10 - fix bug, that kept nodes offline when readding
11 - when verifying DRBD versions, ignore unavailable nodes
12 - DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
18 *(Released Tue, 5 Nov 2013)*
20 Incompatible/important changes
21 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
23 - hroller now also plans for capacity to move non-redundant instances off
24 any node to be rebooted; the old behavior of completely ignoring any
25 non-redundant instances can be restored by adding the --ignore-non-redundant
27 - The cluster option '--no-lvm-storage' was removed in favor of the new option
28 '--enabled-disk-templates'.
29 - On instance creation, disk templates no longer need to be specified
30 with '-t'. The default disk template will be taken from the list of
31 enabled disk templates.
32 - The monitoring daemon is now running as root, in order to be able to collect
33 information only available to root (such as the state of Xen instances).
34 - The ConfD client is now IPv6 compatible.
35 - File and shared file storage is no longer dis/enabled at configure time,
36 but using the option '--enabled-disk-templates' at cluster initialization and
38 - The default directories for file and shared file storage are not anymore
39 specified at configure time, but taken from the cluster's configuration.
40 They can be set at cluster initialization and modification with
41 '--file-storage-dir' and '--shared-file-storage-dir'.
42 - Cluster verification now includes stricter checks regarding the
43 default file and shared file storage directories. It now checks that
44 the directories are explicitely allowed in the 'file-storage-paths' file and
45 that the directories exist on all nodes.
46 - The list of allowed disk templates in the instance policy and the list
47 of cluster-wide enabled disk templates is now checked for consistency
48 on cluster or group modification. On cluster initialization, the ipolicy
49 disk templates are ensured to be a subset of the cluster-wide enabled
55 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
56 the correct command syntax. It is possible to use different DRBD versions
57 on different nodes as long as they are compatible to each other. This
58 enables rolling upgrades of DRBD with no downtime. As permanent operation
59 of different DRBD versions within a node group is discouraged,
60 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
61 - New "inst-status-xen" data collector for the monitoring daemon, providing
62 information about the state of the xen instances on the nodes.
63 - New "lv" data collector for the monitoring daemon, collecting data about the
64 logical volumes on the nodes, and pairing them with the name of the instances
66 - New "diskstats" data collector, collecting the data from /proc/diskstats and
67 presenting them over the monitoring daemon interface.
68 - The ConfD client is now IPv6 compatible.
72 The following new dependencies have been added.
76 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
77 for the unit tests (and only used for testing).
81 - ``hslogger`` (http://software.complete.org/hslogger) is now always
82 required, even if confd is not enabled.
87 - Correctly start/stop luxid during gnt-cluster master-failover (inherited
89 - Improved error messsages (inherited from stable-2.8)
95 *(Released Tue, 15 Oct 2013)*
97 The third release candidate in the 2.9 series. Since 2.9.0 rc2:
99 - in implicit configuration upgrade, match ipolicy with enabled disk templates
100 - improved harep documentation (inherited from stable-2.8)
106 *(Released Wed, 9 Oct 2013)*
108 The second release candidate in the 2.9 series. Since 2.9.0 rc1:
110 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
111 at least one DRBD instance.
112 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
119 *(Released Tue, 1 Oct 2013)*
121 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
124 - update of the documentation, in particular installation instructions
125 - merging of LD_* constants into DT_* constants
126 - python style changes to be compatible with newer versions of pylint
132 *(Released Thu, 29 Aug 2013)*
134 This was the first beta release of the 2.9 series. All important changes
135 are listed in the latest 2.9 entry.
141 *(Released Thu, 07 Nov 2013)*
143 - DRBD: ensure peers are UpToDate for dual-primary
144 - Improve error message for replace-disks
145 - More dependency checks at configure time
146 - Placate warnings on ganeti.outils_unittest.py
152 *(Released Thu, 17 Oct 2013)*
154 - Correctly start/stop luxid during gnt-cluster master-failover
155 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
156 - Fix path for the job queue serial file
157 - Improved harep man page
158 - Minor documentation improvements
164 *(Released Mon, 30 Sep 2013)*
166 Incompatible/important changes
167 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
169 - Instance policy can contain multiple instance specs, as described in
170 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
171 <design-partitioned>`. As a consequence, it's not possible to partially change
172 or override instance specs. Bounding specs (min and max) can be specified as a
173 whole using the new option ``--ipolicy-bounds-specs``, while standard
174 specs use the new option ``--ipolicy-std-specs``.
175 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
176 gnt-instance is a valid YAML object.
177 - hail now honors network restrictions when allocating nodes. This led to an
178 update of the IAllocator protocol. See the IAllocator documentation for
180 - confd now only answers static configuration request over the network. luxid
181 was extracted, listens on the local LUXI socket and responds to live queries.
182 This allows finer grained permissions if using separate users.
187 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
188 to always require authentication, ``--require-authentication``. It can
189 be specified in ``$sysconfdir/default/ganeti``.
190 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
191 be used to manage the disk templates to be used by instances in the cluster.
192 Initially, it will be set to a list that includes plain, drbd, if they were
193 enabled by specifying a volume group name, and file and sharedfile, if those
194 were enabled at configure time. Additionally, it will include all disk
195 templates that are currently used by instances. The order of disk templates
196 will be based on Ganeti's history of supporting them. In the future, the
197 first entry of the list will be used as a default disk template on instance
199 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
200 configuration back to the previous stable version.
201 - Disk templates in group ipolicy can be restored to the default value.
202 - Initial support for diskless instances and virtual clusters in QA.
203 - More QA and unit tests for instance policies.
204 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
205 describing why the opcode itself was executed.
206 - The monitoring daemon is now available. It allows users to query the cluster
207 for obtaining information about the status of the system. The daemon is only
208 responsible for providing the information over the network: the actual data
209 gathering is performed by data collectors (currently, only the DRBD status
210 collector is available).
211 - In order to help developers work on Ganeti, a new script
212 (``devel/build_chroot``) is provided, for building a chroot that contains all
213 the required development libraries and tools for compiling Ganeti on a Debian
215 - A new tool, ``harep``, for performing self-repair and recreation of instances
216 in Ganeti has been added.
217 - Split queries are enabled for tags, network, exports, cluster info, groups,
219 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
220 It prints the command line to set the current policies, to ease
222 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
223 devices for KVM virtio-net interfaces will get created with VNET_HDR
224 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
225 virtio-net interfaces, which prevents host kernel tainting and log
226 flooding, when dealing with broken or malicious virtio-net drivers.
227 It's set to true by default.
228 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
230 - Support 'viridian' parameter in Xen HVM
231 - Support DSA SSH keys in bootstrap
232 - To simplify the work of packaging frameworks that want to add the needed users
233 and groups in a split-user setup themselves, at build time three files in
234 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
235 the groups to be generated, the ``users`` file contains, one per line, the
236 users to be generated, optionally followed by their primary group, where
237 important. The ``groupmemberships`` file contains, one per line, additional
238 user-group membership relations that need to be established. The syntax of
239 these files will remain stable in all future versions.
244 The following new dependencies have been added:
247 - The ``curl`` library is not optional anymore for compiling the Haskell code.
248 - ``snap-server`` library (if monitoring is enabled).
251 - The minimum Python version needed to run Ganeti is now 2.6.
252 - ``yaml`` library (only for running the QA).
256 - Perform proper cleanup on termination of Haskell daemons
257 - Fix corner-case in handling of remaining retry time
263 *(Released Tue, 17 Sep 2013)*
265 - To simplify the work of packaging frameworks that want to add the needed users
266 and groups in a split-user setup themselves, at build time three files in
267 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
268 the groups to be generated, the ``users`` file contains, one per line, the
269 users to be generated, optionally followed by their primary group, where
270 important. The ``groupmemberships`` file contains, one per line, additional
271 user-group membership relations that need to be established. The syntax of
272 these files will remain stable in all future versions.
273 - Add a default to file-driver when unspecified over RAPI (Issue 571)
274 - Mark the DSA host pubkey as optional, and remove it during config downgrade
276 - Some documentation fixes
282 *(Released Tue, 27 Aug 2013)*
284 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
286 - Support 'viridian' parameter in Xen HVM (Issue 233)
287 - Include VCS version in ``gnt-cluster version``
288 - Support DSA SSH keys in bootstrap (Issue 338)
289 - Fix batch creation of instances
290 - Use FQDN to check master node status (Issue 551)
291 - Make the DRBD collector more failure-resilient
297 *(Released Fri, 2 Aug 2013)*
299 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
301 - Fix upgrading/downgrading from 2.7
302 - Increase maximum RAPI message size
303 - Documentation updates
304 - Split ``confd`` between ``luxid`` and ``confd``
305 - Merge 2.7 series up to the 2.7.1 release
306 - Allow the ``modify_etc_hosts`` option to be changed
307 - Add better debugging for ``luxid`` queries
308 - Expose bulk parameter for GetJobs in RAPI client
309 - Expose missing ``network`` fields in RAPI
310 - Add some ``cluster verify`` tests
311 - Some unittest fixes
312 - Fix a malfunction in ``hspace``'s tiered allocation
313 - Fix query compatibility between haskell and python implementations
314 - Add the ``vnet_hdr`` HV parameter for KVM
315 - Add ``--cleanup`` to instance failover
316 - Change the connected groups format in ``gnt-network info`` output; it
317 was previously displayed as a raw list by mistake. (Merged from 2.7)
323 *(Released Mon, 24 Jun 2013)*
325 This was the first beta release of the 2.8 series. All important changes
326 are listed in the latest 2.8 entry.
332 *(Released Thu, 26 Sep 2013)*
334 - Change the connected groups format in ``gnt-network info`` output; it
335 was previously displayed as a raw list by mistake
336 - Check disk template in right dict when copying
337 - Support multi-instance allocs without iallocator
338 - Fix some errors in the documentation
339 - Fix formatting of tuple in an error message
345 *(Released Thu, 25 Jul 2013)*
347 - Add logrotate functionality in daemon-util
348 - Add logrotate example file
349 - Add missing fields to network queries over rapi
350 - Fix network object timestamps
351 - Add support for querying network timestamps
352 - Fix a typo in the example crontab
353 - Fix a documentation typo
359 *(Released Thu, 04 Jul 2013)*
361 Incompatible/important changes
362 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
364 - Instance policies for disk size were documented to be on a per-disk
365 basis, but hail applied them to the sum of all disks. This has been
367 - ``hbal`` will now exit with status 0 if, during job execution over
368 LUXI, early exit has been requested and all jobs are successful;
369 before, exit status 1 was used, which cannot be differentiated from
371 - Compatibility with newer versions of rbd has been fixed
372 - ``gnt-instance batch-create`` has been changed to use the bulk create
373 opcode from Ganeti. This lead to incompatible changes in the format of
374 the JSON file. It's now not a custom dict anymore but a dict
375 compatible with the ``OpInstanceCreate`` opcode.
376 - Parent directories for file storage need to be listed in
377 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
378 write the file automatically based on old configuration values, but it
379 can not distribute it across all nodes and the file contents should be
380 verified. Use ``gnt-cluster copyfile
381 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
382 upgraded. The reason for requiring this list of paths now is that
383 before it would have been possible to inject new paths via RPC,
384 allowing files to be created in arbitrary locations. The RPC protocol
385 is protected using SSL/X.509 certificates, but as a design principle
386 Ganeti does not permit arbitrary paths to be passed.
387 - The parsing of the variants file for OSes (see
388 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
389 lines and comment lines (starting with ``#``) are ignored for better
391 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
392 longer available. ``gnt-node add`` now invokes a new tool on the
393 destination node, named ``prepare-node-join``, to configure the SSH
394 daemon. Paramiko is no longer necessary to configure nodes' SSH
395 daemons via ``gnt-node add``.
396 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
397 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
398 the flag is not reset after a master failover.
399 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
400 is recommended. 2.6 will be mandatory from the 2.8 series.
406 - New network management functionality to support automatic allocation
407 of IP addresses and managing of network parameters. See
408 :manpage:`gnt-network(8)` for more details.
409 - New external storage backend, to allow managing arbitrary storage
410 systems external to the cluster. See
411 :manpage:`ganeti-extstorage-interface(7)`.
412 - New ``exclusive-storage`` node parameter added, restricted to
413 nodegroup level. When it's set to true, physical disks are assigned in
414 an exclusive fashion to instances, as documented in :doc:`Partitioned
415 Ganeti <design-partitioned>`. Currently, only instances using the
416 ``plain`` disk template are supported.
417 - The KVM hypervisor has been updated with many new hypervisor
418 parameters, including a generic one for passing arbitrary command line
419 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
420 compatible up to qemu 1.4.
421 - A new tool, called ``mon-collector``, is the stand-alone executor of
422 the data collectors for a monitoring system. As of this version, it
423 just includes the DRBD data collector, that can be executed by calling
424 ``mon-collector`` using the ``drbd`` parameter. See
425 :manpage:`mon-collector(7)`.
426 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
427 for RAPI users. It allows granting permissions to query for
428 information to a specific user without giving
429 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
430 - A new tool named ``node-cleanup`` has been added. It cleans remains of
431 a cluster from a machine by stopping all daemons, removing
432 certificates and ssconf files. Unless the ``--no-backup`` option is
433 given, copies of the certificates are made.
434 - Instance creations now support the use of opportunistic locking,
435 potentially speeding up the (parallel) creation of multiple instances.
436 This feature is currently only available via the :doc:`RAPI
437 <rapi>` interface and when an instance allocator is used. If the
438 ``opportunistic_locking`` parameter is set the opcode will try to
439 acquire as many locks as possible, but will not wait for any locks
440 held by other opcodes. If not enough resources can be found to
441 allocate the instance, the temporary error code
442 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
443 retried thereafter, with or without opportunistic locking.
444 - New experimental linux-ha resource scripts.
445 - Restricted-commands support: ganeti can now be asked (via command line
446 or rapi) to perform commands on a node. These are passed via ganeti
447 RPC rather than ssh. This functionality is restricted to commands
448 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
449 reasons. The file is not copied automatically.
455 - Diskless instances are now externally mirrored (Issue 237). This for
456 now has only been tested in conjunction with explicit target nodes for
458 - Queries not needing locks or RPC access to the node can now be
459 performed by the confd daemon, making them independent from jobs, and
460 thus faster to execute. This is selectable at configure time.
461 - The functionality for allocating multiple instances at once has been
462 overhauled and is now also available through :doc:`RAPI <rapi>`.
464 There are no significant changes from version 2.7.0~rc3.
470 *(Released Tue, 25 Jun 2013)*
472 - Fix permissions on the confd query socket (Issue 477)
473 - Fix permissions on the job archive dir (Issue 498)
474 - Fix handling of an internal exception in replace-disks (Issue 472)
475 - Fix gnt-node info handling of shortened names (Issue 497)
476 - Fix gnt-instance grow-disk when wiping is enabled
477 - Documentation improvements, and support for newer pandoc
478 - Fix hspace honoring ipolicy for disks (Issue 484)
479 - Improve handling of the ``kvm_extra`` HV parameter
485 *(Released Fri, 24 May 2013)*
487 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
489 - Disks added through ``gnt-instance modify`` or created through
490 ``gnt-instance recreate-disks`` are wiped, if the
491 ``prealloc_wipe_disks`` flag is set.
492 - If wiping newly created disks fails, the disks are removed. Also,
493 partial failures in creating disks through ``gnt-instance modify``
494 triggers a cleanup of the partially-created disks.
495 - Removing the master IP address doesn't fail if the address has been
497 - Fix ownership of the OS log dir
498 - Workaround missing SO_PEERCRED constant (Issue 191)
504 *(Released Fri, 3 May 2013)*
506 This was the first release candidate of the 2.7 series. Since beta3:
508 - Fix kvm compatibility with qemu 1.4 (Issue 389)
509 - Documentation updates (admin guide, upgrade notes, install
510 instructions) (Issue 372)
511 - Fix gnt-group list nodes and instances count (Issue 436)
512 - Fix compilation without non-mandatory libraries (Issue 441)
513 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
514 - Make confd logging more verbose at INFO level (Issue 435)
515 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
516 - Fix failure path for instance storage type conversion (Issue 229)
517 - Update htools text backend documentation
518 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
519 - Disable inter-cluster instance move for file-based instances, because
520 it is dependant on instance export, which is not supported for
521 file-based instances. (Issue 414)
522 - Fix gnt-job crashes on non-ascii characters (Issue 427)
523 - Fix volume group checks on non-vm-capable nodes (Issue 432)
529 *(Released Mon, 22 Apr 2013)*
531 This was the third beta release of the 2.7 series. Since beta2:
533 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
534 - Fix data loss on wrong usage of ``gnt-instance move``
535 - Properly export errors in confd-based job queries
536 - Add ``users-setup`` tool
537 - Fix iallocator protocol to report 0 as a disk size for diskless
538 instances. This avoids hail breaking when a diskless instance is
540 - Fix job queue directory permission problem that made confd job queries
541 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
542 for access to archived jobs (Issue 406).
543 - Limit the sizes of networks supported by ``gnt-network`` to something
544 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
545 - Fix bugs in instance disk template conversion
546 - Fix GHC 7 compatibility
547 - Fix ``burnin`` install path (Issue 426).
548 - Allow very small disk grows (Issue 347).
549 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
550 that noded doesn't import masterd code (Issue 419).
551 - Make sure the default metavg at cluster init is the same as the vg, if
552 unspecified (Issue 358).
553 - Fix cleanup of partially created disks (part of Issue 416)
559 *(Released Tue, 2 Apr 2013)*
561 This was the second beta release of the 2.7 series. Since beta1:
563 - Networks no longer have a "type" slot, since this information was
564 unused in Ganeti: instead of it tags should be used.
565 - The rapi client now has a ``target_node`` option to MigrateInstance.
566 - Fix early exit return code for hbal (Issue 386).
567 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
568 - Fix ``rbd showmapped`` output parsing (Issue 312).
569 - Networks are now referenced indexed by UUID, rather than name. This
570 will require running cfgupgrade, from 2.7.0beta1, if networks are in
572 - The OS environment now includes network information.
573 - Deleting of a network is now disallowed if any instance nic is using
574 it, to prevent dangling references.
575 - External storage is now documented in man pages.
576 - The exclusive_storage flag can now only be set at nodegroup level.
577 - Hbal can now submit an explicit priority with its jobs.
578 - Many network related locking fixes.
579 - Bump up the required pylint version to 0.25.1.
580 - Fix the ``no_remember`` option in RAPI client.
581 - Many ipolicy related tests, qa, and fixes.
582 - Many documentation improvements and fixes.
583 - Fix building with ``--disable-file-storage``.
584 - Fix ``-q`` option in htools, which was broken if passed more than
586 - Some haskell/python interaction improvements and fixes.
587 - Fix iallocator in case of missing LVM storage.
588 - Fix confd config load in case of ``--no-lvm-storage``.
589 - The confd/query functionality is now mentioned in the security
596 *(Released Wed, 6 Feb 2013)*
598 This was the first beta release of the 2.7 series. All important changes
599 are listed in the latest 2.7 entry.
605 *(Released Fri, 21 Dec 2012)*
607 Important behaviour change: hbal won't rebalance anymore instances which
608 have the ``auto_balance`` attribute set to false. This was the intention
609 all along, but until now it only skipped those from the N+1 memory
610 reservation (DRBD-specific).
612 A significant number of bug fixes in this release:
614 - Fixed disk adoption interaction with ipolicy checks.
615 - Fixed networking issues when instances are started, stopped or
616 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
617 - Fixed the warning in cluster verify for shared storage instances not
619 - Fixed removal of storage directory on shared file storage (issue 262).
620 - Fixed validation of LVM volume group name in OpClusterSetParams
621 (``gnt-cluster modify``) (issue 285).
622 - Fixed runtime memory increases (``gnt-instance modify -m``).
623 - Fixed live migration under Xen's ``xl`` mode.
624 - Fixed ``gnt-instance console`` with ``xl``.
625 - Fixed building with newer Haskell compiler/libraries.
626 - Fixed PID file writing in Haskell daemons (confd); this prevents
627 restart issues if confd was launched manually (outside of
628 ``daemon-util``) while another copy of it was running
629 - Fixed a type error when doing live migrations with KVM (issue 297) and
630 the error messages for failing migrations have been improved.
631 - Fixed opcode validation for the out-of-band commands (``gnt-node
633 - Fixed a type error when unsetting OS hypervisor parameters (issue
634 311); now it's possible to unset all OS-specific hypervisor
636 - Fixed the ``dry-run`` mode for many operations: verification of
637 results was over-zealous but didn't take into account the ``dry-run``
638 operation, resulting in "wrong" failures.
639 - Fixed bash completion in ``gnt-job list`` when the job queue has
640 hundreds of entries; especially with older ``bash`` versions, this
641 results in significant CPU usage.
643 And lastly, a few other improvements have been made:
645 - Added option to force master-failover without voting (issue 282).
646 - Clarified error message on lock conflict (issue 287).
647 - Logging of newly submitted jobs has been improved (issue 290).
648 - Hostname checks have been made uniform between instance rename and
650 - The ``--submit`` option is now supported by ``gnt-debug delay``.
651 - Shutting down the master daemon by sending SIGTERM now stops it from
652 processing jobs waiting for locks; instead, those jobs will be started
653 once again after the master daemon is started the next time (issue
655 - Support for Xen's ``xl`` program has been improved (besides the fixes
657 - Reduced logging noise in the Haskell confd daemon (only show one log
658 entry for each config reload, instead of two).
659 - Several man page updates and typo fixes.
665 *(Released Fri, 12 Oct 2012)*
667 A small bugfix release. Among the bugs fixed:
669 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
670 made the command unusable.
671 - Commands that issue many jobs don't fail anymore just because some jobs
672 take so long that other jobs are archived.
673 - Failures during ``gnt-instance reinstall`` are reflected by the exit
675 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
677 - When ``always_failover`` is set, ``--allow-failover`` is not required
678 in migrate commands anymore.
679 - ``bash_completion`` works even if extglob is disabled.
680 - Fixed bug with locks that made failover for RDB-based instances fail.
681 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
682 a random node instead of one based on the allocator metric.
683 - Support for newer versions of pylint and pep8.
684 - Hail doesn't fail anymore when trying to add an instance of type
685 ``file``, ``sharedfile`` or ``rbd``.
686 - Added new Makefile target to rebuild the whole distribution, so that
687 all files are included.
693 *(Released Fri, 27 Jul 2012)*
696 .. attention:: The ``LUXI`` protocol has been made more consistent
697 regarding its handling of command arguments. This, however, leads to
698 incompatibility issues with previous versions. Please ensure that you
699 restart Ganeti daemons soon after the upgrade, otherwise most
700 ``LUXI`` calls (job submission, setting/resetting the drain flag,
701 pausing/resuming the watcher, cancelling and archiving jobs, querying
702 the cluster configuration) will fail.
711 The current ``admin_up`` field, which used to denote whether an instance
712 should be running or not, has been removed. Instead, ``admin_state`` is
713 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
715 The rational behind this is that an instance being “down” can have
718 - it could be down during a reboot
719 - it could be temporarily be down for a reinstall
720 - or it could be down because it is deprecated and kept just for its
723 The previous Boolean state was making it difficult to do capacity
724 calculations: should Ganeti reserve memory for a down instance? Now, the
725 tri-state field makes it clear:
727 - in ``up`` and ``down`` state, all resources are reserved for the
728 instance, and it can be at any time brought up if it is down
729 - in ``offline`` state, only disk space is reserved for it, but not
732 The field can have an extra use: since the transition between ``up`` and
733 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
734 transition between ``offline`` and ``down`` is done via ``gnt-instance
735 modify``, it is possible to given different rights to users. For
736 example, owners of an instance could be allowed to start/stop it, but
737 not transition it out of the offline state.
739 Instance policies and specs
740 +++++++++++++++++++++++++++
742 In previous Ganeti versions, an instance creation request was not
743 limited on the minimum size and on the maximum size just by the cluster
744 resources. As such, any policy could be implemented only in third-party
745 clients (RAPI clients, or shell wrappers over ``gnt-*``
746 tools). Furthermore, calculating cluster capacity via ``hspace`` again
747 required external input with regards to instance sizes.
749 In order to improve these workflows and to allow for example better
750 per-node group differentiation, we introduced instance specs, which
753 - minimum instance disk size, disk count, memory size, cpu count
754 - maximum values for the above metrics
755 - and “standard” values (used in ``hspace`` to calculate the standard
758 The minimum/maximum values can be also customised at node-group level,
759 for example allowing more powerful hardware to support bigger instance
762 Beside the instance specs, there are a few other settings belonging to
763 the instance policy framework. It is possible now to customise, per
764 cluster and node-group:
766 - the list of allowed disk templates
767 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
768 - the maximum ratio of instance to spindles (see below for more
769 information) for local storage
771 All these together should allow all tools that talk to Ganeti to know
772 what are the ranges of allowed values for instances and the
773 over-subscription that is allowed.
775 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
776 instance configuration, and the physical CPU configuration from the
777 node. For the spindle ratios however, we didn't track before these
778 values, so new parameters have been added:
780 - a new node parameter ``spindle_count``, defaults to 1, customisable at
781 node group or node level
782 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
784 Note that spindles in this context doesn't need to mean actual
785 mechanical hard-drives; it's just a relative number for both the node
786 I/O capacity and instance I/O consumption.
788 Instance migration behaviour
789 ++++++++++++++++++++++++++++
791 While live-migration is in general desirable over failover, it is
792 possible that for some workloads it is actually worse, due to the
793 variable time of the “suspend” phase during live migration.
795 To allow the tools to work consistently over such instances (without
796 having to hard-code instance names), a new backend parameter
797 ``always_failover`` has been added to control the migration/failover
798 behaviour. When set to True, all migration requests for an instance will
799 instead fall-back to failover.
801 Instance memory ballooning
802 ++++++++++++++++++++++++++
804 Initial support for memory ballooning has been added. The memory for an
805 instance is no longer fixed (backend parameter ``memory``), but instead
806 can vary between minimum and maximum values (backend parameters
807 ``minmem`` and ``maxmem``). Currently we only change an instance's
810 - live migrating or failing over and instance and the target node
811 doesn't have enough memory
812 - user requests changing the memory via ``gnt-instance modify
818 In order to control the use of specific CPUs by instance, support for
819 controlling CPU pinning has been added for the Xen, HVM and LXC
820 hypervisors. This is controlled by a new hypervisor parameter
821 ``cpu_mask``; details about possible values for this are in the
822 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
823 VCPU-to-CPU mapping) form will work well only when all nodes in your
824 cluster have the same amount of CPUs.
829 Another area in which Ganeti was not customisable were the parameters
830 used for storage configuration, e.g. how many stripes to use for LVM,
831 DRBD resync configuration, etc.
833 To improve this area, we've added disks parameters, which are
834 customisable at cluster and node group level, and which allow to
835 specify various parameters for disks (DRBD has the most parameters
836 currently), for example:
838 - DRBD resync algorithm and parameters (e.g. speed)
839 - the default VG for meta-data volumes for DRBD
840 - number of stripes for LVM (plain disk template)
843 These parameters can be modified via ``gnt-cluster modify -D …`` and
844 ``gnt-group modify -D …``, and are used at either instance creation (in
845 case of LVM stripes, for example) or at disk “activation” time
848 Rados block device support
849 ++++++++++++++++++++++++++
851 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
852 denoted by the ``rbd`` disk template type. This is considered
853 experimental, feedback is welcome. For details on configuring it, see
854 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
859 The existing master IP functionality works well only in simple setups (a
860 single network shared by all nodes); however, if nodes belong to
861 different networks, then the ``/32`` setup and lack of routing
862 information is not enough.
864 To allow the master IP to function well in more complex cases, the
865 system was reworked as follows:
867 - a master IP netmask setting has been added
868 - the master IP activation/turn-down code was moved from the node daemon
870 - whether to run the Ganeti-supplied master IP script or a user-supplied
871 on is a ``gnt-cluster init`` setting
873 Details about the location of the standard and custom setup scripts are
874 in the man page :manpage:`gnt-cluster(8)`; for information about the
875 setup script protocol, look at the Ganeti-supplied script.
880 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
883 It is now possible to use TLS-protected connections, and when renewing
884 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
885 it is now possible to specify spice or spice CA certificates. Also, it
886 is possible to configure a password for SPICE sessions via the
887 hypervisor parameter ``spice_password_file``.
889 There are also new parameters to control the compression and streaming
890 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
891 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
892 for the spice parameters.
894 Lastly, it is now possible to see the SPICE connection information via
895 ``gnt-instance console``.
900 A new tool (``tools/ovfconverter``) has been added that supports
901 conversion between Ganeti and the `Open Virtualization Format
902 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
905 This relies on the ``qemu-img`` tool to convert the disk formats, so the
906 actual compatibility with other virtualization solutions depends on it.
911 The configuration query daemon (``ganeti-confd``) is now optional, and
912 has been rewritten in Haskell; whether to use the daemon at all, use the
913 Python (default) or the Haskell version is selectable at configure time
914 via the ``--enable-confd`` parameter, which can take one of the
915 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
916 daemon will result in a smaller footprint; for larger systems, we
917 welcome feedback on the Haskell version which might become the default
920 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
921 daemon running. The Python version doesn't implement the new call.
924 User interface changes
925 ~~~~~~~~~~~~~~~~~~~~~~
927 We have replaced the ``--disks`` option of ``gnt-instance
928 replace-disks`` with a more flexible ``--disk`` option, which allows
929 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
930 disk size and mode can be changed upon recreation (via ``gnt-instance
931 recreate-disks``, which accepts the same ``--disk`` option).
933 As many people are used to a ``show`` command, we have added that as an
934 alias to ``info`` on all ``gnt-*`` commands.
936 The ``gnt-instance grow-disk`` command has a new mode in which it can
937 accept the target size of the disk, instead of the delta; this can be
938 more safe since two runs in absolute mode will be idempotent, and
939 sometimes it's also easier to specify the desired size directly.
941 Also the handling of instances with regard to offline secondaries has
942 been improved. Instance operations should not fail because one of it's
943 secondary nodes is offline, even though it's safe to proceed.
945 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
946 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
947 minors to instance name.
952 RAPI coverage has improved, with (for example) new resources for
953 recreate-disks, node power-cycle, etc.
958 There is partial support for ``xl`` in the Xen hypervisor; feedback is
961 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
962 whether to still support Python 2.4 or move to Python 2.6 as minimum
965 Support for Fedora has been slightly improved; the provided example
966 init.d script should work better on it and the INSTALL file should
967 document the needed dependencies.
972 The deprecated ``QueryLocks`` LUXI request has been removed. Use
973 ``Query(what=QR_LOCK, ...)`` instead.
975 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
976 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
977 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
978 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
979 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
981 RAPI client: ``CertificateError`` now derives from
982 ``GanetiApiError``. This should make it more easy to handle Ganeti
985 Deprecation warnings due to PyCrypto/paramiko import in
986 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
987 make sure to run an up-to-date paramiko version, if you use this tool.
989 The QA scripts now depend on Python 2.5 or above (the main code base
990 still works with Python 2.4).
992 The configuration file (``config.data``) is now written without
993 indentation for performance reasons; if you want to edit it, it can be
994 re-formatted via ``tools/fmtjson``.
996 A number of bugs has been fixed in the cluster merge tool.
998 ``x509`` certification verification (used in import-export) has been
999 changed to allow the same clock skew as permitted by the cluster
1000 verification. This will remove some rare but hard to diagnose errors in
1007 *(Released Thu, 19 Jul 2012)*
1009 Very few changes from rc4 to the final release, only bugfixes:
1011 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
1012 instance, fix CDROM booting for KVM instances)
1013 - fixed node group modification of node parameters
1014 - fixed issue in LUClusterVerifyGroup with multi-group clusters
1015 - fixed generation of bash completion to ensure a stable ordering
1022 *(Released Fri, 13 Jul 2012)*
1024 Third release candidate for 2.6. The following changes were done from
1027 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1028 - Fixed an inconsistency in the LUXI protocol with the provided
1029 arguments (NOT backwards compatible)
1030 - Fixed a bug with node groups ipolicy where ``min`` was greater than
1031 the cluster ``std`` value
1032 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1033 easier instance debugging on nodes (requires ``hconfd`` to work)
1039 *(Released Tue, 03 Jul 2012)*
1041 Second release candidate for 2.6. The following changes were done from
1044 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1046 - Fixed a RAPI regression on missing beparams/memory
1047 - Fixed redistribution of files on offline nodes
1048 - Added possibility to run activate-disks even though secondaries are
1049 offline. With this change it relaxes also the strictness on some other
1050 commands which use activate disks internally:
1051 * ``gnt-instance start|reboot|rename|backup|export``
1052 - Made it possible to remove safely an instance if its secondaries are
1054 - Made it possible to reinstall even though secondaries are offline
1060 *(Released Mon, 25 Jun 2012)*
1062 First release candidate for 2.6. The following changes were done from
1065 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1067 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1068 behaviour. It supports now the modify operation
1069 - ``hcheck`` implemented to analyze cluster health and possibility of
1070 improving health by rebalance
1071 - ``hbal`` has been improved in dealing with split instances
1077 *(Released Mon, 11 Jun 2012)*
1079 Second beta release of 2.6. The following changes were done from beta2
1082 - Fixed ``daemon-util`` with non-root user models
1083 - Fixed creation of plain instances with ``--no-wait-for-sync``
1084 - Fix wrong iv_names when running ``cfgupgrade``
1085 - Export more information in RAPI group queries
1086 - Fixed bug when changing instance network interfaces
1087 - Extended burnin to do NIC changes
1088 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1089 - Changed default for DRBD barriers
1090 - Fixed DRBD error reporting for syncer rate
1091 - Verify the options on disk parameters
1093 And of course various fixes to documentation and improved unittests and
1100 *(Released Wed, 23 May 2012)*
1102 First beta release of 2.6. The following changes were done from beta1 to
1105 - integrated patch for distributions without ``start-stop-daemon``
1106 - adapted example init.d script to work on Fedora
1107 - fixed log handling in Haskell daemons
1108 - adapted checks in the watcher for pycurl linked against libnss
1109 - add partial support for ``xl`` instead of ``xm`` for Xen
1110 - fixed a type issue in cluster verification
1111 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1114 Plus integrated fixes from the 2.5 branch:
1116 - fixed ``kvm-ifup`` to use ``/bin/bash``
1117 - fixed parallel build failures
1118 - KVM live migration when using a custom keymap
1124 *(Released Tue, 24 Jul 2012)*
1126 A small bugfix release, with no new features:
1128 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1129 different default shell (e.g. Debian, Ubuntu)
1130 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1131 243 and Debian bug #650664)
1132 - fixed compatibility with KVM versions that don't support multiple boot
1133 devices (fixes Issue 230 and Debian bug #624256)
1135 Additionally, a few fixes were done to the build system (fixed parallel
1136 build failures) and to the unittests (fixed race condition in test for
1137 FileID functions, and the default enable/disable mode for QA test is now
1144 *(Released Fri, 11 May 2012)*
1146 A small bugfix release.
1148 The main issues solved are on the topic of compatibility with newer LVM
1151 - fixed parsing of ``lv_attr`` field
1152 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1153 the ``--force`` flag is needed
1155 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1156 to accept kernel 3.x too (was hardcoded to 2.6.*).
1158 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1159 of hook results and that also made display of other errors suboptimal
1160 was fixed; the code behaves now like 2.4 and earlier.
1162 Another change in 2.5, the cleanup of the OS scripts environment, is too
1163 aggressive: it removed even the ``PATH`` variable, which requires the OS
1164 scripts to *always* need to export it. Since this is a bit too strict,
1165 we now export a minimal PATH, the same that we export for hooks.
1167 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1168 integrated into this release.
1170 Finally, a few other miscellaneous changes were done (no new features,
1171 just small improvements):
1173 - Fix ``gnt-group --help`` display
1174 - Fix hardcoded Xen kernel path
1175 - Fix grow-disk handling of invalid units
1176 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1177 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1183 *(Released Thu, 12 Apr 2012)*
1185 Incompatible/important changes and bugfixes
1186 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1188 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1189 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1190 to match the underlying LUXI interface.
1191 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1192 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1193 not maintain backwards-compatibility as the underlying operation
1194 changed in an incompatible way. The RAPI client can talk to old
1195 servers, but it needs to be told so as the return value changed.
1196 - When creating file-based instances via RAPI, the ``file_driver``
1197 parameter no longer defaults to ``loop`` and must be specified.
1198 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1200 - Support for the undocumented and deprecated RAPI instance creation
1201 request format version 0 has been dropped. Use version 1, supported
1202 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1203 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1204 documentation <install>`.
1205 - The "cluster-verify" hooks are now executed per group by the
1206 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1207 if you just run ``gnt-cluster verify``, which generates one opcode per
1209 - The environment as passed to the OS scripts is cleared, and thus no
1210 environment variables defined in the node daemon's environment will be
1211 inherited by the scripts.
1212 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1214 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1215 support operations involving multiple node groups.
1216 - Offline nodes are ignored when failing over an instance.
1217 - Support for KVM version 1.0, which changed the version reporting format
1219 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1221 - ``Makefile`` is now compatible with Automake 1.11.2
1222 - Includes all bugfixes made in the 2.4 series
1227 - The ganeti-htools project has been merged into the ganeti-core source
1228 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1229 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1230 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1231 GPT-style partition tables (requires `parted
1232 <http://www.gnu.org/s/parted/>`_).
1233 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1234 - Allowed adding tags on instance creation.
1235 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1236 - Allow instances to be started in a paused state, enabling the user to
1237 see the complete console output on boot using the console.
1238 - Added new hypervisor flag to control default reboot behaviour
1239 (``reboot_behavior``).
1240 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1241 - Improved out-of-band management support:
1243 - Added ``gnt-node health`` command reporting the health status of
1245 - Added ``gnt-node power`` command to manage power status of nodes.
1246 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1248 - Instance migration can fall back to failover if instance is not
1250 - Filters can be used when listing nodes, instances, groups and locks;
1251 see :manpage:`ganeti(7)` manpage.
1252 - Added post-execution status as variables to :doc:`hooks <hooks>`
1254 - Instance tags are exported/imported together with the instance.
1255 - When given an explicit job ID, ``gnt-job info`` will work for archived
1257 - Jobs can define dependencies on other jobs (not yet supported via
1258 RAPI or command line, but used by internal commands and usable via
1261 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1264 - Instance failover is now available as a RAPI resource
1265 (``/2/instances/[instance_name]/failover``).
1266 - ``gnt-instance info`` defaults to static information if primary node
1268 - Opcodes have a new ``comment`` attribute.
1269 - Added basic SPICE support to KVM hypervisor.
1270 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1272 Node group improvements
1273 ~~~~~~~~~~~~~~~~~~~~~~~
1275 - ``gnt-cluster verify`` has been modified to check groups separately,
1276 thereby improving performance.
1277 - Node group support has been added to ``gnt-cluster verify-disks``,
1278 which now operates per node group.
1279 - Watcher has been changed to work better with node groups.
1281 - One process and state file per node group.
1282 - Slow watcher in one group doesn't block other group's watcher.
1284 - Added new command, ``gnt-group evacuate``, to move all instances in a
1285 node group to other groups.
1286 - Added ``gnt-instance change-group`` to move an instance to another
1288 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1289 per-group operations.
1290 - Node groups can be tagged.
1291 - Some operations switch from an exclusive to a shared lock as soon as
1293 - Instance's primary and secondary nodes' groups are now available as
1294 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1295 and ``snodes.group.uuid``).
1300 - Numerous updates to documentation and manpages.
1302 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1304 - Some opcode/job results are now also documented, see :doc:`RAPI
1307 - A lockset's internal lock is now also visible in lock monitor.
1308 - Log messages from job queue workers now contain information about the
1309 opcode they're processing.
1310 - ``gnt-instance console`` no longer requires the instance lock.
1311 - A short delay when waiting for job changes reduces the number of LUXI
1312 requests significantly.
1313 - DRBD metadata volumes are overwritten with zeros during disk creation.
1314 - Out-of-band commands no longer acquire the cluster lock in exclusive
1316 - ``devel/upload`` now uses correct permissions for directories.
1322 *(Released Fri, 23 Mar 2012)*
1324 This was the sixth release candidate of the 2.5 series.
1330 *(Released Mon, 9 Jan 2012)*
1332 This was the fifth release candidate of the 2.5 series.
1338 *(Released Thu, 27 Oct 2011)*
1340 This was the fourth release candidate of the 2.5 series.
1346 *(Released Wed, 26 Oct 2011)*
1348 This was the third release candidate of the 2.5 series.
1354 *(Released Tue, 18 Oct 2011)*
1356 This was the second release candidate of the 2.5 series.
1362 *(Released Tue, 4 Oct 2011)*
1364 This was the first release candidate of the 2.5 series.
1370 *(Released Wed, 31 Aug 2011)*
1372 This was the third beta release of the 2.5 series.
1378 *(Released Mon, 22 Aug 2011)*
1380 This was the second beta release of the 2.5 series.
1386 *(Released Fri, 12 Aug 2011)*
1388 This was the first beta release of the 2.5 series.
1394 *(Released Thu, 27 Oct 2011)*
1396 - Fixed bug when parsing command line parameter values ending in
1398 - Fixed assertion error after unclean master shutdown
1399 - Disable HTTP client pool for RPC, significantly reducing memory usage
1401 - Fixed queue archive creation with wrong permissions
1407 *(Released Tue, 23 Aug 2011)*
1411 - Fixed documentation for importing with ``--src-dir`` option
1412 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1413 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1419 *(Released Fri, 5 Aug 2011)*
1421 Many bug-fixes and a few small features:
1423 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1424 issues when you tried to add an instance with two MAC addresses in one
1426 - KVM: fixed per-instance stored UID value
1427 - KVM: configure bridged NICs at migration start
1428 - KVM: Fix a bug where instance will not start with never KVM versions
1430 - Added OS search path to ``gnt-cluster info``
1431 - Fixed an issue with ``file_storage_dir`` where you were forced to
1432 provide an absolute path, but the documentation states it is a
1433 relative path, the documentation was right
1434 - Added a new parameter to instance stop/start called ``--no-remember``
1435 that will make the state change to not be remembered
1436 - Implemented ``no_remember`` at RAPI level
1437 - Improved the documentation
1438 - Node evacuation: don't call IAllocator if node is already empty
1439 - Fixed bug in DRBD8 replace disks on current nodes
1440 - Fixed bug in recreate-disks for DRBD instances
1441 - Moved assertion checking locks in ``gnt-instance replace-disks``
1442 causing it to abort with not owning the right locks for some situation
1443 - Job queue: Fixed potential race condition when cancelling queued jobs
1444 - Fixed off-by-one bug in job serial generation
1445 - ``gnt-node volumes``: Fix instance names
1446 - Fixed aliases in bash completion
1447 - Fixed a bug in reopening log files after being sent a SIGHUP
1448 - Added a flag to burnin to allow specifying VCPU count
1449 - Bugfixes to non-root Ganeti configuration
1455 *(Released Thu, 12 May 2011)*
1457 Many bug-fixes and a few new small features:
1459 - Fixed a bug related to log opening failures
1460 - Fixed a bug in instance listing with orphan instances
1461 - Fixed a bug which prevented resetting the cluster-level node parameter
1462 ``oob_program`` to the default
1463 - Many fixes related to the ``cluster-merge`` tool
1464 - Fixed a race condition in the lock monitor, which caused failures
1465 during (at least) creation of many instances in parallel
1466 - Improved output for gnt-job info
1467 - Removed the quiet flag on some ssh calls which prevented debugging
1469 - Improved the N+1 failure messages in cluster verify by actually
1470 showing the memory values (needed and available)
1471 - Increased lock attempt timeouts so that when executing long operations
1472 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1473 too early and thus prevent the use of the 'fair' mechanism
1474 - Changed instance query data (``gnt-instance info``) to not acquire
1475 locks unless needed, thus allowing its use on locked instance if only
1476 static information is asked for
1477 - Improved behaviour with filesystems that do not support rename on an
1479 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1480 kept locks on all nodes during the wipe, which is unneeded
1481 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1482 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1483 10GiB) which caused the wipe to fail right at the end in some cases
1484 - Fixed master IP activation when doing master failover with no-voting
1485 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1486 the master node itself
1487 - Fixed potential data-loss in under disk full conditions, where Ganeti
1488 wouldn't check correctly the return code and would consider
1489 partially-written files 'correct'
1490 - Fixed bug related to multiple VGs and DRBD disk replacing
1491 - Added new disk parameter ``metavg`` that allows placement of the meta
1492 device for DRBD in a different volume group
1493 - Fixed error handling in the node daemon when the system libc doesn't
1494 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1495 - Fixed lock release during replace-disks, which kept cluster-wide locks
1496 when doing disk replaces with an iallocator script
1497 - Added check for missing bridges in cluster verify
1498 - Handle EPIPE errors while writing to the terminal better, so that
1499 piping the output to e.g. ``less`` doesn't cause a backtrace
1500 - Fixed rare case where a ^C during Luxi calls could have been
1501 interpreted as server errors, instead of simply terminating
1502 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1504 - Added a few more parameters to the KVM hypervisor, allowing a second
1505 CDROM, custom disk type for CDROMs and a floppy image
1506 - Removed redundant message in instance rename when the name is given
1508 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1509 disks on new nodes, allowing recreation when the original instance
1510 nodes are completely gone
1511 - Added option when converting disk templates to DRBD to skip waiting
1512 for the resync, in order to make the instance available sooner
1513 - Added two new variables to the OS scripts environment (containing the
1515 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1516 to allow use of ``pvgrub`` as bootloader
1517 - Changed the instance memory modifications to only check out-of-memory
1518 conditions on memory increases, and turned the secondary node warnings
1519 into errors (they can still be overridden via ``--force``)
1520 - Fixed the handling of a corner case when the Python installation gets
1521 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1522 to execute a command that doesn't exist
1523 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1524 node of the instance returned failures during instance shutdown; this
1525 adds the option ``--ignore-consistency`` to gnt-instance move
1527 And as usual, various improvements to the error messages, documentation
1534 *(Released Wed, 09 Mar 2011)*
1536 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1537 the RAPI users file if run twice (even with ``--dry-run``).
1539 The release fixes that bug (nothing else changed).
1545 *(Released Mon, 07 Mar 2011)*
1547 Final 2.4.0 release. Just a few small fixes:
1549 - Fixed RAPI node evacuate
1550 - Fixed the kvm-ifup script
1551 - Fixed internal error handling for special job cases
1552 - Updated man page to specify the escaping feature for options
1558 *(Released Mon, 28 Feb 2011)*
1560 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1561 that this feature wiped the disks of the wrong instance, leading to loss
1566 - Fixed title of query field containing instance name
1567 - Expanded the glossary in the documentation
1568 - Fixed one unittest (internal issue)
1574 *(Released Mon, 21 Feb 2011)*
1576 A number of bug fixes plus just a couple functionality changes.
1578 On the user-visible side, the ``gnt-* list`` command output has changed
1579 with respect to "special" field states. The current rc1 style of display
1580 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1581 the default output mode special fields states are displayed as follows:
1583 - Offline resource: ``*``
1584 - Unavailable/not applicable: ``-``
1585 - Data missing (RPC failure): ``?``
1586 - Unknown field: ``??``
1588 Another user-visible change is the addition of ``--force-join`` to
1593 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1594 - Fixed regression in RAPI/instance reinstall where all parameters were
1595 required (instead of optional)
1596 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1597 - Fixed iallocator usage (offline nodes were not considered offline)
1598 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1599 - Fixed hypervisor and OS parameter validation with respect to
1600 non-vm_capable nodes
1601 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1603 - Fixed ``tools/listrunner`` with respect to agent-based usage
1609 *(Released Fri, 4 Feb 2011)*
1611 Many changes and fixes since the beta1 release. While there were some
1612 internal changes, the code has been mostly stabilised for the RC
1615 Note: the dumb allocator was removed in this release, as it was not kept
1616 up-to-date with the IAllocator protocol changes. It is recommended to
1617 use the ``hail`` command from the ganeti-htools package.
1619 Note: the 2.4 and up versions of Ganeti are not compatible with the
1620 0.2.x branch of ganeti-htools. You need to upgrade to
1621 ganeti-htools-0.3.0 (or later).
1623 Regressions fixed from 2.3
1624 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1626 - Fixed the ``gnt-cluster verify-disks`` command
1627 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1629 - Fixed disk adoption breakage
1630 - Fixed wrong headers in instance listing for field aliases
1635 - Fixed corner case in KVM handling of NICs
1636 - Fixed many cases of wrong handling of non-vm_capable nodes
1637 - Fixed a bug where a missing instance symlink was not possible to
1638 recreate with any ``gnt-*`` command (now ``gnt-instance
1639 activate-disks`` does it)
1640 - Fixed the volume group name as reported by ``gnt-cluster
1642 - Increased timeouts for the import-export code, hopefully leading to
1643 fewer aborts due network or instance timeouts
1644 - Fixed bug in ``gnt-node list-storage``
1645 - Fixed bug where not all daemons were started on cluster
1646 initialisation, but only at the first watcher run
1647 - Fixed many bugs in the OOB implementation
1648 - Fixed watcher behaviour in presence of instances with offline
1650 - Fixed instance list output for instances running on the wrong node
1651 - a few fixes to the cluster-merge tool, but it still cannot merge
1652 multi-node groups (currently it is not recommended to use this tool)
1658 - Improved network configuration for the KVM hypervisor
1659 - Added e1000 as a supported NIC for Xen-HVM
1660 - Improved the lvmstrap tool to also be able to use partitions, as
1661 opposed to full disks
1662 - Improved speed of disk wiping (the cluster parameter
1663 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1664 of instance creations
1665 - Added documentation for the OS parameters
1666 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1667 hypervisor is not responding
1668 - Added display of blacklisted and hidden OS information in
1669 ``gnt-cluster info``
1670 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1671 NIC and node parameters, which might create problems with currently
1672 invalid (but undetected) configuration files, but prevents validation
1673 failures when unrelated parameters are modified
1674 - Changed cluster initialisation to wait for the master daemon to become
1676 - Expanded the RAPI interface:
1678 - Added config redistribution resource
1679 - Added activation/deactivation of instance disks
1680 - Added export of console information
1682 - Implemented log file reopening on SIGHUP, which allows using
1683 logrotate(8) for the Ganeti log files
1684 - Added a basic OOB helper script as an example
1690 *(Released Fri, 14 Jan 2011)*
1695 - Fixed timezone issues when formatting timestamps
1696 - Added support for node groups, available via ``gnt-group`` and other
1698 - Added out-of-band framework and management, see :doc:`design
1699 document <design-oob>`
1700 - Removed support for roman numbers from ``gnt-node list`` and
1701 ``gnt-instance list``.
1702 - Allowed modification of master network interface via ``gnt-cluster
1703 modify --master-netdev``
1704 - Accept offline secondaries while shutting down instance disks
1705 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1706 - Added support for multiple LVM volume groups
1707 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1709 - Added commands to list available fields:
1711 - ``gnt-node list-fields``
1712 - ``gnt-group list-fields``
1713 - ``gnt-instance list-fields``
1715 - Updated documentation and man pages
1720 - Moved ``rapi_users`` file into separate directory, now named
1721 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1723 - Added new tool for running commands on many machines,
1724 ``tools/ganeti-listrunner``
1725 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1726 improving the ``gnt-instance console`` output
1727 - Allowed customisation of disk index separator at ``configure`` time
1728 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1729 - Added support for non-partitioned md disks in ``lvmstrap``
1730 - Added script to gracefully power off KVM instances
1731 - Split ``utils`` module into smaller parts
1732 - Changed query operations to return more detailed information, e.g.
1733 whether an information is unavailable due to an offline node. To use
1734 this new functionality, the LUXI call ``Query`` must be used. Field
1735 information is now stored by the master daemon and can be retrieved
1736 using ``QueryFields``. Instances, nodes and groups can also be queried
1737 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1738 exposed via RAPI). The following commands make use of this
1739 infrastructure change:
1741 - ``gnt-group list``
1742 - ``gnt-group list-fields``
1744 - ``gnt-node list-fields``
1745 - ``gnt-instance list``
1746 - ``gnt-instance list-fields``
1747 - ``gnt-debug locks``
1752 - New RAPI resources (see :doc:`rapi`):
1756 - ``/2/groups/[group_name]``
1757 - ``/2/groups/[group_name]/assign-nodes``
1758 - ``/2/groups/[group_name]/modify``
1759 - ``/2/groups/[group_name]/rename``
1760 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1764 - Implemented ``no_install`` for instance creation
1765 - Implemented OS parameters for instance reinstallation, allowing
1766 use of special settings on reinstallation (e.g. for preserving data)
1771 - Added IPv6 support in import/export
1772 - Pause DRBD synchronization while wiping disks on instance creation
1773 - Updated unittests and QA scripts
1774 - Improved network parameters passed to KVM
1775 - Converted man pages from docbook to reStructuredText
1781 *(Released Mon, 20 Dec 2010)*
1783 Released version 2.3.1~rc1 without any changes.
1789 *(Released Wed, 1 Dec 2010)*
1791 - impexpd: Disable OpenSSL compression in socat if possible (backport
1792 from master, commit e90739d625b, see :doc:`installation guide
1793 <install-quick>` for details)
1794 - Changed unittest coverage report to exclude test scripts
1795 - Added script to check version format
1801 *(Released Wed, 1 Dec 2010)*
1803 Released version 2.3.0~rc1 without any changes.
1809 *(Released Fri, 19 Nov 2010)*
1811 A number of bugfixes and documentation updates:
1813 - Update ganeti-os-interface documentation
1814 - Fixed a bug related to duplicate MACs or similar items which should be
1816 - Fix breakage in OS state modify
1817 - Reinstall instance: disallow offline secondaries (fixes bug related to
1818 OS changing but reinstall failing)
1819 - plus all the other fixes between 2.2.1 and 2.2.2
1825 *(Released Tue, 2 Nov 2010)*
1827 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1828 - Fixed master failover race with watcher
1829 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1831 - Accept previously stopped instance for export with instance removal
1832 - Simplify and extend the environment variables for instance OS scripts
1833 - Added new node flags, ``master_capable`` and ``vm_capable``
1834 - Added optional instance disk wiping prior during allocation. This is a
1835 cluster-wide option and can be set/modified using
1836 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1837 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1838 :doc:`install-quick`
1839 - Added a new watcher option (``--ignore-pause``)
1840 - Added option to ignore offline node on instance start/stop
1841 (``--ignore-offline``)
1842 - Allow overriding OS parameters with ``gnt-instance reinstall``
1843 - Added ability to change node's secondary IP address using ``gnt-node
1845 - Implemented privilege separation for all daemons except
1846 ``ganeti-noded``, see ``configure`` options
1847 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1849 - Implemented job priorities (see ``ganeti(7)`` manpage)
1850 - Ignore failures while shutting down instances during failover from
1852 - Exit daemon's bootstrap process only once daemon is ready
1853 - Export more information via ``LUInstanceQuery``/remote API
1854 - Improved documentation, QA and unittests
1855 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1856 restart if the file was created or changed
1857 - Added LUXI protocol version sent with each request and response,
1858 allowing detection of server/client mismatches
1859 - Moved the Python scripts among gnt-* and ganeti-* into modules
1860 - Moved all code related to setting up SSH to an external script,
1862 - Infrastructure changes for node group support in future versions
1868 *(Released Fri, 19 Nov 2010)*
1870 A few small bugs fixed, and some improvements to the build system:
1872 - Fix documentation regarding conversion to drbd
1873 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1875 - Fix error handling in node modify with multiple changes
1876 - Allow remote imports without checked names
1882 *(Released Tue, 19 Oct 2010)*
1884 - Disable SSL session ID cache in RPC client
1890 *(Released Thu, 14 Oct 2010)*
1892 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1893 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1895 - Fix problems with interaction between Curl and Python's HTTP server,
1896 resulting in increased speed in many RPC calls
1897 - Improve our release script to prevent breakage with older aclocal and
1904 *(Released Thu, 7 Oct 2010)*
1906 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1907 value retrieved from master
1908 - Added support for blacklisted or hidden OS definitions
1909 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1910 - Added support for -mem-path in KVM hypervisor abstraction layer
1911 - Allow overriding instance parameters in tool for inter-cluster
1912 instance moves (``tools/move-instance``)
1913 - Improved opcode summaries (e.g. in ``gnt-job list``)
1914 - Improve consistency of OS listing by sorting it
1915 - Documentation updates
1921 *(Released Fri, 8 Oct 2010)*
1923 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1929 *(Released Mon, 4 Oct 2010)*
1931 - Fixed regression in ``gnt-instance rename``
1937 *(Released Wed, 22 Sep 2010)*
1939 - Fixed OS_VARIANT variable for OS scripts
1940 - Fixed cluster tag operations via RAPI
1941 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1942 - Disabled RAPI CA checks in watcher
1948 *(Released Mon, 23 Aug 2010)*
1950 - Support DRBD versions of the format "a.b.c.d"
1952 - Re-introduce support for usage from multiple threads in RAPI client
1953 - Instance renames and modify via RAPI
1954 - Work around race condition between processing and archival in job
1956 - Mark opcodes following failed one as failed, too
1957 - Job field ``lock_status`` was removed due to difficulties making it
1958 work with the changed job queue in Ganeti 2.2; a better way to monitor
1959 locks is expected for a later 2.2.x release
1960 - Fixed dry-run behaviour with many commands
1961 - Support ``ssh-agent`` again when adding nodes
1962 - Many additional bugfixes
1968 *(Released Fri, 30 Jul 2010)*
1970 Important change: the internal RPC mechanism between Ganeti nodes has
1971 changed from using a home-grown http library (based on the Python base
1972 libraries) to use the PycURL library. This requires that PycURL is
1973 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1974 against GnuTLS by default. cURL's support for GnuTLS had known issues
1975 before cURL 7.21.0 and we recommend using the latest cURL release or
1976 linking against OpenSSL. Most other distributions already link PycURL
1977 and cURL against OpenSSL. The command::
1979 python -c 'import pycurl; print pycurl.version'
1981 can be used to determine the libraries PycURL and cURL are linked
1984 Other significant changes:
1986 - Rewrote much of the internals of the job queue, in order to achieve
1987 better parallelism; this decouples job query operations from the job
1988 processing, and it should allow much nicer behaviour of the master
1989 daemon under load, and it also has uncovered some long-standing bugs
1990 related to the job serialisation (now fixed)
1991 - Added a default iallocator setting to the cluster parameters,
1992 eliminating the need to always pass nodes or an iallocator for
1993 operations that require selection of new node(s)
1994 - Added experimental support for the LXC virtualization method
1995 - Added support for OS parameters, which allows the installation of
1996 instances to pass parameter to OS scripts in order to customise the
1998 - Added a hypervisor parameter controlling the migration type (live or
1999 non-live), since hypervisors have various levels of reliability; this
2000 has renamed the 'live' parameter to 'mode'
2001 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
2002 logical volumes, meaning that cluster verify will ignore them and not
2003 flag their presence as errors
2004 - The watcher will now reset the error count for failed instances after
2005 8 hours, thus allowing self-healing if the problem that caused the
2006 instances to be down/fail to start has cleared in the meantime
2007 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2008 check for, and warn, if the drbd module parameter ``usermode_helper``
2009 is not consistent with the cluster-wide setting; this is needed to
2010 make diagnose easier of failed drbd creations
2011 - Started adding base IPv6 support, but this is not yet
2012 enabled/available for use
2013 - Rename operations (cluster, instance) will now return the new name,
2014 which is especially useful if a short name was passed in
2015 - Added support for instance migration in RAPI
2016 - Added a tool to pre-configure nodes for the SSH setup, before joining
2017 them to the cluster; this will allow in the future a simplified model
2018 for node joining (but not yet fully enabled in 2.2); this needs the
2019 paramiko python library
2020 - Fixed handling of name-resolving errors
2021 - Fixed consistency of job results on the error path
2022 - Fixed master-failover race condition when executed multiple times in
2024 - Fixed many bugs related to the job queue (mostly introduced during the
2025 2.2 development cycle, so not all are impacting 2.1)
2026 - Fixed instance migration with missing disk symlinks
2027 - Fixed handling of unknown jobs in ``gnt-job archive``
2028 - And many other small fixes/improvements
2032 - Enhanced both the unittest and the QA coverage
2033 - Switched the opcode validation to a generic model, and extended the
2034 validation to all opcode parameters
2035 - Changed more parts of the code that write shell scripts to use the
2037 - Switched the master daemon to use the asyncore library for the Luxi
2044 *(Released Thu, 17 Jun 2010)*
2046 - Added tool (``move-instance``) and infrastructure to move instances
2047 between separate clusters (see :doc:`separate documentation
2048 <move-instance>` and :doc:`design document <design-2.2>`)
2049 - Added per-request RPC timeout
2050 - RAPI now requires a Content-Type header for requests with a body (e.g.
2051 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2052 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2053 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2055 - Implemented initial support for running Ganeti daemons as separate
2056 users, see configure-time flags ``--with-user-prefix`` and
2057 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2059 - Instances can be removed after export (``gnt-backup export
2060 --remove-instance``)
2061 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2062 key (instead of 1024 bit)
2063 - Added new cluster configuration file for cluster domain secret
2064 - Import/export now use SSL instead of SSH
2065 - Added support for showing estimated time when exporting an instance,
2066 see the ``ganeti-os-interface(7)`` manpage and look for
2073 *(Released Tue, 16 Nov 2010)*
2075 Some more bugfixes. Unless critical bugs occur, this will be the last
2078 - Fix case of MAC special-values
2079 - Fix mac checker regex
2080 - backend: Fix typo causing "out of range" error
2081 - Add missing --units in gnt-instance list man page
2087 *(Released Tue, 24 Aug 2010)*
2090 - Don't ignore secondary node silently on non-mirrored disk templates
2092 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2093 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2094 - Properly document the usb_mouse parameter
2095 - Fix path in ganeti-rapi(8) (issue 116)
2096 - Adjust error message when the ganeti user's .ssh directory is
2098 - Add same-node-check when changing the disk template to drbd
2104 *(Released Fri, 16 Jul 2010)*
2107 - Add an option to only select some reboot types during qa/burnin.
2108 (on some hypervisors consequent reboots are not supported)
2109 - Fix infrequent race condition in master failover. Sometimes the old
2110 master ip address would be still detected as up for a short time
2111 after it was removed, causing failover to fail.
2112 - Decrease mlockall warnings when the ctypes module is missing. On
2113 Python 2.4 we support running even if no ctypes module is installed,
2114 but we were too verbose about this issue.
2115 - Fix building on old distributions, on which man doesn't have a
2117 - Fix RAPI not to ignore the MAC address on instance creation
2118 - Implement the old instance creation format in the RAPI client.
2124 *(Released Thu, 01 Jul 2010)*
2126 A small bugfix release:
2127 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2128 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2129 a non-existing option
2130 - Fix instance create: the --force-variant option was ignored
2131 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2132 - Fix modify node storage with non-FQDN arguments
2133 - Fix RAPI client to authenticate under Python 2.6 when used
2134 for more than 5 requests needing authentication
2135 - Fix gnt-instance modify -t (storage) giving a wrong error message
2136 when converting a non-shutdown drbd instance to plain
2142 *(Released Fri, 18 Jun 2010)*
2144 A small bugfix release:
2146 - Fix live migration of KVM instances started with older Ganeti
2147 versions which had fewer hypervisor parameters
2148 - Fix gnt-instance grow-disk on down instances
2149 - Fix an error-reporting bug during instance migration
2150 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2151 silently ignoring broken ones
2152 - Fix an RPC error reporting bug affecting, for example, RAPI client
2154 - Fix bug triggered by different API version os-es on different nodes
2155 - Fix a bug in instance startup with custom hvparams: OS level
2156 parameters would fail to be applied.
2157 - Fix the RAPI client under Python 2.6 (but more work is needed to
2158 make it work completely well with OpenSSL)
2159 - Fix handling of errors when resolving names from DNS
2165 *(Released Thu, 3 Jun 2010)*
2167 A medium sized development cycle. Some new features, and some
2168 fixes/small improvements/cleanups.
2170 Significant features
2171 ~~~~~~~~~~~~~~~~~~~~
2173 The node deamon now tries to mlock itself into memory, unless the
2174 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2175 its logs, and falls back to console logging. This allows emergency
2176 features such as ``gnt-node powercycle`` to work even in the event of a
2177 broken node disk (tested offlining the disk hosting the node's
2178 filesystem and dropping its memory caches; don't try this at home)
2180 KVM: add vhost-net acceleration support. It can be tested with a new
2181 enough version of the kernel and of qemu-kvm.
2183 KVM: Add instance chrooting feature. If you use privilege dropping for
2184 your VMs you can also now force them to chroot to an empty directory,
2185 before starting the emulated guest.
2187 KVM: Add maximum migration bandwith and maximum downtime tweaking
2188 support (requires a new-enough version of qemu-kvm).
2190 Cluster verify will now warn if the master node doesn't have the master
2191 ip configured on it.
2193 Add a new (incompatible) instance creation request format to RAPI which
2194 supports all parameters (previously only a subset was supported, and it
2195 wasn't possible to extend the old format to accomodate all the new
2196 features. The old format is still supported, and a client can check for
2197 this feature, before using it, by checking for its presence in the
2198 ``features`` RAPI resource.
2200 Now with ancient latin support. Try it passing the ``--roman`` option to
2201 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2202 (requires the python-roman module to be installed, in order to work).
2207 As usual many internal code refactorings, documentation updates, and
2210 - Lots of improvements and cleanups to the experimental Remote API
2211 (RAPI) client library.
2212 - A new unit test suite for the core daemon libraries.
2213 - A fix to creating missing directories makes sure the umask is not
2214 applied anymore. This enforces the same directory permissions
2216 - Better handling terminating daemons with ctrl+c (used when running
2217 them in debugging mode).
2218 - Fix a race condition in live migrating a KVM instance, when stat()
2219 on the old proc status file returned EINVAL, which is an unexpected
2221 - Fixed manpage checking with newer man and utf-8 charachters. But now
2222 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2228 *(Released Fri, 7 May 2010)*
2230 Fix a bug which prevented untagged KVM instances from starting.
2236 *(Released Fri, 7 May 2010)*
2238 Another release with a long development cycle, during which many
2239 different features were added.
2241 Significant features
2242 ~~~~~~~~~~~~~~~~~~~~
2244 The KVM hypervisor now can run the individual instances as non-root, to
2245 reduce the impact of a VM being hijacked due to bugs in the
2246 hypervisor. It is possible to run all instances as a single (non-root)
2247 user, to manually specify a user for each instance, or to dynamically
2248 allocate a user out of a cluster-wide pool to each instance, with the
2249 guarantee that no two instances will run under the same user ID on any
2252 An experimental RAPI client library, that can be used standalone
2253 (without the other Ganeti libraries), is provided in the source tree as
2254 ``lib/rapi/client.py``. Note this client might change its interface in
2255 the future, as we iterate on its capabilities.
2257 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2258 replace the cluster's certificates and crypto keys. This might help in
2259 case they have been compromised, or have simply expired.
2261 A new disk option for instance creation has been added that allows one
2262 to "adopt" currently existing logical volumes, with data
2263 preservation. This should allow easier migration to Ganeti from
2264 unmanaged (or managed via other software) instances.
2266 Another disk improvement is the possibility to convert between redundant
2267 (DRBD) and plain (LVM) disk configuration for an instance. This should
2268 allow better scalability (starting with one node and growing the
2269 cluster, or shrinking a two-node cluster to one node).
2271 A new feature that could help with automated node failovers has been
2272 implemented: if a node sees itself as offline (by querying the master
2273 candidates), it will try to shutdown (hard) all instances and any active
2274 DRBD devices. This reduces the risk of duplicate instances if an
2275 external script automatically failovers the instances on such nodes. To
2276 enable this, the cluster parameter ``maintain_node_health`` should be
2277 enabled; in the future this option (per the name) will enable other
2278 automatic maintenance features.
2280 Instance export/import now will reuse the original instance
2281 specifications for all parameters; that means exporting an instance,
2282 deleting it and the importing it back should give an almost identical
2283 instance. Note that the default import behaviour has changed from
2284 before, where it created only one NIC; now it recreates the original
2287 Cluster verify has added a few new checks: SSL certificates validity,
2288 /etc/hosts consistency across the cluster, etc.
2293 As usual, many internal changes were done, documentation fixes,
2296 - Fixed cluster initialization with disabled cluster storage (regression
2297 introduced in 2.1.1)
2298 - File-based storage supports growing the disks
2299 - Fixed behaviour of node role changes
2300 - Fixed cluster verify for some corner cases, plus a general rewrite of
2301 cluster verify to allow future extension with more checks
2302 - Fixed log spamming by watcher and node daemon (regression introduced
2304 - Fixed possible validation issues when changing the list of enabled
2306 - Fixed cleanup of /etc/hosts during node removal
2307 - Fixed RAPI response for invalid methods
2308 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2309 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2310 from ide disks, etc.)
2311 - Allow OS changes without re-installation (to record a changed OS
2312 outside of Ganeti, or to allow OS renames)
2313 - Allow instance creation without OS installation (useful for example if
2314 the OS will be installed manually, or restored from a backup not in
2316 - Implemented option to make cluster ``copyfile`` use the replication
2318 - Added list of enabled hypervisors to ssconf (possibly useful for
2320 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2322 - A partial form of node re-IP is possible via node readd, which now
2323 allows changed node primary IP
2324 - Command line utilities now show an informational message if the job is
2326 - The logs of the master daemon now show the PID/UID/GID of the
2333 *(Released Fri, 12 Mar 2010)*
2335 During the 2.1.0 long release candidate cycle, a lot of improvements and
2336 changes have accumulated with were released later as 2.1.1.
2341 The node evacuate command (``gnt-node evacuate``) was significantly
2342 rewritten, and as such the IAllocator protocol was changed - a new
2343 request type has been added. This unfortunate change during a stable
2344 series is designed to improve performance of node evacuations; on
2345 clusters with more than about five nodes and which are well-balanced,
2346 evacuation should proceed in parallel for all instances of the node
2347 being evacuated. As such, any existing IAllocator scripts need to be
2348 updated, otherwise the above command will fail due to the unknown
2349 request. The provided "dumb" allocator has not been updated; but the
2350 ganeti-htools package supports the new protocol since version 0.2.4.
2352 Another important change is increased validation of node and instance
2353 names. This might create problems in special cases, if invalid host
2354 names are being used.
2356 Also, a new layer of hypervisor parameters has been added, that sits at
2357 OS level between the cluster defaults and the instance ones. This allows
2358 customisation of virtualization parameters depending on the installed
2359 OS. For example instances with OS 'X' may have a different KVM kernel
2360 (or any other parameter) than the cluster defaults. This is intended to
2361 help managing a multiple OSes on the same cluster, without manual
2362 modification of each instance's parameters.
2364 A tool for merging clusters, ``cluster-merge``, has been added in the
2365 tools sub-directory.
2370 - Improved the int/float conversions that should make the code more
2371 robust in face of errors from the node daemons
2372 - Fixed the remove node code in case of internal configuration errors
2373 - Fixed the node daemon behaviour in face of inconsistent queue
2374 directory (e.g. read-only file-system where we can't open the files
2376 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2377 now it either aborts cleanly or, if given the new "auto_promote"
2378 parameter, will automatically promote other nodes as needed
2379 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2380 completely prevent Ganeti from working
2381 - Fixed bug for instance export when not all disks were successfully
2383 - Fixed behaviour of node add when the new node is slow in starting up
2385 - Fixed handling of signals in the LUXI client, which should improve
2386 behaviour of command-line scripts
2387 - Added checks for invalid node/instance names in the configuration (now
2388 flagged during cluster verify)
2389 - Fixed watcher behaviour for disk activation errors
2390 - Fixed two potentially endless loops in http library, which led to the
2391 RAPI daemon hanging and consuming 100% CPU in some cases
2392 - Fixed bug in RAPI daemon related to hashed passwords
2393 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2395 - Enhanced compatibility with non-Debian OSes, but not using absolute
2396 path in some commands and allowing customisation of the ssh
2397 configuration directory
2398 - Fixed possible future issue with new Python versions by abiding to the
2399 proper use of ``__slots__`` attribute on classes
2400 - Added checks that should prevent directory traversal attacks
2401 - Many documentation fixes based on feedback from users
2406 - Added an "early_release" more for instance replace disks and node
2407 evacuate, where we release locks earlier and thus allow higher
2408 parallelism within the cluster
2409 - Added watcher hooks, intended to allow the watcher to restart other
2410 daemons (e.g. from the ganeti-nbma project), but they can be used of
2411 course for any other purpose
2412 - Added a compile-time disable for DRBD barriers, to increase
2413 performance if the administrator trusts the power supply or the
2414 storage system to not lose writes
2415 - Added the option of using syslog for logging instead of, or in
2416 addition to, Ganeti's own log files
2417 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2418 can indeed boot from a paravirtual NIC
2419 - Added a generic debug level for many operations; while this is not
2420 used widely yet, it allows one to pass the debug value all the way to
2422 - Enhanced the hooks environment for instance moves (failovers,
2423 migrations) where the primary/secondary nodes changed during the
2424 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2425 - Enhanced data validations for many user-supplied values; one important
2426 item is the restrictions imposed on instance and node names, which
2427 might reject some (invalid) host names
2428 - Add a configure-time option to disable file-based storage, if it's not
2429 needed; this allows greater security separation between the master
2430 node and the other nodes from the point of view of the inter-node RPC
2432 - Added user notification in interactive tools if job is waiting in the
2433 job queue or trying to acquire locks
2434 - Added log messages when a job is waiting for locks
2435 - Added filtering by node tags in instance operations which admit
2436 multiple instances (start, stop, reboot, reinstall)
2437 - Added a new tool for cluster mergers, ``cluster-merge``
2438 - Parameters from command line which are of the form ``a=b,c=d`` can now
2439 use backslash escapes to pass in values which contain commas,
2440 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2442 - For KVM, the instance name is the first parameter passed to KVM, so
2443 that it's more visible in the process list
2449 *(Released Tue, 2 Mar 2010)*
2451 Ganeti 2.1 brings many improvements with it. Major changes:
2453 - Added infrastructure to ease automated disk repairs
2454 - Added new daemon to export configuration data in a cheaper way than
2455 using the remote API
2456 - Instance NICs can now be routed instead of being associated with a
2458 - Improved job locking logic to reduce impact of jobs acquiring multiple
2459 locks waiting for other long-running jobs
2461 In-depth implementation details can be found in the Ganeti 2.1 design
2467 - Added chroot hypervisor
2468 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2470 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2471 ``bootloader_path`` and ``bootloader_args``)
2472 - Added the ``use_localtime`` option for the xen-hvm and kvm
2473 hypervisors, and the default value for this has changed to false (in
2474 2.0 xen-hvm always enabled it)
2475 - Added luxi call to submit multiple jobs in one go
2476 - Added cluster initialization option to not modify ``/etc/hosts``
2478 - Added network interface parameters
2479 - Added dry run mode to some LUs
2480 - Added RAPI resources:
2482 - ``/2/instances/[instance_name]/info``
2483 - ``/2/instances/[instance_name]/replace-disks``
2484 - ``/2/nodes/[node_name]/evacuate``
2485 - ``/2/nodes/[node_name]/migrate``
2486 - ``/2/nodes/[node_name]/role``
2487 - ``/2/nodes/[node_name]/storage``
2488 - ``/2/nodes/[node_name]/storage/modify``
2489 - ``/2/nodes/[node_name]/storage/repair``
2491 - Added OpCodes to evacuate or migrate all instances on a node
2492 - Added new command to list storage elements on nodes (``gnt-node
2493 list-storage``) and modify them (``gnt-node modify-storage``)
2494 - Added new ssconf files with master candidate IP address
2495 (``ssconf_master_candidates_ips``), node primary IP address
2496 (``ssconf_node_primary_ips``) and node secondary IP address
2497 (``ssconf_node_secondary_ips``)
2498 - Added ``ganeti-confd`` and a client library to query the Ganeti
2499 configuration via UDP
2500 - Added ability to run hooks after cluster initialization and before
2502 - Added automatic mode for disk replace (``gnt-instance replace-disks
2504 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2505 after catastrophic data-loss
2506 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2508 - Added ``gnt-instance move`` command to move instances
2509 - Added ``gnt-cluster watcher`` command to control watcher
2510 - Added ``gnt-node powercycle`` command to powercycle nodes
2511 - Added new job status field ``lock_status``
2512 - Added parseable error codes to cluster verification (``gnt-cluster
2513 verify --error-codes``) and made output less verbose (use
2514 ``--verbose`` to restore previous behaviour)
2515 - Added UUIDs to the main config entities (cluster, nodes, instances)
2516 - Added support for OS variants
2517 - Added support for hashed passwords in the Ganeti remote API users file
2519 - Added option to specify maximum timeout on instance shutdown
2520 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2521 - Added new helper script to start and stop Ganeti daemons
2522 (``daemon-util``), with the intent to reduce the work necessary to
2523 adjust Ganeti for non-Debian distributions and to start/stop daemons
2525 - Added more unittests
2526 - Fixed critical bug in ganeti-masterd startup
2527 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2528 now customisable at the cluster level for both the KVM and Xen
2529 hypervisors using the new ``migration_port`` parameter
2530 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2531 reinstalling an instance
2532 - Allowed ``@`` in tag names
2533 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2534 - Many documentation updates
2535 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2536 - ``gnt-instance reinstall`` can now reinstall multiple instances
2537 - Updated many command line parameters
2538 - Introduced new OS API version 15
2539 - No longer support a default hypervisor
2540 - Treat virtual LVs as inexistent
2541 - Improved job locking logic to reduce lock contention
2542 - Match instance and node names case insensitively
2543 - Reimplemented bash completion script to be more complete
2550 *(Released Thu, 4 Feb 2010)*
2552 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2553 - Fix a string formatting bug
2554 - Improve safety of the code in some error paths
2555 - Improve data validation in the master of values returned from nodes
2561 *(Released Thu, 17 Dec 2009)*
2563 - Fix security issue due to missing validation of iallocator names; this
2564 allows local and remote execution of arbitrary executables
2565 - Fix failure of gnt-node list during instance removal
2566 - Ship the RAPI documentation in the archive
2572 *(Released Wed, 30 Sep 2009)*
2574 - Fixed many wrong messages
2575 - Fixed a few bugs related to the locking library
2576 - Fixed MAC checking at instance creation time
2577 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2578 - Fixed a few issues related to signal handling in both daemons and
2580 - Fixed the example startup script provided
2581 - Fixed insserv dependencies in the example startup script (patch from
2583 - Fixed handling of drained nodes in the iallocator framework
2584 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2586 - Fixed error related to invalid job IDs in job polling
2587 - Fixed job/opcode persistence on unclean master shutdown
2588 - Fixed handling of partial job processing after unclean master
2590 - Fixed error reporting from LUs, previously all errors were converted
2591 into execution errors
2592 - Fixed error reporting from burnin
2593 - Decreased significantly the memory usage of the job queue
2594 - Optimised slightly multi-job submission
2595 - Optimised slightly opcode loading
2596 - Backported the multi-job submit framework from the development
2597 branch; multi-instance start and stop should be faster
2598 - Added script to clean archived jobs after 21 days; this will reduce
2599 the size of the queue directory
2600 - Added some extra checks in disk size tracking
2601 - Added an example ethers hook script
2602 - Added a cluster parameter that prevents Ganeti from modifying of
2604 - Added more node information to RAPI responses
2605 - Added a ``gnt-job watch`` command that allows following the ouput of a
2607 - Added a bind-address option to ganeti-rapi
2608 - Added more checks to the configuration verify
2609 - Enhanced the burnin script such that some operations can be retried
2611 - Converted instance reinstall to multi-instance model
2617 *(Released Fri, 7 Aug 2009)*
2619 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2620 to allow using the pre-2.0.2 behaviour in activation, if any existing
2621 instances have mismatched disk sizes in the configuration
2622 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2623 any configuration mismatches for disk sizes
2624 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2625 failover to work on two-node clusters
2626 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2628 - Fixed detection of OS script errors in ``gnt-backup export``
2629 - Fixed exit code of ``gnt-backup export``
2635 *(Released Fri, 17 Jul 2009)*
2637 - Added experimental support for stripped logical volumes; this should
2638 enhance performance but comes with a higher complexity in the block
2639 device handling; stripping is only enabled when passing
2640 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2641 affected even in the non-stripped mode
2642 - Improved resiliency against transient failures at the end of DRBD
2643 resyncs, and in general of DRBD resync checks
2644 - Fixed a couple of issues with exports and snapshot errors
2645 - Fixed a couple of issues in instance listing
2646 - Added display of the disk size in ``gnt-instance info``
2647 - Fixed checking for valid OSes in instance creation
2648 - Fixed handling of the "vcpus" parameter in instance listing and in
2649 general of invalid parameters
2650 - Fixed http server library, and thus RAPI, to handle invalid
2651 username/password combinations correctly; this means that now they
2652 report unauthorized for queries too, not only for modifications,
2653 allowing earlier detect of configuration problems
2654 - Added a new "role" node list field, equivalent to the master/master
2655 candidate/drained/offline flags combinations
2656 - Fixed cluster modify and changes of candidate pool size
2657 - Fixed cluster verify error messages for wrong files on regular nodes
2658 - Fixed a couple of issues with node demotion from master candidate role
2659 - Fixed node readd issues
2660 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2661 - Added a new ``--no-voting`` option for masterfailover to fix failover
2662 on two-nodes clusters when the former master node is unreachable
2663 - Added instance reinstall over RAPI
2669 *(Released Tue, 16 Jun 2009)*
2671 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2672 allow re-adding the start in single-user option (regression from 1.2)
2673 - the watcher writes the instance status to a file, to allow monitoring
2674 to report the instance status (from the master) based on cached
2675 results of the watcher's queries; while this can get stale if the
2676 watcher is being locked due to other work on the cluster, this is
2677 still an improvement
2678 - the watcher now also restarts the node daemon and the rapi daemon if
2680 - fixed the watcher to handle full and drained queue cases
2681 - hooks export more instance data in the environment, which helps if
2682 hook scripts need to take action based on the instance's properties
2683 (no longer need to query back into ganeti)
2684 - instance failovers when the instance is stopped do not check for free
2685 RAM, so that failing over a stopped instance is possible in low memory
2687 - rapi uses queries for tags instead of jobs (for less job traffic), and
2688 for cluster tags it won't talk to masterd at all but read them from
2690 - a couple of error handling fixes in RAPI
2691 - drbd handling: improved the error handling of inconsistent disks after
2692 resync to reduce the frequency of "there are some degraded disks for
2693 this instance" messages
2694 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2695 error handling path called a wrong function name)
2701 *(Released Wed, 27 May 2009)*
2703 - no changes from rc5
2709 *(Released Wed, 20 May 2009)*
2711 - fix a couple of bugs (validation, argument checks)
2712 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2713 - some small improvements to RAPI and IAllocator
2714 - make watcher automatically start the master daemon if down
2720 *(Released Mon, 27 Apr 2009)*
2722 - change the OS list to not require locks; this helps with big clusters
2723 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2724 volume group is broken
2725 - ``gnt-instance info``, without any arguments, doesn't run for all
2726 instances anymore; either pass ``--all`` or pass the desired
2727 instances; this helps against mistakes on big clusters where listing
2728 the information for all instances takes a long time
2729 - miscellaneous doc and man pages fixes
2735 *(Released Wed, 8 Apr 2009)*
2737 - Change the internal locking model of some ``gnt-node`` commands, in
2738 order to reduce contention (and blocking of master daemon) when
2739 batching many creation/reinstall jobs
2740 - Fixes to Xen soft reboot
2741 - No longer build documentation at build time, instead distribute it in
2742 the archive, in order to reduce the need for the whole docbook/rst
2749 *(Released Fri, 27 Mar 2009)*
2751 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2752 - Fix watcher startup sequence, improves the behaviour of busy clusters
2753 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2754 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2755 bind address and other places
2756 - Some documentation fixes and updates
2762 *(Released Mon, 2 Mar 2009)*
2764 - More documentation updates, now all docs should be more-or-less
2766 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2768 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2770 - Fix an issue related to $libdir/run/ganeti and cluster creation
2776 *(Released Thu, 19 Feb 2009)*
2778 - Xen PVM and KVM have switched the default value for the instance root
2779 disk to the first partition on the first drive, instead of the whole
2780 drive; this means that the OS installation scripts must be changed
2782 - Man pages have been updated
2783 - RAPI has been switched by default to HTTPS, and the exported functions
2784 should all work correctly
2785 - RAPI v1 has been removed
2786 - Many improvements to the KVM hypervisor
2787 - Block device errors are now better reported
2788 - Many other bugfixes and small improvements
2794 *(Released Mon, 26 Jan 2009)*
2796 - Version 2 is a general rewrite of the code and therefore the
2797 differences are too many to list, see the design document for 2.0 in
2798 the ``doc/`` subdirectory for more details
2799 - In this beta version there is not yet a migration path from 1.2 (there
2800 will be one in the final 2.0 release)
2801 - A few significant changes are:
2803 - all commands are executed by a daemon (``ganeti-masterd``) and the
2804 various ``gnt-*`` commands are just front-ends to it
2805 - all the commands are entered into, and executed from a job queue,
2806 see the ``gnt-job(8)`` manpage
2807 - the RAPI daemon supports read-write operations, secured by basic
2808 HTTP authentication on top of HTTPS
2809 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2810 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2811 to migrate to DRBD 8 first while still running Ganeti 1.2)
2812 - DRBD devices are using statically allocated minor numbers, which
2813 will be assigned to existing instances during the migration process
2814 - there is support for both Xen PVM and Xen HVM instances running on
2816 - KVM virtualization is supported too
2817 - file-based storage has been implemented, which means that it is
2818 possible to run the cluster without LVM and DRBD storage, for
2819 example using a shared filesystem exported from shared storage (and
2820 still have live migration)
2826 *(Released Tue, 13 Jan 2009)*
2828 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2829 - Reuse the old instance mac address by default on instance import, if
2830 the instance name is the same.
2831 - Handle situations in which the node info rpc returns incomplete
2833 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2834 - Improved version of batcher:
2836 - state file support
2837 - instance mac address support
2838 - support for HVM clusters/instances
2840 - Add an option to show the number of cpu sockets and nodes in
2842 - Support OSes that handle more than one version of the OS api (but do
2843 not change the current API in any other way)
2844 - Fix ``gnt-node migrate``
2845 - ``gnt-debug`` man page
2846 - Fixes various more typos and small issues
2847 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2853 *(Released Wed, 24 Sep 2008)*
2855 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2856 type of disk exported to fully virtualized instances.
2857 - provide access to the serial console of HVM instances
2858 - instance auto_balance flag, set by default. If turned off it will
2859 avoid warnings on cluster verify if there is not enough memory to fail
2860 over an instance. in the future it will prevent automatically failing
2861 it over when we will support that.
2862 - batcher tool for instance creation, see ``tools/README.batcher``
2863 - ``gnt-instance reinstall --select-os`` to interactively select a new
2864 operating system when reinstalling an instance.
2865 - when changing the memory amount on instance modify a check has been
2866 added that the instance will be able to start. also warnings are
2867 emitted if the instance will not be able to fail over, if auto_balance
2869 - documentation fixes
2870 - sync fields between ``gnt-instance list/modify/add/import``
2871 - fix a race condition in drbd when the sync speed was set after giving
2872 the device a remote peer.
2878 *(Released Tue, 22 Jul 2008)*
2880 - note: the allowed size and number of tags per object were reduced
2881 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2882 - fixed twisted 8.x compatibility
2883 - fixed ``gnt-instance replace-disks`` with iallocator
2884 - add TCP keepalives on twisted connections to detect restarted nodes
2885 - disk increase support, see ``gnt-instance grow-disk``
2886 - implement bulk node/instance query for RAPI
2887 - add tags in node/instance listing (optional)
2888 - experimental migration (and live migration) support, read the man page
2889 for ``gnt-instance migrate``
2890 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2891 has some small improvements in handling its state file
2897 *(Released Fri, 13 Jun 2008)*
2899 - Experimental readonly, REST-based remote API implementation;
2900 automatically started on master node, TCP port 5080, if enabled by
2901 ``--enable-rapi`` parameter to configure script.
2902 - Instance allocator support. Add and import instance accept a
2903 ``--iallocator`` parameter, and call that instance allocator to decide
2904 which node to use for the instance. The iallocator document describes
2905 what's expected from an allocator script.
2906 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2907 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2908 node is lost there is still enough memory to fail over the instances
2910 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2911 ``gnt-cluster verify``, to check for site-specific compliance. All the
2912 hooks will run, and their output, if any, will be displayed. Any
2913 failing hook will make the verification return an error value.
2914 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2915 primary and secondary interfaces
2916 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2917 that is still declared as part of the cluster and has failed.
2918 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2919 specifying output fields, that just adds the chosen fields to the
2921 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2922 export from the filesystem.
2923 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2924 have been added. Using them you can enable/disable acpi and pae
2925 support, and specify a path for a cd image to be exported to the
2926 instance. These parameters as the name suggest only work on HVM
2928 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2929 PAE support will be set to the previously hardcoded values, but the
2930 (previously hardcoded) path to the CDROM ISO image will be unset and
2931 if required, needs to be set manually with ``gnt-instance modify``
2933 - The address to which an instance's VNC console is bound is now
2934 selectable per-instance, rather than being cluster wide. Of course
2935 this only applies to instances controlled via VNC, so currently just
2936 applies to HVM clusters.
2942 *(Released Mon, 18 Feb 2008)*
2944 - more tweaks to the disk activation code (especially helpful for DRBD)
2945 - change the default ``gnt-instance list`` output format, now there is
2946 one combined status field (see the manpage for the exact values this
2948 - some more fixes for the mac export to hooks change
2949 - make Ganeti not break with DRBD 8.2.x (which changed the version
2950 format in ``/proc/drbd``) (issue 24)
2951 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2952 template, allowing migration from DRBD0.7+MD to DRBD8
2958 *(Released Wed, 30 Jan 2008)*
2960 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2962 - add command aliases infrastructure and a few aliases
2963 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2964 man pages and the ``--help`` option of ``gnt-node
2965 list``/``gnt-instance list``
2966 - fix ``gnt-backup list`` with down nodes (issue 21)
2967 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2968 - fix the dist archive and add a check for including svn/git files in
2970 - some developer-related changes: improve the burnin and the QA suite,
2971 add an upload script for testing during development
2977 *(Released Wed, 16 Jan 2008)*
2979 - experimental HVM support, read the install document, section
2980 "Initializing the cluster"
2981 - allow for the PVM hypervisor per-instance kernel and initrd paths
2982 - add a new command ``gnt-cluster verify-disks`` which uses a new
2983 algorithm to improve the reconnection of the DRBD pairs if the device
2984 on the secondary node has gone away
2985 - make logical volume code auto-activate LVs at disk activation time
2986 - slightly improve the speed of activating disks
2987 - allow specification of the MAC address at instance creation time, and
2988 changing it later via ``gnt-instance modify``
2989 - fix handling of external commands that generate lots of output on
2991 - update documentation with regard to minimum version of DRBD8 supported
2997 *(Released Tue, 4 Dec 2007)*
2999 - Log the ``xm create`` output to the node daemon log on failure (to
3000 help diagnosing the error)
3001 - In debug mode, log all external commands output if failed to the logs
3002 - Change parsing of lvm commands to ignore stderr
3008 *(Released Wed, 28 Nov 2007)*
3010 - Another round of updates to the DRBD 8 code to deal with more failures
3011 in the replace secondary node operation
3012 - Some more logging of failures in disk operations (lvm, drbd)
3013 - A few documentation updates
3020 *(Released Tue, 13 Nov 2007)*
3022 - Change configuration file format from Python's Pickle to JSON.
3023 Upgrading is possible using the cfgupgrade utility.
3024 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3025 faster replace disks and is more stable (DRBD 8 has many improvements
3026 compared to DRBD 0.7)
3027 - Added command line tags support (see man pages for ``gnt-instance``,
3028 ``gnt-node``, ``gnt-cluster``)
3029 - Added instance rename support
3030 - Added multi-instance startup/shutdown
3031 - Added cluster rename support
3032 - Added ``gnt-node evacuate`` to simplify some node operations
3033 - Added instance reboot operation that can speedup reboot as compared to
3035 - Soften the requirement that hostnames are in FQDN format
3036 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3038 - Removed dependency on debian's patched fping that uses the
3039 non-standard ``-S`` option
3040 - Now the OS definitions are searched for in multiple, configurable
3041 paths (easier for distros to package)
3042 - Some changes to the hooks infrastructure (especially the new
3043 post-configuration update hook)
3044 - Other small bugfixes
3046 .. vim: set textwidth=72 syntax=rst :