields states are displayed as follows:
386 - Offline resource: ``*``
387 - Unavailable/not applicable: ``-``
388 - Data missing (RPC failure): ``?``
389 - Unknown field: ``??``
391 Another user-visible change is the addition of ``--force-join`` to
396 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
397 - Fixed regression in RAPI/instance reinstall where all parameters were
398 required (instead of optional)
399 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
400 - Fixed iallocator usage (offline nodes were not considered offline)
401 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
402 - Fixed hypervisor and OS parameter validation with respect to
404 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
406 - Fixed ``tools/listrunner`` with respect to agent-based usage
412 *(Released Fri, 4 Feb 2011)*
414 Many changes and fixes since the beta1 release. While there were some
415 internal changes, the code has been mostly stabilised for the RC
418 Note: the dumb allocator was removed in this release, as it was not kept
419 up-to-date with the IAllocator protocol changes. It is recommended to
420 use the ``hail`` command from the ganeti-htools package.
422 Note: the 2.4 and up versions of Ganeti are not compatible with the
423 0.2.x branch of ganeti-htools. You need to upgrade to
424 ganeti-htools-0.3.0 (or later).
426 Regressions fixed from 2.3
427 ~~~~~~~~~~~~~~~~~~~~~~~~~~
429 - Fixed the ``gnt-cluster verify-disks`` command
430 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
432 - Fixed disk adoption breakage
433 - Fixed wrong headers in instance listing for field aliases
438 - Fixed corner case in KVM handling of NICs
439 - Fixed many cases of wrong handling of non-vm_capable nodes
440 - Fixed a bug where a missing instance symlink was not possible to
441 recreate with any ``gnt-*`` command (now ``gnt-instance
442 activate-disks`` does it)
443 - Fixed the volume group name as reported by ``gnt-cluster
445 - Increased timeouts for the import-export code, hopefully leading to
446 fewer aborts due network or instance timeouts
447 - Fixed bug in ``gnt-node list-storage``
448 - Fixed bug where not all daemons were started on cluster
449 initialisation, but only at the first watcher run
450 - Fixed many bugs in the OOB implementation
451 - Fixed watcher behaviour in presence of instances with offline
453 - Fixed instance list output for instances running on the wrong node
454 - a few fixes to the cluster-merge tool, but it still cannot merge
455 multi-node groups (currently it is not recommended to use this tool)
461 - Improved network configuration for the KVM hypervisor
462 - Added e1000 as a supported NIC for Xen-HVM
463 - Improved the lvmstrap tool to also be able to use partitions, as
464 opposed to full disks
465 - Improved speed of disk wiping (the cluster parameter
466 ``prealloc_wipe_disks``, so that it has a low impact on the total time
467 of instance creations
468 - Added documentation for the OS parameters
469 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
470 hypervisor is not responding
471 - Added display of blacklisted and hidden OS information in
473 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
474 NIC and node parameters, which might create problems with currently
475 invalid (but undetected) configuration files, but prevents validation
476 failures when unrelated parameters are modified
477 - Changed cluster initialisation to wait for the master daemon to become
479 - Expanded the RAPI interface:
481 - Added config redistribution resource
482 - Added activation/deactivation of instance disks
483 - Added export of console information
485 - Implemented log file reopening on SIGHUP, which allows using
486 logrotate(8) for the Ganeti log files
487 - Added a basic OOB helper script as an example
493 *(Released Fri, 14 Jan 2011)*
498 - Fixed timezone issues when formatting timestamps
499 - Added support for node groups, available via ``gnt-group`` and other
501 - Added out-of-band framework and management, see :doc:`design
502 document <design-oob>`
503 - Removed support for roman numbers from ``gnt-node list`` and
504 ``gnt-instance list``.
505 - Allowed modification of master network interface via ``gnt-cluster
506 modify --master-netdev``
507 - Accept offline secondaries while shutting down instance disks
508 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
509 - Added support for multiple LVM volume groups
510 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
512 - Added commands to list available fields:
514 - ``gnt-node list-fields``
515 - ``gnt-group list-fields``
516 - ``gnt-instance list-fields``
518 - Updated documentation and man pages
523 - Moved ``rapi_users`` file into separate directory, now named
524 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
526 - Added new tool for running commands on many machines,
527 ``tools/ganeti-listrunner``
528 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
529 improving the ``gnt-instance console`` output
530 - Allowed customisation of disk index separator at ``configure`` time
531 - Export node group allocation policy to :doc:`iallocator <iallocator>`
532 - Added support for non-partitioned md disks in ``lvmstrap``
533 - Added script to gracefully power off KVM instances
534 - Split ``utils`` module into smaller parts
535 - Changed query operations to return more detailed information, e.g.
536 whether an information is unavailable due to an offline node. To use
537 this new functionality, the LUXI call ``Query`` must be used. Field
538 information is now stored by the master daemon and can be retrieved
539 using ``QueryFields``. Instances, nodes and groups can also be queried
540 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
541 exposed via RAPI). The following commands make use of this
542 infrastructure change:
545 - ``gnt-group list-fields``
547 - ``gnt-node list-fields``
548 - ``gnt-instance list``
549 - ``gnt-instance list-fields``
550 - ``gnt-debug locks``
555 - New RAPI resources (see :doc:`rapi`):
559 - ``/2/groups/[group_name]``
560 - ``/2/groups/[group_name]/assign-nodes``
561 - ``/2/groups/[group_name]/modify``
562 - ``/2/groups/[group_name]/rename``
563 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
567 - Implemented ``no_install`` for instance creation
568 - Implemented OS parameters for instance reinstallation, allowing
569 use of special settings on reinstallation (e.g. for preserving data)
574 - Added IPv6 support in import/export
575 - Pause DRBD synchronization while wiping disks on instance creation
576 - Updated unittests and QA scripts
577 - Improved network parameters passed to KVM
578 - Converted man pages from docbook to reStructuredText
584 *(Released Mon, 20 Dec 2010)*
586 Released version 2.3.1~rc1 without any changes.
592 *(Released Wed, 1 Dec 2010)*
594 - impexpd: Disable OpenSSL compression in socat if possible (backport
595 from master, commit e90739d625b, see :doc:`installation guide
596 <install-quick>` for details)
597 - Changed unittest coverage report to exclude test scripts
598 - Added script to check version format
604 *(Released Wed, 1 Dec 2010)*
606 Released version 2.3.0~rc1 without any changes.
612 *(Released Fri, 19 Nov 2010)*
614 A number of bugfixes and documentation updates:
616 - Update ganeti-os-interface documentation
617 - Fixed a bug related to duplicate MACs or similar items which should be
619 - Fix breakage in OS state modify
620 - Reinstall instance: disallow offline secondaries (fixes bug related to
621 OS changing but reinstall failing)
622 - plus all the other fixes between 2.2.1 and 2.2.2
628 *(Released Tue, 2 Nov 2010)*
630 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
631 - Fixed master failover race with watcher
632 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
634 - Accept previously stopped instance for export with instance removal
635 - Simplify and extend the environment variables for instance OS scripts
636 - Added new node flags, ``master_capable`` and ``vm_capable``
637 - Added optional instance disk wiping prior during allocation. This is a
638 cluster-wide option and can be set/modified using
639 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
640 - Added IPv6 support, see :doc:`design document <design-2.3>` and
642 - Added a new watcher option (``--ignore-pause``)
643 - Added option to ignore offline node on instance start/stop
644 (``--ignore-offline``)
645 - Allow overriding OS parameters with ``gnt-instance reinstall``
646 - Added ability to change node's secondary IP address using ``gnt-node
648 - Implemented privilege separation for all daemons except
649 ``ganeti-noded``, see ``configure`` options
650 - Complain if an instance's disk is marked faulty in ``gnt-cluster
652 - Implemented job priorities (see ``ganeti(7)`` manpage)
653 - Ignore failures while shutting down instances during failover from
655 - Exit daemon's bootstrap process only once daemon is ready
656 - Export more information via ``LUInstanceQuery``/remote API
657 - Improved documentation, QA and unittests
658 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
659 restart if the file was created or changed
660 - Added LUXI protocol version sent with each request and response,
661 allowing detection of server/client mismatches
662 - Moved the Python scripts among gnt-* and ganeti-* into modules
663 - Moved all code related to setting up SSH to an external script,
665 - Infrastructure changes for node group support in future versions
671 *(Released Fri, 19 Nov 2010)*
673 A few small bugs fixed, and some improvements to the build system:
675 - Fix documentation regarding conversion to drbd
676 - Fix validation of parameters in cluster modify (``gnt-cluster modify
678 - Fix error handling in node modify with multiple changes
679 - Allow remote imports without checked names
685 *(Released Tue, 19 Oct 2010)*
687 - Disable SSL session ID cache in RPC client
693 *(Released Thu, 14 Oct 2010)*
695 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
696 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
698 - Fix problems with interaction between Curl and Python's HTTP server,
699 resulting in increased speed in many RPC calls
700 - Improve our release script to prevent breakage with older aclocal and
707 *(Released Thu, 7 Oct 2010)*
709 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
710 value retrieved from master
711 - Added support for blacklisted or hidden OS definitions
712 - Added simple lock monitor (accessible via (``gnt-debug locks``)
713 - Added support for -mem-path in KVM hypervisor abstraction layer
714 - Allow overriding instance parameters in tool for inter-cluster
715 instance moves (``tools/move-instance``)
716 - Improved opcode summaries (e.g. in ``gnt-job list``)
717 - Improve consistency of OS listing by sorting it
718 - Documentation updates
724 *(Released Fri, 8 Oct 2010)*
726 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
732 *(Released Mon, 4 Oct 2010)*
734 - Fixed regression in ``gnt-instance rename``
740 *(Released Wed, 22 Sep 2010)*
742 - Fixed OS_VARIANT variable for OS scripts
743 - Fixed cluster tag operations via RAPI
744 - Made ``setup-ssh`` exit with non-zero code if an error occurred
745 - Disabled RAPI CA checks in watcher
751 *(Released Mon, 23 Aug 2010)*
753 - Support DRBD versions of the format "a.b.c.d"
755 - Re-introduce support for usage from multiple threads in RAPI client
756 - Instance renames and modify via RAPI
757 - Work around race condition between processing and archival in job
759 - Mark opcodes following failed one as failed, too
760 - Job field ``lock_status`` was removed due to difficulties making it
761 work with the changed job queue in Ganeti 2.2; a better way to monitor
762 locks is expected for a later 2.2.x release
763 - Fixed dry-run behaviour with many commands
764 - Support ``ssh-agent`` again when adding nodes
765 - Many additional bugfixes
771 *(Released Fri, 30 Jul 2010)*
773 Important change: the internal RPC mechanism between Ganeti nodes has
774 changed from using a home-grown http library (based on the Python base
775 libraries) to use the PycURL library. This requires that PycURL is
776 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
777 against GnuTLS by default. cURL's support for GnuTLS had known issues
778 before cURL 7.21.0 and we recommend using the latest cURL release or
779 linking against OpenSSL. Most other distributions already link PycURL
780 and cURL against OpenSSL. The command::
782 python -c 'import pycurl; print pycurl.version'
784 can be used to determine the libraries PycURL and cURL are linked
787 Other significant changes:
789 - Rewrote much of the internals of the job queue, in order to achieve
790 better parallelism; this decouples job query operations from the job
791 processing, and it should allow much nicer behaviour of the master
792 daemon under load, and it also has uncovered some long-standing bugs
793 related to the job serialisation (now fixed)
794 - Added a default iallocator setting to the cluster parameters,
795 eliminating the need to always pass nodes or an iallocator for
796 operations that require selection of new node(s)
797 - Added experimental support for the LXC virtualization method
798 - Added support for OS parameters, which allows the installation of
799 instances to pass parameter to OS scripts in order to customise the
801 - Added a hypervisor parameter controlling the migration type (live or
802 non-live), since hypervisors have various levels of reliability; this
803 has renamed the 'live' parameter to 'mode'
804 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
805 logical volumes, meaning that cluster verify will ignore them and not
806 flag their presence as errors
807 - The watcher will now reset the error count for failed instances after
808 8 hours, thus allowing self-healing if the problem that caused the
809 instances to be down/fail to start has cleared in the meantime
810 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
811 check for, and warn, if the drbd module parameter ``usermode_helper``
812 is not consistent with the cluster-wide setting; this is needed to
813 make diagnose easier of failed drbd creations
814 - Started adding base IPv6 support, but this is not yet
815 enabled/available for use
816 - Rename operations (cluster, instance) will now return the new name,
817 which is especially useful if a short name was passed in
818 - Added support for instance migration in RAPI
819 - Added a tool to pre-configure nodes for the SSH setup, before joining
820 them to the cluster; this will allow in the future a simplified model
821 for node joining (but not yet fully enabled in 2.2); this needs the
822 paramiko python library
823 - Fixed handling of name-resolving errors
824 - Fixed consistency of job results on the error path
825 - Fixed master-failover race condition when executed multiple times in
827 - Fixed many bugs related to the job queue (mostly introduced during the
828 2.2 development cycle, so not all are impacting 2.1)
829 - Fixed instance migration with missing disk symlinks
830 - Fixed handling of unknown jobs in ``gnt-job archive``
831 - And many other small fixes/improvements
835 - Enhanced both the unittest and the QA coverage
836 - Switched the opcode validation to a generic model, and extended the
837 validation to all opcode parameters
838 - Changed more parts of the code that write shell scripts to use the
840 - Switched the master daemon to use the asyncore library for the Luxi
847 *(Released Thu, 17 Jun 2010)*
849 - Added tool (``move-instance``) and infrastructure to move instances
850 between separate clusters (see :doc:`separate documentation
851 <move-instance>` and :doc:`design document <design-2.2>`)
852 - Added per-request RPC timeout
853 - RAPI now requires a Content-Type header for requests with a body (e.g.
854 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
855 :rfc:`2616` (HTTP/1.1), section 7.2.1)
856 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
858 - Implemented initial support for running Ganeti daemons as separate
859 users, see configure-time flags ``--with-user-prefix`` and
860 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
862 - Instances can be removed after export (``gnt-backup export
864 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
865 key (instead of 1024 bit)
866 - Added new cluster configuration file for cluster domain secret
867 - Import/export now use SSL instead of SSH
868 - Added support for showing estimated time when exporting an instance,
869 see the ``ganeti-os-interface(7)`` manpage and look for
876 *(Released Tue, 16 Nov 2010)*
878 Some more bugfixes. Unless critical bugs occur, this will be the last
881 - Fix case of MAC special-values
882 - Fix mac checker regex
883 - backend: Fix typo causing "out of range" error
884 - Add missing --units in gnt-instance list man page
890 *(Released Tue, 24 Aug 2010)*
893 - Don't ignore secondary node silently on non-mirrored disk templates
895 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
896 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
897 - Properly document the usb_mouse parameter
898 - Fix path in ganeti-rapi(8) (issue 116)
899 - Adjust error message when the ganeti user's .ssh directory is
901 - Add same-node-check when changing the disk template to drbd
907 *(Released Fri, 16 Jul 2010)*
910 - Add an option to only select some reboot types during qa/burnin.
911 (on some hypervisors consequent reboots are not supported)
912 - Fix infrequent race condition in master failover. Sometimes the old
913 master ip address would be still detected as up for a short time
914 after it was removed, causing failover to fail.
915 - Decrease mlockall warnings when the ctypes module is missing. On
916 Python 2.4 we support running even if no ctypes module is installed,
917 but we were too verbose about this issue.
918 - Fix building on old distributions, on which man doesn't have a
920 - Fix RAPI not to ignore the MAC address on instance creation
921 - Implement the old instance creation format in the RAPI client.
927 *(Released Thu, 01 Jul 2010)*
929 A small bugfix release:
930 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
931 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
932 a non-existing option
933 - Fix instance create: the --force-variant option was ignored
934 - Improve pylint 0.21 compatibility and warnings with Python 2.6
935 - Fix modify node storage with non-FQDN arguments
936 - Fix RAPI client to authenticate under Python 2.6 when used
937 for more than 5 requests needing authentication
938 - Fix gnt-instance modify -t (storage) giving a wrong error message
939 when converting a non-shutdown drbd instance to plain
945 *(Released Fri, 18 Jun 2010)*
947 A small bugfix release:
949 - Fix live migration of KVM instances started with older Ganeti
950 versions which had fewer hypervisor parameters
951 - Fix gnt-instance grow-disk on down instances
952 - Fix an error-reporting bug during instance migration
953 - Better checking of the ``--net`` and ``--disk`` values, to avoid
954 silently ignoring broken ones
955 - Fix an RPC error reporting bug affecting, for example, RAPI client
957 - Fix bug triggered by different API version os-es on different nodes
958 - Fix a bug in instance startup with custom hvparams: OS level
959 parameters would fail to be applied.
960 - Fix the RAPI client under Python 2.6 (but more work is needed to
961 make it work completely well with OpenSSL)
962 - Fix handling of errors when resolving names from DNS
968 *(Released Thu, 3 Jun 2010)*
970 A medium sized development cycle. Some new features, and some
971 fixes/small improvements/cleanups.
976 The node deamon now tries to mlock itself into memory, unless the
977 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
978 its logs, and falls back to console logging. This allows emergency
979 features such as ``gnt-node powercycle`` to work even in the event of a
980 broken node disk (tested offlining the disk hosting the node's
981 filesystem and dropping its memory caches; don't try this at home)
983 KVM: add vhost-net acceleration support. It can be tested with a new
984 enough version of the kernel and of qemu-kvm.
986 KVM: Add instance chrooting feature. If you use privilege dropping for
987 your VMs you can also now force them to chroot to an empty directory,
988 before starting the emulated guest.
990 KVM: Add maximum migration bandwith and maximum downtime tweaking
991 support (requires a new-enough version of qemu-kvm).
993 Cluster verify will now warn if the master node doesn't have the master
996 Add a new (incompatible) instance creation request format to RAPI which
997 supports all parameters (previously only a subset was supported, and it
998 wasn't possible to extend the old format to accomodate all the new
999 features. The old format is still supported, and a client can check for
1000 this feature, before using it, by checking for its presence in the
1001 ``features`` RAPI resource.
1003 Now with ancient latin support. Try it passing the ``--roman`` option to
1004 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1005 (requires the python-roman module to be installed, in order to work).
1010 As usual many internal code refactorings, documentation updates, and
1013 - Lots of improvements and cleanups to the experimental Remote API
1014 (RAPI) client library.
1015 - A new unit test suite for the core daemon libraries.
1016 - A fix to creating missing directories makes sure the umask is not
1017 applied anymore. This enforces the same directory permissions
1019 - Better handling terminating daemons with ctrl+c (used when running
1020 them in debugging mode).
1021 - Fix a race condition in live migrating a KVM instance, when stat()
1022 on the old proc status file returned EINVAL, which is an unexpected
1024 - Fixed manpage checking with newer man and utf-8 charachters. But now
1025 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1031 *(Released Fri, 7 May 2010)*
1033 Fix a bug which prevented untagged KVM instances from starting.
1039 *(Released Fri, 7 May 2010)*
1041 Another release with a long development cycle, during which many
1042 different features were added.
1044 Significant features
1045 ~~~~~~~~~~~~~~~~~~~~
1047 The KVM hypervisor now can run the individual instances as non-root, to
1048 reduce the impact of a VM being hijacked due to bugs in the
1049 hypervisor. It is possible to run all instances as a single (non-root)
1050 user, to manually specify a user for each instance, or to dynamically
1051 allocate a user out of a cluster-wide pool to each instance, with the
1052 guarantee that no two instances will run under the same user ID on any
1055 An experimental RAPI client library, that can be used standalone
1056 (without the other Ganeti libraries), is provided in the source tree as
1057 ``lib/rapi/client.py``. Note this client might change its interface in
1058 the future, as we iterate on its capabilities.
1060 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1061 replace the cluster's certificates and crypto keys. This might help in
1062 case they have been compromised, or have simply expired.
1064 A new disk option for instance creation has been added that allows one
1065 to "adopt" currently existing logical volumes, with data
1066 preservation. This should allow easier migration to Ganeti from
1067 unmanaged (or managed via other software) instances.
1069 Another disk improvement is the possibility to convert between redundant
1070 (DRBD) and plain (LVM) disk configuration for an instance. This should
1071 allow better scalability (starting with one node and growing the
1072 cluster, or shrinking a two-node cluster to one node).
1074 A new feature that could help with automated node failovers has been
1075 implemented: if a node sees itself as offline (by querying the master
1076 candidates), it will try to shutdown (hard) all instances and any active
1077 DRBD devices. This reduces the risk of duplicate instances if an
1078 external script automatically failovers the instances on such nodes. To
1079 enable this, the cluster parameter ``maintain_node_health`` should be
1080 enabled; in the future this option (per the name) will enable other
1081 automatic maintenance features.
1083 Instance export/import now will reuse the original instance
1084 specifications for all parameters; that means exporting an instance,
1085 deleting it and the importing it back should give an almost identical
1086 instance. Note that the default import behaviour has changed from
1087 before, where it created only one NIC; now it recreates the original
1090 Cluster verify has added a few new checks: SSL certificates validity,
1091 /etc/hosts consistency across the cluster, etc.
1096 As usual, many internal changes were done, documentation fixes,
1099 - Fixed cluster initialization with disabled cluster storage (regression
1100 introduced in 2.1.1)
1101 - File-based storage supports growing the disks
1102 - Fixed behaviour of node role changes
1103 - Fixed cluster verify for some corner cases, plus a general rewrite of
1104 cluster verify to allow future extension with more checks
1105 - Fixed log spamming by watcher and node daemon (regression introduced
1107 - Fixed possible validation issues when changing the list of enabled
1109 - Fixed cleanup of /etc/hosts during node removal
1110 - Fixed RAPI response for invalid methods
1111 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1112 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1113 from ide disks, etc.)
1114 - Allow OS changes without re-installation (to record a changed OS
1115 outside of Ganeti, or to allow OS renames)
1116 - Allow instance creation without OS installation (useful for example if
1117 the OS will be installed manually, or restored from a backup not in
1119 - Implemented option to make cluster ``copyfile`` use the replication
1121 - Added list of enabled hypervisors to ssconf (possibly useful for
1123 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1125 - A partial form of node re-IP is possible via node readd, which now
1126 allows changed node primary IP
1127 - Command line utilities now show an informational message if the job is
1129 - The logs of the master daemon now show the PID/UID/GID of the
1136 *(Released Fri, 12 Mar 2010)*
1138 During the 2.1.0 long release candidate cycle, a lot of improvements and
1139 changes have accumulated with were released later as 2.1.1.
1144 The node evacuate command (``gnt-node evacuate``) was significantly
1145 rewritten, and as such the IAllocator protocol was changed - a new
1146 request type has been added. This unfortunate change during a stable
1147 series is designed to improve performance of node evacuations; on
1148 clusters with more than about five nodes and which are well-balanced,
1149 evacuation should proceed in parallel for all instances of the node
1150 being evacuated. As such, any existing IAllocator scripts need to be
1151 updated, otherwise the above command will fail due to the unknown
1152 request. The provided "dumb" allocator has not been updated; but the
1153 ganeti-htools package supports the new protocol since version 0.2.4.
1155 Another important change is increased validation of node and instance
1156 names. This might create problems in special cases, if invalid host
1157 names are being used.
1159 Also, a new layer of hypervisor parameters has been added, that sits at
1160 OS level between the cluster defaults and the instance ones. This allows
1161 customisation of virtualization parameters depending on the installed
1162 OS. For example instances with OS 'X' may have a different KVM kernel
1163 (or any other parameter) than the cluster defaults. This is intended to
1164 help managing a multiple OSes on the same cluster, without manual
1165 modification of each instance's parameters.
1167 A tool for merging clusters, ``cluster-merge``, has been added in the
1168 tools sub-directory.
1173 - Improved the int/float conversions that should make the code more
1174 robust in face of errors from the node daemons
1175 - Fixed the remove node code in case of internal configuration errors
1176 - Fixed the node daemon behaviour in face of inconsistent queue
1177 directory (e.g. read-only file-system where we can't open the files
1179 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1180 now it either aborts cleanly or, if given the new "auto_promote"
1181 parameter, will automatically promote other nodes as needed
1182 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1183 completely prevent Ganeti from working
1184 - Fixed bug for instance export when not all disks were successfully
1186 - Fixed behaviour of node add when the new node is slow in starting up
1188 - Fixed handling of signals in the LUXI client, which should improve
1189 behaviour of command-line scripts
1190 - Added checks for invalid node/instance names in the configuration (now
1191 flagged during cluster verify)
1192 - Fixed watcher behaviour for disk activation errors
1193 - Fixed two potentially endless loops in http library, which led to the
1194 RAPI daemon hanging and consuming 100% CPU in some cases
1195 - Fixed bug in RAPI daemon related to hashed passwords
1196 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1198 - Enhanced compatibility with non-Debian OSes, but not using absolute
1199 path in some commands and allowing customisation of the ssh
1200 configuration directory
1201 - Fixed possible future issue with new Python versions by abiding to the
1202 proper use of ``__slots__`` attribute on classes
1203 - Added checks that should prevent directory traversal attacks
1204 - Many documentation fixes based on feedback from users
1209 - Added an "early_release" more for instance replace disks and node
1210 evacuate, where we release locks earlier and thus allow higher
1211 parallelism within the cluster
1212 - Added watcher hooks, intended to allow the watcher to restart other
1213 daemons (e.g. from the ganeti-nbma project), but they can be used of
1214 course for any other purpose
1215 - Added a compile-time disable for DRBD barriers, to increase
1216 performance if the administrator trusts the power supply or the
1217 storage system to not lose writes
1218 - Added the option of using syslog for logging instead of, or in
1219 addition to, Ganeti's own log files
1220 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1221 can indeed boot from a paravirtual NIC
1222 - Added a generic debug level for many operations; while this is not
1223 used widely yet, it allows one to pass the debug value all the way to
1225 - Enhanced the hooks environment for instance moves (failovers,
1226 migrations) where the primary/secondary nodes changed during the
1227 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1228 - Enhanced data validations for many user-supplied values; one important
1229 item is the restrictions imposed on instance and node names, which
1230 might reject some (invalid) host names
1231 - Add a configure-time option to disable file-based storage, if it's not
1232 needed; this allows greater security separation between the master
1233 node and the other nodes from the point of view of the inter-node RPC
1235 - Added user notification in interactive tools if job is waiting in the
1236 job queue or trying to acquire locks
1237 - Added log messages when a job is waiting for locks
1238 - Added filtering by node tags in instance operations which admit
1239 multiple instances (start, stop, reboot, reinstall)
1240 - Added a new tool for cluster mergers, ``cluster-merge``
1241 - Parameters from command line which are of the form ``a=b,c=d`` can now
1242 use backslash escapes to pass in values which contain commas,
1243 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1245 - For KVM, the instance name is the first parameter passed to KVM, so
1246 that it's more visible in the process list
1252 *(Released Tue, 2 Mar 2010)*
1254 Ganeti 2.1 brings many improvements with it. Major changes:
1256 - Added infrastructure to ease automated disk repairs
1257 - Added new daemon to export configuration data in a cheaper way than
1258 using the remote API
1259 - Instance NICs can now be routed instead of being associated with a
1261 - Improved job locking logic to reduce impact of jobs acquiring multiple
1262 locks waiting for other long-running jobs
1264 In-depth implementation details can be found in the Ganeti 2.1 design
1270 - Added chroot hypervisor
1271 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1273 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1274 ``bootloader_path`` and ``bootloader_args``)
1275 - Added the ``use_localtime`` option for the xen-hvm and kvm
1276 hypervisors, and the default value for this has changed to false (in
1277 2.0 xen-hvm always enabled it)
1278 - Added luxi call to submit multiple jobs in one go
1279 - Added cluster initialization option to not modify ``/etc/hosts``
1281 - Added network interface parameters
1282 - Added dry run mode to some LUs
1283 - Added RAPI resources:
1285 - ``/2/instances/[instance_name]/info``
1286 - ``/2/instances/[instance_name]/replace-disks``
1287 - ``/2/nodes/[node_name]/evacuate``
1288 - ``/2/nodes/[node_name]/migrate``
1289 - ``/2/nodes/[node_name]/role``
1290 - ``/2/nodes/[node_name]/storage``
1291 - ``/2/nodes/[node_name]/storage/modify``
1292 - ``/2/nodes/[node_name]/storage/repair``
1294 - Added OpCodes to evacuate or migrate all instances on a node
1295 - Added new command to list storage elements on nodes (``gnt-node
1296 list-storage``) and modify them (``gnt-node modify-storage``)
1297 - Added new ssconf files with master candidate IP address
1298 (``ssconf_master_candidates_ips``), node primary IP address
1299 (``ssconf_node_primary_ips``) and node secondary IP address
1300 (``ssconf_node_secondary_ips``)
1301 - Added ``ganeti-confd`` and a client library to query the Ganeti
1302 configuration via UDP
1303 - Added ability to run hooks after cluster initialization and before
1305 - Added automatic mode for disk replace (``gnt-instance replace-disks
1307 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1308 after catastrophic data-loss
1309 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1311 - Added ``gnt-instance move`` command to move instances
1312 - Added ``gnt-cluster watcher`` command to control watcher
1313 - Added ``gnt-node powercycle`` command to powercycle nodes
1314 - Added new job status field ``lock_status``
1315 - Added parseable error codes to cluster verification (``gnt-cluster
1316 verify --error-codes``) and made output less verbose (use
1317 ``--verbose`` to restore previous behaviour)
1318 - Added UUIDs to the main config entities (cluster, nodes, instances)
1319 - Added support for OS variants
1320 - Added support for hashed passwords in the Ganeti remote API users file
1322 - Added option to specify maximum timeout on instance shutdown
1323 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1324 - Added new helper script to start and stop Ganeti daemons
1325 (``daemon-util``), with the intent to reduce the work necessary to
1326 adjust Ganeti for non-Debian distributions and to start/stop daemons
1328 - Added more unittests
1329 - Fixed critical bug in ganeti-masterd startup
1330 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1331 now customisable at the cluster level for both the KVM and Xen
1332 hypervisors using the new ``migration_port`` parameter
1333 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1334 reinstalling an instance
1335 - Allowed ``@`` in tag names
1336 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1337 - Many documentation updates
1338 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1339 - ``gnt-instance reinstall`` can now reinstall multiple instances
1340 - Updated many command line parameters
1341 - Introduced new OS API version 15
1342 - No longer support a default hypervisor
1343 - Treat virtual LVs as inexistent
1344 - Improved job locking logic to reduce lock contention
1345 - Match instance and node names case insensitively
1346 - Reimplemented bash completion script to be more complete
1353 *(Released Thu, 4 Feb 2010)*
1355 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1356 - Fix a string formatting bug
1357 - Improve safety of the code in some error paths
1358 - Improve data validation in the master of values returned from nodes
1364 *(Released Thu, 17 Dec 2009)*
1366 - Fix security issue due to missing validation of iallocator names; this
1367 allows local and remote execution of arbitrary executables
1368 - Fix failure of gnt-node list during instance removal
1369 - Ship the RAPI documentation in the archive
1375 *(Released Wed, 30 Sep 2009)*
1377 - Fixed many wrong messages
1378 - Fixed a few bugs related to the locking library
1379 - Fixed MAC checking at instance creation time
1380 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1381 - Fixed a few issues related to signal handling in both daemons and
1383 - Fixed the example startup script provided
1384 - Fixed insserv dependencies in the example startup script (patch from
1386 - Fixed handling of drained nodes in the iallocator framework
1387 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1389 - Fixed error related to invalid job IDs in job polling
1390 - Fixed job/opcode persistence on unclean master shutdown
1391 - Fixed handling of partial job processing after unclean master
1393 - Fixed error reporting from LUs, previously all errors were converted
1394 into execution errors
1395 - Fixed error reporting from burnin
1396 - Decreased significantly the memory usage of the job queue
1397 - Optimised slightly multi-job submission
1398 - Optimised slightly opcode loading
1399 - Backported the multi-job submit framework from the development
1400 branch; multi-instance start and stop should be faster
1401 - Added script to clean archived jobs after 21 days; this will reduce
1402 the size of the queue directory
1403 - Added some extra checks in disk size tracking
1404 - Added an example ethers hook script
1405 - Added a cluster parameter that prevents Ganeti from modifying of
1407 - Added more node information to RAPI responses
1408 - Added a ``gnt-job watch`` command that allows following the ouput of a
1410 - Added a bind-address option to ganeti-rapi
1411 - Added more checks to the configuration verify
1412 - Enhanced the burnin script such that some operations can be retried
1414 - Converted instance reinstall to multi-instance model
1420 *(Released Fri, 7 Aug 2009)*
1422 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1423 to allow using the pre-2.0.2 behaviour in activation, if any existing
1424 instances have mismatched disk sizes in the configuration
1425 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
1426 any configuration mismatches for disk sizes
1427 - Added ``gnt-master cluste-failover --no-voting`` to allow master
1428 failover to work on two-node clusters
1429 - Fixed the ``--net`` option of ``gnt-backup import``, which was
1431 - Fixed detection of OS script errors in ``gnt-backup export``
1432 - Fixed exit code of ``gnt-backup export``
1438 *(Released Fri, 17 Jul 2009)*
1440 - Added experimental support for stripped logical volumes; this should
1441 enhance performance but comes with a higher complexity in the block
1442 device handling; stripping is only enabled when passing
1443 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
1444 affected even in the non-stripped mode
1445 - Improved resiliency against transient failures at the end of DRBD
1446 resyncs, and in general of DRBD resync checks
1447 - Fixed a couple of issues with exports and snapshot errors
1448 - Fixed a couple of issues in instance listing
1449 - Added display of the disk size in ``gnt-instance info``
1450 - Fixed checking for valid OSes in instance creation
1451 - Fixed handling of the "vcpus" parameter in instance listing and in
1452 general of invalid parameters
1453 - Fixed http server library, and thus RAPI, to handle invalid
1454 username/password combinations correctly; this means that now they
1455 report unauthorized for queries too, not only for modifications,
1456 allowing earlier detect of configuration problems
1457 - Added a new "role" node list field, equivalent to the master/master
1458 candidate/drained/offline flags combinations
1459 - Fixed cluster modify and changes of candidate pool size
1460 - Fixed cluster verify error messages for wrong files on regular nodes
1461 - Fixed a couple of issues with node demotion from master candidate role
1462 - Fixed node readd issues
1463 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
1464 - Added a new ``--no-voting`` option for masterfailover to fix failover
1465 on two-nodes clusters when the former master node is unreachable
1466 - Added instance reinstall over RAPI
1472 *(Released Tue, 16 Jun 2009)*
1474 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
1475 allow re-adding the start in single-user option (regression from 1.2)
1476 - the watcher writes the instance status to a file, to allow monitoring
1477 to report the instance status (from the master) based on cached
1478 results of the watcher's queries; while this can get stale if the
1479 watcher is being locked due to other work on the cluster, this is
1480 still an improvement
1481 - the watcher now also restarts the node daemon and the rapi daemon if
1483 - fixed the watcher to handle full and drained queue cases
1484 - hooks export more instance data in the environment, which helps if
1485 hook scripts need to take action based on the instance's properties
1486 (no longer need to query back into ganeti)
1487 - instance failovers when the instance is stopped do not check for free
1488 RAM, so that failing over a stopped instance is possible in low memory
1490 - rapi uses queries for tags instead of jobs (for less job traffic), and
1491 for cluster tags it won't talk to masterd at all but read them from
1493 - a couple of error handling fixes in RAPI
1494 - drbd handling: improved the error handling of inconsistent disks after
1495 resync to reduce the frequency of "there are some degraded disks for
1496 this instance" messages
1497 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
1498 error handling path called a wrong function name)
1504 *(Released Wed, 27 May 2009)*
1506 - no changes from rc5
1509 Version 2.0 release candidate 5
1510 -------------------------------
1512 *(Released Wed, 20 May 2009)*
1514 - fix a couple of bugs (validation, argument checks)
1515 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
1516 - some small improvements to RAPI and IAllocator
1517 - make watcher automatically start the master daemon if down
1520 Version 2.0 release candidate 4
1521 -------------------------------
1523 *(Released Mon, 27 Apr 2009)*
1525 - change the OS list to not require locks; this helps with big clusters
1526 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
1527 volume group is broken
1528 - ``gnt-instance info``, without any arguments, doesn't run for all
1529 instances anymore; either pass ``--all`` or pass the desired
1530 instances; this helps against mistakes on big clusters where listing
1531 the information for all instances takes a long time
1532 - miscellaneous doc and man pages fixes
1535 Version 2.0 release candidate 3
1536 -------------------------------
1538 *(Released Wed, 8 Apr 2009)*
1540 - Change the internal locking model of some ``gnt-node`` commands, in
1541 order to reduce contention (and blocking of master daemon) when
1542 batching many creation/reinstall jobs
1543 - Fixes to Xen soft reboot
1544 - No longer build documentation at build time, instead distribute it in
1545 the archive, in order to reduce the need for the whole docbook/rst
1549 Version 2.0 release candidate 2
1550 -------------------------------
1552 *(Released Fri, 27 Mar 2009)*
1554 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
1555 - Fix watcher startup sequence, improves the behaviour of busy clusters
1556 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
1557 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
1558 bind address and other places
1559 - Some documentation fixes and updates
1562 Version 2.0 release candidate 1
1563 -------------------------------
1565 *(Released Mon, 2 Mar 2009)*
1567 - More documentation updates, now all docs should be more-or-less
1569 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
1571 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
1573 - Fix an issue related to $libdir/run/ganeti and cluster creation
1579 *(Released Thu, 19 Feb 2009)*
1581 - Xen PVM and KVM have switched the default value for the instance root
1582 disk to the first partition on the first drive, instead of the whole
1583 drive; this means that the OS installation scripts must be changed
1585 - Man pages have been updated
1586 - RAPI has been switched by default to HTTPS, and the exported functions
1587 should all work correctly
1588 - RAPI v1 has been removed
1589 - Many improvements to the KVM hypervisor
1590 - Block device errors are now better reported
1591 - Many other bugfixes and small improvements
1597 *(Released Mon, 26 Jan 2009)*
1599 - Version 2 is a general rewrite of the code and therefore the
1600 differences are too many to list, see the design document for 2.0 in
1601 the ``doc/`` subdirectory for more details
1602 - In this beta version there is not yet a migration path from 1.2 (there
1603 will be one in the final 2.0 release)
1604 - A few significant changes are:
1606 - all commands are executed by a daemon (``ganeti-masterd``) and the
1607 various ``gnt-*`` commands are just front-ends to it
1608 - all the commands are entered into, and executed from a job queue,
1609 see the ``gnt-job(8)`` manpage
1610 - the RAPI daemon supports read-write operations, secured by basic
1611 HTTP authentication on top of HTTPS
1612 - DRBD version 0.7 support has been removed, DRBD 8 is the only
1613 supported version (when migrating from Ganeti 1.2 to 2.0, you need
1614 to migrate to DRBD 8 first while still running Ganeti 1.2)
1615 - DRBD devices are using statically allocated minor numbers, which
1616 will be assigned to existing instances during the migration process
1617 - there is support for both Xen PVM and Xen HVM instances running on
1619 - KVM virtualization is supported too
1620 - file-based storage has been implemented, which means that it is
1621 possible to run the cluster without LVM and DRBD storage, for
1622 example using a shared filesystem exported from shared storage (and
1623 still have live migration)
1629 *(Released Tue, 13 Jan 2009)*
1631 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
1632 - Reuse the old instance mac address by default on instance import, if
1633 the instance name is the same.
1634 - Handle situations in which the node info rpc returns incomplete
1636 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
1637 - Improved version of batcher:
1639 - state file support
1640 - instance mac address support
1641 - support for HVM clusters/instances
1643 - Add an option to show the number of cpu sockets and nodes in
1645 - Support OSes that handle more than one version of the OS api (but do
1646 not change the current API in any other way)
1647 - Fix ``gnt-node migrate``
1648 - ``gnt-debug`` man page
1649 - Fixes various more typos and small issues
1650 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
1656 *(Released Wed, 24 Sep 2008)*
1658 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
1659 type of disk exported to fully virtualized instances.
1660 - provide access to the serial console of HVM instances
1661 - instance auto_balance flag, set by default. If turned off it will
1662 avoid warnings on cluster verify if there is not enough memory to fail
1663 over an instance. in the future it will prevent automatically failing
1664 it over when we will support that.
1665 - batcher tool for instance creation, see ``tools/README.batcher``
1666 - ``gnt-instance reinstall --select-os`` to interactively select a new
1667 operating system when reinstalling an instance.
1668 - when changing the memory amount on instance modify a check has been
1669 added that the instance will be able to start. also warnings are
1670 emitted if the instance will not be able to fail over, if auto_balance
1672 - documentation fixes
1673 - sync fields between ``gnt-instance list/modify/add/import``
1674 - fix a race condition in drbd when the sync speed was set after giving
1675 the device a remote peer.
1681 *(Released Tue, 22 Jul 2008)*
1683 - note: the allowed size and number of tags per object were reduced
1684 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1685 - fixed twisted 8.x compatibility
1686 - fixed ``gnt-instance replace-disks`` with iallocator
1687 - add TCP keepalives on twisted connections to detect restarted nodes
1688 - disk increase support, see ``gnt-instance grow-disk``
1689 - implement bulk node/instance query for RAPI
1690 - add tags in node/instance listing (optional)
1691 - experimental migration (and live migration) support, read the man page
1692 for ``gnt-instance migrate``
1693 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1694 has some small improvements in handling its state file
1700 *(Released Fri, 13 Jun 2008)*
1702 - Experimental readonly, REST-based remote API implementation;
1703 automatically started on master node, TCP port 5080, if enabled by
1704 ``--enable-rapi`` parameter to configure script.
1705 - Instance allocator support. Add and import instance accept a
1706 ``--iallocator`` parameter, and call that instance allocator to decide
1707 which node to use for the instance. The iallocator document describes
1708 what's expected from an allocator script.
1709 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1710 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1711 node is lost there is still enough memory to fail over the instances
1713 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1714 ``gnt-cluster verify``, to check for site-specific compliance. All the
1715 hooks will run, and their output, if any, will be displayed. Any
1716 failing hook will make the verification return an error value.
1717 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1718 primary and secondary interfaces
1719 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1720 that is still declared as part of the cluster and has failed.
1721 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1722 specifying output fields, that just adds the chosen fields to the
1724 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1725 export from the filesystem.
1726 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1727 have been added. Using them you can enable/disable acpi and pae
1728 support, and specify a path for a cd image to be exported to the
1729 instance. These parameters as the name suggest only work on HVM
1731 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1732 PAE support will be set to the previously hardcoded values, but the
1733 (previously hardcoded) path to the CDROM ISO image will be unset and
1734 if required, needs to be set manually with ``gnt-instance modify``
1736 - The address to which an instance's VNC console is bound is now
1737 selectable per-instance, rather than being cluster wide. Of course
1738 this only applies to instances controlled via VNC, so currently just
1739 applies to HVM clusters.
1745 *(Released Mon, 18 Feb 2008)*
1747 - more tweaks to the disk activation code (especially helpful for DRBD)
1748 - change the default ``gnt-instance list`` output format, now there is
1749 one combined status field (see the manpage for the exact values this
1751 - some more fixes for the mac export to hooks change
1752 - make Ganeti not break with DRBD 8.2.x (which changed the version
1753 format in ``/proc/drbd``) (issue 24)
1754 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1755 template, allowing migration from DRBD0.7+MD to DRBD8
1761 *(Released Wed, 30 Jan 2008)*
1763 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1765 - add command aliases infrastructure and a few aliases
1766 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1767 man pages and the ``--help`` option of ``gnt-node
1768 list``/``gnt-instance list``
1769 - fix ``gnt-backup list`` with down nodes (issue 21)
1770 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1771 - fix the dist archive and add a check for including svn/git files in
1773 - some developer-related changes: improve the burnin and the QA suite,
1774 add an upload script for testing during development
1780 *(Released Wed, 16 Jan 2008)*
1782 - experimental HVM support, read the install document, section
1783 "Initializing the cluster"
1784 - allow for the PVM hypervisor per-instance kernel and initrd paths
1785 - add a new command ``gnt-cluster verify-disks`` which uses a new
1786 algorithm to improve the reconnection of the DRBD pairs if the device
1787 on the secondary node has gone away
1788 - make logical volume code auto-activate LVs at disk activation time
1789 - slightly improve the speed of activating disks
1790 - allow specification of the MAC address at instance creation time, and
1791 changing it later via ``gnt-instance modify``
1792 - fix handling of external commands that generate lots of output on
1794 - update documentation with regard to minimum version of DRBD8 supported
1800 *(Released Tue, 4 Dec 2007)*
1802 - Log the ``xm create`` output to the node daemon log on failure (to
1803 help diagnosing the error)
1804 - In debug mode, log all external commands output if failed to the logs
1805 - Change parsing of lvm commands to ignore stderr
1811 *(Released Wed, 28 Nov 2007)*
1813 - Another round of updates to the DRBD 8 code to deal with more failures
1814 in the replace secondary node operation
1815 - Some more logging of failures in disk operations (lvm, drbd)
1816 - A few documentation updates
1823 *(Released Tue, 13 Nov 2007)*
1825 - Change configuration file format from Python's Pickle to JSON.
1826 Upgrading is possible using the cfgupgrade utility.
1827 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1828 faster replace disks and is more stable (DRBD 8 has many improvements
1829 compared to DRBD 0.7)
1830 - Added command line tags support (see man pages for ``gnt-instance``,
1831 ``gnt-node``, ``gnt-cluster``)
1832 - Added instance rename support
1833 - Added multi-instance startup/shutdown
1834 - Added cluster rename support
1835 - Added ``gnt-node evacuate`` to simplify some node operations
1836 - Added instance reboot operation that can speedup reboot as compared to
1838 - Soften the requirement that hostnames are in FQDN format
1839 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1841 - Removed dependency on debian's patched fping that uses the
1842 non-standard ``-S`` option
1843 - Now the OS definitions are searched for in multiple, configurable
1844 paths (easier for distros to package)
1845 - Some changes to the hooks infrastructure (especially the new
1846 post-configuration update hook)
1847 - Other small bugfixes
1849 .. vim: set textwidth=72 syntax=rst :