10 - Ensure that all the hypervisors exist in the config file (Issue 640)
16 *(Released Fri, 13 Dec 2013)*
18 - use custom KVM path if set for version checking
19 - SingleNotifyPipeCondition: don't share pollers
21 Inherited from the 2.8 branch:
23 - Fixed Luxi daemon socket permissions after master-failover
24 - Improve IP version detection code directly checking for colons rather than
25 passing the family from the cluster object
26 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
27 opportunistically anymore (Issue 622)
28 - Allow link local IPv6 gateways (Issue 624)
29 - Fix error printing (Issue 616)
30 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
31 disk modifications, keep the old one. If name=none then set disk name to
33 - Update build_chroot script to work with the latest hackage packages
34 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
35 - Fix evacuation out of drained node (Issue 615)
36 - Add default file_driver if missing (Issue 571)
37 - Fix job error message after unclean master shutdown (Issue 618)
38 - Lock group(s) when creating instances (Issue 621)
39 - SetDiskID() before accepting an instance (Issue 633)
40 - Allow the ext template disks to receive arbitrary parameters, both at creation
41 time and while being modified
42 - Xen handle domain shutdown (future proofing cherry-pick)
43 - Refactor reading live data in htools (future proofing cherry-pick)
49 *(Released Wed, 13 Nov 2013)*
51 - fix bug, that kept nodes offline when readding
52 - when verifying DRBD versions, ignore unavailable nodes
53 - fix bug that made the console unavailable on kvm in split-user
55 - DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
61 *(Released Tue, 5 Nov 2013)*
63 Incompatible/important changes
64 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
66 - hroller now also plans for capacity to move non-redundant instances off
67 any node to be rebooted; the old behavior of completely ignoring any
68 non-redundant instances can be restored by adding the --ignore-non-redundant
70 - The cluster option '--no-lvm-storage' was removed in favor of the new option
71 '--enabled-disk-templates'.
72 - On instance creation, disk templates no longer need to be specified
73 with '-t'. The default disk template will be taken from the list of
74 enabled disk templates.
75 - The monitoring daemon is now running as root, in order to be able to collect
76 information only available to root (such as the state of Xen instances).
77 - The ConfD client is now IPv6 compatible.
78 - File and shared file storage is no longer dis/enabled at configure time,
79 but using the option '--enabled-disk-templates' at cluster initialization and
81 - The default directories for file and shared file storage are not anymore
82 specified at configure time, but taken from the cluster's configuration.
83 They can be set at cluster initialization and modification with
84 '--file-storage-dir' and '--shared-file-storage-dir'.
85 - Cluster verification now includes stricter checks regarding the
86 default file and shared file storage directories. It now checks that
87 the directories are explicitely allowed in the 'file-storage-paths' file and
88 that the directories exist on all nodes.
89 - The list of allowed disk templates in the instance policy and the list
90 of cluster-wide enabled disk templates is now checked for consistency
91 on cluster or group modification. On cluster initialization, the ipolicy
92 disk templates are ensured to be a subset of the cluster-wide enabled
98 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
99 the correct command syntax. It is possible to use different DRBD versions
100 on different nodes as long as they are compatible to each other. This
101 enables rolling upgrades of DRBD with no downtime. As permanent operation
102 of different DRBD versions within a node group is discouraged,
103 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
104 - New "inst-status-xen" data collector for the monitoring daemon, providing
105 information about the state of the xen instances on the nodes.
106 - New "lv" data collector for the monitoring daemon, collecting data about the
107 logical volumes on the nodes, and pairing them with the name of the instances
109 - New "diskstats" data collector, collecting the data from /proc/diskstats and
110 presenting them over the monitoring daemon interface.
111 - The ConfD client is now IPv6 compatible.
115 The following new dependencies have been added.
119 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
120 for the unit tests (and only used for testing).
124 - ``hslogger`` (http://software.complete.org/hslogger) is now always
125 required, even if confd is not enabled.
130 - Correctly start/stop luxid during gnt-cluster master-failover (inherited
132 - Improved error messsages (inherited from stable-2.8)
138 *(Released Tue, 15 Oct 2013)*
140 The third release candidate in the 2.9 series. Since 2.9.0 rc2:
142 - in implicit configuration upgrade, match ipolicy with enabled disk templates
143 - improved harep documentation (inherited from stable-2.8)
149 *(Released Wed, 9 Oct 2013)*
151 The second release candidate in the 2.9 series. Since 2.9.0 rc1:
153 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
154 at least one DRBD instance.
155 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
162 *(Released Tue, 1 Oct 2013)*
164 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
167 - update of the documentation, in particular installation instructions
168 - merging of LD_* constants into DT_* constants
169 - python style changes to be compatible with newer versions of pylint
175 *(Released Thu, 29 Aug 2013)*
177 This was the first beta release of the 2.9 series. All important changes
178 are listed in the latest 2.9 entry.
184 *(Released Thu, 12 Dec 2013)*
186 - Fixed Luxi daemon socket permissions after master-failover
187 - Improve IP version detection code directly checking for colons rather than
188 passing the family from the cluster object
189 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
190 opportunistically anymore (Issue 622)
191 - Allow link local IPv6 gateways (Issue 624)
192 - Fix error printing (Issue 616)
193 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
194 disk modifications, keep the old one. If name=none then set disk name to
196 - Update build_chroot script to work with the latest hackage packages
197 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
198 - Fix evacuation out of drained node (Issue 615)
199 - Add default file_driver if missing (Issue 571)
200 - Fix job error message after unclean master shutdown (Issue 618)
201 - Lock group(s) when creating instances (Issue 621)
202 - SetDiskID() before accepting an instance (Issue 633)
203 - Allow the ext template disks to receive arbitrary parameters, both at creation
204 time and while being modified
205 - Xen handle domain shutdown (future proofing cherry-pick)
206 - Refactor reading live data in htools (future proofing cherry-pick)
212 *(Released Thu, 07 Nov 2013)*
214 - DRBD: ensure peers are UpToDate for dual-primary
215 - Improve error message for replace-disks
216 - More dependency checks at configure time
217 - Placate warnings on ganeti.outils_unittest.py
223 *(Released Thu, 17 Oct 2013)*
225 - Correctly start/stop luxid during gnt-cluster master-failover
226 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
227 - Fix path for the job queue serial file
228 - Improved harep man page
229 - Minor documentation improvements
235 *(Released Mon, 30 Sep 2013)*
237 Incompatible/important changes
238 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
240 - Instance policy can contain multiple instance specs, as described in
241 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
242 <design-partitioned>`. As a consequence, it's not possible to partially change
243 or override instance specs. Bounding specs (min and max) can be specified as a
244 whole using the new option ``--ipolicy-bounds-specs``, while standard
245 specs use the new option ``--ipolicy-std-specs``.
246 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
247 gnt-instance is a valid YAML object.
248 - hail now honors network restrictions when allocating nodes. This led to an
249 update of the IAllocator protocol. See the IAllocator documentation for
251 - confd now only answers static configuration request over the network. luxid
252 was extracted, listens on the local LUXI socket and responds to live queries.
253 This allows finer grained permissions if using separate users.
258 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
259 to always require authentication, ``--require-authentication``. It can
260 be specified in ``$sysconfdir/default/ganeti``.
261 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
262 be used to manage the disk templates to be used by instances in the cluster.
263 Initially, it will be set to a list that includes plain, drbd, if they were
264 enabled by specifying a volume group name, and file and sharedfile, if those
265 were enabled at configure time. Additionally, it will include all disk
266 templates that are currently used by instances. The order of disk templates
267 will be based on Ganeti's history of supporting them. In the future, the
268 first entry of the list will be used as a default disk template on instance
270 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
271 configuration back to the previous stable version.
272 - Disk templates in group ipolicy can be restored to the default value.
273 - Initial support for diskless instances and virtual clusters in QA.
274 - More QA and unit tests for instance policies.
275 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
276 describing why the opcode itself was executed.
277 - The monitoring daemon is now available. It allows users to query the cluster
278 for obtaining information about the status of the system. The daemon is only
279 responsible for providing the information over the network: the actual data
280 gathering is performed by data collectors (currently, only the DRBD status
281 collector is available).
282 - In order to help developers work on Ganeti, a new script
283 (``devel/build_chroot``) is provided, for building a chroot that contains all
284 the required development libraries and tools for compiling Ganeti on a Debian
286 - A new tool, ``harep``, for performing self-repair and recreation of instances
287 in Ganeti has been added.
288 - Split queries are enabled for tags, network, exports, cluster info, groups,
290 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
291 It prints the command line to set the current policies, to ease
293 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
294 devices for KVM virtio-net interfaces will get created with VNET_HDR
295 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
296 virtio-net interfaces, which prevents host kernel tainting and log
297 flooding, when dealing with broken or malicious virtio-net drivers.
298 It's set to true by default.
299 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
301 - Support 'viridian' parameter in Xen HVM
302 - Support DSA SSH keys in bootstrap
303 - To simplify the work of packaging frameworks that want to add the needed users
304 and groups in a split-user setup themselves, at build time three files in
305 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
306 the groups to be generated, the ``users`` file contains, one per line, the
307 users to be generated, optionally followed by their primary group, where
308 important. The ``groupmemberships`` file contains, one per line, additional
309 user-group membership relations that need to be established. The syntax of
310 these files will remain stable in all future versions.
315 The following new dependencies have been added:
318 - The ``curl`` library is not optional anymore for compiling the Haskell code.
319 - ``snap-server`` library (if monitoring is enabled).
322 - The minimum Python version needed to run Ganeti is now 2.6.
323 - ``yaml`` library (only for running the QA).
327 - Perform proper cleanup on termination of Haskell daemons
328 - Fix corner-case in handling of remaining retry time
334 *(Released Tue, 17 Sep 2013)*
336 - To simplify the work of packaging frameworks that want to add the needed users
337 and groups in a split-user setup themselves, at build time three files in
338 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
339 the groups to be generated, the ``users`` file contains, one per line, the
340 users to be generated, optionally followed by their primary group, where
341 important. The ``groupmemberships`` file contains, one per line, additional
342 user-group membership relations that need to be established. The syntax of
343 these files will remain stable in all future versions.
344 - Add a default to file-driver when unspecified over RAPI (Issue 571)
345 - Mark the DSA host pubkey as optional, and remove it during config downgrade
347 - Some documentation fixes
353 *(Released Tue, 27 Aug 2013)*
355 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
357 - Support 'viridian' parameter in Xen HVM (Issue 233)
358 - Include VCS version in ``gnt-cluster version``
359 - Support DSA SSH keys in bootstrap (Issue 338)
360 - Fix batch creation of instances
361 - Use FQDN to check master node status (Issue 551)
362 - Make the DRBD collector more failure-resilient
368 *(Released Fri, 2 Aug 2013)*
370 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
372 - Fix upgrading/downgrading from 2.7
373 - Increase maximum RAPI message size
374 - Documentation updates
375 - Split ``confd`` between ``luxid`` and ``confd``
376 - Merge 2.7 series up to the 2.7.1 release
377 - Allow the ``modify_etc_hosts`` option to be changed
378 - Add better debugging for ``luxid`` queries
379 - Expose bulk parameter for GetJobs in RAPI client
380 - Expose missing ``network`` fields in RAPI
381 - Add some ``cluster verify`` tests
382 - Some unittest fixes
383 - Fix a malfunction in ``hspace``'s tiered allocation
384 - Fix query compatibility between haskell and python implementations
385 - Add the ``vnet_hdr`` HV parameter for KVM
386 - Add ``--cleanup`` to instance failover
387 - Change the connected groups format in ``gnt-network info`` output; it
388 was previously displayed as a raw list by mistake. (Merged from 2.7)
394 *(Released Mon, 24 Jun 2013)*
396 This was the first beta release of the 2.8 series. All important changes
397 are listed in the latest 2.8 entry.
403 *(Released Thu, 26 Sep 2013)*
405 - Change the connected groups format in ``gnt-network info`` output; it
406 was previously displayed as a raw list by mistake
407 - Check disk template in right dict when copying
408 - Support multi-instance allocs without iallocator
409 - Fix some errors in the documentation
410 - Fix formatting of tuple in an error message
416 *(Released Thu, 25 Jul 2013)*
418 - Add logrotate functionality in daemon-util
419 - Add logrotate example file
420 - Add missing fields to network queries over rapi
421 - Fix network object timestamps
422 - Add support for querying network timestamps
423 - Fix a typo in the example crontab
424 - Fix a documentation typo
430 *(Released Thu, 04 Jul 2013)*
432 Incompatible/important changes
433 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
435 - Instance policies for disk size were documented to be on a per-disk
436 basis, but hail applied them to the sum of all disks. This has been
438 - ``hbal`` will now exit with status 0 if, during job execution over
439 LUXI, early exit has been requested and all jobs are successful;
440 before, exit status 1 was used, which cannot be differentiated from
442 - Compatibility with newer versions of rbd has been fixed
443 - ``gnt-instance batch-create`` has been changed to use the bulk create
444 opcode from Ganeti. This lead to incompatible changes in the format of
445 the JSON file. It's now not a custom dict anymore but a dict
446 compatible with the ``OpInstanceCreate`` opcode.
447 - Parent directories for file storage need to be listed in
448 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
449 write the file automatically based on old configuration values, but it
450 can not distribute it across all nodes and the file contents should be
451 verified. Use ``gnt-cluster copyfile
452 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
453 upgraded. The reason for requiring this list of paths now is that
454 before it would have been possible to inject new paths via RPC,
455 allowing files to be created in arbitrary locations. The RPC protocol
456 is protected using SSL/X.509 certificates, but as a design principle
457 Ganeti does not permit arbitrary paths to be passed.
458 - The parsing of the variants file for OSes (see
459 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
460 lines and comment lines (starting with ``#``) are ignored for better
462 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
463 longer available. ``gnt-node add`` now invokes a new tool on the
464 destination node, named ``prepare-node-join``, to configure the SSH
465 daemon. Paramiko is no longer necessary to configure nodes' SSH
466 daemons via ``gnt-node add``.
467 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
468 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
469 the flag is not reset after a master failover.
470 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
471 is recommended. 2.6 will be mandatory from the 2.8 series.
477 - New network management functionality to support automatic allocation
478 of IP addresses and managing of network parameters. See
479 :manpage:`gnt-network(8)` for more details.
480 - New external storage backend, to allow managing arbitrary storage
481 systems external to the cluster. See
482 :manpage:`ganeti-extstorage-interface(7)`.
483 - New ``exclusive-storage`` node parameter added, restricted to
484 nodegroup level. When it's set to true, physical disks are assigned in
485 an exclusive fashion to instances, as documented in :doc:`Partitioned
486 Ganeti <design-partitioned>`. Currently, only instances using the
487 ``plain`` disk template are supported.
488 - The KVM hypervisor has been updated with many new hypervisor
489 parameters, including a generic one for passing arbitrary command line
490 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
491 compatible up to qemu 1.4.
492 - A new tool, called ``mon-collector``, is the stand-alone executor of
493 the data collectors for a monitoring system. As of this version, it
494 just includes the DRBD data collector, that can be executed by calling
495 ``mon-collector`` using the ``drbd`` parameter. See
496 :manpage:`mon-collector(7)`.
497 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
498 for RAPI users. It allows granting permissions to query for
499 information to a specific user without giving
500 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
501 - A new tool named ``node-cleanup`` has been added. It cleans remains of
502 a cluster from a machine by stopping all daemons, removing
503 certificates and ssconf files. Unless the ``--no-backup`` option is
504 given, copies of the certificates are made.
505 - Instance creations now support the use of opportunistic locking,
506 potentially speeding up the (parallel) creation of multiple instances.
507 This feature is currently only available via the :doc:`RAPI
508 <rapi>` interface and when an instance allocator is used. If the
509 ``opportunistic_locking`` parameter is set the opcode will try to
510 acquire as many locks as possible, but will not wait for any locks
511 held by other opcodes. If not enough resources can be found to
512 allocate the instance, the temporary error code
513 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
514 retried thereafter, with or without opportunistic locking.
515 - New experimental linux-ha resource scripts.
516 - Restricted-commands support: ganeti can now be asked (via command line
517 or rapi) to perform commands on a node. These are passed via ganeti
518 RPC rather than ssh. This functionality is restricted to commands
519 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
520 reasons. The file is not copied automatically.
526 - Diskless instances are now externally mirrored (Issue 237). This for
527 now has only been tested in conjunction with explicit target nodes for
529 - Queries not needing locks or RPC access to the node can now be
530 performed by the confd daemon, making them independent from jobs, and
531 thus faster to execute. This is selectable at configure time.
532 - The functionality for allocating multiple instances at once has been
533 overhauled and is now also available through :doc:`RAPI <rapi>`.
535 There are no significant changes from version 2.7.0~rc3.
541 *(Released Tue, 25 Jun 2013)*
543 - Fix permissions on the confd query socket (Issue 477)
544 - Fix permissions on the job archive dir (Issue 498)
545 - Fix handling of an internal exception in replace-disks (Issue 472)
546 - Fix gnt-node info handling of shortened names (Issue 497)
547 - Fix gnt-instance grow-disk when wiping is enabled
548 - Documentation improvements, and support for newer pandoc
549 - Fix hspace honoring ipolicy for disks (Issue 484)
550 - Improve handling of the ``kvm_extra`` HV parameter
556 *(Released Fri, 24 May 2013)*
558 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
560 - Disks added through ``gnt-instance modify`` or created through
561 ``gnt-instance recreate-disks`` are wiped, if the
562 ``prealloc_wipe_disks`` flag is set.
563 - If wiping newly created disks fails, the disks are removed. Also,
564 partial failures in creating disks through ``gnt-instance modify``
565 triggers a cleanup of the partially-created disks.
566 - Removing the master IP address doesn't fail if the address has been
568 - Fix ownership of the OS log dir
569 - Workaround missing SO_PEERCRED constant (Issue 191)
575 *(Released Fri, 3 May 2013)*
577 This was the first release candidate of the 2.7 series. Since beta3:
579 - Fix kvm compatibility with qemu 1.4 (Issue 389)
580 - Documentation updates (admin guide, upgrade notes, install
581 instructions) (Issue 372)
582 - Fix gnt-group list nodes and instances count (Issue 436)
583 - Fix compilation without non-mandatory libraries (Issue 441)
584 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
585 - Make confd logging more verbose at INFO level (Issue 435)
586 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
587 - Fix failure path for instance storage type conversion (Issue 229)
588 - Update htools text backend documentation
589 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
590 - Disable inter-cluster instance move for file-based instances, because
591 it is dependant on instance export, which is not supported for
592 file-based instances. (Issue 414)
593 - Fix gnt-job crashes on non-ascii characters (Issue 427)
594 - Fix volume group checks on non-vm-capable nodes (Issue 432)
600 *(Released Mon, 22 Apr 2013)*
602 This was the third beta release of the 2.7 series. Since beta2:
604 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
605 - Fix data loss on wrong usage of ``gnt-instance move``
606 - Properly export errors in confd-based job queries
607 - Add ``users-setup`` tool
608 - Fix iallocator protocol to report 0 as a disk size for diskless
609 instances. This avoids hail breaking when a diskless instance is
611 - Fix job queue directory permission problem that made confd job queries
612 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
613 for access to archived jobs (Issue 406).
614 - Limit the sizes of networks supported by ``gnt-network`` to something
615 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
616 - Fix bugs in instance disk template conversion
617 - Fix GHC 7 compatibility
618 - Fix ``burnin`` install path (Issue 426).
619 - Allow very small disk grows (Issue 347).
620 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
621 that noded doesn't import masterd code (Issue 419).
622 - Make sure the default metavg at cluster init is the same as the vg, if
623 unspecified (Issue 358).
624 - Fix cleanup of partially created disks (part of Issue 416)
630 *(Released Tue, 2 Apr 2013)*
632 This was the second beta release of the 2.7 series. Since beta1:
634 - Networks no longer have a "type" slot, since this information was
635 unused in Ganeti: instead of it tags should be used.
636 - The rapi client now has a ``target_node`` option to MigrateInstance.
637 - Fix early exit return code for hbal (Issue 386).
638 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
639 - Fix ``rbd showmapped`` output parsing (Issue 312).
640 - Networks are now referenced indexed by UUID, rather than name. This
641 will require running cfgupgrade, from 2.7.0beta1, if networks are in
643 - The OS environment now includes network information.
644 - Deleting of a network is now disallowed if any instance nic is using
645 it, to prevent dangling references.
646 - External storage is now documented in man pages.
647 - The exclusive_storage flag can now only be set at nodegroup level.
648 - Hbal can now submit an explicit priority with its jobs.
649 - Many network related locking fixes.
650 - Bump up the required pylint version to 0.25.1.
651 - Fix the ``no_remember`` option in RAPI client.
652 - Many ipolicy related tests, qa, and fixes.
653 - Many documentation improvements and fixes.
654 - Fix building with ``--disable-file-storage``.
655 - Fix ``-q`` option in htools, which was broken if passed more than
657 - Some haskell/python interaction improvements and fixes.
658 - Fix iallocator in case of missing LVM storage.
659 - Fix confd config load in case of ``--no-lvm-storage``.
660 - The confd/query functionality is now mentioned in the security
667 *(Released Wed, 6 Feb 2013)*
669 This was the first beta release of the 2.7 series. All important changes
670 are listed in the latest 2.7 entry.
676 *(Released Fri, 21 Dec 2012)*
678 Important behaviour change: hbal won't rebalance anymore instances which
679 have the ``auto_balance`` attribute set to false. This was the intention
680 all along, but until now it only skipped those from the N+1 memory
681 reservation (DRBD-specific).
683 A significant number of bug fixes in this release:
685 - Fixed disk adoption interaction with ipolicy checks.
686 - Fixed networking issues when instances are started, stopped or
687 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
688 - Fixed the warning in cluster verify for shared storage instances not
690 - Fixed removal of storage directory on shared file storage (issue 262).
691 - Fixed validation of LVM volume group name in OpClusterSetParams
692 (``gnt-cluster modify``) (issue 285).
693 - Fixed runtime memory increases (``gnt-instance modify -m``).
694 - Fixed live migration under Xen's ``xl`` mode.
695 - Fixed ``gnt-instance console`` with ``xl``.
696 - Fixed building with newer Haskell compiler/libraries.
697 - Fixed PID file writing in Haskell daemons (confd); this prevents
698 restart issues if confd was launched manually (outside of
699 ``daemon-util``) while another copy of it was running
700 - Fixed a type error when doing live migrations with KVM (issue 297) and
701 the error messages for failing migrations have been improved.
702 - Fixed opcode validation for the out-of-band commands (``gnt-node
704 - Fixed a type error when unsetting OS hypervisor parameters (issue
705 311); now it's possible to unset all OS-specific hypervisor
707 - Fixed the ``dry-run`` mode for many operations: verification of
708 results was over-zealous but didn't take into account the ``dry-run``
709 operation, resulting in "wrong" failures.
710 - Fixed bash completion in ``gnt-job list`` when the job queue has
711 hundreds of entries; especially with older ``bash`` versions, this
712 results in significant CPU usage.
714 And lastly, a few other improvements have been made:
716 - Added option to force master-failover without voting (issue 282).
717 - Clarified error message on lock conflict (issue 287).
718 - Logging of newly submitted jobs has been improved (issue 290).
719 - Hostname checks have been made uniform between instance rename and
721 - The ``--submit`` option is now supported by ``gnt-debug delay``.
722 - Shutting down the master daemon by sending SIGTERM now stops it from
723 processing jobs waiting for locks; instead, those jobs will be started
724 once again after the master daemon is started the next time (issue
726 - Support for Xen's ``xl`` program has been improved (besides the fixes
728 - Reduced logging noise in the Haskell confd daemon (only show one log
729 entry for each config reload, instead of two).
730 - Several man page updates and typo fixes.
736 *(Released Fri, 12 Oct 2012)*
738 A small bugfix release. Among the bugs fixed:
740 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
741 made the command unusable.
742 - Commands that issue many jobs don't fail anymore just because some jobs
743 take so long that other jobs are archived.
744 - Failures during ``gnt-instance reinstall`` are reflected by the exit
746 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
748 - When ``always_failover`` is set, ``--allow-failover`` is not required
749 in migrate commands anymore.
750 - ``bash_completion`` works even if extglob is disabled.
751 - Fixed bug with locks that made failover for RDB-based instances fail.
752 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
753 a random node instead of one based on the allocator metric.
754 - Support for newer versions of pylint and pep8.
755 - Hail doesn't fail anymore when trying to add an instance of type
756 ``file``, ``sharedfile`` or ``rbd``.
757 - Added new Makefile target to rebuild the whole distribution, so that
758 all files are included.
764 *(Released Fri, 27 Jul 2012)*
767 .. attention:: The ``LUXI`` protocol has been made more consistent
768 regarding its handling of command arguments. This, however, leads to
769 incompatibility issues with previous versions. Please ensure that you
770 restart Ganeti daemons soon after the upgrade, otherwise most
771 ``LUXI`` calls (job submission, setting/resetting the drain flag,
772 pausing/resuming the watcher, cancelling and archiving jobs, querying
773 the cluster configuration) will fail.
782 The current ``admin_up`` field, which used to denote whether an instance
783 should be running or not, has been removed. Instead, ``admin_state`` is
784 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
786 The rational behind this is that an instance being “down” can have
789 - it could be down during a reboot
790 - it could be temporarily be down for a reinstall
791 - or it could be down because it is deprecated and kept just for its
794 The previous Boolean state was making it difficult to do capacity
795 calculations: should Ganeti reserve memory for a down instance? Now, the
796 tri-state field makes it clear:
798 - in ``up`` and ``down`` state, all resources are reserved for the
799 instance, and it can be at any time brought up if it is down
800 - in ``offline`` state, only disk space is reserved for it, but not
803 The field can have an extra use: since the transition between ``up`` and
804 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
805 transition between ``offline`` and ``down`` is done via ``gnt-instance
806 modify``, it is possible to given different rights to users. For
807 example, owners of an instance could be allowed to start/stop it, but
808 not transition it out of the offline state.
810 Instance policies and specs
811 +++++++++++++++++++++++++++
813 In previous Ganeti versions, an instance creation request was not
814 limited on the minimum size and on the maximum size just by the cluster
815 resources. As such, any policy could be implemented only in third-party
816 clients (RAPI clients, or shell wrappers over ``gnt-*``
817 tools). Furthermore, calculating cluster capacity via ``hspace`` again
818 required external input with regards to instance sizes.
820 In order to improve these workflows and to allow for example better
821 per-node group differentiation, we introduced instance specs, which
824 - minimum instance disk size, disk count, memory size, cpu count
825 - maximum values for the above metrics
826 - and “standard” values (used in ``hspace`` to calculate the standard
829 The minimum/maximum values can be also customised at node-group level,
830 for example allowing more powerful hardware to support bigger instance
833 Beside the instance specs, there are a few other settings belonging to
834 the instance policy framework. It is possible now to customise, per
835 cluster and node-group:
837 - the list of allowed disk templates
838 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
839 - the maximum ratio of instance to spindles (see below for more
840 information) for local storage
842 All these together should allow all tools that talk to Ganeti to know
843 what are the ranges of allowed values for instances and the
844 over-subscription that is allowed.
846 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
847 instance configuration, and the physical CPU configuration from the
848 node. For the spindle ratios however, we didn't track before these
849 values, so new parameters have been added:
851 - a new node parameter ``spindle_count``, defaults to 1, customisable at
852 node group or node level
853 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
855 Note that spindles in this context doesn't need to mean actual
856 mechanical hard-drives; it's just a relative number for both the node
857 I/O capacity and instance I/O consumption.
859 Instance migration behaviour
860 ++++++++++++++++++++++++++++
862 While live-migration is in general desirable over failover, it is
863 possible that for some workloads it is actually worse, due to the
864 variable time of the “suspend” phase during live migration.
866 To allow the tools to work consistently over such instances (without
867 having to hard-code instance names), a new backend parameter
868 ``always_failover`` has been added to control the migration/failover
869 behaviour. When set to True, all migration requests for an instance will
870 instead fall-back to failover.
872 Instance memory ballooning
873 ++++++++++++++++++++++++++
875 Initial support for memory ballooning has been added. The memory for an
876 instance is no longer fixed (backend parameter ``memory``), but instead
877 can vary between minimum and maximum values (backend parameters
878 ``minmem`` and ``maxmem``). Currently we only change an instance's
881 - live migrating or failing over and instance and the target node
882 doesn't have enough memory
883 - user requests changing the memory via ``gnt-instance modify
889 In order to control the use of specific CPUs by instance, support for
890 controlling CPU pinning has been added for the Xen, HVM and LXC
891 hypervisors. This is controlled by a new hypervisor parameter
892 ``cpu_mask``; details about possible values for this are in the
893 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
894 VCPU-to-CPU mapping) form will work well only when all nodes in your
895 cluster have the same amount of CPUs.
900 Another area in which Ganeti was not customisable were the parameters
901 used for storage configuration, e.g. how many stripes to use for LVM,
902 DRBD resync configuration, etc.
904 To improve this area, we've added disks parameters, which are
905 customisable at cluster and node group level, and which allow to
906 specify various parameters for disks (DRBD has the most parameters
907 currently), for example:
909 - DRBD resync algorithm and parameters (e.g. speed)
910 - the default VG for meta-data volumes for DRBD
911 - number of stripes for LVM (plain disk template)
914 These parameters can be modified via ``gnt-cluster modify -D …`` and
915 ``gnt-group modify -D …``, and are used at either instance creation (in
916 case of LVM stripes, for example) or at disk “activation” time
919 Rados block device support
920 ++++++++++++++++++++++++++
922 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
923 denoted by the ``rbd`` disk template type. This is considered
924 experimental, feedback is welcome. For details on configuring it, see
925 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
930 The existing master IP functionality works well only in simple setups (a
931 single network shared by all nodes); however, if nodes belong to
932 different networks, then the ``/32`` setup and lack of routing
933 information is not enough.
935 To allow the master IP to function well in more complex cases, the
936 system was reworked as follows:
938 - a master IP netmask setting has been added
939 - the master IP activation/turn-down code was moved from the node daemon
941 - whether to run the Ganeti-supplied master IP script or a user-supplied
942 on is a ``gnt-cluster init`` setting
944 Details about the location of the standard and custom setup scripts are
945 in the man page :manpage:`gnt-cluster(8)`; for information about the
946 setup script protocol, look at the Ganeti-supplied script.
951 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
954 It is now possible to use TLS-protected connections, and when renewing
955 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
956 it is now possible to specify spice or spice CA certificates. Also, it
957 is possible to configure a password for SPICE sessions via the
958 hypervisor parameter ``spice_password_file``.
960 There are also new parameters to control the compression and streaming
961 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
962 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
963 for the spice parameters.
965 Lastly, it is now possible to see the SPICE connection information via
966 ``gnt-instance console``.
971 A new tool (``tools/ovfconverter``) has been added that supports
972 conversion between Ganeti and the `Open Virtualization Format
973 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
976 This relies on the ``qemu-img`` tool to convert the disk formats, so the
977 actual compatibility with other virtualization solutions depends on it.
982 The configuration query daemon (``ganeti-confd``) is now optional, and
983 has been rewritten in Haskell; whether to use the daemon at all, use the
984 Python (default) or the Haskell version is selectable at configure time
985 via the ``--enable-confd`` parameter, which can take one of the
986 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
987 daemon will result in a smaller footprint; for larger systems, we
988 welcome feedback on the Haskell version which might become the default
991 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
992 daemon running. The Python version doesn't implement the new call.
995 User interface changes
996 ~~~~~~~~~~~~~~~~~~~~~~
998 We have replaced the ``--disks`` option of ``gnt-instance
999 replace-disks`` with a more flexible ``--disk`` option, which allows
1000 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
1001 disk size and mode can be changed upon recreation (via ``gnt-instance
1002 recreate-disks``, which accepts the same ``--disk`` option).
1004 As many people are used to a ``show`` command, we have added that as an
1005 alias to ``info`` on all ``gnt-*`` commands.
1007 The ``gnt-instance grow-disk`` command has a new mode in which it can
1008 accept the target size of the disk, instead of the delta; this can be
1009 more safe since two runs in absolute mode will be idempotent, and
1010 sometimes it's also easier to specify the desired size directly.
1012 Also the handling of instances with regard to offline secondaries has
1013 been improved. Instance operations should not fail because one of it's
1014 secondary nodes is offline, even though it's safe to proceed.
1016 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
1017 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1018 minors to instance name.
1023 RAPI coverage has improved, with (for example) new resources for
1024 recreate-disks, node power-cycle, etc.
1029 There is partial support for ``xl`` in the Xen hypervisor; feedback is
1032 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1033 whether to still support Python 2.4 or move to Python 2.6 as minimum
1036 Support for Fedora has been slightly improved; the provided example
1037 init.d script should work better on it and the INSTALL file should
1038 document the needed dependencies.
1043 The deprecated ``QueryLocks`` LUXI request has been removed. Use
1044 ``Query(what=QR_LOCK, ...)`` instead.
1046 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1047 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1048 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1049 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1050 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1052 RAPI client: ``CertificateError`` now derives from
1053 ``GanetiApiError``. This should make it more easy to handle Ganeti
1056 Deprecation warnings due to PyCrypto/paramiko import in
1057 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
1058 make sure to run an up-to-date paramiko version, if you use this tool.
1060 The QA scripts now depend on Python 2.5 or above (the main code base
1061 still works with Python 2.4).
1063 The configuration file (``config.data``) is now written without
1064 indentation for performance reasons; if you want to edit it, it can be
1065 re-formatted via ``tools/fmtjson``.
1067 A number of bugs has been fixed in the cluster merge tool.
1069 ``x509`` certification verification (used in import-export) has been
1070 changed to allow the same clock skew as permitted by the cluster
1071 verification. This will remove some rare but hard to diagnose errors in
1078 *(Released Thu, 19 Jul 2012)*
1080 Very few changes from rc4 to the final release, only bugfixes:
1082 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
1083 instance, fix CDROM booting for KVM instances)
1084 - fixed node group modification of node parameters
1085 - fixed issue in LUClusterVerifyGroup with multi-group clusters
1086 - fixed generation of bash completion to ensure a stable ordering
1093 *(Released Fri, 13 Jul 2012)*
1095 Third release candidate for 2.6. The following changes were done from
1098 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1099 - Fixed an inconsistency in the LUXI protocol with the provided
1100 arguments (NOT backwards compatible)
1101 - Fixed a bug with node groups ipolicy where ``min`` was greater than
1102 the cluster ``std`` value
1103 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1104 easier instance debugging on nodes (requires ``hconfd`` to work)
1110 *(Released Tue, 03 Jul 2012)*
1112 Second release candidate for 2.6. The following changes were done from
1115 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1117 - Fixed a RAPI regression on missing beparams/memory
1118 - Fixed redistribution of files on offline nodes
1119 - Added possibility to run activate-disks even though secondaries are
1120 offline. With this change it relaxes also the strictness on some other
1121 commands which use activate disks internally:
1122 * ``gnt-instance start|reboot|rename|backup|export``
1123 - Made it possible to remove safely an instance if its secondaries are
1125 - Made it possible to reinstall even though secondaries are offline
1131 *(Released Mon, 25 Jun 2012)*
1133 First release candidate for 2.6. The following changes were done from
1136 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1138 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1139 behaviour. It supports now the modify operation
1140 - ``hcheck`` implemented to analyze cluster health and possibility of
1141 improving health by rebalance
1142 - ``hbal`` has been improved in dealing with split instances
1148 *(Released Mon, 11 Jun 2012)*
1150 Second beta release of 2.6. The following changes were done from beta2
1153 - Fixed ``daemon-util`` with non-root user models
1154 - Fixed creation of plain instances with ``--no-wait-for-sync``
1155 - Fix wrong iv_names when running ``cfgupgrade``
1156 - Export more information in RAPI group queries
1157 - Fixed bug when changing instance network interfaces
1158 - Extended burnin to do NIC changes
1159 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1160 - Changed default for DRBD barriers
1161 - Fixed DRBD error reporting for syncer rate
1162 - Verify the options on disk parameters
1164 And of course various fixes to documentation and improved unittests and
1171 *(Released Wed, 23 May 2012)*
1173 First beta release of 2.6. The following changes were done from beta1 to
1176 - integrated patch for distributions without ``start-stop-daemon``
1177 - adapted example init.d script to work on Fedora
1178 - fixed log handling in Haskell daemons
1179 - adapted checks in the watcher for pycurl linked against libnss
1180 - add partial support for ``xl`` instead of ``xm`` for Xen
1181 - fixed a type issue in cluster verification
1182 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1185 Plus integrated fixes from the 2.5 branch:
1187 - fixed ``kvm-ifup`` to use ``/bin/bash``
1188 - fixed parallel build failures
1189 - KVM live migration when using a custom keymap
1195 *(Released Tue, 24 Jul 2012)*
1197 A small bugfix release, with no new features:
1199 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1200 different default shell (e.g. Debian, Ubuntu)
1201 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1202 243 and Debian bug #650664)
1203 - fixed compatibility with KVM versions that don't support multiple boot
1204 devices (fixes Issue 230 and Debian bug #624256)
1206 Additionally, a few fixes were done to the build system (fixed parallel
1207 build failures) and to the unittests (fixed race condition in test for
1208 FileID functions, and the default enable/disable mode for QA test is now
1215 *(Released Fri, 11 May 2012)*
1217 A small bugfix release.
1219 The main issues solved are on the topic of compatibility with newer LVM
1222 - fixed parsing of ``lv_attr`` field
1223 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1224 the ``--force`` flag is needed
1226 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1227 to accept kernel 3.x too (was hardcoded to 2.6.*).
1229 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1230 of hook results and that also made display of other errors suboptimal
1231 was fixed; the code behaves now like 2.4 and earlier.
1233 Another change in 2.5, the cleanup of the OS scripts environment, is too
1234 aggressive: it removed even the ``PATH`` variable, which requires the OS
1235 scripts to *always* need to export it. Since this is a bit too strict,
1236 we now export a minimal PATH, the same that we export for hooks.
1238 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1239 integrated into this release.
1241 Finally, a few other miscellaneous changes were done (no new features,
1242 just small improvements):
1244 - Fix ``gnt-group --help`` display
1245 - Fix hardcoded Xen kernel path
1246 - Fix grow-disk handling of invalid units
1247 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1248 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1254 *(Released Thu, 12 Apr 2012)*
1256 Incompatible/important changes and bugfixes
1257 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1259 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1260 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1261 to match the underlying LUXI interface.
1262 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1263 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1264 not maintain backwards-compatibility as the underlying operation
1265 changed in an incompatible way. The RAPI client can talk to old
1266 servers, but it needs to be told so as the return value changed.
1267 - When creating file-based instances via RAPI, the ``file_driver``
1268 parameter no longer defaults to ``loop`` and must be specified.
1269 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1271 - Support for the undocumented and deprecated RAPI instance creation
1272 request format version 0 has been dropped. Use version 1, supported
1273 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1274 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1275 documentation <install>`.
1276 - The "cluster-verify" hooks are now executed per group by the
1277 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1278 if you just run ``gnt-cluster verify``, which generates one opcode per
1280 - The environment as passed to the OS scripts is cleared, and thus no
1281 environment variables defined in the node daemon's environment will be
1282 inherited by the scripts.
1283 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1285 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1286 support operations involving multiple node groups.
1287 - Offline nodes are ignored when failing over an instance.
1288 - Support for KVM version 1.0, which changed the version reporting format
1290 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1292 - ``Makefile`` is now compatible with Automake 1.11.2
1293 - Includes all bugfixes made in the 2.4 series
1298 - The ganeti-htools project has been merged into the ganeti-core source
1299 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1300 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1301 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1302 GPT-style partition tables (requires `parted
1303 <http://www.gnu.org/s/parted/>`_).
1304 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1305 - Allowed adding tags on instance creation.
1306 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1307 - Allow instances to be started in a paused state, enabling the user to
1308 see the complete console output on boot using the console.
1309 - Added new hypervisor flag to control default reboot behaviour
1310 (``reboot_behavior``).
1311 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1312 - Improved out-of-band management support:
1314 - Added ``gnt-node health`` command reporting the health status of
1316 - Added ``gnt-node power`` command to manage power status of nodes.
1317 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1319 - Instance migration can fall back to failover if instance is not
1321 - Filters can be used when listing nodes, instances, groups and locks;
1322 see :manpage:`ganeti(7)` manpage.
1323 - Added post-execution status as variables to :doc:`hooks <hooks>`
1325 - Instance tags are exported/imported together with the instance.
1326 - When given an explicit job ID, ``gnt-job info`` will work for archived
1328 - Jobs can define dependencies on other jobs (not yet supported via
1329 RAPI or command line, but used by internal commands and usable via
1332 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1335 - Instance failover is now available as a RAPI resource
1336 (``/2/instances/[instance_name]/failover``).
1337 - ``gnt-instance info`` defaults to static information if primary node
1339 - Opcodes have a new ``comment`` attribute.
1340 - Added basic SPICE support to KVM hypervisor.
1341 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1343 Node group improvements
1344 ~~~~~~~~~~~~~~~~~~~~~~~
1346 - ``gnt-cluster verify`` has been modified to check groups separately,
1347 thereby improving performance.
1348 - Node group support has been added to ``gnt-cluster verify-disks``,
1349 which now operates per node group.
1350 - Watcher has been changed to work better with node groups.
1352 - One process and state file per node group.
1353 - Slow watcher in one group doesn't block other group's watcher.
1355 - Added new command, ``gnt-group evacuate``, to move all instances in a
1356 node group to other groups.
1357 - Added ``gnt-instance change-group`` to move an instance to another
1359 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1360 per-group operations.
1361 - Node groups can be tagged.
1362 - Some operations switch from an exclusive to a shared lock as soon as
1364 - Instance's primary and secondary nodes' groups are now available as
1365 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1366 and ``snodes.group.uuid``).
1371 - Numerous updates to documentation and manpages.
1373 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1375 - Some opcode/job results are now also documented, see :doc:`RAPI
1378 - A lockset's internal lock is now also visible in lock monitor.
1379 - Log messages from job queue workers now contain information about the
1380 opcode they're processing.
1381 - ``gnt-instance console`` no longer requires the instance lock.
1382 - A short delay when waiting for job changes reduces the number of LUXI
1383 requests significantly.
1384 - DRBD metadata volumes are overwritten with zeros during disk creation.
1385 - Out-of-band commands no longer acquire the cluster lock in exclusive
1387 - ``devel/upload`` now uses correct permissions for directories.
1393 *(Released Fri, 23 Mar 2012)*
1395 This was the sixth release candidate of the 2.5 series.
1401 *(Released Mon, 9 Jan 2012)*
1403 This was the fifth release candidate of the 2.5 series.
1409 *(Released Thu, 27 Oct 2011)*
1411 This was the fourth release candidate of the 2.5 series.
1417 *(Released Wed, 26 Oct 2011)*
1419 This was the third release candidate of the 2.5 series.
1425 *(Released Tue, 18 Oct 2011)*
1427 This was the second release candidate of the 2.5 series.
1433 *(Released Tue, 4 Oct 2011)*
1435 This was the first release candidate of the 2.5 series.
1441 *(Released Wed, 31 Aug 2011)*
1443 This was the third beta release of the 2.5 series.
1449 *(Released Mon, 22 Aug 2011)*
1451 This was the second beta release of the 2.5 series.
1457 *(Released Fri, 12 Aug 2011)*
1459 This was the first beta release of the 2.5 series.
1465 *(Released Thu, 27 Oct 2011)*
1467 - Fixed bug when parsing command line parameter values ending in
1469 - Fixed assertion error after unclean master shutdown
1470 - Disable HTTP client pool for RPC, significantly reducing memory usage
1472 - Fixed queue archive creation with wrong permissions
1478 *(Released Tue, 23 Aug 2011)*
1482 - Fixed documentation for importing with ``--src-dir`` option
1483 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1484 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1490 *(Released Fri, 5 Aug 2011)*
1492 Many bug-fixes and a few small features:
1494 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1495 issues when you tried to add an instance with two MAC addresses in one
1497 - KVM: fixed per-instance stored UID value
1498 - KVM: configure bridged NICs at migration start
1499 - KVM: Fix a bug where instance will not start with never KVM versions
1501 - Added OS search path to ``gnt-cluster info``
1502 - Fixed an issue with ``file_storage_dir`` where you were forced to
1503 provide an absolute path, but the documentation states it is a
1504 relative path, the documentation was right
1505 - Added a new parameter to instance stop/start called ``--no-remember``
1506 that will make the state change to not be remembered
1507 - Implemented ``no_remember`` at RAPI level
1508 - Improved the documentation
1509 - Node evacuation: don't call IAllocator if node is already empty
1510 - Fixed bug in DRBD8 replace disks on current nodes
1511 - Fixed bug in recreate-disks for DRBD instances
1512 - Moved assertion checking locks in ``gnt-instance replace-disks``
1513 causing it to abort with not owning the right locks for some situation
1514 - Job queue: Fixed potential race condition when cancelling queued jobs
1515 - Fixed off-by-one bug in job serial generation
1516 - ``gnt-node volumes``: Fix instance names
1517 - Fixed aliases in bash completion
1518 - Fixed a bug in reopening log files after being sent a SIGHUP
1519 - Added a flag to burnin to allow specifying VCPU count
1520 - Bugfixes to non-root Ganeti configuration
1526 *(Released Thu, 12 May 2011)*
1528 Many bug-fixes and a few new small features:
1530 - Fixed a bug related to log opening failures
1531 - Fixed a bug in instance listing with orphan instances
1532 - Fixed a bug which prevented resetting the cluster-level node parameter
1533 ``oob_program`` to the default
1534 - Many fixes related to the ``cluster-merge`` tool
1535 - Fixed a race condition in the lock monitor, which caused failures
1536 during (at least) creation of many instances in parallel
1537 - Improved output for gnt-job info
1538 - Removed the quiet flag on some ssh calls which prevented debugging
1540 - Improved the N+1 failure messages in cluster verify by actually
1541 showing the memory values (needed and available)
1542 - Increased lock attempt timeouts so that when executing long operations
1543 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1544 too early and thus prevent the use of the 'fair' mechanism
1545 - Changed instance query data (``gnt-instance info``) to not acquire
1546 locks unless needed, thus allowing its use on locked instance if only
1547 static information is asked for
1548 - Improved behaviour with filesystems that do not support rename on an
1550 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1551 kept locks on all nodes during the wipe, which is unneeded
1552 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1553 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1554 10GiB) which caused the wipe to fail right at the end in some cases
1555 - Fixed master IP activation when doing master failover with no-voting
1556 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1557 the master node itself
1558 - Fixed potential data-loss in under disk full conditions, where Ganeti
1559 wouldn't check correctly the return code and would consider
1560 partially-written files 'correct'
1561 - Fixed bug related to multiple VGs and DRBD disk replacing
1562 - Added new disk parameter ``metavg`` that allows placement of the meta
1563 device for DRBD in a different volume group
1564 - Fixed error handling in the node daemon when the system libc doesn't
1565 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1566 - Fixed lock release during replace-disks, which kept cluster-wide locks
1567 when doing disk replaces with an iallocator script
1568 - Added check for missing bridges in cluster verify
1569 - Handle EPIPE errors while writing to the terminal better, so that
1570 piping the output to e.g. ``less`` doesn't cause a backtrace
1571 - Fixed rare case where a ^C during Luxi calls could have been
1572 interpreted as server errors, instead of simply terminating
1573 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1575 - Added a few more parameters to the KVM hypervisor, allowing a second
1576 CDROM, custom disk type for CDROMs and a floppy image
1577 - Removed redundant message in instance rename when the name is given
1579 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1580 disks on new nodes, allowing recreation when the original instance
1581 nodes are completely gone
1582 - Added option when converting disk templates to DRBD to skip waiting
1583 for the resync, in order to make the instance available sooner
1584 - Added two new variables to the OS scripts environment (containing the
1586 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1587 to allow use of ``pvgrub`` as bootloader
1588 - Changed the instance memory modifications to only check out-of-memory
1589 conditions on memory increases, and turned the secondary node warnings
1590 into errors (they can still be overridden via ``--force``)
1591 - Fixed the handling of a corner case when the Python installation gets
1592 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1593 to execute a command that doesn't exist
1594 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1595 node of the instance returned failures during instance shutdown; this
1596 adds the option ``--ignore-consistency`` to gnt-instance move
1598 And as usual, various improvements to the error messages, documentation
1605 *(Released Wed, 09 Mar 2011)*
1607 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1608 the RAPI users file if run twice (even with ``--dry-run``).
1610 The release fixes that bug (nothing else changed).
1616 *(Released Mon, 07 Mar 2011)*
1618 Final 2.4.0 release. Just a few small fixes:
1620 - Fixed RAPI node evacuate
1621 - Fixed the kvm-ifup script
1622 - Fixed internal error handling for special job cases
1623 - Updated man page to specify the escaping feature for options
1629 *(Released Mon, 28 Feb 2011)*
1631 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1632 that this feature wiped the disks of the wrong instance, leading to loss
1637 - Fixed title of query field containing instance name
1638 - Expanded the glossary in the documentation
1639 - Fixed one unittest (internal issue)
1645 *(Released Mon, 21 Feb 2011)*
1647 A number of bug fixes plus just a couple functionality changes.
1649 On the user-visible side, the ``gnt-* list`` command output has changed
1650 with respect to "special" field states. The current rc1 style of display
1651 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1652 the default output mode special fields states are displayed as follows:
1654 - Offline resource: ``*``
1655 - Unavailable/not applicable: ``-``
1656 - Data missing (RPC failure): ``?``
1657 - Unknown field: ``??``
1659 Another user-visible change is the addition of ``--force-join`` to
1664 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1665 - Fixed regression in RAPI/instance reinstall where all parameters were
1666 required (instead of optional)
1667 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1668 - Fixed iallocator usage (offline nodes were not considered offline)
1669 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1670 - Fixed hypervisor and OS parameter validation with respect to
1671 non-vm_capable nodes
1672 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1674 - Fixed ``tools/listrunner`` with respect to agent-based usage
1680 *(Released Fri, 4 Feb 2011)*
1682 Many changes and fixes since the beta1 release. While there were some
1683 internal changes, the code has been mostly stabilised for the RC
1686 Note: the dumb allocator was removed in this release, as it was not kept
1687 up-to-date with the IAllocator protocol changes. It is recommended to
1688 use the ``hail`` command from the ganeti-htools package.
1690 Note: the 2.4 and up versions of Ganeti are not compatible with the
1691 0.2.x branch of ganeti-htools. You need to upgrade to
1692 ganeti-htools-0.3.0 (or later).
1694 Regressions fixed from 2.3
1695 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1697 - Fixed the ``gnt-cluster verify-disks`` command
1698 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1700 - Fixed disk adoption breakage
1701 - Fixed wrong headers in instance listing for field aliases
1706 - Fixed corner case in KVM handling of NICs
1707 - Fixed many cases of wrong handling of non-vm_capable nodes
1708 - Fixed a bug where a missing instance symlink was not possible to
1709 recreate with any ``gnt-*`` command (now ``gnt-instance
1710 activate-disks`` does it)
1711 - Fixed the volume group name as reported by ``gnt-cluster
1713 - Increased timeouts for the import-export code, hopefully leading to
1714 fewer aborts due network or instance timeouts
1715 - Fixed bug in ``gnt-node list-storage``
1716 - Fixed bug where not all daemons were started on cluster
1717 initialisation, but only at the first watcher run
1718 - Fixed many bugs in the OOB implementation
1719 - Fixed watcher behaviour in presence of instances with offline
1721 - Fixed instance list output for instances running on the wrong node
1722 - a few fixes to the cluster-merge tool, but it still cannot merge
1723 multi-node groups (currently it is not recommended to use this tool)
1729 - Improved network configuration for the KVM hypervisor
1730 - Added e1000 as a supported NIC for Xen-HVM
1731 - Improved the lvmstrap tool to also be able to use partitions, as
1732 opposed to full disks
1733 - Improved speed of disk wiping (the cluster parameter
1734 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1735 of instance creations
1736 - Added documentation for the OS parameters
1737 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1738 hypervisor is not responding
1739 - Added display of blacklisted and hidden OS information in
1740 ``gnt-cluster info``
1741 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1742 NIC and node parameters, which might create problems with currently
1743 invalid (but undetected) configuration files, but prevents validation
1744 failures when unrelated parameters are modified
1745 - Changed cluster initialisation to wait for the master daemon to become
1747 - Expanded the RAPI interface:
1749 - Added config redistribution resource
1750 - Added activation/deactivation of instance disks
1751 - Added export of console information
1753 - Implemented log file reopening on SIGHUP, which allows using
1754 logrotate(8) for the Ganeti log files
1755 - Added a basic OOB helper script as an example
1761 *(Released Fri, 14 Jan 2011)*
1766 - Fixed timezone issues when formatting timestamps
1767 - Added support for node groups, available via ``gnt-group`` and other
1769 - Added out-of-band framework and management, see :doc:`design
1770 document <design-oob>`
1771 - Removed support for roman numbers from ``gnt-node list`` and
1772 ``gnt-instance list``.
1773 - Allowed modification of master network interface via ``gnt-cluster
1774 modify --master-netdev``
1775 - Accept offline secondaries while shutting down instance disks
1776 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1777 - Added support for multiple LVM volume groups
1778 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1780 - Added commands to list available fields:
1782 - ``gnt-node list-fields``
1783 - ``gnt-group list-fields``
1784 - ``gnt-instance list-fields``
1786 - Updated documentation and man pages
1791 - Moved ``rapi_users`` file into separate directory, now named
1792 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1794 - Added new tool for running commands on many machines,
1795 ``tools/ganeti-listrunner``
1796 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1797 improving the ``gnt-instance console`` output
1798 - Allowed customisation of disk index separator at ``configure`` time
1799 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1800 - Added support for non-partitioned md disks in ``lvmstrap``
1801 - Added script to gracefully power off KVM instances
1802 - Split ``utils`` module into smaller parts
1803 - Changed query operations to return more detailed information, e.g.
1804 whether an information is unavailable due to an offline node. To use
1805 this new functionality, the LUXI call ``Query`` must be used. Field
1806 information is now stored by the master daemon and can be retrieved
1807 using ``QueryFields``. Instances, nodes and groups can also be queried
1808 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1809 exposed via RAPI). The following commands make use of this
1810 infrastructure change:
1812 - ``gnt-group list``
1813 - ``gnt-group list-fields``
1815 - ``gnt-node list-fields``
1816 - ``gnt-instance list``
1817 - ``gnt-instance list-fields``
1818 - ``gnt-debug locks``
1823 - New RAPI resources (see :doc:`rapi`):
1827 - ``/2/groups/[group_name]``
1828 - ``/2/groups/[group_name]/assign-nodes``
1829 - ``/2/groups/[group_name]/modify``
1830 - ``/2/groups/[group_name]/rename``
1831 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1835 - Implemented ``no_install`` for instance creation
1836 - Implemented OS parameters for instance reinstallation, allowing
1837 use of special settings on reinstallation (e.g. for preserving data)
1842 - Added IPv6 support in import/export
1843 - Pause DRBD synchronization while wiping disks on instance creation
1844 - Updated unittests and QA scripts
1845 - Improved network parameters passed to KVM
1846 - Converted man pages from docbook to reStructuredText
1852 *(Released Mon, 20 Dec 2010)*
1854 Released version 2.3.1~rc1 without any changes.
1860 *(Released Wed, 1 Dec 2010)*
1862 - impexpd: Disable OpenSSL compression in socat if possible (backport
1863 from master, commit e90739d625b, see :doc:`installation guide
1864 <install-quick>` for details)
1865 - Changed unittest coverage report to exclude test scripts
1866 - Added script to check version format
1872 *(Released Wed, 1 Dec 2010)*
1874 Released version 2.3.0~rc1 without any changes.
1880 *(Released Fri, 19 Nov 2010)*
1882 A number of bugfixes and documentation updates:
1884 - Update ganeti-os-interface documentation
1885 - Fixed a bug related to duplicate MACs or similar items which should be
1887 - Fix breakage in OS state modify
1888 - Reinstall instance: disallow offline secondaries (fixes bug related to
1889 OS changing but reinstall failing)
1890 - plus all the other fixes between 2.2.1 and 2.2.2
1896 *(Released Tue, 2 Nov 2010)*
1898 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1899 - Fixed master failover race with watcher
1900 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1902 - Accept previously stopped instance for export with instance removal
1903 - Simplify and extend the environment variables for instance OS scripts
1904 - Added new node flags, ``master_capable`` and ``vm_capable``
1905 - Added optional instance disk wiping prior during allocation. This is a
1906 cluster-wide option and can be set/modified using
1907 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1908 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1909 :doc:`install-quick`
1910 - Added a new watcher option (``--ignore-pause``)
1911 - Added option to ignore offline node on instance start/stop
1912 (``--ignore-offline``)
1913 - Allow overriding OS parameters with ``gnt-instance reinstall``
1914 - Added ability to change node's secondary IP address using ``gnt-node
1916 - Implemented privilege separation for all daemons except
1917 ``ganeti-noded``, see ``configure`` options
1918 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1920 - Implemented job priorities (see ``ganeti(7)`` manpage)
1921 - Ignore failures while shutting down instances during failover from
1923 - Exit daemon's bootstrap process only once daemon is ready
1924 - Export more information via ``LUInstanceQuery``/remote API
1925 - Improved documentation, QA and unittests
1926 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1927 restart if the file was created or changed
1928 - Added LUXI protocol version sent with each request and response,
1929 allowing detection of server/client mismatches
1930 - Moved the Python scripts among gnt-* and ganeti-* into modules
1931 - Moved all code related to setting up SSH to an external script,
1933 - Infrastructure changes for node group support in future versions
1939 *(Released Fri, 19 Nov 2010)*
1941 A few small bugs fixed, and some improvements to the build system:
1943 - Fix documentation regarding conversion to drbd
1944 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1946 - Fix error handling in node modify with multiple changes
1947 - Allow remote imports without checked names
1953 *(Released Tue, 19 Oct 2010)*
1955 - Disable SSL session ID cache in RPC client
1961 *(Released Thu, 14 Oct 2010)*
1963 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1964 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1966 - Fix problems with interaction between Curl and Python's HTTP server,
1967 resulting in increased speed in many RPC calls
1968 - Improve our release script to prevent breakage with older aclocal and
1975 *(Released Thu, 7 Oct 2010)*
1977 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1978 value retrieved from master
1979 - Added support for blacklisted or hidden OS definitions
1980 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1981 - Added support for -mem-path in KVM hypervisor abstraction layer
1982 - Allow overriding instance parameters in tool for inter-cluster
1983 instance moves (``tools/move-instance``)
1984 - Improved opcode summaries (e.g. in ``gnt-job list``)
1985 - Improve consistency of OS listing by sorting it
1986 - Documentation updates
1992 *(Released Fri, 8 Oct 2010)*
1994 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
2000 *(Released Mon, 4 Oct 2010)*
2002 - Fixed regression in ``gnt-instance rename``
2008 *(Released Wed, 22 Sep 2010)*
2010 - Fixed OS_VARIANT variable for OS scripts
2011 - Fixed cluster tag operations via RAPI
2012 - Made ``setup-ssh`` exit with non-zero code if an error occurred
2013 - Disabled RAPI CA checks in watcher
2019 *(Released Mon, 23 Aug 2010)*
2021 - Support DRBD versions of the format "a.b.c.d"
2023 - Re-introduce support for usage from multiple threads in RAPI client
2024 - Instance renames and modify via RAPI
2025 - Work around race condition between processing and archival in job
2027 - Mark opcodes following failed one as failed, too
2028 - Job field ``lock_status`` was removed due to difficulties making it
2029 work with the changed job queue in Ganeti 2.2; a better way to monitor
2030 locks is expected for a later 2.2.x release
2031 - Fixed dry-run behaviour with many commands
2032 - Support ``ssh-agent`` again when adding nodes
2033 - Many additional bugfixes
2039 *(Released Fri, 30 Jul 2010)*
2041 Important change: the internal RPC mechanism between Ganeti nodes has
2042 changed from using a home-grown http library (based on the Python base
2043 libraries) to use the PycURL library. This requires that PycURL is
2044 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2045 against GnuTLS by default. cURL's support for GnuTLS had known issues
2046 before cURL 7.21.0 and we recommend using the latest cURL release or
2047 linking against OpenSSL. Most other distributions already link PycURL
2048 and cURL against OpenSSL. The command::
2050 python -c 'import pycurl; print pycurl.version'
2052 can be used to determine the libraries PycURL and cURL are linked
2055 Other significant changes:
2057 - Rewrote much of the internals of the job queue, in order to achieve
2058 better parallelism; this decouples job query operations from the job
2059 processing, and it should allow much nicer behaviour of the master
2060 daemon under load, and it also has uncovered some long-standing bugs
2061 related to the job serialisation (now fixed)
2062 - Added a default iallocator setting to the cluster parameters,
2063 eliminating the need to always pass nodes or an iallocator for
2064 operations that require selection of new node(s)
2065 - Added experimental support for the LXC virtualization method
2066 - Added support for OS parameters, which allows the installation of
2067 instances to pass parameter to OS scripts in order to customise the
2069 - Added a hypervisor parameter controlling the migration type (live or
2070 non-live), since hypervisors have various levels of reliability; this
2071 has renamed the 'live' parameter to 'mode'
2072 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
2073 logical volumes, meaning that cluster verify will ignore them and not
2074 flag their presence as errors
2075 - The watcher will now reset the error count for failed instances after
2076 8 hours, thus allowing self-healing if the problem that caused the
2077 instances to be down/fail to start has cleared in the meantime
2078 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2079 check for, and warn, if the drbd module parameter ``usermode_helper``
2080 is not consistent with the cluster-wide setting; this is needed to
2081 make diagnose easier of failed drbd creations
2082 - Started adding base IPv6 support, but this is not yet
2083 enabled/available for use
2084 - Rename operations (cluster, instance) will now return the new name,
2085 which is especially useful if a short name was passed in
2086 - Added support for instance migration in RAPI
2087 - Added a tool to pre-configure nodes for the SSH setup, before joining
2088 them to the cluster; this will allow in the future a simplified model
2089 for node joining (but not yet fully enabled in 2.2); this needs the
2090 paramiko python library
2091 - Fixed handling of name-resolving errors
2092 - Fixed consistency of job results on the error path
2093 - Fixed master-failover race condition when executed multiple times in
2095 - Fixed many bugs related to the job queue (mostly introduced during the
2096 2.2 development cycle, so not all are impacting 2.1)
2097 - Fixed instance migration with missing disk symlinks
2098 - Fixed handling of unknown jobs in ``gnt-job archive``
2099 - And many other small fixes/improvements
2103 - Enhanced both the unittest and the QA coverage
2104 - Switched the opcode validation to a generic model, and extended the
2105 validation to all opcode parameters
2106 - Changed more parts of the code that write shell scripts to use the
2108 - Switched the master daemon to use the asyncore library for the Luxi
2115 *(Released Thu, 17 Jun 2010)*
2117 - Added tool (``move-instance``) and infrastructure to move instances
2118 between separate clusters (see :doc:`separate documentation
2119 <move-instance>` and :doc:`design document <design-2.2>`)
2120 - Added per-request RPC timeout
2121 - RAPI now requires a Content-Type header for requests with a body (e.g.
2122 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2123 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2124 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2126 - Implemented initial support for running Ganeti daemons as separate
2127 users, see configure-time flags ``--with-user-prefix`` and
2128 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2130 - Instances can be removed after export (``gnt-backup export
2131 --remove-instance``)
2132 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2133 key (instead of 1024 bit)
2134 - Added new cluster configuration file for cluster domain secret
2135 - Import/export now use SSL instead of SSH
2136 - Added support for showing estimated time when exporting an instance,
2137 see the ``ganeti-os-interface(7)`` manpage and look for
2144 *(Released Tue, 16 Nov 2010)*
2146 Some more bugfixes. Unless critical bugs occur, this will be the last
2149 - Fix case of MAC special-values
2150 - Fix mac checker regex
2151 - backend: Fix typo causing "out of range" error
2152 - Add missing --units in gnt-instance list man page
2158 *(Released Tue, 24 Aug 2010)*
2161 - Don't ignore secondary node silently on non-mirrored disk templates
2163 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2164 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2165 - Properly document the usb_mouse parameter
2166 - Fix path in ganeti-rapi(8) (issue 116)
2167 - Adjust error message when the ganeti user's .ssh directory is
2169 - Add same-node-check when changing the disk template to drbd
2175 *(Released Fri, 16 Jul 2010)*
2178 - Add an option to only select some reboot types during qa/burnin.
2179 (on some hypervisors consequent reboots are not supported)
2180 - Fix infrequent race condition in master failover. Sometimes the old
2181 master ip address would be still detected as up for a short time
2182 after it was removed, causing failover to fail.
2183 - Decrease mlockall warnings when the ctypes module is missing. On
2184 Python 2.4 we support running even if no ctypes module is installed,
2185 but we were too verbose about this issue.
2186 - Fix building on old distributions, on which man doesn't have a
2188 - Fix RAPI not to ignore the MAC address on instance creation
2189 - Implement the old instance creation format in the RAPI client.
2195 *(Released Thu, 01 Jul 2010)*
2197 A small bugfix release:
2198 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2199 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2200 a non-existing option
2201 - Fix instance create: the --force-variant option was ignored
2202 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2203 - Fix modify node storage with non-FQDN arguments
2204 - Fix RAPI client to authenticate under Python 2.6 when used
2205 for more than 5 requests needing authentication
2206 - Fix gnt-instance modify -t (storage) giving a wrong error message
2207 when converting a non-shutdown drbd instance to plain
2213 *(Released Fri, 18 Jun 2010)*
2215 A small bugfix release:
2217 - Fix live migration of KVM instances started with older Ganeti
2218 versions which had fewer hypervisor parameters
2219 - Fix gnt-instance grow-disk on down instances
2220 - Fix an error-reporting bug during instance migration
2221 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2222 silently ignoring broken ones
2223 - Fix an RPC error reporting bug affecting, for example, RAPI client
2225 - Fix bug triggered by different API version os-es on different nodes
2226 - Fix a bug in instance startup with custom hvparams: OS level
2227 parameters would fail to be applied.
2228 - Fix the RAPI client under Python 2.6 (but more work is needed to
2229 make it work completely well with OpenSSL)
2230 - Fix handling of errors when resolving names from DNS
2236 *(Released Thu, 3 Jun 2010)*
2238 A medium sized development cycle. Some new features, and some
2239 fixes/small improvements/cleanups.
2241 Significant features
2242 ~~~~~~~~~~~~~~~~~~~~
2244 The node deamon now tries to mlock itself into memory, unless the
2245 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2246 its logs, and falls back to console logging. This allows emergency
2247 features such as ``gnt-node powercycle`` to work even in the event of a
2248 broken node disk (tested offlining the disk hosting the node's
2249 filesystem and dropping its memory caches; don't try this at home)
2251 KVM: add vhost-net acceleration support. It can be tested with a new
2252 enough version of the kernel and of qemu-kvm.
2254 KVM: Add instance chrooting feature. If you use privilege dropping for
2255 your VMs you can also now force them to chroot to an empty directory,
2256 before starting the emulated guest.
2258 KVM: Add maximum migration bandwith and maximum downtime tweaking
2259 support (requires a new-enough version of qemu-kvm).
2261 Cluster verify will now warn if the master node doesn't have the master
2262 ip configured on it.
2264 Add a new (incompatible) instance creation request format to RAPI which
2265 supports all parameters (previously only a subset was supported, and it
2266 wasn't possible to extend the old format to accomodate all the new
2267 features. The old format is still supported, and a client can check for
2268 this feature, before using it, by checking for its presence in the
2269 ``features`` RAPI resource.
2271 Now with ancient latin support. Try it passing the ``--roman`` option to
2272 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2273 (requires the python-roman module to be installed, in order to work).
2278 As usual many internal code refactorings, documentation updates, and
2281 - Lots of improvements and cleanups to the experimental Remote API
2282 (RAPI) client library.
2283 - A new unit test suite for the core daemon libraries.
2284 - A fix to creating missing directories makes sure the umask is not
2285 applied anymore. This enforces the same directory permissions
2287 - Better handling terminating daemons with ctrl+c (used when running
2288 them in debugging mode).
2289 - Fix a race condition in live migrating a KVM instance, when stat()
2290 on the old proc status file returned EINVAL, which is an unexpected
2292 - Fixed manpage checking with newer man and utf-8 charachters. But now
2293 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2299 *(Released Fri, 7 May 2010)*
2301 Fix a bug which prevented untagged KVM instances from starting.
2307 *(Released Fri, 7 May 2010)*
2309 Another release with a long development cycle, during which many
2310 different features were added.
2312 Significant features
2313 ~~~~~~~~~~~~~~~~~~~~
2315 The KVM hypervisor now can run the individual instances as non-root, to
2316 reduce the impact of a VM being hijacked due to bugs in the
2317 hypervisor. It is possible to run all instances as a single (non-root)
2318 user, to manually specify a user for each instance, or to dynamically
2319 allocate a user out of a cluster-wide pool to each instance, with the
2320 guarantee that no two instances will run under the same user ID on any
2323 An experimental RAPI client library, that can be used standalone
2324 (without the other Ganeti libraries), is provided in the source tree as
2325 ``lib/rapi/client.py``. Note this client might change its interface in
2326 the future, as we iterate on its capabilities.
2328 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2329 replace the cluster's certificates and crypto keys. This might help in
2330 case they have been compromised, or have simply expired.
2332 A new disk option for instance creation has been added that allows one
2333 to "adopt" currently existing logical volumes, with data
2334 preservation. This should allow easier migration to Ganeti from
2335 unmanaged (or managed via other software) instances.
2337 Another disk improvement is the possibility to convert between redundant
2338 (DRBD) and plain (LVM) disk configuration for an instance. This should
2339 allow better scalability (starting with one node and growing the
2340 cluster, or shrinking a two-node cluster to one node).
2342 A new feature that could help with automated node failovers has been
2343 implemented: if a node sees itself as offline (by querying the master
2344 candidates), it will try to shutdown (hard) all instances and any active
2345 DRBD devices. This reduces the risk of duplicate instances if an
2346 external script automatically failovers the instances on such nodes. To
2347 enable this, the cluster parameter ``maintain_node_health`` should be
2348 enabled; in the future this option (per the name) will enable other
2349 automatic maintenance features.
2351 Instance export/import now will reuse the original instance
2352 specifications for all parameters; that means exporting an instance,
2353 deleting it and the importing it back should give an almost identical
2354 instance. Note that the default import behaviour has changed from
2355 before, where it created only one NIC; now it recreates the original
2358 Cluster verify has added a few new checks: SSL certificates validity,
2359 /etc/hosts consistency across the cluster, etc.
2364 As usual, many internal changes were done, documentation fixes,
2367 - Fixed cluster initialization with disabled cluster storage (regression
2368 introduced in 2.1.1)
2369 - File-based storage supports growing the disks
2370 - Fixed behaviour of node role changes
2371 - Fixed cluster verify for some corner cases, plus a general rewrite of
2372 cluster verify to allow future extension with more checks
2373 - Fixed log spamming by watcher and node daemon (regression introduced
2375 - Fixed possible validation issues when changing the list of enabled
2377 - Fixed cleanup of /etc/hosts during node removal
2378 - Fixed RAPI response for invalid methods
2379 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2380 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2381 from ide disks, etc.)
2382 - Allow OS changes without re-installation (to record a changed OS
2383 outside of Ganeti, or to allow OS renames)
2384 - Allow instance creation without OS installation (useful for example if
2385 the OS will be installed manually, or restored from a backup not in
2387 - Implemented option to make cluster ``copyfile`` use the replication
2389 - Added list of enabled hypervisors to ssconf (possibly useful for
2391 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2393 - A partial form of node re-IP is possible via node readd, which now
2394 allows changed node primary IP
2395 - Command line utilities now show an informational message if the job is
2397 - The logs of the master daemon now show the PID/UID/GID of the
2404 *(Released Fri, 12 Mar 2010)*
2406 During the 2.1.0 long release candidate cycle, a lot of improvements and
2407 changes have accumulated with were released later as 2.1.1.
2412 The node evacuate command (``gnt-node evacuate``) was significantly
2413 rewritten, and as such the IAllocator protocol was changed - a new
2414 request type has been added. This unfortunate change during a stable
2415 series is designed to improve performance of node evacuations; on
2416 clusters with more than about five nodes and which are well-balanced,
2417 evacuation should proceed in parallel for all instances of the node
2418 being evacuated. As such, any existing IAllocator scripts need to be
2419 updated, otherwise the above command will fail due to the unknown
2420 request. The provided "dumb" allocator has not been updated; but the
2421 ganeti-htools package supports the new protocol since version 0.2.4.
2423 Another important change is increased validation of node and instance
2424 names. This might create problems in special cases, if invalid host
2425 names are being used.
2427 Also, a new layer of hypervisor parameters has been added, that sits at
2428 OS level between the cluster defaults and the instance ones. This allows
2429 customisation of virtualization parameters depending on the installed
2430 OS. For example instances with OS 'X' may have a different KVM kernel
2431 (or any other parameter) than the cluster defaults. This is intended to
2432 help managing a multiple OSes on the same cluster, without manual
2433 modification of each instance's parameters.
2435 A tool for merging clusters, ``cluster-merge``, has been added in the
2436 tools sub-directory.
2441 - Improved the int/float conversions that should make the code more
2442 robust in face of errors from the node daemons
2443 - Fixed the remove node code in case of internal configuration errors
2444 - Fixed the node daemon behaviour in face of inconsistent queue
2445 directory (e.g. read-only file-system where we can't open the files
2447 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2448 now it either aborts cleanly or, if given the new "auto_promote"
2449 parameter, will automatically promote other nodes as needed
2450 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2451 completely prevent Ganeti from working
2452 - Fixed bug for instance export when not all disks were successfully
2454 - Fixed behaviour of node add when the new node is slow in starting up
2456 - Fixed handling of signals in the LUXI client, which should improve
2457 behaviour of command-line scripts
2458 - Added checks for invalid node/instance names in the configuration (now
2459 flagged during cluster verify)
2460 - Fixed watcher behaviour for disk activation errors
2461 - Fixed two potentially endless loops in http library, which led to the
2462 RAPI daemon hanging and consuming 100% CPU in some cases
2463 - Fixed bug in RAPI daemon related to hashed passwords
2464 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2466 - Enhanced compatibility with non-Debian OSes, but not using absolute
2467 path in some commands and allowing customisation of the ssh
2468 configuration directory
2469 - Fixed possible future issue with new Python versions by abiding to the
2470 proper use of ``__slots__`` attribute on classes
2471 - Added checks that should prevent directory traversal attacks
2472 - Many documentation fixes based on feedback from users
2477 - Added an "early_release" more for instance replace disks and node
2478 evacuate, where we release locks earlier and thus allow higher
2479 parallelism within the cluster
2480 - Added watcher hooks, intended to allow the watcher to restart other
2481 daemons (e.g. from the ganeti-nbma project), but they can be used of
2482 course for any other purpose
2483 - Added a compile-time disable for DRBD barriers, to increase
2484 performance if the administrator trusts the power supply or the
2485 storage system to not lose writes
2486 - Added the option of using syslog for logging instead of, or in
2487 addition to, Ganeti's own log files
2488 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2489 can indeed boot from a paravirtual NIC
2490 - Added a generic debug level for many operations; while this is not
2491 used widely yet, it allows one to pass the debug value all the way to
2493 - Enhanced the hooks environment for instance moves (failovers,
2494 migrations) where the primary/secondary nodes changed during the
2495 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2496 - Enhanced data validations for many user-supplied values; one important
2497 item is the restrictions imposed on instance and node names, which
2498 might reject some (invalid) host names
2499 - Add a configure-time option to disable file-based storage, if it's not
2500 needed; this allows greater security separation between the master
2501 node and the other nodes from the point of view of the inter-node RPC
2503 - Added user notification in interactive tools if job is waiting in the
2504 job queue or trying to acquire locks
2505 - Added log messages when a job is waiting for locks
2506 - Added filtering by node tags in instance operations which admit
2507 multiple instances (start, stop, reboot, reinstall)
2508 - Added a new tool for cluster mergers, ``cluster-merge``
2509 - Parameters from command line which are of the form ``a=b,c=d`` can now
2510 use backslash escapes to pass in values which contain commas,
2511 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2513 - For KVM, the instance name is the first parameter passed to KVM, so
2514 that it's more visible in the process list
2520 *(Released Tue, 2 Mar 2010)*
2522 Ganeti 2.1 brings many improvements with it. Major changes:
2524 - Added infrastructure to ease automated disk repairs
2525 - Added new daemon to export configuration data in a cheaper way than
2526 using the remote API
2527 - Instance NICs can now be routed instead of being associated with a
2529 - Improved job locking logic to reduce impact of jobs acquiring multiple
2530 locks waiting for other long-running jobs
2532 In-depth implementation details can be found in the Ganeti 2.1 design
2538 - Added chroot hypervisor
2539 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2541 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2542 ``bootloader_path`` and ``bootloader_args``)
2543 - Added the ``use_localtime`` option for the xen-hvm and kvm
2544 hypervisors, and the default value for this has changed to false (in
2545 2.0 xen-hvm always enabled it)
2546 - Added luxi call to submit multiple jobs in one go
2547 - Added cluster initialization option to not modify ``/etc/hosts``
2549 - Added network interface parameters
2550 - Added dry run mode to some LUs
2551 - Added RAPI resources:
2553 - ``/2/instances/[instance_name]/info``
2554 - ``/2/instances/[instance_name]/replace-disks``
2555 - ``/2/nodes/[node_name]/evacuate``
2556 - ``/2/nodes/[node_name]/migrate``
2557 - ``/2/nodes/[node_name]/role``
2558 - ``/2/nodes/[node_name]/storage``
2559 - ``/2/nodes/[node_name]/storage/modify``
2560 - ``/2/nodes/[node_name]/storage/repair``
2562 - Added OpCodes to evacuate or migrate all instances on a node
2563 - Added new command to list storage elements on nodes (``gnt-node
2564 list-storage``) and modify them (``gnt-node modify-storage``)
2565 - Added new ssconf files with master candidate IP address
2566 (``ssconf_master_candidates_ips``), node primary IP address
2567 (``ssconf_node_primary_ips``) and node secondary IP address
2568 (``ssconf_node_secondary_ips``)
2569 - Added ``ganeti-confd`` and a client library to query the Ganeti
2570 configuration via UDP
2571 - Added ability to run hooks after cluster initialization and before
2573 - Added automatic mode for disk replace (``gnt-instance replace-disks
2575 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2576 after catastrophic data-loss
2577 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2579 - Added ``gnt-instance move`` command to move instances
2580 - Added ``gnt-cluster watcher`` command to control watcher
2581 - Added ``gnt-node powercycle`` command to powercycle nodes
2582 - Added new job status field ``lock_status``
2583 - Added parseable error codes to cluster verification (``gnt-cluster
2584 verify --error-codes``) and made output less verbose (use
2585 ``--verbose`` to restore previous behaviour)
2586 - Added UUIDs to the main config entities (cluster, nodes, instances)
2587 - Added support for OS variants
2588 - Added support for hashed passwords in the Ganeti remote API users file
2590 - Added option to specify maximum timeout on instance shutdown
2591 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2592 - Added new helper script to start and stop Ganeti daemons
2593 (``daemon-util``), with the intent to reduce the work necessary to
2594 adjust Ganeti for non-Debian distributions and to start/stop daemons
2596 - Added more unittests
2597 - Fixed critical bug in ganeti-masterd startup
2598 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2599 now customisable at the cluster level for both the KVM and Xen
2600 hypervisors using the new ``migration_port`` parameter
2601 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2602 reinstalling an instance
2603 - Allowed ``@`` in tag names
2604 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2605 - Many documentation updates
2606 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2607 - ``gnt-instance reinstall`` can now reinstall multiple instances
2608 - Updated many command line parameters
2609 - Introduced new OS API version 15
2610 - No longer support a default hypervisor
2611 - Treat virtual LVs as inexistent
2612 - Improved job locking logic to reduce lock contention
2613 - Match instance and node names case insensitively
2614 - Reimplemented bash completion script to be more complete
2621 *(Released Thu, 4 Feb 2010)*
2623 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2624 - Fix a string formatting bug
2625 - Improve safety of the code in some error paths
2626 - Improve data validation in the master of values returned from nodes
2632 *(Released Thu, 17 Dec 2009)*
2634 - Fix security issue due to missing validation of iallocator names; this
2635 allows local and remote execution of arbitrary executables
2636 - Fix failure of gnt-node list during instance removal
2637 - Ship the RAPI documentation in the archive
2643 *(Released Wed, 30 Sep 2009)*
2645 - Fixed many wrong messages
2646 - Fixed a few bugs related to the locking library
2647 - Fixed MAC checking at instance creation time
2648 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2649 - Fixed a few issues related to signal handling in both daemons and
2651 - Fixed the example startup script provided
2652 - Fixed insserv dependencies in the example startup script (patch from
2654 - Fixed handling of drained nodes in the iallocator framework
2655 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2657 - Fixed error related to invalid job IDs in job polling
2658 - Fixed job/opcode persistence on unclean master shutdown
2659 - Fixed handling of partial job processing after unclean master
2661 - Fixed error reporting from LUs, previously all errors were converted
2662 into execution errors
2663 - Fixed error reporting from burnin
2664 - Decreased significantly the memory usage of the job queue
2665 - Optimised slightly multi-job submission
2666 - Optimised slightly opcode loading
2667 - Backported the multi-job submit framework from the development
2668 branch; multi-instance start and stop should be faster
2669 - Added script to clean archived jobs after 21 days; this will reduce
2670 the size of the queue directory
2671 - Added some extra checks in disk size tracking
2672 - Added an example ethers hook script
2673 - Added a cluster parameter that prevents Ganeti from modifying of
2675 - Added more node information to RAPI responses
2676 - Added a ``gnt-job watch`` command that allows following the ouput of a
2678 - Added a bind-address option to ganeti-rapi
2679 - Added more checks to the configuration verify
2680 - Enhanced the burnin script such that some operations can be retried
2682 - Converted instance reinstall to multi-instance model
2688 *(Released Fri, 7 Aug 2009)*
2690 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2691 to allow using the pre-2.0.2 behaviour in activation, if any existing
2692 instances have mismatched disk sizes in the configuration
2693 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2694 any configuration mismatches for disk sizes
2695 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2696 failover to work on two-node clusters
2697 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2699 - Fixed detection of OS script errors in ``gnt-backup export``
2700 - Fixed exit code of ``gnt-backup export``
2706 *(Released Fri, 17 Jul 2009)*
2708 - Added experimental support for stripped logical volumes; this should
2709 enhance performance but comes with a higher complexity in the block
2710 device handling; stripping is only enabled when passing
2711 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2712 affected even in the non-stripped mode
2713 - Improved resiliency against transient failures at the end of DRBD
2714 resyncs, and in general of DRBD resync checks
2715 - Fixed a couple of issues with exports and snapshot errors
2716 - Fixed a couple of issues in instance listing
2717 - Added display of the disk size in ``gnt-instance info``
2718 - Fixed checking for valid OSes in instance creation
2719 - Fixed handling of the "vcpus" parameter in instance listing and in
2720 general of invalid parameters
2721 - Fixed http server library, and thus RAPI, to handle invalid
2722 username/password combinations correctly; this means that now they
2723 report unauthorized for queries too, not only for modifications,
2724 allowing earlier detect of configuration problems
2725 - Added a new "role" node list field, equivalent to the master/master
2726 candidate/drained/offline flags combinations
2727 - Fixed cluster modify and changes of candidate pool size
2728 - Fixed cluster verify error messages for wrong files on regular nodes
2729 - Fixed a couple of issues with node demotion from master candidate role
2730 - Fixed node readd issues
2731 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2732 - Added a new ``--no-voting`` option for masterfailover to fix failover
2733 on two-nodes clusters when the former master node is unreachable
2734 - Added instance reinstall over RAPI
2740 *(Released Tue, 16 Jun 2009)*
2742 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2743 allow re-adding the start in single-user option (regression from 1.2)
2744 - the watcher writes the instance status to a file, to allow monitoring
2745 to report the instance status (from the master) based on cached
2746 results of the watcher's queries; while this can get stale if the
2747 watcher is being locked due to other work on the cluster, this is
2748 still an improvement
2749 - the watcher now also restarts the node daemon and the rapi daemon if
2751 - fixed the watcher to handle full and drained queue cases
2752 - hooks export more instance data in the environment, which helps if
2753 hook scripts need to take action based on the instance's properties
2754 (no longer need to query back into ganeti)
2755 - instance failovers when the instance is stopped do not check for free
2756 RAM, so that failing over a stopped instance is possible in low memory
2758 - rapi uses queries for tags instead of jobs (for less job traffic), and
2759 for cluster tags it won't talk to masterd at all but read them from
2761 - a couple of error handling fixes in RAPI
2762 - drbd handling: improved the error handling of inconsistent disks after
2763 resync to reduce the frequency of "there are some degraded disks for
2764 this instance" messages
2765 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2766 error handling path called a wrong function name)
2772 *(Released Wed, 27 May 2009)*
2774 - no changes from rc5
2780 *(Released Wed, 20 May 2009)*
2782 - fix a couple of bugs (validation, argument checks)
2783 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2784 - some small improvements to RAPI and IAllocator
2785 - make watcher automatically start the master daemon if down
2791 *(Released Mon, 27 Apr 2009)*
2793 - change the OS list to not require locks; this helps with big clusters
2794 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2795 volume group is broken
2796 - ``gnt-instance info``, without any arguments, doesn't run for all
2797 instances anymore; either pass ``--all`` or pass the desired
2798 instances; this helps against mistakes on big clusters where listing
2799 the information for all instances takes a long time
2800 - miscellaneous doc and man pages fixes
2806 *(Released Wed, 8 Apr 2009)*
2808 - Change the internal locking model of some ``gnt-node`` commands, in
2809 order to reduce contention (and blocking of master daemon) when
2810 batching many creation/reinstall jobs
2811 - Fixes to Xen soft reboot
2812 - No longer build documentation at build time, instead distribute it in
2813 the archive, in order to reduce the need for the whole docbook/rst
2820 *(Released Fri, 27 Mar 2009)*
2822 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2823 - Fix watcher startup sequence, improves the behaviour of busy clusters
2824 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2825 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2826 bind address and other places
2827 - Some documentation fixes and updates
2833 *(Released Mon, 2 Mar 2009)*
2835 - More documentation updates, now all docs should be more-or-less
2837 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2839 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2841 - Fix an issue related to $libdir/run/ganeti and cluster creation
2847 *(Released Thu, 19 Feb 2009)*
2849 - Xen PVM and KVM have switched the default value for the instance root
2850 disk to the first partition on the first drive, instead of the whole
2851 drive; this means that the OS installation scripts must be changed
2853 - Man pages have been updated
2854 - RAPI has been switched by default to HTTPS, and the exported functions
2855 should all work correctly
2856 - RAPI v1 has been removed
2857 - Many improvements to the KVM hypervisor
2858 - Block device errors are now better reported
2859 - Many other bugfixes and small improvements
2865 *(Released Mon, 26 Jan 2009)*
2867 - Version 2 is a general rewrite of the code and therefore the
2868 differences are too many to list, see the design document for 2.0 in
2869 the ``doc/`` subdirectory for more details
2870 - In this beta version there is not yet a migration path from 1.2 (there
2871 will be one in the final 2.0 release)
2872 - A few significant changes are:
2874 - all commands are executed by a daemon (``ganeti-masterd``) and the
2875 various ``gnt-*`` commands are just front-ends to it
2876 - all the commands are entered into, and executed from a job queue,
2877 see the ``gnt-job(8)`` manpage
2878 - the RAPI daemon supports read-write operations, secured by basic
2879 HTTP authentication on top of HTTPS
2880 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2881 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2882 to migrate to DRBD 8 first while still running Ganeti 1.2)
2883 - DRBD devices are using statically allocated minor numbers, which
2884 will be assigned to existing instances during the migration process
2885 - there is support for both Xen PVM and Xen HVM instances running on
2887 - KVM virtualization is supported too
2888 - file-based storage has been implemented, which means that it is
2889 possible to run the cluster without LVM and DRBD storage, for
2890 example using a shared filesystem exported from shared storage (and
2891 still have live migration)
2897 *(Released Tue, 13 Jan 2009)*
2899 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2900 - Reuse the old instance mac address by default on instance import, if
2901 the instance name is the same.
2902 - Handle situations in which the node info rpc returns incomplete
2904 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2905 - Improved version of batcher:
2907 - state file support
2908 - instance mac address support
2909 - support for HVM clusters/instances
2911 - Add an option to show the number of cpu sockets and nodes in
2913 - Support OSes that handle more than one version of the OS api (but do
2914 not change the current API in any other way)
2915 - Fix ``gnt-node migrate``
2916 - ``gnt-debug`` man page
2917 - Fixes various more typos and small issues
2918 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2924 *(Released Wed, 24 Sep 2008)*
2926 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2927 type of disk exported to fully virtualized instances.
2928 - provide access to the serial console of HVM instances
2929 - instance auto_balance flag, set by default. If turned off it will
2930 avoid warnings on cluster verify if there is not enough memory to fail
2931 over an instance. in the future it will prevent automatically failing
2932 it over when we will support that.
2933 - batcher tool for instance creation, see ``tools/README.batcher``
2934 - ``gnt-instance reinstall --select-os`` to interactively select a new
2935 operating system when reinstalling an instance.
2936 - when changing the memory amount on instance modify a check has been
2937 added that the instance will be able to start. also warnings are
2938 emitted if the instance will not be able to fail over, if auto_balance
2940 - documentation fixes
2941 - sync fields between ``gnt-instance list/modify/add/import``
2942 - fix a race condition in drbd when the sync speed was set after giving
2943 the device a remote peer.
2949 *(Released Tue, 22 Jul 2008)*
2951 - note: the allowed size and number of tags per object were reduced
2952 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2953 - fixed twisted 8.x compatibility
2954 - fixed ``gnt-instance replace-disks`` with iallocator
2955 - add TCP keepalives on twisted connections to detect restarted nodes
2956 - disk increase support, see ``gnt-instance grow-disk``
2957 - implement bulk node/instance query for RAPI
2958 - add tags in node/instance listing (optional)
2959 - experimental migration (and live migration) support, read the man page
2960 for ``gnt-instance migrate``
2961 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2962 has some small improvements in handling its state file
2968 *(Released Fri, 13 Jun 2008)*
2970 - Experimental readonly, REST-based remote API implementation;
2971 automatically started on master node, TCP port 5080, if enabled by
2972 ``--enable-rapi`` parameter to configure script.
2973 - Instance allocator support. Add and import instance accept a
2974 ``--iallocator`` parameter, and call that instance allocator to decide
2975 which node to use for the instance. The iallocator document describes
2976 what's expected from an allocator script.
2977 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2978 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2979 node is lost there is still enough memory to fail over the instances
2981 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2982 ``gnt-cluster verify``, to check for site-specific compliance. All the
2983 hooks will run, and their output, if any, will be displayed. Any
2984 failing hook will make the verification return an error value.
2985 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2986 primary and secondary interfaces
2987 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2988 that is still declared as part of the cluster and has failed.
2989 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2990 specifying output fields, that just adds the chosen fields to the
2992 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2993 export from the filesystem.
2994 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2995 have been added. Using them you can enable/disable acpi and pae
2996 support, and specify a path for a cd image to be exported to the
2997 instance. These parameters as the name suggest only work on HVM
2999 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
3000 PAE support will be set to the previously hardcoded values, but the
3001 (previously hardcoded) path to the CDROM ISO image will be unset and
3002 if required, needs to be set manually with ``gnt-instance modify``
3004 - The address to which an instance's VNC console is bound is now
3005 selectable per-instance, rather than being cluster wide. Of course
3006 this only applies to instances controlled via VNC, so currently just
3007 applies to HVM clusters.
3013 *(Released Mon, 18 Feb 2008)*
3015 - more tweaks to the disk activation code (especially helpful for DRBD)
3016 - change the default ``gnt-instance list`` output format, now there is
3017 one combined status field (see the manpage for the exact values this
3019 - some more fixes for the mac export to hooks change
3020 - make Ganeti not break with DRBD 8.2.x (which changed the version
3021 format in ``/proc/drbd``) (issue 24)
3022 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3023 template, allowing migration from DRBD0.7+MD to DRBD8
3029 *(Released Wed, 30 Jan 2008)*
3031 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3033 - add command aliases infrastructure and a few aliases
3034 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
3035 man pages and the ``--help`` option of ``gnt-node
3036 list``/``gnt-instance list``
3037 - fix ``gnt-backup list`` with down nodes (issue 21)
3038 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3039 - fix the dist archive and add a check for including svn/git files in
3041 - some developer-related changes: improve the burnin and the QA suite,
3042 add an upload script for testing during development
3048 *(Released Wed, 16 Jan 2008)*
3050 - experimental HVM support, read the install document, section
3051 "Initializing the cluster"
3052 - allow for the PVM hypervisor per-instance kernel and initrd paths
3053 - add a new command ``gnt-cluster verify-disks`` which uses a new
3054 algorithm to improve the reconnection of the DRBD pairs if the device
3055 on the secondary node has gone away
3056 - make logical volume code auto-activate LVs at disk activation time
3057 - slightly improve the speed of activating disks
3058 - allow specification of the MAC address at instance creation time, and
3059 changing it later via ``gnt-instance modify``
3060 - fix handling of external commands that generate lots of output on
3062 - update documentation with regard to minimum version of DRBD8 supported
3068 *(Released Tue, 4 Dec 2007)*
3070 - Log the ``xm create`` output to the node daemon log on failure (to
3071 help diagnosing the error)
3072 - In debug mode, log all external commands output if failed to the logs
3073 - Change parsing of lvm commands to ignore stderr
3079 *(Released Wed, 28 Nov 2007)*
3081 - Another round of updates to the DRBD 8 code to deal with more failures
3082 in the replace secondary node operation
3083 - Some more logging of failures in disk operations (lvm, drbd)
3084 - A few documentation updates
3091 *(Released Tue, 13 Nov 2007)*
3093 - Change configuration file format from Python's Pickle to JSON.
3094 Upgrading is possible using the cfgupgrade utility.
3095 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3096 faster replace disks and is more stable (DRBD 8 has many improvements
3097 compared to DRBD 0.7)
3098 - Added command line tags support (see man pages for ``gnt-instance``,
3099 ``gnt-node``, ``gnt-cluster``)
3100 - Added instance rename support
3101 - Added multi-instance startup/shutdown
3102 - Added cluster rename support
3103 - Added ``gnt-node evacuate`` to simplify some node operations
3104 - Added instance reboot operation that can speedup reboot as compared to
3106 - Soften the requirement that hostnames are in FQDN format
3107 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3109 - Removed dependency on debian's patched fping that uses the
3110 non-standard ``-S`` option
3111 - Now the OS definitions are searched for in multiple, configurable
3112 paths (easier for distros to package)
3113 - Some changes to the hooks infrastructure (especially the new
3114 post-configuration update hook)
3115 - Other small bugfixes
3117 .. vim: set textwidth=72 syntax=rst :