8 *(Released Wed, 13 Nov 2013)*
10 - fix bug, that kept nodes offline when readding
11 - when verifying DRBD versions, ignore unavailable nodes
12 - fix bug that made the console unavailable on kvm in split-user
14 - DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
20 *(Released Tue, 5 Nov 2013)*
22 Incompatible/important changes
23 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
25 - hroller now also plans for capacity to move non-redundant instances off
26 any node to be rebooted; the old behavior of completely ignoring any
27 non-redundant instances can be restored by adding the --ignore-non-redundant
29 - The cluster option '--no-lvm-storage' was removed in favor of the new option
30 '--enabled-disk-templates'.
31 - On instance creation, disk templates no longer need to be specified
32 with '-t'. The default disk template will be taken from the list of
33 enabled disk templates.
34 - The monitoring daemon is now running as root, in order to be able to collect
35 information only available to root (such as the state of Xen instances).
36 - The ConfD client is now IPv6 compatible.
37 - File and shared file storage is no longer dis/enabled at configure time,
38 but using the option '--enabled-disk-templates' at cluster initialization and
40 - The default directories for file and shared file storage are not anymore
41 specified at configure time, but taken from the cluster's configuration.
42 They can be set at cluster initialization and modification with
43 '--file-storage-dir' and '--shared-file-storage-dir'.
44 - Cluster verification now includes stricter checks regarding the
45 default file and shared file storage directories. It now checks that
46 the directories are explicitely allowed in the 'file-storage-paths' file and
47 that the directories exist on all nodes.
48 - The list of allowed disk templates in the instance policy and the list
49 of cluster-wide enabled disk templates is now checked for consistency
50 on cluster or group modification. On cluster initialization, the ipolicy
51 disk templates are ensured to be a subset of the cluster-wide enabled
57 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
58 the correct command syntax. It is possible to use different DRBD versions
59 on different nodes as long as they are compatible to each other. This
60 enables rolling upgrades of DRBD with no downtime. As permanent operation
61 of different DRBD versions within a node group is discouraged,
62 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
63 - New "inst-status-xen" data collector for the monitoring daemon, providing
64 information about the state of the xen instances on the nodes.
65 - New "lv" data collector for the monitoring daemon, collecting data about the
66 logical volumes on the nodes, and pairing them with the name of the instances
68 - New "diskstats" data collector, collecting the data from /proc/diskstats and
69 presenting them over the monitoring daemon interface.
70 - The ConfD client is now IPv6 compatible.
74 The following new dependencies have been added.
78 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
79 for the unit tests (and only used for testing).
83 - ``hslogger`` (http://software.complete.org/hslogger) is now always
84 required, even if confd is not enabled.
89 - Correctly start/stop luxid during gnt-cluster master-failover (inherited
91 - Improved error messsages (inherited from stable-2.8)
97 *(Released Tue, 15 Oct 2013)*
99 The third release candidate in the 2.9 series. Since 2.9.0 rc2:
101 - in implicit configuration upgrade, match ipolicy with enabled disk templates
102 - improved harep documentation (inherited from stable-2.8)
108 *(Released Wed, 9 Oct 2013)*
110 The second release candidate in the 2.9 series. Since 2.9.0 rc1:
112 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
113 at least one DRBD instance.
114 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
121 *(Released Tue, 1 Oct 2013)*
123 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
126 - update of the documentation, in particular installation instructions
127 - merging of LD_* constants into DT_* constants
128 - python style changes to be compatible with newer versions of pylint
134 *(Released Thu, 29 Aug 2013)*
136 This was the first beta release of the 2.9 series. All important changes
137 are listed in the latest 2.9 entry.
143 *(Released Thu, 12 Dec 2013)*
145 - Fixed Luxi daemon socket permissions after master-failover
146 - Improve IP version detection code directly checking for colons rather than
147 passing the family from the cluster object
148 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
149 opportunistically anymore (Issue 622)
150 - Allow link local IPv6 gateways (Issue 624)
151 - Fix error printing (Issue 616)
152 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
153 disk modifications, keep the old one. If name=none then set disk name to
155 - Update build_chroot script to work with the latest hackage packages
156 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
157 - Fix evacuation out of drained node (Issue 615)
158 - Add default file_driver if missing (Issue 571)
159 - Fix job error message after unclean master shutdown (Issue 618)
160 - Lock group(s) when creating instances (Issue 621)
161 - SetDiskID() before accepting an instance (Issue 633)
162 - Allow the ext template disks to receive arbitrary parameters, both at creation
163 time and while being modified
164 - Xen handle domain shutdown (future proofing cherry-pick)
165 - Refactor reading live data in htools (future proofing cherry-pick)
171 *(Released Thu, 07 Nov 2013)*
173 - DRBD: ensure peers are UpToDate for dual-primary
174 - Improve error message for replace-disks
175 - More dependency checks at configure time
176 - Placate warnings on ganeti.outils_unittest.py
182 *(Released Thu, 17 Oct 2013)*
184 - Correctly start/stop luxid during gnt-cluster master-failover
185 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
186 - Fix path for the job queue serial file
187 - Improved harep man page
188 - Minor documentation improvements
194 *(Released Mon, 30 Sep 2013)*
196 Incompatible/important changes
197 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
199 - Instance policy can contain multiple instance specs, as described in
200 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
201 <design-partitioned>`. As a consequence, it's not possible to partially change
202 or override instance specs. Bounding specs (min and max) can be specified as a
203 whole using the new option ``--ipolicy-bounds-specs``, while standard
204 specs use the new option ``--ipolicy-std-specs``.
205 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
206 gnt-instance is a valid YAML object.
207 - hail now honors network restrictions when allocating nodes. This led to an
208 update of the IAllocator protocol. See the IAllocator documentation for
210 - confd now only answers static configuration request over the network. luxid
211 was extracted, listens on the local LUXI socket and responds to live queries.
212 This allows finer grained permissions if using separate users.
217 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
218 to always require authentication, ``--require-authentication``. It can
219 be specified in ``$sysconfdir/default/ganeti``.
220 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
221 be used to manage the disk templates to be used by instances in the cluster.
222 Initially, it will be set to a list that includes plain, drbd, if they were
223 enabled by specifying a volume group name, and file and sharedfile, if those
224 were enabled at configure time. Additionally, it will include all disk
225 templates that are currently used by instances. The order of disk templates
226 will be based on Ganeti's history of supporting them. In the future, the
227 first entry of the list will be used as a default disk template on instance
229 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
230 configuration back to the previous stable version.
231 - Disk templates in group ipolicy can be restored to the default value.
232 - Initial support for diskless instances and virtual clusters in QA.
233 - More QA and unit tests for instance policies.
234 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
235 describing why the opcode itself was executed.
236 - The monitoring daemon is now available. It allows users to query the cluster
237 for obtaining information about the status of the system. The daemon is only
238 responsible for providing the information over the network: the actual data
239 gathering is performed by data collectors (currently, only the DRBD status
240 collector is available).
241 - In order to help developers work on Ganeti, a new script
242 (``devel/build_chroot``) is provided, for building a chroot that contains all
243 the required development libraries and tools for compiling Ganeti on a Debian
245 - A new tool, ``harep``, for performing self-repair and recreation of instances
246 in Ganeti has been added.
247 - Split queries are enabled for tags, network, exports, cluster info, groups,
249 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
250 It prints the command line to set the current policies, to ease
252 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
253 devices for KVM virtio-net interfaces will get created with VNET_HDR
254 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
255 virtio-net interfaces, which prevents host kernel tainting and log
256 flooding, when dealing with broken or malicious virtio-net drivers.
257 It's set to true by default.
258 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
260 - Support 'viridian' parameter in Xen HVM
261 - Support DSA SSH keys in bootstrap
262 - To simplify the work of packaging frameworks that want to add the needed users
263 and groups in a split-user setup themselves, at build time three files in
264 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
265 the groups to be generated, the ``users`` file contains, one per line, the
266 users to be generated, optionally followed by their primary group, where
267 important. The ``groupmemberships`` file contains, one per line, additional
268 user-group membership relations that need to be established. The syntax of
269 these files will remain stable in all future versions.
274 The following new dependencies have been added:
277 - The ``curl`` library is not optional anymore for compiling the Haskell code.
278 - ``snap-server`` library (if monitoring is enabled).
281 - The minimum Python version needed to run Ganeti is now 2.6.
282 - ``yaml`` library (only for running the QA).
286 - Perform proper cleanup on termination of Haskell daemons
287 - Fix corner-case in handling of remaining retry time
293 *(Released Tue, 17 Sep 2013)*
295 - To simplify the work of packaging frameworks that want to add the needed users
296 and groups in a split-user setup themselves, at build time three files in
297 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
298 the groups to be generated, the ``users`` file contains, one per line, the
299 users to be generated, optionally followed by their primary group, where
300 important. The ``groupmemberships`` file contains, one per line, additional
301 user-group membership relations that need to be established. The syntax of
302 these files will remain stable in all future versions.
303 - Add a default to file-driver when unspecified over RAPI (Issue 571)
304 - Mark the DSA host pubkey as optional, and remove it during config downgrade
306 - Some documentation fixes
312 *(Released Tue, 27 Aug 2013)*
314 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
316 - Support 'viridian' parameter in Xen HVM (Issue 233)
317 - Include VCS version in ``gnt-cluster version``
318 - Support DSA SSH keys in bootstrap (Issue 338)
319 - Fix batch creation of instances
320 - Use FQDN to check master node status (Issue 551)
321 - Make the DRBD collector more failure-resilient
327 *(Released Fri, 2 Aug 2013)*
329 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
331 - Fix upgrading/downgrading from 2.7
332 - Increase maximum RAPI message size
333 - Documentation updates
334 - Split ``confd`` between ``luxid`` and ``confd``
335 - Merge 2.7 series up to the 2.7.1 release
336 - Allow the ``modify_etc_hosts`` option to be changed
337 - Add better debugging for ``luxid`` queries
338 - Expose bulk parameter for GetJobs in RAPI client
339 - Expose missing ``network`` fields in RAPI
340 - Add some ``cluster verify`` tests
341 - Some unittest fixes
342 - Fix a malfunction in ``hspace``'s tiered allocation
343 - Fix query compatibility between haskell and python implementations
344 - Add the ``vnet_hdr`` HV parameter for KVM
345 - Add ``--cleanup`` to instance failover
346 - Change the connected groups format in ``gnt-network info`` output; it
347 was previously displayed as a raw list by mistake. (Merged from 2.7)
353 *(Released Mon, 24 Jun 2013)*
355 This was the first beta release of the 2.8 series. All important changes
356 are listed in the latest 2.8 entry.
362 *(Released Thu, 26 Sep 2013)*
364 - Change the connected groups format in ``gnt-network info`` output; it
365 was previously displayed as a raw list by mistake
366 - Check disk template in right dict when copying
367 - Support multi-instance allocs without iallocator
368 - Fix some errors in the documentation
369 - Fix formatting of tuple in an error message
375 *(Released Thu, 25 Jul 2013)*
377 - Add logrotate functionality in daemon-util
378 - Add logrotate example file
379 - Add missing fields to network queries over rapi
380 - Fix network object timestamps
381 - Add support for querying network timestamps
382 - Fix a typo in the example crontab
383 - Fix a documentation typo
389 *(Released Thu, 04 Jul 2013)*
391 Incompatible/important changes
392 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
394 - Instance policies for disk size were documented to be on a per-disk
395 basis, but hail applied them to the sum of all disks. This has been
397 - ``hbal`` will now exit with status 0 if, during job execution over
398 LUXI, early exit has been requested and all jobs are successful;
399 before, exit status 1 was used, which cannot be differentiated from
401 - Compatibility with newer versions of rbd has been fixed
402 - ``gnt-instance batch-create`` has been changed to use the bulk create
403 opcode from Ganeti. This lead to incompatible changes in the format of
404 the JSON file. It's now not a custom dict anymore but a dict
405 compatible with the ``OpInstanceCreate`` opcode.
406 - Parent directories for file storage need to be listed in
407 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
408 write the file automatically based on old configuration values, but it
409 can not distribute it across all nodes and the file contents should be
410 verified. Use ``gnt-cluster copyfile
411 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
412 upgraded. The reason for requiring this list of paths now is that
413 before it would have been possible to inject new paths via RPC,
414 allowing files to be created in arbitrary locations. The RPC protocol
415 is protected using SSL/X.509 certificates, but as a design principle
416 Ganeti does not permit arbitrary paths to be passed.
417 - The parsing of the variants file for OSes (see
418 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
419 lines and comment lines (starting with ``#``) are ignored for better
421 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
422 longer available. ``gnt-node add`` now invokes a new tool on the
423 destination node, named ``prepare-node-join``, to configure the SSH
424 daemon. Paramiko is no longer necessary to configure nodes' SSH
425 daemons via ``gnt-node add``.
426 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
427 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
428 the flag is not reset after a master failover.
429 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
430 is recommended. 2.6 will be mandatory from the 2.8 series.
436 - New network management functionality to support automatic allocation
437 of IP addresses and managing of network parameters. See
438 :manpage:`gnt-network(8)` for more details.
439 - New external storage backend, to allow managing arbitrary storage
440 systems external to the cluster. See
441 :manpage:`ganeti-extstorage-interface(7)`.
442 - New ``exclusive-storage`` node parameter added, restricted to
443 nodegroup level. When it's set to true, physical disks are assigned in
444 an exclusive fashion to instances, as documented in :doc:`Partitioned
445 Ganeti <design-partitioned>`. Currently, only instances using the
446 ``plain`` disk template are supported.
447 - The KVM hypervisor has been updated with many new hypervisor
448 parameters, including a generic one for passing arbitrary command line
449 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
450 compatible up to qemu 1.4.
451 - A new tool, called ``mon-collector``, is the stand-alone executor of
452 the data collectors for a monitoring system. As of this version, it
453 just includes the DRBD data collector, that can be executed by calling
454 ``mon-collector`` using the ``drbd`` parameter. See
455 :manpage:`mon-collector(7)`.
456 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
457 for RAPI users. It allows granting permissions to query for
458 information to a specific user without giving
459 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
460 - A new tool named ``node-cleanup`` has been added. It cleans remains of
461 a cluster from a machine by stopping all daemons, removing
462 certificates and ssconf files. Unless the ``--no-backup`` option is
463 given, copies of the certificates are made.
464 - Instance creations now support the use of opportunistic locking,
465 potentially speeding up the (parallel) creation of multiple instances.
466 This feature is currently only available via the :doc:`RAPI
467 <rapi>` interface and when an instance allocator is used. If the
468 ``opportunistic_locking`` parameter is set the opcode will try to
469 acquire as many locks as possible, but will not wait for any locks
470 held by other opcodes. If not enough resources can be found to
471 allocate the instance, the temporary error code
472 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
473 retried thereafter, with or without opportunistic locking.
474 - New experimental linux-ha resource scripts.
475 - Restricted-commands support: ganeti can now be asked (via command line
476 or rapi) to perform commands on a node. These are passed via ganeti
477 RPC rather than ssh. This functionality is restricted to commands
478 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
479 reasons. The file is not copied automatically.
485 - Diskless instances are now externally mirrored (Issue 237). This for
486 now has only been tested in conjunction with explicit target nodes for
488 - Queries not needing locks or RPC access to the node can now be
489 performed by the confd daemon, making them independent from jobs, and
490 thus faster to execute. This is selectable at configure time.
491 - The functionality for allocating multiple instances at once has been
492 overhauled and is now also available through :doc:`RAPI <rapi>`.
494 There are no significant changes from version 2.7.0~rc3.
500 *(Released Tue, 25 Jun 2013)*
502 - Fix permissions on the confd query socket (Issue 477)
503 - Fix permissions on the job archive dir (Issue 498)
504 - Fix handling of an internal exception in replace-disks (Issue 472)
505 - Fix gnt-node info handling of shortened names (Issue 497)
506 - Fix gnt-instance grow-disk when wiping is enabled
507 - Documentation improvements, and support for newer pandoc
508 - Fix hspace honoring ipolicy for disks (Issue 484)
509 - Improve handling of the ``kvm_extra`` HV parameter
515 *(Released Fri, 24 May 2013)*
517 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
519 - Disks added through ``gnt-instance modify`` or created through
520 ``gnt-instance recreate-disks`` are wiped, if the
521 ``prealloc_wipe_disks`` flag is set.
522 - If wiping newly created disks fails, the disks are removed. Also,
523 partial failures in creating disks through ``gnt-instance modify``
524 triggers a cleanup of the partially-created disks.
525 - Removing the master IP address doesn't fail if the address has been
527 - Fix ownership of the OS log dir
528 - Workaround missing SO_PEERCRED constant (Issue 191)
534 *(Released Fri, 3 May 2013)*
536 This was the first release candidate of the 2.7 series. Since beta3:
538 - Fix kvm compatibility with qemu 1.4 (Issue 389)
539 - Documentation updates (admin guide, upgrade notes, install
540 instructions) (Issue 372)
541 - Fix gnt-group list nodes and instances count (Issue 436)
542 - Fix compilation without non-mandatory libraries (Issue 441)
543 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
544 - Make confd logging more verbose at INFO level (Issue 435)
545 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
546 - Fix failure path for instance storage type conversion (Issue 229)
547 - Update htools text backend documentation
548 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
549 - Disable inter-cluster instance move for file-based instances, because
550 it is dependant on instance export, which is not supported for
551 file-based instances. (Issue 414)
552 - Fix gnt-job crashes on non-ascii characters (Issue 427)
553 - Fix volume group checks on non-vm-capable nodes (Issue 432)
559 *(Released Mon, 22 Apr 2013)*
561 This was the third beta release of the 2.7 series. Since beta2:
563 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
564 - Fix data loss on wrong usage of ``gnt-instance move``
565 - Properly export errors in confd-based job queries
566 - Add ``users-setup`` tool
567 - Fix iallocator protocol to report 0 as a disk size for diskless
568 instances. This avoids hail breaking when a diskless instance is
570 - Fix job queue directory permission problem that made confd job queries
571 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
572 for access to archived jobs (Issue 406).
573 - Limit the sizes of networks supported by ``gnt-network`` to something
574 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
575 - Fix bugs in instance disk template conversion
576 - Fix GHC 7 compatibility
577 - Fix ``burnin`` install path (Issue 426).
578 - Allow very small disk grows (Issue 347).
579 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
580 that noded doesn't import masterd code (Issue 419).
581 - Make sure the default metavg at cluster init is the same as the vg, if
582 unspecified (Issue 358).
583 - Fix cleanup of partially created disks (part of Issue 416)
589 *(Released Tue, 2 Apr 2013)*
591 This was the second beta release of the 2.7 series. Since beta1:
593 - Networks no longer have a "type" slot, since this information was
594 unused in Ganeti: instead of it tags should be used.
595 - The rapi client now has a ``target_node`` option to MigrateInstance.
596 - Fix early exit return code for hbal (Issue 386).
597 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
598 - Fix ``rbd showmapped`` output parsing (Issue 312).
599 - Networks are now referenced indexed by UUID, rather than name. This
600 will require running cfgupgrade, from 2.7.0beta1, if networks are in
602 - The OS environment now includes network information.
603 - Deleting of a network is now disallowed if any instance nic is using
604 it, to prevent dangling references.
605 - External storage is now documented in man pages.
606 - The exclusive_storage flag can now only be set at nodegroup level.
607 - Hbal can now submit an explicit priority with its jobs.
608 - Many network related locking fixes.
609 - Bump up the required pylint version to 0.25.1.
610 - Fix the ``no_remember`` option in RAPI client.
611 - Many ipolicy related tests, qa, and fixes.
612 - Many documentation improvements and fixes.
613 - Fix building with ``--disable-file-storage``.
614 - Fix ``-q`` option in htools, which was broken if passed more than
616 - Some haskell/python interaction improvements and fixes.
617 - Fix iallocator in case of missing LVM storage.
618 - Fix confd config load in case of ``--no-lvm-storage``.
619 - The confd/query functionality is now mentioned in the security
626 *(Released Wed, 6 Feb 2013)*
628 This was the first beta release of the 2.7 series. All important changes
629 are listed in the latest 2.7 entry.
635 *(Released Fri, 21 Dec 2012)*
637 Important behaviour change: hbal won't rebalance anymore instances which
638 have the ``auto_balance`` attribute set to false. This was the intention
639 all along, but until now it only skipped those from the N+1 memory
640 reservation (DRBD-specific).
642 A significant number of bug fixes in this release:
644 - Fixed disk adoption interaction with ipolicy checks.
645 - Fixed networking issues when instances are started, stopped or
646 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
647 - Fixed the warning in cluster verify for shared storage instances not
649 - Fixed removal of storage directory on shared file storage (issue 262).
650 - Fixed validation of LVM volume group name in OpClusterSetParams
651 (``gnt-cluster modify``) (issue 285).
652 - Fixed runtime memory increases (``gnt-instance modify -m``).
653 - Fixed live migration under Xen's ``xl`` mode.
654 - Fixed ``gnt-instance console`` with ``xl``.
655 - Fixed building with newer Haskell compiler/libraries.
656 - Fixed PID file writing in Haskell daemons (confd); this prevents
657 restart issues if confd was launched manually (outside of
658 ``daemon-util``) while another copy of it was running
659 - Fixed a type error when doing live migrations with KVM (issue 297) and
660 the error messages for failing migrations have been improved.
661 - Fixed opcode validation for the out-of-band commands (``gnt-node
663 - Fixed a type error when unsetting OS hypervisor parameters (issue
664 311); now it's possible to unset all OS-specific hypervisor
666 - Fixed the ``dry-run`` mode for many operations: verification of
667 results was over-zealous but didn't take into account the ``dry-run``
668 operation, resulting in "wrong" failures.
669 - Fixed bash completion in ``gnt-job list`` when the job queue has
670 hundreds of entries; especially with older ``bash`` versions, this
671 results in significant CPU usage.
673 And lastly, a few other improvements have been made:
675 - Added option to force master-failover without voting (issue 282).
676 - Clarified error message on lock conflict (issue 287).
677 - Logging of newly submitted jobs has been improved (issue 290).
678 - Hostname checks have been made uniform between instance rename and
680 - The ``--submit`` option is now supported by ``gnt-debug delay``.
681 - Shutting down the master daemon by sending SIGTERM now stops it from
682 processing jobs waiting for locks; instead, those jobs will be started
683 once again after the master daemon is started the next time (issue
685 - Support for Xen's ``xl`` program has been improved (besides the fixes
687 - Reduced logging noise in the Haskell confd daemon (only show one log
688 entry for each config reload, instead of two).
689 - Several man page updates and typo fixes.
695 *(Released Fri, 12 Oct 2012)*
697 A small bugfix release. Among the bugs fixed:
699 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
700 made the command unusable.
701 - Commands that issue many jobs don't fail anymore just because some jobs
702 take so long that other jobs are archived.
703 - Failures during ``gnt-instance reinstall`` are reflected by the exit
705 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
707 - When ``always_failover`` is set, ``--allow-failover`` is not required
708 in migrate commands anymore.
709 - ``bash_completion`` works even if extglob is disabled.
710 - Fixed bug with locks that made failover for RDB-based instances fail.
711 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
712 a random node instead of one based on the allocator metric.
713 - Support for newer versions of pylint and pep8.
714 - Hail doesn't fail anymore when trying to add an instance of type
715 ``file``, ``sharedfile`` or ``rbd``.
716 - Added new Makefile target to rebuild the whole distribution, so that
717 all files are included.
723 *(Released Fri, 27 Jul 2012)*
726 .. attention:: The ``LUXI`` protocol has been made more consistent
727 regarding its handling of command arguments. This, however, leads to
728 incompatibility issues with previous versions. Please ensure that you
729 restart Ganeti daemons soon after the upgrade, otherwise most
730 ``LUXI`` calls (job submission, setting/resetting the drain flag,
731 pausing/resuming the watcher, cancelling and archiving jobs, querying
732 the cluster configuration) will fail.
741 The current ``admin_up`` field, which used to denote whether an instance
742 should be running or not, has been removed. Instead, ``admin_state`` is
743 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
745 The rational behind this is that an instance being “down” can have
748 - it could be down during a reboot
749 - it could be temporarily be down for a reinstall
750 - or it could be down because it is deprecated and kept just for its
753 The previous Boolean state was making it difficult to do capacity
754 calculations: should Ganeti reserve memory for a down instance? Now, the
755 tri-state field makes it clear:
757 - in ``up`` and ``down`` state, all resources are reserved for the
758 instance, and it can be at any time brought up if it is down
759 - in ``offline`` state, only disk space is reserved for it, but not
762 The field can have an extra use: since the transition between ``up`` and
763 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
764 transition between ``offline`` and ``down`` is done via ``gnt-instance
765 modify``, it is possible to given different rights to users. For
766 example, owners of an instance could be allowed to start/stop it, but
767 not transition it out of the offline state.
769 Instance policies and specs
770 +++++++++++++++++++++++++++
772 In previous Ganeti versions, an instance creation request was not
773 limited on the minimum size and on the maximum size just by the cluster
774 resources. As such, any policy could be implemented only in third-party
775 clients (RAPI clients, or shell wrappers over ``gnt-*``
776 tools). Furthermore, calculating cluster capacity via ``hspace`` again
777 required external input with regards to instance sizes.
779 In order to improve these workflows and to allow for example better
780 per-node group differentiation, we introduced instance specs, which
783 - minimum instance disk size, disk count, memory size, cpu count
784 - maximum values for the above metrics
785 - and “standard” values (used in ``hspace`` to calculate the standard
788 The minimum/maximum values can be also customised at node-group level,
789 for example allowing more powerful hardware to support bigger instance
792 Beside the instance specs, there are a few other settings belonging to
793 the instance policy framework. It is possible now to customise, per
794 cluster and node-group:
796 - the list of allowed disk templates
797 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
798 - the maximum ratio of instance to spindles (see below for more
799 information) for local storage
801 All these together should allow all tools that talk to Ganeti to know
802 what are the ranges of allowed values for instances and the
803 over-subscription that is allowed.
805 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
806 instance configuration, and the physical CPU configuration from the
807 node. For the spindle ratios however, we didn't track before these
808 values, so new parameters have been added:
810 - a new node parameter ``spindle_count``, defaults to 1, customisable at
811 node group or node level
812 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
814 Note that spindles in this context doesn't need to mean actual
815 mechanical hard-drives; it's just a relative number for both the node
816 I/O capacity and instance I/O consumption.
818 Instance migration behaviour
819 ++++++++++++++++++++++++++++
821 While live-migration is in general desirable over failover, it is
822 possible that for some workloads it is actually worse, due to the
823 variable time of the “suspend” phase during live migration.
825 To allow the tools to work consistently over such instances (without
826 having to hard-code instance names), a new backend parameter
827 ``always_failover`` has been added to control the migration/failover
828 behaviour. When set to True, all migration requests for an instance will
829 instead fall-back to failover.
831 Instance memory ballooning
832 ++++++++++++++++++++++++++
834 Initial support for memory ballooning has been added. The memory for an
835 instance is no longer fixed (backend parameter ``memory``), but instead
836 can vary between minimum and maximum values (backend parameters
837 ``minmem`` and ``maxmem``). Currently we only change an instance's
840 - live migrating or failing over and instance and the target node
841 doesn't have enough memory
842 - user requests changing the memory via ``gnt-instance modify
848 In order to control the use of specific CPUs by instance, support for
849 controlling CPU pinning has been added for the Xen, HVM and LXC
850 hypervisors. This is controlled by a new hypervisor parameter
851 ``cpu_mask``; details about possible values for this are in the
852 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
853 VCPU-to-CPU mapping) form will work well only when all nodes in your
854 cluster have the same amount of CPUs.
859 Another area in which Ganeti was not customisable were the parameters
860 used for storage configuration, e.g. how many stripes to use for LVM,
861 DRBD resync configuration, etc.
863 To improve this area, we've added disks parameters, which are
864 customisable at cluster and node group level, and which allow to
865 specify various parameters for disks (DRBD has the most parameters
866 currently), for example:
868 - DRBD resync algorithm and parameters (e.g. speed)
869 - the default VG for meta-data volumes for DRBD
870 - number of stripes for LVM (plain disk template)
873 These parameters can be modified via ``gnt-cluster modify -D …`` and
874 ``gnt-group modify -D …``, and are used at either instance creation (in
875 case of LVM stripes, for example) or at disk “activation” time
878 Rados block device support
879 ++++++++++++++++++++++++++
881 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
882 denoted by the ``rbd`` disk template type. This is considered
883 experimental, feedback is welcome. For details on configuring it, see
884 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
889 The existing master IP functionality works well only in simple setups (a
890 single network shared by all nodes); however, if nodes belong to
891 different networks, then the ``/32`` setup and lack of routing
892 information is not enough.
894 To allow the master IP to function well in more complex cases, the
895 system was reworked as follows:
897 - a master IP netmask setting has been added
898 - the master IP activation/turn-down code was moved from the node daemon
900 - whether to run the Ganeti-supplied master IP script or a user-supplied
901 on is a ``gnt-cluster init`` setting
903 Details about the location of the standard and custom setup scripts are
904 in the man page :manpage:`gnt-cluster(8)`; for information about the
905 setup script protocol, look at the Ganeti-supplied script.
910 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
913 It is now possible to use TLS-protected connections, and when renewing
914 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
915 it is now possible to specify spice or spice CA certificates. Also, it
916 is possible to configure a password for SPICE sessions via the
917 hypervisor parameter ``spice_password_file``.
919 There are also new parameters to control the compression and streaming
920 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
921 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
922 for the spice parameters.
924 Lastly, it is now possible to see the SPICE connection information via
925 ``gnt-instance console``.
930 A new tool (``tools/ovfconverter``) has been added that supports
931 conversion between Ganeti and the `Open Virtualization Format
932 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
935 This relies on the ``qemu-img`` tool to convert the disk formats, so the
936 actual compatibility with other virtualization solutions depends on it.
941 The configuration query daemon (``ganeti-confd``) is now optional, and
942 has been rewritten in Haskell; whether to use the daemon at all, use the
943 Python (default) or the Haskell version is selectable at configure time
944 via the ``--enable-confd`` parameter, which can take one of the
945 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
946 daemon will result in a smaller footprint; for larger systems, we
947 welcome feedback on the Haskell version which might become the default
950 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
951 daemon running. The Python version doesn't implement the new call.
954 User interface changes
955 ~~~~~~~~~~~~~~~~~~~~~~
957 We have replaced the ``--disks`` option of ``gnt-instance
958 replace-disks`` with a more flexible ``--disk`` option, which allows
959 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
960 disk size and mode can be changed upon recreation (via ``gnt-instance
961 recreate-disks``, which accepts the same ``--disk`` option).
963 As many people are used to a ``show`` command, we have added that as an
964 alias to ``info`` on all ``gnt-*`` commands.
966 The ``gnt-instance grow-disk`` command has a new mode in which it can
967 accept the target size of the disk, instead of the delta; this can be
968 more safe since two runs in absolute mode will be idempotent, and
969 sometimes it's also easier to specify the desired size directly.
971 Also the handling of instances with regard to offline secondaries has
972 been improved. Instance operations should not fail because one of it's
973 secondary nodes is offline, even though it's safe to proceed.
975 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
976 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
977 minors to instance name.
982 RAPI coverage has improved, with (for example) new resources for
983 recreate-disks, node power-cycle, etc.
988 There is partial support for ``xl`` in the Xen hypervisor; feedback is
991 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
992 whether to still support Python 2.4 or move to Python 2.6 as minimum
995 Support for Fedora has been slightly improved; the provided example
996 init.d script should work better on it and the INSTALL file should
997 document the needed dependencies.
1002 The deprecated ``QueryLocks`` LUXI request has been removed. Use
1003 ``Query(what=QR_LOCK, ...)`` instead.
1005 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1006 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1007 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1008 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1009 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1011 RAPI client: ``CertificateError`` now derives from
1012 ``GanetiApiError``. This should make it more easy to handle Ganeti
1015 Deprecation warnings due to PyCrypto/paramiko import in
1016 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
1017 make sure to run an up-to-date paramiko version, if you use this tool.
1019 The QA scripts now depend on Python 2.5 or above (the main code base
1020 still works with Python 2.4).
1022 The configuration file (``config.data``) is now written without
1023 indentation for performance reasons; if you want to edit it, it can be
1024 re-formatted via ``tools/fmtjson``.
1026 A number of bugs has been fixed in the cluster merge tool.
1028 ``x509`` certification verification (used in import-export) has been
1029 changed to allow the same clock skew as permitted by the cluster
1030 verification. This will remove some rare but hard to diagnose errors in
1037 *(Released Thu, 19 Jul 2012)*
1039 Very few changes from rc4 to the final release, only bugfixes:
1041 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
1042 instance, fix CDROM booting for KVM instances)
1043 - fixed node group modification of node parameters
1044 - fixed issue in LUClusterVerifyGroup with multi-group clusters
1045 - fixed generation of bash completion to ensure a stable ordering
1052 *(Released Fri, 13 Jul 2012)*
1054 Third release candidate for 2.6. The following changes were done from
1057 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1058 - Fixed an inconsistency in the LUXI protocol with the provided
1059 arguments (NOT backwards compatible)
1060 - Fixed a bug with node groups ipolicy where ``min`` was greater than
1061 the cluster ``std`` value
1062 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1063 easier instance debugging on nodes (requires ``hconfd`` to work)
1069 *(Released Tue, 03 Jul 2012)*
1071 Second release candidate for 2.6. The following changes were done from
1074 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1076 - Fixed a RAPI regression on missing beparams/memory
1077 - Fixed redistribution of files on offline nodes
1078 - Added possibility to run activate-disks even though secondaries are
1079 offline. With this change it relaxes also the strictness on some other
1080 commands which use activate disks internally:
1081 * ``gnt-instance start|reboot|rename|backup|export``
1082 - Made it possible to remove safely an instance if its secondaries are
1084 - Made it possible to reinstall even though secondaries are offline
1090 *(Released Mon, 25 Jun 2012)*
1092 First release candidate for 2.6. The following changes were done from
1095 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1097 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1098 behaviour. It supports now the modify operation
1099 - ``hcheck`` implemented to analyze cluster health and possibility of
1100 improving health by rebalance
1101 - ``hbal`` has been improved in dealing with split instances
1107 *(Released Mon, 11 Jun 2012)*
1109 Second beta release of 2.6. The following changes were done from beta2
1112 - Fixed ``daemon-util`` with non-root user models
1113 - Fixed creation of plain instances with ``--no-wait-for-sync``
1114 - Fix wrong iv_names when running ``cfgupgrade``
1115 - Export more information in RAPI group queries
1116 - Fixed bug when changing instance network interfaces
1117 - Extended burnin to do NIC changes
1118 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1119 - Changed default for DRBD barriers
1120 - Fixed DRBD error reporting for syncer rate
1121 - Verify the options on disk parameters
1123 And of course various fixes to documentation and improved unittests and
1130 *(Released Wed, 23 May 2012)*
1132 First beta release of 2.6. The following changes were done from beta1 to
1135 - integrated patch for distributions without ``start-stop-daemon``
1136 - adapted example init.d script to work on Fedora
1137 - fixed log handling in Haskell daemons
1138 - adapted checks in the watcher for pycurl linked against libnss
1139 - add partial support for ``xl`` instead of ``xm`` for Xen
1140 - fixed a type issue in cluster verification
1141 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1144 Plus integrated fixes from the 2.5 branch:
1146 - fixed ``kvm-ifup`` to use ``/bin/bash``
1147 - fixed parallel build failures
1148 - KVM live migration when using a custom keymap
1154 *(Released Tue, 24 Jul 2012)*
1156 A small bugfix release, with no new features:
1158 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1159 different default shell (e.g. Debian, Ubuntu)
1160 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1161 243 and Debian bug #650664)
1162 - fixed compatibility with KVM versions that don't support multiple boot
1163 devices (fixes Issue 230 and Debian bug #624256)
1165 Additionally, a few fixes were done to the build system (fixed parallel
1166 build failures) and to the unittests (fixed race condition in test for
1167 FileID functions, and the default enable/disable mode for QA test is now
1174 *(Released Fri, 11 May 2012)*
1176 A small bugfix release.
1178 The main issues solved are on the topic of compatibility with newer LVM
1181 - fixed parsing of ``lv_attr`` field
1182 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1183 the ``--force`` flag is needed
1185 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1186 to accept kernel 3.x too (was hardcoded to 2.6.*).
1188 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1189 of hook results and that also made display of other errors suboptimal
1190 was fixed; the code behaves now like 2.4 and earlier.
1192 Another change in 2.5, the cleanup of the OS scripts environment, is too
1193 aggressive: it removed even the ``PATH`` variable, which requires the OS
1194 scripts to *always* need to export it. Since this is a bit too strict,
1195 we now export a minimal PATH, the same that we export for hooks.
1197 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1198 integrated into this release.
1200 Finally, a few other miscellaneous changes were done (no new features,
1201 just small improvements):
1203 - Fix ``gnt-group --help`` display
1204 - Fix hardcoded Xen kernel path
1205 - Fix grow-disk handling of invalid units
1206 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1207 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1213 *(Released Thu, 12 Apr 2012)*
1215 Incompatible/important changes and bugfixes
1216 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1218 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1219 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1220 to match the underlying LUXI interface.
1221 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1222 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1223 not maintain backwards-compatibility as the underlying operation
1224 changed in an incompatible way. The RAPI client can talk to old
1225 servers, but it needs to be told so as the return value changed.
1226 - When creating file-based instances via RAPI, the ``file_driver``
1227 parameter no longer defaults to ``loop`` and must be specified.
1228 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1230 - Support for the undocumented and deprecated RAPI instance creation
1231 request format version 0 has been dropped. Use version 1, supported
1232 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1233 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1234 documentation <install>`.
1235 - The "cluster-verify" hooks are now executed per group by the
1236 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1237 if you just run ``gnt-cluster verify``, which generates one opcode per
1239 - The environment as passed to the OS scripts is cleared, and thus no
1240 environment variables defined in the node daemon's environment will be
1241 inherited by the scripts.
1242 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1244 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1245 support operations involving multiple node groups.
1246 - Offline nodes are ignored when failing over an instance.
1247 - Support for KVM version 1.0, which changed the version reporting format
1249 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1251 - ``Makefile`` is now compatible with Automake 1.11.2
1252 - Includes all bugfixes made in the 2.4 series
1257 - The ganeti-htools project has been merged into the ganeti-core source
1258 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1259 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1260 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1261 GPT-style partition tables (requires `parted
1262 <http://www.gnu.org/s/parted/>`_).
1263 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1264 - Allowed adding tags on instance creation.
1265 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1266 - Allow instances to be started in a paused state, enabling the user to
1267 see the complete console output on boot using the console.
1268 - Added new hypervisor flag to control default reboot behaviour
1269 (``reboot_behavior``).
1270 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1271 - Improved out-of-band management support:
1273 - Added ``gnt-node health`` command reporting the health status of
1275 - Added ``gnt-node power`` command to manage power status of nodes.
1276 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1278 - Instance migration can fall back to failover if instance is not
1280 - Filters can be used when listing nodes, instances, groups and locks;
1281 see :manpage:`ganeti(7)` manpage.
1282 - Added post-execution status as variables to :doc:`hooks <hooks>`
1284 - Instance tags are exported/imported together with the instance.
1285 - When given an explicit job ID, ``gnt-job info`` will work for archived
1287 - Jobs can define dependencies on other jobs (not yet supported via
1288 RAPI or command line, but used by internal commands and usable via
1291 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1294 - Instance failover is now available as a RAPI resource
1295 (``/2/instances/[instance_name]/failover``).
1296 - ``gnt-instance info`` defaults to static information if primary node
1298 - Opcodes have a new ``comment`` attribute.
1299 - Added basic SPICE support to KVM hypervisor.
1300 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1302 Node group improvements
1303 ~~~~~~~~~~~~~~~~~~~~~~~
1305 - ``gnt-cluster verify`` has been modified to check groups separately,
1306 thereby improving performance.
1307 - Node group support has been added to ``gnt-cluster verify-disks``,
1308 which now operates per node group.
1309 - Watcher has been changed to work better with node groups.
1311 - One process and state file per node group.
1312 - Slow watcher in one group doesn't block other group's watcher.
1314 - Added new command, ``gnt-group evacuate``, to move all instances in a
1315 node group to other groups.
1316 - Added ``gnt-instance change-group`` to move an instance to another
1318 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1319 per-group operations.
1320 - Node groups can be tagged.
1321 - Some operations switch from an exclusive to a shared lock as soon as
1323 - Instance's primary and secondary nodes' groups are now available as
1324 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1325 and ``snodes.group.uuid``).
1330 - Numerous updates to documentation and manpages.
1332 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1334 - Some opcode/job results are now also documented, see :doc:`RAPI
1337 - A lockset's internal lock is now also visible in lock monitor.
1338 - Log messages from job queue workers now contain information about the
1339 opcode they're processing.
1340 - ``gnt-instance console`` no longer requires the instance lock.
1341 - A short delay when waiting for job changes reduces the number of LUXI
1342 requests significantly.
1343 - DRBD metadata volumes are overwritten with zeros during disk creation.
1344 - Out-of-band commands no longer acquire the cluster lock in exclusive
1346 - ``devel/upload`` now uses correct permissions for directories.
1352 *(Released Fri, 23 Mar 2012)*
1354 This was the sixth release candidate of the 2.5 series.
1360 *(Released Mon, 9 Jan 2012)*
1362 This was the fifth release candidate of the 2.5 series.
1368 *(Released Thu, 27 Oct 2011)*
1370 This was the fourth release candidate of the 2.5 series.
1376 *(Released Wed, 26 Oct 2011)*
1378 This was the third release candidate of the 2.5 series.
1384 *(Released Tue, 18 Oct 2011)*
1386 This was the second release candidate of the 2.5 series.
1392 *(Released Tue, 4 Oct 2011)*
1394 This was the first release candidate of the 2.5 series.
1400 *(Released Wed, 31 Aug 2011)*
1402 This was the third beta release of the 2.5 series.
1408 *(Released Mon, 22 Aug 2011)*
1410 This was the second beta release of the 2.5 series.
1416 *(Released Fri, 12 Aug 2011)*
1418 This was the first beta release of the 2.5 series.
1424 *(Released Thu, 27 Oct 2011)*
1426 - Fixed bug when parsing command line parameter values ending in
1428 - Fixed assertion error after unclean master shutdown
1429 - Disable HTTP client pool for RPC, significantly reducing memory usage
1431 - Fixed queue archive creation with wrong permissions
1437 *(Released Tue, 23 Aug 2011)*
1441 - Fixed documentation for importing with ``--src-dir`` option
1442 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1443 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1449 *(Released Fri, 5 Aug 2011)*
1451 Many bug-fixes and a few small features:
1453 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1454 issues when you tried to add an instance with two MAC addresses in one
1456 - KVM: fixed per-instance stored UID value
1457 - KVM: configure bridged NICs at migration start
1458 - KVM: Fix a bug where instance will not start with never KVM versions
1460 - Added OS search path to ``gnt-cluster info``
1461 - Fixed an issue with ``file_storage_dir`` where you were forced to
1462 provide an absolute path, but the documentation states it is a
1463 relative path, the documentation was right
1464 - Added a new parameter to instance stop/start called ``--no-remember``
1465 that will make the state change to not be remembered
1466 - Implemented ``no_remember`` at RAPI level
1467 - Improved the documentation
1468 - Node evacuation: don't call IAllocator if node is already empty
1469 - Fixed bug in DRBD8 replace disks on current nodes
1470 - Fixed bug in recreate-disks for DRBD instances
1471 - Moved assertion checking locks in ``gnt-instance replace-disks``
1472 causing it to abort with not owning the right locks for some situation
1473 - Job queue: Fixed potential race condition when cancelling queued jobs
1474 - Fixed off-by-one bug in job serial generation
1475 - ``gnt-node volumes``: Fix instance names
1476 - Fixed aliases in bash completion
1477 - Fixed a bug in reopening log files after being sent a SIGHUP
1478 - Added a flag to burnin to allow specifying VCPU count
1479 - Bugfixes to non-root Ganeti configuration
1485 *(Released Thu, 12 May 2011)*
1487 Many bug-fixes and a few new small features:
1489 - Fixed a bug related to log opening failures
1490 - Fixed a bug in instance listing with orphan instances
1491 - Fixed a bug which prevented resetting the cluster-level node parameter
1492 ``oob_program`` to the default
1493 - Many fixes related to the ``cluster-merge`` tool
1494 - Fixed a race condition in the lock monitor, which caused failures
1495 during (at least) creation of many instances in parallel
1496 - Improved output for gnt-job info
1497 - Removed the quiet flag on some ssh calls which prevented debugging
1499 - Improved the N+1 failure messages in cluster verify by actually
1500 showing the memory values (needed and available)
1501 - Increased lock attempt timeouts so that when executing long operations
1502 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1503 too early and thus prevent the use of the 'fair' mechanism
1504 - Changed instance query data (``gnt-instance info``) to not acquire
1505 locks unless needed, thus allowing its use on locked instance if only
1506 static information is asked for
1507 - Improved behaviour with filesystems that do not support rename on an
1509 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1510 kept locks on all nodes during the wipe, which is unneeded
1511 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1512 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1513 10GiB) which caused the wipe to fail right at the end in some cases
1514 - Fixed master IP activation when doing master failover with no-voting
1515 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1516 the master node itself
1517 - Fixed potential data-loss in under disk full conditions, where Ganeti
1518 wouldn't check correctly the return code and would consider
1519 partially-written files 'correct'
1520 - Fixed bug related to multiple VGs and DRBD disk replacing
1521 - Added new disk parameter ``metavg`` that allows placement of the meta
1522 device for DRBD in a different volume group
1523 - Fixed error handling in the node daemon when the system libc doesn't
1524 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1525 - Fixed lock release during replace-disks, which kept cluster-wide locks
1526 when doing disk replaces with an iallocator script
1527 - Added check for missing bridges in cluster verify
1528 - Handle EPIPE errors while writing to the terminal better, so that
1529 piping the output to e.g. ``less`` doesn't cause a backtrace
1530 - Fixed rare case where a ^C during Luxi calls could have been
1531 interpreted as server errors, instead of simply terminating
1532 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1534 - Added a few more parameters to the KVM hypervisor, allowing a second
1535 CDROM, custom disk type for CDROMs and a floppy image
1536 - Removed redundant message in instance rename when the name is given
1538 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1539 disks on new nodes, allowing recreation when the original instance
1540 nodes are completely gone
1541 - Added option when converting disk templates to DRBD to skip waiting
1542 for the resync, in order to make the instance available sooner
1543 - Added two new variables to the OS scripts environment (containing the
1545 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1546 to allow use of ``pvgrub`` as bootloader
1547 - Changed the instance memory modifications to only check out-of-memory
1548 conditions on memory increases, and turned the secondary node warnings
1549 into errors (they can still be overridden via ``--force``)
1550 - Fixed the handling of a corner case when the Python installation gets
1551 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1552 to execute a command that doesn't exist
1553 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1554 node of the instance returned failures during instance shutdown; this
1555 adds the option ``--ignore-consistency`` to gnt-instance move
1557 And as usual, various improvements to the error messages, documentation
1564 *(Released Wed, 09 Mar 2011)*
1566 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1567 the RAPI users file if run twice (even with ``--dry-run``).
1569 The release fixes that bug (nothing else changed).
1575 *(Released Mon, 07 Mar 2011)*
1577 Final 2.4.0 release. Just a few small fixes:
1579 - Fixed RAPI node evacuate
1580 - Fixed the kvm-ifup script
1581 - Fixed internal error handling for special job cases
1582 - Updated man page to specify the escaping feature for options
1588 *(Released Mon, 28 Feb 2011)*
1590 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1591 that this feature wiped the disks of the wrong instance, leading to loss
1596 - Fixed title of query field containing instance name
1597 - Expanded the glossary in the documentation
1598 - Fixed one unittest (internal issue)
1604 *(Released Mon, 21 Feb 2011)*
1606 A number of bug fixes plus just a couple functionality changes.
1608 On the user-visible side, the ``gnt-* list`` command output has changed
1609 with respect to "special" field states. The current rc1 style of display
1610 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1611 the default output mode special fields states are displayed as follows:
1613 - Offline resource: ``*``
1614 - Unavailable/not applicable: ``-``
1615 - Data missing (RPC failure): ``?``
1616 - Unknown field: ``??``
1618 Another user-visible change is the addition of ``--force-join`` to
1623 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1624 - Fixed regression in RAPI/instance reinstall where all parameters were
1625 required (instead of optional)
1626 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1627 - Fixed iallocator usage (offline nodes were not considered offline)
1628 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1629 - Fixed hypervisor and OS parameter validation with respect to
1630 non-vm_capable nodes
1631 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1633 - Fixed ``tools/listrunner`` with respect to agent-based usage
1639 *(Released Fri, 4 Feb 2011)*
1641 Many changes and fixes since the beta1 release. While there were some
1642 internal changes, the code has been mostly stabilised for the RC
1645 Note: the dumb allocator was removed in this release, as it was not kept
1646 up-to-date with the IAllocator protocol changes. It is recommended to
1647 use the ``hail`` command from the ganeti-htools package.
1649 Note: the 2.4 and up versions of Ganeti are not compatible with the
1650 0.2.x branch of ganeti-htools. You need to upgrade to
1651 ganeti-htools-0.3.0 (or later).
1653 Regressions fixed from 2.3
1654 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1656 - Fixed the ``gnt-cluster verify-disks`` command
1657 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1659 - Fixed disk adoption breakage
1660 - Fixed wrong headers in instance listing for field aliases
1665 - Fixed corner case in KVM handling of NICs
1666 - Fixed many cases of wrong handling of non-vm_capable nodes
1667 - Fixed a bug where a missing instance symlink was not possible to
1668 recreate with any ``gnt-*`` command (now ``gnt-instance
1669 activate-disks`` does it)
1670 - Fixed the volume group name as reported by ``gnt-cluster
1672 - Increased timeouts for the import-export code, hopefully leading to
1673 fewer aborts due network or instance timeouts
1674 - Fixed bug in ``gnt-node list-storage``
1675 - Fixed bug where not all daemons were started on cluster
1676 initialisation, but only at the first watcher run
1677 - Fixed many bugs in the OOB implementation
1678 - Fixed watcher behaviour in presence of instances with offline
1680 - Fixed instance list output for instances running on the wrong node
1681 - a few fixes to the cluster-merge tool, but it still cannot merge
1682 multi-node groups (currently it is not recommended to use this tool)
1688 - Improved network configuration for the KVM hypervisor
1689 - Added e1000 as a supported NIC for Xen-HVM
1690 - Improved the lvmstrap tool to also be able to use partitions, as
1691 opposed to full disks
1692 - Improved speed of disk wiping (the cluster parameter
1693 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1694 of instance creations
1695 - Added documentation for the OS parameters
1696 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1697 hypervisor is not responding
1698 - Added display of blacklisted and hidden OS information in
1699 ``gnt-cluster info``
1700 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1701 NIC and node parameters, which might create problems with currently
1702 invalid (but undetected) configuration files, but prevents validation
1703 failures when unrelated parameters are modified
1704 - Changed cluster initialisation to wait for the master daemon to become
1706 - Expanded the RAPI interface:
1708 - Added config redistribution resource
1709 - Added activation/deactivation of instance disks
1710 - Added export of console information
1712 - Implemented log file reopening on SIGHUP, which allows using
1713 logrotate(8) for the Ganeti log files
1714 - Added a basic OOB helper script as an example
1720 *(Released Fri, 14 Jan 2011)*
1725 - Fixed timezone issues when formatting timestamps
1726 - Added support for node groups, available via ``gnt-group`` and other
1728 - Added out-of-band framework and management, see :doc:`design
1729 document <design-oob>`
1730 - Removed support for roman numbers from ``gnt-node list`` and
1731 ``gnt-instance list``.
1732 - Allowed modification of master network interface via ``gnt-cluster
1733 modify --master-netdev``
1734 - Accept offline secondaries while shutting down instance disks
1735 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1736 - Added support for multiple LVM volume groups
1737 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1739 - Added commands to list available fields:
1741 - ``gnt-node list-fields``
1742 - ``gnt-group list-fields``
1743 - ``gnt-instance list-fields``
1745 - Updated documentation and man pages
1750 - Moved ``rapi_users`` file into separate directory, now named
1751 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1753 - Added new tool for running commands on many machines,
1754 ``tools/ganeti-listrunner``
1755 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1756 improving the ``gnt-instance console`` output
1757 - Allowed customisation of disk index separator at ``configure`` time
1758 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1759 - Added support for non-partitioned md disks in ``lvmstrap``
1760 - Added script to gracefully power off KVM instances
1761 - Split ``utils`` module into smaller parts
1762 - Changed query operations to return more detailed information, e.g.
1763 whether an information is unavailable due to an offline node. To use
1764 this new functionality, the LUXI call ``Query`` must be used. Field
1765 information is now stored by the master daemon and can be retrieved
1766 using ``QueryFields``. Instances, nodes and groups can also be queried
1767 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1768 exposed via RAPI). The following commands make use of this
1769 infrastructure change:
1771 - ``gnt-group list``
1772 - ``gnt-group list-fields``
1774 - ``gnt-node list-fields``
1775 - ``gnt-instance list``
1776 - ``gnt-instance list-fields``
1777 - ``gnt-debug locks``
1782 - New RAPI resources (see :doc:`rapi`):
1786 - ``/2/groups/[group_name]``
1787 - ``/2/groups/[group_name]/assign-nodes``
1788 - ``/2/groups/[group_name]/modify``
1789 - ``/2/groups/[group_name]/rename``
1790 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1794 - Implemented ``no_install`` for instance creation
1795 - Implemented OS parameters for instance reinstallation, allowing
1796 use of special settings on reinstallation (e.g. for preserving data)
1801 - Added IPv6 support in import/export
1802 - Pause DRBD synchronization while wiping disks on instance creation
1803 - Updated unittests and QA scripts
1804 - Improved network parameters passed to KVM
1805 - Converted man pages from docbook to reStructuredText
1811 *(Released Mon, 20 Dec 2010)*
1813 Released version 2.3.1~rc1 without any changes.
1819 *(Released Wed, 1 Dec 2010)*
1821 - impexpd: Disable OpenSSL compression in socat if possible (backport
1822 from master, commit e90739d625b, see :doc:`installation guide
1823 <install-quick>` for details)
1824 - Changed unittest coverage report to exclude test scripts
1825 - Added script to check version format
1831 *(Released Wed, 1 Dec 2010)*
1833 Released version 2.3.0~rc1 without any changes.
1839 *(Released Fri, 19 Nov 2010)*
1841 A number of bugfixes and documentation updates:
1843 - Update ganeti-os-interface documentation
1844 - Fixed a bug related to duplicate MACs or similar items which should be
1846 - Fix breakage in OS state modify
1847 - Reinstall instance: disallow offline secondaries (fixes bug related to
1848 OS changing but reinstall failing)
1849 - plus all the other fixes between 2.2.1 and 2.2.2
1855 *(Released Tue, 2 Nov 2010)*
1857 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1858 - Fixed master failover race with watcher
1859 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1861 - Accept previously stopped instance for export with instance removal
1862 - Simplify and extend the environment variables for instance OS scripts
1863 - Added new node flags, ``master_capable`` and ``vm_capable``
1864 - Added optional instance disk wiping prior during allocation. This is a
1865 cluster-wide option and can be set/modified using
1866 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1867 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1868 :doc:`install-quick`
1869 - Added a new watcher option (``--ignore-pause``)
1870 - Added option to ignore offline node on instance start/stop
1871 (``--ignore-offline``)
1872 - Allow overriding OS parameters with ``gnt-instance reinstall``
1873 - Added ability to change node's secondary IP address using ``gnt-node
1875 - Implemented privilege separation for all daemons except
1876 ``ganeti-noded``, see ``configure`` options
1877 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1879 - Implemented job priorities (see ``ganeti(7)`` manpage)
1880 - Ignore failures while shutting down instances during failover from
1882 - Exit daemon's bootstrap process only once daemon is ready
1883 - Export more information via ``LUInstanceQuery``/remote API
1884 - Improved documentation, QA and unittests
1885 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1886 restart if the file was created or changed
1887 - Added LUXI protocol version sent with each request and response,
1888 allowing detection of server/client mismatches
1889 - Moved the Python scripts among gnt-* and ganeti-* into modules
1890 - Moved all code related to setting up SSH to an external script,
1892 - Infrastructure changes for node group support in future versions
1898 *(Released Fri, 19 Nov 2010)*
1900 A few small bugs fixed, and some improvements to the build system:
1902 - Fix documentation regarding conversion to drbd
1903 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1905 - Fix error handling in node modify with multiple changes
1906 - Allow remote imports without checked names
1912 *(Released Tue, 19 Oct 2010)*
1914 - Disable SSL session ID cache in RPC client
1920 *(Released Thu, 14 Oct 2010)*
1922 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1923 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1925 - Fix problems with interaction between Curl and Python's HTTP server,
1926 resulting in increased speed in many RPC calls
1927 - Improve our release script to prevent breakage with older aclocal and
1934 *(Released Thu, 7 Oct 2010)*
1936 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1937 value retrieved from master
1938 - Added support for blacklisted or hidden OS definitions
1939 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1940 - Added support for -mem-path in KVM hypervisor abstraction layer
1941 - Allow overriding instance parameters in tool for inter-cluster
1942 instance moves (``tools/move-instance``)
1943 - Improved opcode summaries (e.g. in ``gnt-job list``)
1944 - Improve consistency of OS listing by sorting it
1945 - Documentation updates
1951 *(Released Fri, 8 Oct 2010)*
1953 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1959 *(Released Mon, 4 Oct 2010)*
1961 - Fixed regression in ``gnt-instance rename``
1967 *(Released Wed, 22 Sep 2010)*
1969 - Fixed OS_VARIANT variable for OS scripts
1970 - Fixed cluster tag operations via RAPI
1971 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1972 - Disabled RAPI CA checks in watcher
1978 *(Released Mon, 23 Aug 2010)*
1980 - Support DRBD versions of the format "a.b.c.d"
1982 - Re-introduce support for usage from multiple threads in RAPI client
1983 - Instance renames and modify via RAPI
1984 - Work around race condition between processing and archival in job
1986 - Mark opcodes following failed one as failed, too
1987 - Job field ``lock_status`` was removed due to difficulties making it
1988 work with the changed job queue in Ganeti 2.2; a better way to monitor
1989 locks is expected for a later 2.2.x release
1990 - Fixed dry-run behaviour with many commands
1991 - Support ``ssh-agent`` again when adding nodes
1992 - Many additional bugfixes
1998 *(Released Fri, 30 Jul 2010)*
2000 Important change: the internal RPC mechanism between Ganeti nodes has
2001 changed from using a home-grown http library (based on the Python base
2002 libraries) to use the PycURL library. This requires that PycURL is
2003 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2004 against GnuTLS by default. cURL's support for GnuTLS had known issues
2005 before cURL 7.21.0 and we recommend using the latest cURL release or
2006 linking against OpenSSL. Most other distributions already link PycURL
2007 and cURL against OpenSSL. The command::
2009 python -c 'import pycurl; print pycurl.version'
2011 can be used to determine the libraries PycURL and cURL are linked
2014 Other significant changes:
2016 - Rewrote much of the internals of the job queue, in order to achieve
2017 better parallelism; this decouples job query operations from the job
2018 processing, and it should allow much nicer behaviour of the master
2019 daemon under load, and it also has uncovered some long-standing bugs
2020 related to the job serialisation (now fixed)
2021 - Added a default iallocator setting to the cluster parameters,
2022 eliminating the need to always pass nodes or an iallocator for
2023 operations that require selection of new node(s)
2024 - Added experimental support for the LXC virtualization method
2025 - Added support for OS parameters, which allows the installation of
2026 instances to pass parameter to OS scripts in order to customise the
2028 - Added a hypervisor parameter controlling the migration type (live or
2029 non-live), since hypervisors have various levels of reliability; this
2030 has renamed the 'live' parameter to 'mode'
2031 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
2032 logical volumes, meaning that cluster verify will ignore them and not
2033 flag their presence as errors
2034 - The watcher will now reset the error count for failed instances after
2035 8 hours, thus allowing self-healing if the problem that caused the
2036 instances to be down/fail to start has cleared in the meantime
2037 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2038 check for, and warn, if the drbd module parameter ``usermode_helper``
2039 is not consistent with the cluster-wide setting; this is needed to
2040 make diagnose easier of failed drbd creations
2041 - Started adding base IPv6 support, but this is not yet
2042 enabled/available for use
2043 - Rename operations (cluster, instance) will now return the new name,
2044 which is especially useful if a short name was passed in
2045 - Added support for instance migration in RAPI
2046 - Added a tool to pre-configure nodes for the SSH setup, before joining
2047 them to the cluster; this will allow in the future a simplified model
2048 for node joining (but not yet fully enabled in 2.2); this needs the
2049 paramiko python library
2050 - Fixed handling of name-resolving errors
2051 - Fixed consistency of job results on the error path
2052 - Fixed master-failover race condition when executed multiple times in
2054 - Fixed many bugs related to the job queue (mostly introduced during the
2055 2.2 development cycle, so not all are impacting 2.1)
2056 - Fixed instance migration with missing disk symlinks
2057 - Fixed handling of unknown jobs in ``gnt-job archive``
2058 - And many other small fixes/improvements
2062 - Enhanced both the unittest and the QA coverage
2063 - Switched the opcode validation to a generic model, and extended the
2064 validation to all opcode parameters
2065 - Changed more parts of the code that write shell scripts to use the
2067 - Switched the master daemon to use the asyncore library for the Luxi
2074 *(Released Thu, 17 Jun 2010)*
2076 - Added tool (``move-instance``) and infrastructure to move instances
2077 between separate clusters (see :doc:`separate documentation
2078 <move-instance>` and :doc:`design document <design-2.2>`)
2079 - Added per-request RPC timeout
2080 - RAPI now requires a Content-Type header for requests with a body (e.g.
2081 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2082 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2083 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2085 - Implemented initial support for running Ganeti daemons as separate
2086 users, see configure-time flags ``--with-user-prefix`` and
2087 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2089 - Instances can be removed after export (``gnt-backup export
2090 --remove-instance``)
2091 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2092 key (instead of 1024 bit)
2093 - Added new cluster configuration file for cluster domain secret
2094 - Import/export now use SSL instead of SSH
2095 - Added support for showing estimated time when exporting an instance,
2096 see the ``ganeti-os-interface(7)`` manpage and look for
2103 *(Released Tue, 16 Nov 2010)*
2105 Some more bugfixes. Unless critical bugs occur, this will be the last
2108 - Fix case of MAC special-values
2109 - Fix mac checker regex
2110 - backend: Fix typo causing "out of range" error
2111 - Add missing --units in gnt-instance list man page
2117 *(Released Tue, 24 Aug 2010)*
2120 - Don't ignore secondary node silently on non-mirrored disk templates
2122 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2123 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2124 - Properly document the usb_mouse parameter
2125 - Fix path in ganeti-rapi(8) (issue 116)
2126 - Adjust error message when the ganeti user's .ssh directory is
2128 - Add same-node-check when changing the disk template to drbd
2134 *(Released Fri, 16 Jul 2010)*
2137 - Add an option to only select some reboot types during qa/burnin.
2138 (on some hypervisors consequent reboots are not supported)
2139 - Fix infrequent race condition in master failover. Sometimes the old
2140 master ip address would be still detected as up for a short time
2141 after it was removed, causing failover to fail.
2142 - Decrease mlockall warnings when the ctypes module is missing. On
2143 Python 2.4 we support running even if no ctypes module is installed,
2144 but we were too verbose about this issue.
2145 - Fix building on old distributions, on which man doesn't have a
2147 - Fix RAPI not to ignore the MAC address on instance creation
2148 - Implement the old instance creation format in the RAPI client.
2154 *(Released Thu, 01 Jul 2010)*
2156 A small bugfix release:
2157 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2158 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2159 a non-existing option
2160 - Fix instance create: the --force-variant option was ignored
2161 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2162 - Fix modify node storage with non-FQDN arguments
2163 - Fix RAPI client to authenticate under Python 2.6 when used
2164 for more than 5 requests needing authentication
2165 - Fix gnt-instance modify -t (storage) giving a wrong error message
2166 when converting a non-shutdown drbd instance to plain
2172 *(Released Fri, 18 Jun 2010)*
2174 A small bugfix release:
2176 - Fix live migration of KVM instances started with older Ganeti
2177 versions which had fewer hypervisor parameters
2178 - Fix gnt-instance grow-disk on down instances
2179 - Fix an error-reporting bug during instance migration
2180 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2181 silently ignoring broken ones
2182 - Fix an RPC error reporting bug affecting, for example, RAPI client
2184 - Fix bug triggered by different API version os-es on different nodes
2185 - Fix a bug in instance startup with custom hvparams: OS level
2186 parameters would fail to be applied.
2187 - Fix the RAPI client under Python 2.6 (but more work is needed to
2188 make it work completely well with OpenSSL)
2189 - Fix handling of errors when resolving names from DNS
2195 *(Released Thu, 3 Jun 2010)*
2197 A medium sized development cycle. Some new features, and some
2198 fixes/small improvements/cleanups.
2200 Significant features
2201 ~~~~~~~~~~~~~~~~~~~~
2203 The node deamon now tries to mlock itself into memory, unless the
2204 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2205 its logs, and falls back to console logging. This allows emergency
2206 features such as ``gnt-node powercycle`` to work even in the event of a
2207 broken node disk (tested offlining the disk hosting the node's
2208 filesystem and dropping its memory caches; don't try this at home)
2210 KVM: add vhost-net acceleration support. It can be tested with a new
2211 enough version of the kernel and of qemu-kvm.
2213 KVM: Add instance chrooting feature. If you use privilege dropping for
2214 your VMs you can also now force them to chroot to an empty directory,
2215 before starting the emulated guest.
2217 KVM: Add maximum migration bandwith and maximum downtime tweaking
2218 support (requires a new-enough version of qemu-kvm).
2220 Cluster verify will now warn if the master node doesn't have the master
2221 ip configured on it.
2223 Add a new (incompatible) instance creation request format to RAPI which
2224 supports all parameters (previously only a subset was supported, and it
2225 wasn't possible to extend the old format to accomodate all the new
2226 features. The old format is still supported, and a client can check for
2227 this feature, before using it, by checking for its presence in the
2228 ``features`` RAPI resource.
2230 Now with ancient latin support. Try it passing the ``--roman`` option to
2231 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2232 (requires the python-roman module to be installed, in order to work).
2237 As usual many internal code refactorings, documentation updates, and
2240 - Lots of improvements and cleanups to the experimental Remote API
2241 (RAPI) client library.
2242 - A new unit test suite for the core daemon libraries.
2243 - A fix to creating missing directories makes sure the umask is not
2244 applied anymore. This enforces the same directory permissions
2246 - Better handling terminating daemons with ctrl+c (used when running
2247 them in debugging mode).
2248 - Fix a race condition in live migrating a KVM instance, when stat()
2249 on the old proc status file returned EINVAL, which is an unexpected
2251 - Fixed manpage checking with newer man and utf-8 charachters. But now
2252 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2258 *(Released Fri, 7 May 2010)*
2260 Fix a bug which prevented untagged KVM instances from starting.
2266 *(Released Fri, 7 May 2010)*
2268 Another release with a long development cycle, during which many
2269 different features were added.
2271 Significant features
2272 ~~~~~~~~~~~~~~~~~~~~
2274 The KVM hypervisor now can run the individual instances as non-root, to
2275 reduce the impact of a VM being hijacked due to bugs in the
2276 hypervisor. It is possible to run all instances as a single (non-root)
2277 user, to manually specify a user for each instance, or to dynamically
2278 allocate a user out of a cluster-wide pool to each instance, with the
2279 guarantee that no two instances will run under the same user ID on any
2282 An experimental RAPI client library, that can be used standalone
2283 (without the other Ganeti libraries), is provided in the source tree as
2284 ``lib/rapi/client.py``. Note this client might change its interface in
2285 the future, as we iterate on its capabilities.
2287 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2288 replace the cluster's certificates and crypto keys. This might help in
2289 case they have been compromised, or have simply expired.
2291 A new disk option for instance creation has been added that allows one
2292 to "adopt" currently existing logical volumes, with data
2293 preservation. This should allow easier migration to Ganeti from
2294 unmanaged (or managed via other software) instances.
2296 Another disk improvement is the possibility to convert between redundant
2297 (DRBD) and plain (LVM) disk configuration for an instance. This should
2298 allow better scalability (starting with one node and growing the
2299 cluster, or shrinking a two-node cluster to one node).
2301 A new feature that could help with automated node failovers has been
2302 implemented: if a node sees itself as offline (by querying the master
2303 candidates), it will try to shutdown (hard) all instances and any active
2304 DRBD devices. This reduces the risk of duplicate instances if an
2305 external script automatically failovers the instances on such nodes. To
2306 enable this, the cluster parameter ``maintain_node_health`` should be
2307 enabled; in the future this option (per the name) will enable other
2308 automatic maintenance features.
2310 Instance export/import now will reuse the original instance
2311 specifications for all parameters; that means exporting an instance,
2312 deleting it and the importing it back should give an almost identical
2313 instance. Note that the default import behaviour has changed from
2314 before, where it created only one NIC; now it recreates the original
2317 Cluster verify has added a few new checks: SSL certificates validity,
2318 /etc/hosts consistency across the cluster, etc.
2323 As usual, many internal changes were done, documentation fixes,
2326 - Fixed cluster initialization with disabled cluster storage (regression
2327 introduced in 2.1.1)
2328 - File-based storage supports growing the disks
2329 - Fixed behaviour of node role changes
2330 - Fixed cluster verify for some corner cases, plus a general rewrite of
2331 cluster verify to allow future extension with more checks
2332 - Fixed log spamming by watcher and node daemon (regression introduced
2334 - Fixed possible validation issues when changing the list of enabled
2336 - Fixed cleanup of /etc/hosts during node removal
2337 - Fixed RAPI response for invalid methods
2338 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2339 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2340 from ide disks, etc.)
2341 - Allow OS changes without re-installation (to record a changed OS
2342 outside of Ganeti, or to allow OS renames)
2343 - Allow instance creation without OS installation (useful for example if
2344 the OS will be installed manually, or restored from a backup not in
2346 - Implemented option to make cluster ``copyfile`` use the replication
2348 - Added list of enabled hypervisors to ssconf (possibly useful for
2350 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2352 - A partial form of node re-IP is possible via node readd, which now
2353 allows changed node primary IP
2354 - Command line utilities now show an informational message if the job is
2356 - The logs of the master daemon now show the PID/UID/GID of the
2363 *(Released Fri, 12 Mar 2010)*
2365 During the 2.1.0 long release candidate cycle, a lot of improvements and
2366 changes have accumulated with were released later as 2.1.1.
2371 The node evacuate command (``gnt-node evacuate``) was significantly
2372 rewritten, and as such the IAllocator protocol was changed - a new
2373 request type has been added. This unfortunate change during a stable
2374 series is designed to improve performance of node evacuations; on
2375 clusters with more than about five nodes and which are well-balanced,
2376 evacuation should proceed in parallel for all instances of the node
2377 being evacuated. As such, any existing IAllocator scripts need to be
2378 updated, otherwise the above command will fail due to the unknown
2379 request. The provided "dumb" allocator has not been updated; but the
2380 ganeti-htools package supports the new protocol since version 0.2.4.
2382 Another important change is increased validation of node and instance
2383 names. This might create problems in special cases, if invalid host
2384 names are being used.
2386 Also, a new layer of hypervisor parameters has been added, that sits at
2387 OS level between the cluster defaults and the instance ones. This allows
2388 customisation of virtualization parameters depending on the installed
2389 OS. For example instances with OS 'X' may have a different KVM kernel
2390 (or any other parameter) than the cluster defaults. This is intended to
2391 help managing a multiple OSes on the same cluster, without manual
2392 modification of each instance's parameters.
2394 A tool for merging clusters, ``cluster-merge``, has been added in the
2395 tools sub-directory.
2400 - Improved the int/float conversions that should make the code more
2401 robust in face of errors from the node daemons
2402 - Fixed the remove node code in case of internal configuration errors
2403 - Fixed the node daemon behaviour in face of inconsistent queue
2404 directory (e.g. read-only file-system where we can't open the files
2406 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2407 now it either aborts cleanly or, if given the new "auto_promote"
2408 parameter, will automatically promote other nodes as needed
2409 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2410 completely prevent Ganeti from working
2411 - Fixed bug for instance export when not all disks were successfully
2413 - Fixed behaviour of node add when the new node is slow in starting up
2415 - Fixed handling of signals in the LUXI client, which should improve
2416 behaviour of command-line scripts
2417 - Added checks for invalid node/instance names in the configuration (now
2418 flagged during cluster verify)
2419 - Fixed watcher behaviour for disk activation errors
2420 - Fixed two potentially endless loops in http library, which led to the
2421 RAPI daemon hanging and consuming 100% CPU in some cases
2422 - Fixed bug in RAPI daemon related to hashed passwords
2423 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2425 - Enhanced compatibility with non-Debian OSes, but not using absolute
2426 path in some commands and allowing customisation of the ssh
2427 configuration directory
2428 - Fixed possible future issue with new Python versions by abiding to the
2429 proper use of ``__slots__`` attribute on classes
2430 - Added checks that should prevent directory traversal attacks
2431 - Many documentation fixes based on feedback from users
2436 - Added an "early_release" more for instance replace disks and node
2437 evacuate, where we release locks earlier and thus allow higher
2438 parallelism within the cluster
2439 - Added watcher hooks, intended to allow the watcher to restart other
2440 daemons (e.g. from the ganeti-nbma project), but they can be used of
2441 course for any other purpose
2442 - Added a compile-time disable for DRBD barriers, to increase
2443 performance if the administrator trusts the power supply or the
2444 storage system to not lose writes
2445 - Added the option of using syslog for logging instead of, or in
2446 addition to, Ganeti's own log files
2447 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2448 can indeed boot from a paravirtual NIC
2449 - Added a generic debug level for many operations; while this is not
2450 used widely yet, it allows one to pass the debug value all the way to
2452 - Enhanced the hooks environment for instance moves (failovers,
2453 migrations) where the primary/secondary nodes changed during the
2454 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2455 - Enhanced data validations for many user-supplied values; one important
2456 item is the restrictions imposed on instance and node names, which
2457 might reject some (invalid) host names
2458 - Add a configure-time option to disable file-based storage, if it's not
2459 needed; this allows greater security separation between the master
2460 node and the other nodes from the point of view of the inter-node RPC
2462 - Added user notification in interactive tools if job is waiting in the
2463 job queue or trying to acquire locks
2464 - Added log messages when a job is waiting for locks
2465 - Added filtering by node tags in instance operations which admit
2466 multiple instances (start, stop, reboot, reinstall)
2467 - Added a new tool for cluster mergers, ``cluster-merge``
2468 - Parameters from command line which are of the form ``a=b,c=d`` can now
2469 use backslash escapes to pass in values which contain commas,
2470 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2472 - For KVM, the instance name is the first parameter passed to KVM, so
2473 that it's more visible in the process list
2479 *(Released Tue, 2 Mar 2010)*
2481 Ganeti 2.1 brings many improvements with it. Major changes:
2483 - Added infrastructure to ease automated disk repairs
2484 - Added new daemon to export configuration data in a cheaper way than
2485 using the remote API
2486 - Instance NICs can now be routed instead of being associated with a
2488 - Improved job locking logic to reduce impact of jobs acquiring multiple
2489 locks waiting for other long-running jobs
2491 In-depth implementation details can be found in the Ganeti 2.1 design
2497 - Added chroot hypervisor
2498 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2500 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2501 ``bootloader_path`` and ``bootloader_args``)
2502 - Added the ``use_localtime`` option for the xen-hvm and kvm
2503 hypervisors, and the default value for this has changed to false (in
2504 2.0 xen-hvm always enabled it)
2505 - Added luxi call to submit multiple jobs in one go
2506 - Added cluster initialization option to not modify ``/etc/hosts``
2508 - Added network interface parameters
2509 - Added dry run mode to some LUs
2510 - Added RAPI resources:
2512 - ``/2/instances/[instance_name]/info``
2513 - ``/2/instances/[instance_name]/replace-disks``
2514 - ``/2/nodes/[node_name]/evacuate``
2515 - ``/2/nodes/[node_name]/migrate``
2516 - ``/2/nodes/[node_name]/role``
2517 - ``/2/nodes/[node_name]/storage``
2518 - ``/2/nodes/[node_name]/storage/modify``
2519 - ``/2/nodes/[node_name]/storage/repair``
2521 - Added OpCodes to evacuate or migrate all instances on a node
2522 - Added new command to list storage elements on nodes (``gnt-node
2523 list-storage``) and modify them (``gnt-node modify-storage``)
2524 - Added new ssconf files with master candidate IP address
2525 (``ssconf_master_candidates_ips``), node primary IP address
2526 (``ssconf_node_primary_ips``) and node secondary IP address
2527 (``ssconf_node_secondary_ips``)
2528 - Added ``ganeti-confd`` and a client library to query the Ganeti
2529 configuration via UDP
2530 - Added ability to run hooks after cluster initialization and before
2532 - Added automatic mode for disk replace (``gnt-instance replace-disks
2534 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2535 after catastrophic data-loss
2536 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2538 - Added ``gnt-instance move`` command to move instances
2539 - Added ``gnt-cluster watcher`` command to control watcher
2540 - Added ``gnt-node powercycle`` command to powercycle nodes
2541 - Added new job status field ``lock_status``
2542 - Added parseable error codes to cluster verification (``gnt-cluster
2543 verify --error-codes``) and made output less verbose (use
2544 ``--verbose`` to restore previous behaviour)
2545 - Added UUIDs to the main config entities (cluster, nodes, instances)
2546 - Added support for OS variants
2547 - Added support for hashed passwords in the Ganeti remote API users file
2549 - Added option to specify maximum timeout on instance shutdown
2550 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2551 - Added new helper script to start and stop Ganeti daemons
2552 (``daemon-util``), with the intent to reduce the work necessary to
2553 adjust Ganeti for non-Debian distributions and to start/stop daemons
2555 - Added more unittests
2556 - Fixed critical bug in ganeti-masterd startup
2557 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2558 now customisable at the cluster level for both the KVM and Xen
2559 hypervisors using the new ``migration_port`` parameter
2560 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2561 reinstalling an instance
2562 - Allowed ``@`` in tag names
2563 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2564 - Many documentation updates
2565 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2566 - ``gnt-instance reinstall`` can now reinstall multiple instances
2567 - Updated many command line parameters
2568 - Introduced new OS API version 15
2569 - No longer support a default hypervisor
2570 - Treat virtual LVs as inexistent
2571 - Improved job locking logic to reduce lock contention
2572 - Match instance and node names case insensitively
2573 - Reimplemented bash completion script to be more complete
2580 *(Released Thu, 4 Feb 2010)*
2582 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2583 - Fix a string formatting bug
2584 - Improve safety of the code in some error paths
2585 - Improve data validation in the master of values returned from nodes
2591 *(Released Thu, 17 Dec 2009)*
2593 - Fix security issue due to missing validation of iallocator names; this
2594 allows local and remote execution of arbitrary executables
2595 - Fix failure of gnt-node list during instance removal
2596 - Ship the RAPI documentation in the archive
2602 *(Released Wed, 30 Sep 2009)*
2604 - Fixed many wrong messages
2605 - Fixed a few bugs related to the locking library
2606 - Fixed MAC checking at instance creation time
2607 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2608 - Fixed a few issues related to signal handling in both daemons and
2610 - Fixed the example startup script provided
2611 - Fixed insserv dependencies in the example startup script (patch from
2613 - Fixed handling of drained nodes in the iallocator framework
2614 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2616 - Fixed error related to invalid job IDs in job polling
2617 - Fixed job/opcode persistence on unclean master shutdown
2618 - Fixed handling of partial job processing after unclean master
2620 - Fixed error reporting from LUs, previously all errors were converted
2621 into execution errors
2622 - Fixed error reporting from burnin
2623 - Decreased significantly the memory usage of the job queue
2624 - Optimised slightly multi-job submission
2625 - Optimised slightly opcode loading
2626 - Backported the multi-job submit framework from the development
2627 branch; multi-instance start and stop should be faster
2628 - Added script to clean archived jobs after 21 days; this will reduce
2629 the size of the queue directory
2630 - Added some extra checks in disk size tracking
2631 - Added an example ethers hook script
2632 - Added a cluster parameter that prevents Ganeti from modifying of
2634 - Added more node information to RAPI responses
2635 - Added a ``gnt-job watch`` command that allows following the ouput of a
2637 - Added a bind-address option to ganeti-rapi
2638 - Added more checks to the configuration verify
2639 - Enhanced the burnin script such that some operations can be retried
2641 - Converted instance reinstall to multi-instance model
2647 *(Released Fri, 7 Aug 2009)*
2649 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2650 to allow using the pre-2.0.2 behaviour in activation, if any existing
2651 instances have mismatched disk sizes in the configuration
2652 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2653 any configuration mismatches for disk sizes
2654 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2655 failover to work on two-node clusters
2656 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2658 - Fixed detection of OS script errors in ``gnt-backup export``
2659 - Fixed exit code of ``gnt-backup export``
2665 *(Released Fri, 17 Jul 2009)*
2667 - Added experimental support for stripped logical volumes; this should
2668 enhance performance but comes with a higher complexity in the block
2669 device handling; stripping is only enabled when passing
2670 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2671 affected even in the non-stripped mode
2672 - Improved resiliency against transient failures at the end of DRBD
2673 resyncs, and in general of DRBD resync checks
2674 - Fixed a couple of issues with exports and snapshot errors
2675 - Fixed a couple of issues in instance listing
2676 - Added display of the disk size in ``gnt-instance info``
2677 - Fixed checking for valid OSes in instance creation
2678 - Fixed handling of the "vcpus" parameter in instance listing and in
2679 general of invalid parameters
2680 - Fixed http server library, and thus RAPI, to handle invalid
2681 username/password combinations correctly; this means that now they
2682 report unauthorized for queries too, not only for modifications,
2683 allowing earlier detect of configuration problems
2684 - Added a new "role" node list field, equivalent to the master/master
2685 candidate/drained/offline flags combinations
2686 - Fixed cluster modify and changes of candidate pool size
2687 - Fixed cluster verify error messages for wrong files on regular nodes
2688 - Fixed a couple of issues with node demotion from master candidate role
2689 - Fixed node readd issues
2690 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2691 - Added a new ``--no-voting`` option for masterfailover to fix failover
2692 on two-nodes clusters when the former master node is unreachable
2693 - Added instance reinstall over RAPI
2699 *(Released Tue, 16 Jun 2009)*
2701 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2702 allow re-adding the start in single-user option (regression from 1.2)
2703 - the watcher writes the instance status to a file, to allow monitoring
2704 to report the instance status (from the master) based on cached
2705 results of the watcher's queries; while this can get stale if the
2706 watcher is being locked due to other work on the cluster, this is
2707 still an improvement
2708 - the watcher now also restarts the node daemon and the rapi daemon if
2710 - fixed the watcher to handle full and drained queue cases
2711 - hooks export more instance data in the environment, which helps if
2712 hook scripts need to take action based on the instance's properties
2713 (no longer need to query back into ganeti)
2714 - instance failovers when the instance is stopped do not check for free
2715 RAM, so that failing over a stopped instance is possible in low memory
2717 - rapi uses queries for tags instead of jobs (for less job traffic), and
2718 for cluster tags it won't talk to masterd at all but read them from
2720 - a couple of error handling fixes in RAPI
2721 - drbd handling: improved the error handling of inconsistent disks after
2722 resync to reduce the frequency of "there are some degraded disks for
2723 this instance" messages
2724 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2725 error handling path called a wrong function name)
2731 *(Released Wed, 27 May 2009)*
2733 - no changes from rc5
2739 *(Released Wed, 20 May 2009)*
2741 - fix a couple of bugs (validation, argument checks)
2742 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2743 - some small improvements to RAPI and IAllocator
2744 - make watcher automatically start the master daemon if down
2750 *(Released Mon, 27 Apr 2009)*
2752 - change the OS list to not require locks; this helps with big clusters
2753 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2754 volume group is broken
2755 - ``gnt-instance info``, without any arguments, doesn't run for all
2756 instances anymore; either pass ``--all`` or pass the desired
2757 instances; this helps against mistakes on big clusters where listing
2758 the information for all instances takes a long time
2759 - miscellaneous doc and man pages fixes
2765 *(Released Wed, 8 Apr 2009)*
2767 - Change the internal locking model of some ``gnt-node`` commands, in
2768 order to reduce contention (and blocking of master daemon) when
2769 batching many creation/reinstall jobs
2770 - Fixes to Xen soft reboot
2771 - No longer build documentation at build time, instead distribute it in
2772 the archive, in order to reduce the need for the whole docbook/rst
2779 *(Released Fri, 27 Mar 2009)*
2781 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2782 - Fix watcher startup sequence, improves the behaviour of busy clusters
2783 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2784 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2785 bind address and other places
2786 - Some documentation fixes and updates
2792 *(Released Mon, 2 Mar 2009)*
2794 - More documentation updates, now all docs should be more-or-less
2796 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2798 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2800 - Fix an issue related to $libdir/run/ganeti and cluster creation
2806 *(Released Thu, 19 Feb 2009)*
2808 - Xen PVM and KVM have switched the default value for the instance root
2809 disk to the first partition on the first drive, instead of the whole
2810 drive; this means that the OS installation scripts must be changed
2812 - Man pages have been updated
2813 - RAPI has been switched by default to HTTPS, and the exported functions
2814 should all work correctly
2815 - RAPI v1 has been removed
2816 - Many improvements to the KVM hypervisor
2817 - Block device errors are now better reported
2818 - Many other bugfixes and small improvements
2824 *(Released Mon, 26 Jan 2009)*
2826 - Version 2 is a general rewrite of the code and therefore the
2827 differences are too many to list, see the design document for 2.0 in
2828 the ``doc/`` subdirectory for more details
2829 - In this beta version there is not yet a migration path from 1.2 (there
2830 will be one in the final 2.0 release)
2831 - A few significant changes are:
2833 - all commands are executed by a daemon (``ganeti-masterd``) and the
2834 various ``gnt-*`` commands are just front-ends to it
2835 - all the commands are entered into, and executed from a job queue,
2836 see the ``gnt-job(8)`` manpage
2837 - the RAPI daemon supports read-write operations, secured by basic
2838 HTTP authentication on top of HTTPS
2839 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2840 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2841 to migrate to DRBD 8 first while still running Ganeti 1.2)
2842 - DRBD devices are using statically allocated minor numbers, which
2843 will be assigned to existing instances during the migration process
2844 - there is support for both Xen PVM and Xen HVM instances running on
2846 - KVM virtualization is supported too
2847 - file-based storage has been implemented, which means that it is
2848 possible to run the cluster without LVM and DRBD storage, for
2849 example using a shared filesystem exported from shared storage (and
2850 still have live migration)
2856 *(Released Tue, 13 Jan 2009)*
2858 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2859 - Reuse the old instance mac address by default on instance import, if
2860 the instance name is the same.
2861 - Handle situations in which the node info rpc returns incomplete
2863 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2864 - Improved version of batcher:
2866 - state file support
2867 - instance mac address support
2868 - support for HVM clusters/instances
2870 - Add an option to show the number of cpu sockets and nodes in
2872 - Support OSes that handle more than one version of the OS api (but do
2873 not change the current API in any other way)
2874 - Fix ``gnt-node migrate``
2875 - ``gnt-debug`` man page
2876 - Fixes various more typos and small issues
2877 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2883 *(Released Wed, 24 Sep 2008)*
2885 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2886 type of disk exported to fully virtualized instances.
2887 - provide access to the serial console of HVM instances
2888 - instance auto_balance flag, set by default. If turned off it will
2889 avoid warnings on cluster verify if there is not enough memory to fail
2890 over an instance. in the future it will prevent automatically failing
2891 it over when we will support that.
2892 - batcher tool for instance creation, see ``tools/README.batcher``
2893 - ``gnt-instance reinstall --select-os`` to interactively select a new
2894 operating system when reinstalling an instance.
2895 - when changing the memory amount on instance modify a check has been
2896 added that the instance will be able to start. also warnings are
2897 emitted if the instance will not be able to fail over, if auto_balance
2899 - documentation fixes
2900 - sync fields between ``gnt-instance list/modify/add/import``
2901 - fix a race condition in drbd when the sync speed was set after giving
2902 the device a remote peer.
2908 *(Released Tue, 22 Jul 2008)*
2910 - note: the allowed size and number of tags per object were reduced
2911 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2912 - fixed twisted 8.x compatibility
2913 - fixed ``gnt-instance replace-disks`` with iallocator
2914 - add TCP keepalives on twisted connections to detect restarted nodes
2915 - disk increase support, see ``gnt-instance grow-disk``
2916 - implement bulk node/instance query for RAPI
2917 - add tags in node/instance listing (optional)
2918 - experimental migration (and live migration) support, read the man page
2919 for ``gnt-instance migrate``
2920 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2921 has some small improvements in handling its state file
2927 *(Released Fri, 13 Jun 2008)*
2929 - Experimental readonly, REST-based remote API implementation;
2930 automatically started on master node, TCP port 5080, if enabled by
2931 ``--enable-rapi`` parameter to configure script.
2932 - Instance allocator support. Add and import instance accept a
2933 ``--iallocator`` parameter, and call that instance allocator to decide
2934 which node to use for the instance. The iallocator document describes
2935 what's expected from an allocator script.
2936 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2937 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2938 node is lost there is still enough memory to fail over the instances
2940 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2941 ``gnt-cluster verify``, to check for site-specific compliance. All the
2942 hooks will run, and their output, if any, will be displayed. Any
2943 failing hook will make the verification return an error value.
2944 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2945 primary and secondary interfaces
2946 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2947 that is still declared as part of the cluster and has failed.
2948 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2949 specifying output fields, that just adds the chosen fields to the
2951 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2952 export from the filesystem.
2953 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2954 have been added. Using them you can enable/disable acpi and pae
2955 support, and specify a path for a cd image to be exported to the
2956 instance. These parameters as the name suggest only work on HVM
2958 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2959 PAE support will be set to the previously hardcoded values, but the
2960 (previously hardcoded) path to the CDROM ISO image will be unset and
2961 if required, needs to be set manually with ``gnt-instance modify``
2963 - The address to which an instance's VNC console is bound is now
2964 selectable per-instance, rather than being cluster wide. Of course
2965 this only applies to instances controlled via VNC, so currently just
2966 applies to HVM clusters.
2972 *(Released Mon, 18 Feb 2008)*
2974 - more tweaks to the disk activation code (especially helpful for DRBD)
2975 - change the default ``gnt-instance list`` output format, now there is
2976 one combined status field (see the manpage for the exact values this
2978 - some more fixes for the mac export to hooks change
2979 - make Ganeti not break with DRBD 8.2.x (which changed the version
2980 format in ``/proc/drbd``) (issue 24)
2981 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2982 template, allowing migration from DRBD0.7+MD to DRBD8
2988 *(Released Wed, 30 Jan 2008)*
2990 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2992 - add command aliases infrastructure and a few aliases
2993 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2994 man pages and the ``--help`` option of ``gnt-node
2995 list``/``gnt-instance list``
2996 - fix ``gnt-backup list`` with down nodes (issue 21)
2997 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2998 - fix the dist archive and add a check for including svn/git files in
3000 - some developer-related changes: improve the burnin and the QA suite,
3001 add an upload script for testing during development
3007 *(Released Wed, 16 Jan 2008)*
3009 - experimental HVM support, read the install document, section
3010 "Initializing the cluster"
3011 - allow for the PVM hypervisor per-instance kernel and initrd paths
3012 - add a new command ``gnt-cluster verify-disks`` which uses a new
3013 algorithm to improve the reconnection of the DRBD pairs if the device
3014 on the secondary node has gone away
3015 - make logical volume code auto-activate LVs at disk activation time
3016 - slightly improve the speed of activating disks
3017 - allow specification of the MAC address at instance creation time, and
3018 changing it later via ``gnt-instance modify``
3019 - fix handling of external commands that generate lots of output on
3021 - update documentation with regard to minimum version of DRBD8 supported
3027 *(Released Tue, 4 Dec 2007)*
3029 - Log the ``xm create`` output to the node daemon log on failure (to
3030 help diagnosing the error)
3031 - In debug mode, log all external commands output if failed to the logs
3032 - Change parsing of lvm commands to ignore stderr
3038 *(Released Wed, 28 Nov 2007)*
3040 - Another round of updates to the DRBD 8 code to deal with more failures
3041 in the replace secondary node operation
3042 - Some more logging of failures in disk operations (lvm, drbd)
3043 - A few documentation updates
3050 *(Released Tue, 13 Nov 2007)*
3052 - Change configuration file format from Python's Pickle to JSON.
3053 Upgrading is possible using the cfgupgrade utility.
3054 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3055 faster replace disks and is more stable (DRBD 8 has many improvements
3056 compared to DRBD 0.7)
3057 - Added command line tags support (see man pages for ``gnt-instance``,
3058 ``gnt-node``, ``gnt-cluster``)
3059 - Added instance rename support
3060 - Added multi-instance startup/shutdown
3061 - Added cluster rename support
3062 - Added ``gnt-node evacuate`` to simplify some node operations
3063 - Added instance reboot operation that can speedup reboot as compared to
3065 - Soften the requirement that hostnames are in FQDN format
3066 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3068 - Removed dependency on debian's patched fping that uses the
3069 non-standard ``-S`` option
3070 - Now the OS definitions are searched for in multiple, configurable
3071 paths (easier for distros to package)
3072 - Some changes to the hooks infrastructure (especially the new
3073 post-configuration update hook)
3074 - Other small bugfixes
3076 .. vim: set textwidth=72 syntax=rst :