10 - Ensure that all the hypervisors exist in the config file (Issue 640)
11 - Correctly recognise the role as master node (Issue 687)
12 - configure: allow detection of Sphinx 1.2+ (Issue 502)
18 *(Released Fri, 13 Dec 2013)*
20 - use custom KVM path if set for version checking
21 - SingleNotifyPipeCondition: don't share pollers
23 Inherited from the 2.8 branch:
25 - Fixed Luxi daemon socket permissions after master-failover
26 - Improve IP version detection code directly checking for colons rather than
27 passing the family from the cluster object
28 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
29 opportunistically anymore (Issue 622)
30 - Allow link local IPv6 gateways (Issue 624)
31 - Fix error printing (Issue 616)
32 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
33 disk modifications, keep the old one. If name=none then set disk name to
35 - Update build_chroot script to work with the latest hackage packages
36 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
37 - Fix evacuation out of drained node (Issue 615)
38 - Add default file_driver if missing (Issue 571)
39 - Fix job error message after unclean master shutdown (Issue 618)
40 - Lock group(s) when creating instances (Issue 621)
41 - SetDiskID() before accepting an instance (Issue 633)
42 - Allow the ext template disks to receive arbitrary parameters, both at creation
43 time and while being modified
44 - Xen handle domain shutdown (future proofing cherry-pick)
45 - Refactor reading live data in htools (future proofing cherry-pick)
51 *(Released Wed, 13 Nov 2013)*
53 - fix bug, that kept nodes offline when readding
54 - when verifying DRBD versions, ignore unavailable nodes
55 - fix bug that made the console unavailable on kvm in split-user
57 - DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
63 *(Released Tue, 5 Nov 2013)*
65 Incompatible/important changes
66 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
68 - hroller now also plans for capacity to move non-redundant instances off
69 any node to be rebooted; the old behavior of completely ignoring any
70 non-redundant instances can be restored by adding the --ignore-non-redundant
72 - The cluster option '--no-lvm-storage' was removed in favor of the new option
73 '--enabled-disk-templates'.
74 - On instance creation, disk templates no longer need to be specified
75 with '-t'. The default disk template will be taken from the list of
76 enabled disk templates.
77 - The monitoring daemon is now running as root, in order to be able to collect
78 information only available to root (such as the state of Xen instances).
79 - The ConfD client is now IPv6 compatible.
80 - File and shared file storage is no longer dis/enabled at configure time,
81 but using the option '--enabled-disk-templates' at cluster initialization and
83 - The default directories for file and shared file storage are not anymore
84 specified at configure time, but taken from the cluster's configuration.
85 They can be set at cluster initialization and modification with
86 '--file-storage-dir' and '--shared-file-storage-dir'.
87 - Cluster verification now includes stricter checks regarding the
88 default file and shared file storage directories. It now checks that
89 the directories are explicitely allowed in the 'file-storage-paths' file and
90 that the directories exist on all nodes.
91 - The list of allowed disk templates in the instance policy and the list
92 of cluster-wide enabled disk templates is now checked for consistency
93 on cluster or group modification. On cluster initialization, the ipolicy
94 disk templates are ensured to be a subset of the cluster-wide enabled
100 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
101 the correct command syntax. It is possible to use different DRBD versions
102 on different nodes as long as they are compatible to each other. This
103 enables rolling upgrades of DRBD with no downtime. As permanent operation
104 of different DRBD versions within a node group is discouraged,
105 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
106 - New "inst-status-xen" data collector for the monitoring daemon, providing
107 information about the state of the xen instances on the nodes.
108 - New "lv" data collector for the monitoring daemon, collecting data about the
109 logical volumes on the nodes, and pairing them with the name of the instances
111 - New "diskstats" data collector, collecting the data from /proc/diskstats and
112 presenting them over the monitoring daemon interface.
113 - The ConfD client is now IPv6 compatible.
117 The following new dependencies have been added.
121 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
122 for the unit tests (and only used for testing).
126 - ``hslogger`` (http://software.complete.org/hslogger) is now always
127 required, even if confd is not enabled.
132 - Correctly start/stop luxid during gnt-cluster master-failover (inherited
134 - Improved error messsages (inherited from stable-2.8)
140 *(Released Tue, 15 Oct 2013)*
142 The third release candidate in the 2.9 series. Since 2.9.0 rc2:
144 - in implicit configuration upgrade, match ipolicy with enabled disk templates
145 - improved harep documentation (inherited from stable-2.8)
151 *(Released Wed, 9 Oct 2013)*
153 The second release candidate in the 2.9 series. Since 2.9.0 rc1:
155 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
156 at least one DRBD instance.
157 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
164 *(Released Tue, 1 Oct 2013)*
166 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
169 - update of the documentation, in particular installation instructions
170 - merging of LD_* constants into DT_* constants
171 - python style changes to be compatible with newer versions of pylint
177 *(Released Thu, 29 Aug 2013)*
179 This was the first beta release of the 2.9 series. All important changes
180 are listed in the latest 2.9 entry.
186 *(Released Thu, 23 Jan 2014)*
188 - Change the list separator for the usb_devices parameter from comma to space.
189 Commas could not work because they are already the hypervisor option
190 separator (Issue 649)
191 - Add support for blktap2 file-driver (Issue 638)
192 - Add network tag definitions to the haskell codebase (Issue 641)
193 - Fix RAPI network tag handling
194 - Add the network tags to the tags searched by gnt-cluster search-tags
195 - Fix caching bug preventing jobs from being cancelled
196 - Start-master/stop-master was always failing of ConfD was disabled. (Issue 685)
202 *(Released Thu, 12 Dec 2013)*
204 - Fixed Luxi daemon socket permissions after master-failover
205 - Improve IP version detection code directly checking for colons rather than
206 passing the family from the cluster object
207 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
208 opportunistically anymore (Issue 622)
209 - Allow link local IPv6 gateways (Issue 624)
210 - Fix error printing (Issue 616)
211 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
212 disk modifications, keep the old one. If name=none then set disk name to
214 - Update build_chroot script to work with the latest hackage packages
215 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
216 - Fix evacuation out of drained node (Issue 615)
217 - Add default file_driver if missing (Issue 571)
218 - Fix job error message after unclean master shutdown (Issue 618)
219 - Lock group(s) when creating instances (Issue 621)
220 - SetDiskID() before accepting an instance (Issue 633)
221 - Allow the ext template disks to receive arbitrary parameters, both at creation
222 time and while being modified
223 - Xen handle domain shutdown (future proofing cherry-pick)
224 - Refactor reading live data in htools (future proofing cherry-pick)
230 *(Released Thu, 07 Nov 2013)*
232 - DRBD: ensure peers are UpToDate for dual-primary
233 - Improve error message for replace-disks
234 - More dependency checks at configure time
235 - Placate warnings on ganeti.outils_unittest.py
241 *(Released Thu, 17 Oct 2013)*
243 - Correctly start/stop luxid during gnt-cluster master-failover
244 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
245 - Fix path for the job queue serial file
246 - Improved harep man page
247 - Minor documentation improvements
253 *(Released Mon, 30 Sep 2013)*
255 Incompatible/important changes
256 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
258 - Instance policy can contain multiple instance specs, as described in
259 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
260 <design-partitioned>`. As a consequence, it's not possible to partially change
261 or override instance specs. Bounding specs (min and max) can be specified as a
262 whole using the new option ``--ipolicy-bounds-specs``, while standard
263 specs use the new option ``--ipolicy-std-specs``.
264 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
265 gnt-instance is a valid YAML object.
266 - hail now honors network restrictions when allocating nodes. This led to an
267 update of the IAllocator protocol. See the IAllocator documentation for
269 - confd now only answers static configuration request over the network. luxid
270 was extracted, listens on the local LUXI socket and responds to live queries.
271 This allows finer grained permissions if using separate users.
276 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
277 to always require authentication, ``--require-authentication``. It can
278 be specified in ``$sysconfdir/default/ganeti``.
279 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
280 be used to manage the disk templates to be used by instances in the cluster.
281 Initially, it will be set to a list that includes plain, drbd, if they were
282 enabled by specifying a volume group name, and file and sharedfile, if those
283 were enabled at configure time. Additionally, it will include all disk
284 templates that are currently used by instances. The order of disk templates
285 will be based on Ganeti's history of supporting them. In the future, the
286 first entry of the list will be used as a default disk template on instance
288 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
289 configuration back to the previous stable version.
290 - Disk templates in group ipolicy can be restored to the default value.
291 - Initial support for diskless instances and virtual clusters in QA.
292 - More QA and unit tests for instance policies.
293 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
294 describing why the opcode itself was executed.
295 - The monitoring daemon is now available. It allows users to query the cluster
296 for obtaining information about the status of the system. The daemon is only
297 responsible for providing the information over the network: the actual data
298 gathering is performed by data collectors (currently, only the DRBD status
299 collector is available).
300 - In order to help developers work on Ganeti, a new script
301 (``devel/build_chroot``) is provided, for building a chroot that contains all
302 the required development libraries and tools for compiling Ganeti on a Debian
304 - A new tool, ``harep``, for performing self-repair and recreation of instances
305 in Ganeti has been added.
306 - Split queries are enabled for tags, network, exports, cluster info, groups,
308 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
309 It prints the command line to set the current policies, to ease
311 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
312 devices for KVM virtio-net interfaces will get created with VNET_HDR
313 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
314 virtio-net interfaces, which prevents host kernel tainting and log
315 flooding, when dealing with broken or malicious virtio-net drivers.
316 It's set to true by default.
317 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
319 - Support 'viridian' parameter in Xen HVM
320 - Support DSA SSH keys in bootstrap
321 - To simplify the work of packaging frameworks that want to add the needed users
322 and groups in a split-user setup themselves, at build time three files in
323 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
324 the groups to be generated, the ``users`` file contains, one per line, the
325 users to be generated, optionally followed by their primary group, where
326 important. The ``groupmemberships`` file contains, one per line, additional
327 user-group membership relations that need to be established. The syntax of
328 these files will remain stable in all future versions.
333 The following new dependencies have been added:
336 - The ``curl`` library is not optional anymore for compiling the Haskell code.
337 - ``snap-server`` library (if monitoring is enabled).
340 - The minimum Python version needed to run Ganeti is now 2.6.
341 - ``yaml`` library (only for running the QA).
345 - Perform proper cleanup on termination of Haskell daemons
346 - Fix corner-case in handling of remaining retry time
352 *(Released Tue, 17 Sep 2013)*
354 - To simplify the work of packaging frameworks that want to add the needed users
355 and groups in a split-user setup themselves, at build time three files in
356 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
357 the groups to be generated, the ``users`` file contains, one per line, the
358 users to be generated, optionally followed by their primary group, where
359 important. The ``groupmemberships`` file contains, one per line, additional
360 user-group membership relations that need to be established. The syntax of
361 these files will remain stable in all future versions.
362 - Add a default to file-driver when unspecified over RAPI (Issue 571)
363 - Mark the DSA host pubkey as optional, and remove it during config downgrade
365 - Some documentation fixes
371 *(Released Tue, 27 Aug 2013)*
373 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
375 - Support 'viridian' parameter in Xen HVM (Issue 233)
376 - Include VCS version in ``gnt-cluster version``
377 - Support DSA SSH keys in bootstrap (Issue 338)
378 - Fix batch creation of instances
379 - Use FQDN to check master node status (Issue 551)
380 - Make the DRBD collector more failure-resilient
386 *(Released Fri, 2 Aug 2013)*
388 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
390 - Fix upgrading/downgrading from 2.7
391 - Increase maximum RAPI message size
392 - Documentation updates
393 - Split ``confd`` between ``luxid`` and ``confd``
394 - Merge 2.7 series up to the 2.7.1 release
395 - Allow the ``modify_etc_hosts`` option to be changed
396 - Add better debugging for ``luxid`` queries
397 - Expose bulk parameter for GetJobs in RAPI client
398 - Expose missing ``network`` fields in RAPI
399 - Add some ``cluster verify`` tests
400 - Some unittest fixes
401 - Fix a malfunction in ``hspace``'s tiered allocation
402 - Fix query compatibility between haskell and python implementations
403 - Add the ``vnet_hdr`` HV parameter for KVM
404 - Add ``--cleanup`` to instance failover
405 - Change the connected groups format in ``gnt-network info`` output; it
406 was previously displayed as a raw list by mistake. (Merged from 2.7)
412 *(Released Mon, 24 Jun 2013)*
414 This was the first beta release of the 2.8 series. All important changes
415 are listed in the latest 2.8 entry.
421 *(Released Thu, 26 Sep 2013)*
423 - Change the connected groups format in ``gnt-network info`` output; it
424 was previously displayed as a raw list by mistake
425 - Check disk template in right dict when copying
426 - Support multi-instance allocs without iallocator
427 - Fix some errors in the documentation
428 - Fix formatting of tuple in an error message
434 *(Released Thu, 25 Jul 2013)*
436 - Add logrotate functionality in daemon-util
437 - Add logrotate example file
438 - Add missing fields to network queries over rapi
439 - Fix network object timestamps
440 - Add support for querying network timestamps
441 - Fix a typo in the example crontab
442 - Fix a documentation typo
448 *(Released Thu, 04 Jul 2013)*
450 Incompatible/important changes
451 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
453 - Instance policies for disk size were documented to be on a per-disk
454 basis, but hail applied them to the sum of all disks. This has been
456 - ``hbal`` will now exit with status 0 if, during job execution over
457 LUXI, early exit has been requested and all jobs are successful;
458 before, exit status 1 was used, which cannot be differentiated from
460 - Compatibility with newer versions of rbd has been fixed
461 - ``gnt-instance batch-create`` has been changed to use the bulk create
462 opcode from Ganeti. This lead to incompatible changes in the format of
463 the JSON file. It's now not a custom dict anymore but a dict
464 compatible with the ``OpInstanceCreate`` opcode.
465 - Parent directories for file storage need to be listed in
466 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
467 write the file automatically based on old configuration values, but it
468 can not distribute it across all nodes and the file contents should be
469 verified. Use ``gnt-cluster copyfile
470 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
471 upgraded. The reason for requiring this list of paths now is that
472 before it would have been possible to inject new paths via RPC,
473 allowing files to be created in arbitrary locations. The RPC protocol
474 is protected using SSL/X.509 certificates, but as a design principle
475 Ganeti does not permit arbitrary paths to be passed.
476 - The parsing of the variants file for OSes (see
477 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
478 lines and comment lines (starting with ``#``) are ignored for better
480 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
481 longer available. ``gnt-node add`` now invokes a new tool on the
482 destination node, named ``prepare-node-join``, to configure the SSH
483 daemon. Paramiko is no longer necessary to configure nodes' SSH
484 daemons via ``gnt-node add``.
485 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
486 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
487 the flag is not reset after a master failover.
488 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
489 is recommended. 2.6 will be mandatory from the 2.8 series.
495 - New network management functionality to support automatic allocation
496 of IP addresses and managing of network parameters. See
497 :manpage:`gnt-network(8)` for more details.
498 - New external storage backend, to allow managing arbitrary storage
499 systems external to the cluster. See
500 :manpage:`ganeti-extstorage-interface(7)`.
501 - New ``exclusive-storage`` node parameter added, restricted to
502 nodegroup level. When it's set to true, physical disks are assigned in
503 an exclusive fashion to instances, as documented in :doc:`Partitioned
504 Ganeti <design-partitioned>`. Currently, only instances using the
505 ``plain`` disk template are supported.
506 - The KVM hypervisor has been updated with many new hypervisor
507 parameters, including a generic one for passing arbitrary command line
508 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
509 compatible up to qemu 1.4.
510 - A new tool, called ``mon-collector``, is the stand-alone executor of
511 the data collectors for a monitoring system. As of this version, it
512 just includes the DRBD data collector, that can be executed by calling
513 ``mon-collector`` using the ``drbd`` parameter. See
514 :manpage:`mon-collector(7)`.
515 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
516 for RAPI users. It allows granting permissions to query for
517 information to a specific user without giving
518 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
519 - A new tool named ``node-cleanup`` has been added. It cleans remains of
520 a cluster from a machine by stopping all daemons, removing
521 certificates and ssconf files. Unless the ``--no-backup`` option is
522 given, copies of the certificates are made.
523 - Instance creations now support the use of opportunistic locking,
524 potentially speeding up the (parallel) creation of multiple instances.
525 This feature is currently only available via the :doc:`RAPI
526 <rapi>` interface and when an instance allocator is used. If the
527 ``opportunistic_locking`` parameter is set the opcode will try to
528 acquire as many locks as possible, but will not wait for any locks
529 held by other opcodes. If not enough resources can be found to
530 allocate the instance, the temporary error code
531 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
532 retried thereafter, with or without opportunistic locking.
533 - New experimental linux-ha resource scripts.
534 - Restricted-commands support: ganeti can now be asked (via command line
535 or rapi) to perform commands on a node. These are passed via ganeti
536 RPC rather than ssh. This functionality is restricted to commands
537 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
538 reasons. The file is not copied automatically.
544 - Diskless instances are now externally mirrored (Issue 237). This for
545 now has only been tested in conjunction with explicit target nodes for
547 - Queries not needing locks or RPC access to the node can now be
548 performed by the confd daemon, making them independent from jobs, and
549 thus faster to execute. This is selectable at configure time.
550 - The functionality for allocating multiple instances at once has been
551 overhauled and is now also available through :doc:`RAPI <rapi>`.
553 There are no significant changes from version 2.7.0~rc3.
559 *(Released Tue, 25 Jun 2013)*
561 - Fix permissions on the confd query socket (Issue 477)
562 - Fix permissions on the job archive dir (Issue 498)
563 - Fix handling of an internal exception in replace-disks (Issue 472)
564 - Fix gnt-node info handling of shortened names (Issue 497)
565 - Fix gnt-instance grow-disk when wiping is enabled
566 - Documentation improvements, and support for newer pandoc
567 - Fix hspace honoring ipolicy for disks (Issue 484)
568 - Improve handling of the ``kvm_extra`` HV parameter
574 *(Released Fri, 24 May 2013)*
576 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
578 - Disks added through ``gnt-instance modify`` or created through
579 ``gnt-instance recreate-disks`` are wiped, if the
580 ``prealloc_wipe_disks`` flag is set.
581 - If wiping newly created disks fails, the disks are removed. Also,
582 partial failures in creating disks through ``gnt-instance modify``
583 triggers a cleanup of the partially-created disks.
584 - Removing the master IP address doesn't fail if the address has been
586 - Fix ownership of the OS log dir
587 - Workaround missing SO_PEERCRED constant (Issue 191)
593 *(Released Fri, 3 May 2013)*
595 This was the first release candidate of the 2.7 series. Since beta3:
597 - Fix kvm compatibility with qemu 1.4 (Issue 389)
598 - Documentation updates (admin guide, upgrade notes, install
599 instructions) (Issue 372)
600 - Fix gnt-group list nodes and instances count (Issue 436)
601 - Fix compilation without non-mandatory libraries (Issue 441)
602 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
603 - Make confd logging more verbose at INFO level (Issue 435)
604 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
605 - Fix failure path for instance storage type conversion (Issue 229)
606 - Update htools text backend documentation
607 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
608 - Disable inter-cluster instance move for file-based instances, because
609 it is dependant on instance export, which is not supported for
610 file-based instances. (Issue 414)
611 - Fix gnt-job crashes on non-ascii characters (Issue 427)
612 - Fix volume group checks on non-vm-capable nodes (Issue 432)
618 *(Released Mon, 22 Apr 2013)*
620 This was the third beta release of the 2.7 series. Since beta2:
622 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
623 - Fix data loss on wrong usage of ``gnt-instance move``
624 - Properly export errors in confd-based job queries
625 - Add ``users-setup`` tool
626 - Fix iallocator protocol to report 0 as a disk size for diskless
627 instances. This avoids hail breaking when a diskless instance is
629 - Fix job queue directory permission problem that made confd job queries
630 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
631 for access to archived jobs (Issue 406).
632 - Limit the sizes of networks supported by ``gnt-network`` to something
633 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
634 - Fix bugs in instance disk template conversion
635 - Fix GHC 7 compatibility
636 - Fix ``burnin`` install path (Issue 426).
637 - Allow very small disk grows (Issue 347).
638 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
639 that noded doesn't import masterd code (Issue 419).
640 - Make sure the default metavg at cluster init is the same as the vg, if
641 unspecified (Issue 358).
642 - Fix cleanup of partially created disks (part of Issue 416)
648 *(Released Tue, 2 Apr 2013)*
650 This was the second beta release of the 2.7 series. Since beta1:
652 - Networks no longer have a "type" slot, since this information was
653 unused in Ganeti: instead of it tags should be used.
654 - The rapi client now has a ``target_node`` option to MigrateInstance.
655 - Fix early exit return code for hbal (Issue 386).
656 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
657 - Fix ``rbd showmapped`` output parsing (Issue 312).
658 - Networks are now referenced indexed by UUID, rather than name. This
659 will require running cfgupgrade, from 2.7.0beta1, if networks are in
661 - The OS environment now includes network information.
662 - Deleting of a network is now disallowed if any instance nic is using
663 it, to prevent dangling references.
664 - External storage is now documented in man pages.
665 - The exclusive_storage flag can now only be set at nodegroup level.
666 - Hbal can now submit an explicit priority with its jobs.
667 - Many network related locking fixes.
668 - Bump up the required pylint version to 0.25.1.
669 - Fix the ``no_remember`` option in RAPI client.
670 - Many ipolicy related tests, qa, and fixes.
671 - Many documentation improvements and fixes.
672 - Fix building with ``--disable-file-storage``.
673 - Fix ``-q`` option in htools, which was broken if passed more than
675 - Some haskell/python interaction improvements and fixes.
676 - Fix iallocator in case of missing LVM storage.
677 - Fix confd config load in case of ``--no-lvm-storage``.
678 - The confd/query functionality is now mentioned in the security
685 *(Released Wed, 6 Feb 2013)*
687 This was the first beta release of the 2.7 series. All important changes
688 are listed in the latest 2.7 entry.
694 *(Released Fri, 21 Dec 2012)*
696 Important behaviour change: hbal won't rebalance anymore instances which
697 have the ``auto_balance`` attribute set to false. This was the intention
698 all along, but until now it only skipped those from the N+1 memory
699 reservation (DRBD-specific).
701 A significant number of bug fixes in this release:
703 - Fixed disk adoption interaction with ipolicy checks.
704 - Fixed networking issues when instances are started, stopped or
705 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
706 - Fixed the warning in cluster verify for shared storage instances not
708 - Fixed removal of storage directory on shared file storage (issue 262).
709 - Fixed validation of LVM volume group name in OpClusterSetParams
710 (``gnt-cluster modify``) (issue 285).
711 - Fixed runtime memory increases (``gnt-instance modify -m``).
712 - Fixed live migration under Xen's ``xl`` mode.
713 - Fixed ``gnt-instance console`` with ``xl``.
714 - Fixed building with newer Haskell compiler/libraries.
715 - Fixed PID file writing in Haskell daemons (confd); this prevents
716 restart issues if confd was launched manually (outside of
717 ``daemon-util``) while another copy of it was running
718 - Fixed a type error when doing live migrations with KVM (issue 297) and
719 the error messages for failing migrations have been improved.
720 - Fixed opcode validation for the out-of-band commands (``gnt-node
722 - Fixed a type error when unsetting OS hypervisor parameters (issue
723 311); now it's possible to unset all OS-specific hypervisor
725 - Fixed the ``dry-run`` mode for many operations: verification of
726 results was over-zealous but didn't take into account the ``dry-run``
727 operation, resulting in "wrong" failures.
728 - Fixed bash completion in ``gnt-job list`` when the job queue has
729 hundreds of entries; especially with older ``bash`` versions, this
730 results in significant CPU usage.
732 And lastly, a few other improvements have been made:
734 - Added option to force master-failover without voting (issue 282).
735 - Clarified error message on lock conflict (issue 287).
736 - Logging of newly submitted jobs has been improved (issue 290).
737 - Hostname checks have been made uniform between instance rename and
739 - The ``--submit`` option is now supported by ``gnt-debug delay``.
740 - Shutting down the master daemon by sending SIGTERM now stops it from
741 processing jobs waiting for locks; instead, those jobs will be started
742 once again after the master daemon is started the next time (issue
744 - Support for Xen's ``xl`` program has been improved (besides the fixes
746 - Reduced logging noise in the Haskell confd daemon (only show one log
747 entry for each config reload, instead of two).
748 - Several man page updates and typo fixes.
754 *(Released Fri, 12 Oct 2012)*
756 A small bugfix release. Among the bugs fixed:
758 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
759 made the command unusable.
760 - Commands that issue many jobs don't fail anymore just because some jobs
761 take so long that other jobs are archived.
762 - Failures during ``gnt-instance reinstall`` are reflected by the exit
764 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
766 - When ``always_failover`` is set, ``--allow-failover`` is not required
767 in migrate commands anymore.
768 - ``bash_completion`` works even if extglob is disabled.
769 - Fixed bug with locks that made failover for RDB-based instances fail.
770 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
771 a random node instead of one based on the allocator metric.
772 - Support for newer versions of pylint and pep8.
773 - Hail doesn't fail anymore when trying to add an instance of type
774 ``file``, ``sharedfile`` or ``rbd``.
775 - Added new Makefile target to rebuild the whole distribution, so that
776 all files are included.
782 *(Released Fri, 27 Jul 2012)*
785 .. attention:: The ``LUXI`` protocol has been made more consistent
786 regarding its handling of command arguments. This, however, leads to
787 incompatibility issues with previous versions. Please ensure that you
788 restart Ganeti daemons soon after the upgrade, otherwise most
789 ``LUXI`` calls (job submission, setting/resetting the drain flag,
790 pausing/resuming the watcher, cancelling and archiving jobs, querying
791 the cluster configuration) will fail.
800 The current ``admin_up`` field, which used to denote whether an instance
801 should be running or not, has been removed. Instead, ``admin_state`` is
802 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
804 The rational behind this is that an instance being “down” can have
807 - it could be down during a reboot
808 - it could be temporarily be down for a reinstall
809 - or it could be down because it is deprecated and kept just for its
812 The previous Boolean state was making it difficult to do capacity
813 calculations: should Ganeti reserve memory for a down instance? Now, the
814 tri-state field makes it clear:
816 - in ``up`` and ``down`` state, all resources are reserved for the
817 instance, and it can be at any time brought up if it is down
818 - in ``offline`` state, only disk space is reserved for it, but not
821 The field can have an extra use: since the transition between ``up`` and
822 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
823 transition between ``offline`` and ``down`` is done via ``gnt-instance
824 modify``, it is possible to given different rights to users. For
825 example, owners of an instance could be allowed to start/stop it, but
826 not transition it out of the offline state.
828 Instance policies and specs
829 +++++++++++++++++++++++++++
831 In previous Ganeti versions, an instance creation request was not
832 limited on the minimum size and on the maximum size just by the cluster
833 resources. As such, any policy could be implemented only in third-party
834 clients (RAPI clients, or shell wrappers over ``gnt-*``
835 tools). Furthermore, calculating cluster capacity via ``hspace`` again
836 required external input with regards to instance sizes.
838 In order to improve these workflows and to allow for example better
839 per-node group differentiation, we introduced instance specs, which
842 - minimum instance disk size, disk count, memory size, cpu count
843 - maximum values for the above metrics
844 - and “standard” values (used in ``hspace`` to calculate the standard
847 The minimum/maximum values can be also customised at node-group level,
848 for example allowing more powerful hardware to support bigger instance
851 Beside the instance specs, there are a few other settings belonging to
852 the instance policy framework. It is possible now to customise, per
853 cluster and node-group:
855 - the list of allowed disk templates
856 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
857 - the maximum ratio of instance to spindles (see below for more
858 information) for local storage
860 All these together should allow all tools that talk to Ganeti to know
861 what are the ranges of allowed values for instances and the
862 over-subscription that is allowed.
864 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
865 instance configuration, and the physical CPU configuration from the
866 node. For the spindle ratios however, we didn't track before these
867 values, so new parameters have been added:
869 - a new node parameter ``spindle_count``, defaults to 1, customisable at
870 node group or node level
871 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
873 Note that spindles in this context doesn't need to mean actual
874 mechanical hard-drives; it's just a relative number for both the node
875 I/O capacity and instance I/O consumption.
877 Instance migration behaviour
878 ++++++++++++++++++++++++++++
880 While live-migration is in general desirable over failover, it is
881 possible that for some workloads it is actually worse, due to the
882 variable time of the “suspend” phase during live migration.
884 To allow the tools to work consistently over such instances (without
885 having to hard-code instance names), a new backend parameter
886 ``always_failover`` has been added to control the migration/failover
887 behaviour. When set to True, all migration requests for an instance will
888 instead fall-back to failover.
890 Instance memory ballooning
891 ++++++++++++++++++++++++++
893 Initial support for memory ballooning has been added. The memory for an
894 instance is no longer fixed (backend parameter ``memory``), but instead
895 can vary between minimum and maximum values (backend parameters
896 ``minmem`` and ``maxmem``). Currently we only change an instance's
899 - live migrating or failing over and instance and the target node
900 doesn't have enough memory
901 - user requests changing the memory via ``gnt-instance modify
907 In order to control the use of specific CPUs by instance, support for
908 controlling CPU pinning has been added for the Xen, HVM and LXC
909 hypervisors. This is controlled by a new hypervisor parameter
910 ``cpu_mask``; details about possible values for this are in the
911 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
912 VCPU-to-CPU mapping) form will work well only when all nodes in your
913 cluster have the same amount of CPUs.
918 Another area in which Ganeti was not customisable were the parameters
919 used for storage configuration, e.g. how many stripes to use for LVM,
920 DRBD resync configuration, etc.
922 To improve this area, we've added disks parameters, which are
923 customisable at cluster and node group level, and which allow to
924 specify various parameters for disks (DRBD has the most parameters
925 currently), for example:
927 - DRBD resync algorithm and parameters (e.g. speed)
928 - the default VG for meta-data volumes for DRBD
929 - number of stripes for LVM (plain disk template)
932 These parameters can be modified via ``gnt-cluster modify -D …`` and
933 ``gnt-group modify -D …``, and are used at either instance creation (in
934 case of LVM stripes, for example) or at disk “activation” time
937 Rados block device support
938 ++++++++++++++++++++++++++
940 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
941 denoted by the ``rbd`` disk template type. This is considered
942 experimental, feedback is welcome. For details on configuring it, see
943 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
948 The existing master IP functionality works well only in simple setups (a
949 single network shared by all nodes); however, if nodes belong to
950 different networks, then the ``/32`` setup and lack of routing
951 information is not enough.
953 To allow the master IP to function well in more complex cases, the
954 system was reworked as follows:
956 - a master IP netmask setting has been added
957 - the master IP activation/turn-down code was moved from the node daemon
959 - whether to run the Ganeti-supplied master IP script or a user-supplied
960 on is a ``gnt-cluster init`` setting
962 Details about the location of the standard and custom setup scripts are
963 in the man page :manpage:`gnt-cluster(8)`; for information about the
964 setup script protocol, look at the Ganeti-supplied script.
969 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
972 It is now possible to use TLS-protected connections, and when renewing
973 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
974 it is now possible to specify spice or spice CA certificates. Also, it
975 is possible to configure a password for SPICE sessions via the
976 hypervisor parameter ``spice_password_file``.
978 There are also new parameters to control the compression and streaming
979 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
980 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
981 for the spice parameters.
983 Lastly, it is now possible to see the SPICE connection information via
984 ``gnt-instance console``.
989 A new tool (``tools/ovfconverter``) has been added that supports
990 conversion between Ganeti and the `Open Virtualization Format
991 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
994 This relies on the ``qemu-img`` tool to convert the disk formats, so the
995 actual compatibility with other virtualization solutions depends on it.
1000 The configuration query daemon (``ganeti-confd``) is now optional, and
1001 has been rewritten in Haskell; whether to use the daemon at all, use the
1002 Python (default) or the Haskell version is selectable at configure time
1003 via the ``--enable-confd`` parameter, which can take one of the
1004 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
1005 daemon will result in a smaller footprint; for larger systems, we
1006 welcome feedback on the Haskell version which might become the default
1009 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
1010 daemon running. The Python version doesn't implement the new call.
1013 User interface changes
1014 ~~~~~~~~~~~~~~~~~~~~~~
1016 We have replaced the ``--disks`` option of ``gnt-instance
1017 replace-disks`` with a more flexible ``--disk`` option, which allows
1018 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
1019 disk size and mode can be changed upon recreation (via ``gnt-instance
1020 recreate-disks``, which accepts the same ``--disk`` option).
1022 As many people are used to a ``show`` command, we have added that as an
1023 alias to ``info`` on all ``gnt-*`` commands.
1025 The ``gnt-instance grow-disk`` command has a new mode in which it can
1026 accept the target size of the disk, instead of the delta; this can be
1027 more safe since two runs in absolute mode will be idempotent, and
1028 sometimes it's also easier to specify the desired size directly.
1030 Also the handling of instances with regard to offline secondaries has
1031 been improved. Instance operations should not fail because one of it's
1032 secondary nodes is offline, even though it's safe to proceed.
1034 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
1035 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1036 minors to instance name.
1041 RAPI coverage has improved, with (for example) new resources for
1042 recreate-disks, node power-cycle, etc.
1047 There is partial support for ``xl`` in the Xen hypervisor; feedback is
1050 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1051 whether to still support Python 2.4 or move to Python 2.6 as minimum
1054 Support for Fedora has been slightly improved; the provided example
1055 init.d script should work better on it and the INSTALL file should
1056 document the needed dependencies.
1061 The deprecated ``QueryLocks`` LUXI request has been removed. Use
1062 ``Query(what=QR_LOCK, ...)`` instead.
1064 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1065 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1066 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1067 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1068 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1070 RAPI client: ``CertificateError`` now derives from
1071 ``GanetiApiError``. This should make it more easy to handle Ganeti
1074 Deprecation warnings due to PyCrypto/paramiko import in
1075 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
1076 make sure to run an up-to-date paramiko version, if you use this tool.
1078 The QA scripts now depend on Python 2.5 or above (the main code base
1079 still works with Python 2.4).
1081 The configuration file (``config.data``) is now written without
1082 indentation for performance reasons; if you want to edit it, it can be
1083 re-formatted via ``tools/fmtjson``.
1085 A number of bugs has been fixed in the cluster merge tool.
1087 ``x509`` certification verification (used in import-export) has been
1088 changed to allow the same clock skew as permitted by the cluster
1089 verification. This will remove some rare but hard to diagnose errors in
1096 *(Released Thu, 19 Jul 2012)*
1098 Very few changes from rc4 to the final release, only bugfixes:
1100 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
1101 instance, fix CDROM booting for KVM instances)
1102 - fixed node group modification of node parameters
1103 - fixed issue in LUClusterVerifyGroup with multi-group clusters
1104 - fixed generation of bash completion to ensure a stable ordering
1111 *(Released Fri, 13 Jul 2012)*
1113 Third release candidate for 2.6. The following changes were done from
1116 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1117 - Fixed an inconsistency in the LUXI protocol with the provided
1118 arguments (NOT backwards compatible)
1119 - Fixed a bug with node groups ipolicy where ``min`` was greater than
1120 the cluster ``std`` value
1121 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1122 easier instance debugging on nodes (requires ``hconfd`` to work)
1128 *(Released Tue, 03 Jul 2012)*
1130 Second release candidate for 2.6. The following changes were done from
1133 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1135 - Fixed a RAPI regression on missing beparams/memory
1136 - Fixed redistribution of files on offline nodes
1137 - Added possibility to run activate-disks even though secondaries are
1138 offline. With this change it relaxes also the strictness on some other
1139 commands which use activate disks internally:
1140 * ``gnt-instance start|reboot|rename|backup|export``
1141 - Made it possible to remove safely an instance if its secondaries are
1143 - Made it possible to reinstall even though secondaries are offline
1149 *(Released Mon, 25 Jun 2012)*
1151 First release candidate for 2.6. The following changes were done from
1154 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1156 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1157 behaviour. It supports now the modify operation
1158 - ``hcheck`` implemented to analyze cluster health and possibility of
1159 improving health by rebalance
1160 - ``hbal`` has been improved in dealing with split instances
1166 *(Released Mon, 11 Jun 2012)*
1168 Second beta release of 2.6. The following changes were done from beta2
1171 - Fixed ``daemon-util`` with non-root user models
1172 - Fixed creation of plain instances with ``--no-wait-for-sync``
1173 - Fix wrong iv_names when running ``cfgupgrade``
1174 - Export more information in RAPI group queries
1175 - Fixed bug when changing instance network interfaces
1176 - Extended burnin to do NIC changes
1177 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1178 - Changed default for DRBD barriers
1179 - Fixed DRBD error reporting for syncer rate
1180 - Verify the options on disk parameters
1182 And of course various fixes to documentation and improved unittests and
1189 *(Released Wed, 23 May 2012)*
1191 First beta release of 2.6. The following changes were done from beta1 to
1194 - integrated patch for distributions without ``start-stop-daemon``
1195 - adapted example init.d script to work on Fedora
1196 - fixed log handling in Haskell daemons
1197 - adapted checks in the watcher for pycurl linked against libnss
1198 - add partial support for ``xl`` instead of ``xm`` for Xen
1199 - fixed a type issue in cluster verification
1200 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1203 Plus integrated fixes from the 2.5 branch:
1205 - fixed ``kvm-ifup`` to use ``/bin/bash``
1206 - fixed parallel build failures
1207 - KVM live migration when using a custom keymap
1213 *(Released Tue, 24 Jul 2012)*
1215 A small bugfix release, with no new features:
1217 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1218 different default shell (e.g. Debian, Ubuntu)
1219 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1220 243 and Debian bug #650664)
1221 - fixed compatibility with KVM versions that don't support multiple boot
1222 devices (fixes Issue 230 and Debian bug #624256)
1224 Additionally, a few fixes were done to the build system (fixed parallel
1225 build failures) and to the unittests (fixed race condition in test for
1226 FileID functions, and the default enable/disable mode for QA test is now
1233 *(Released Fri, 11 May 2012)*
1235 A small bugfix release.
1237 The main issues solved are on the topic of compatibility with newer LVM
1240 - fixed parsing of ``lv_attr`` field
1241 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1242 the ``--force`` flag is needed
1244 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1245 to accept kernel 3.x too (was hardcoded to 2.6.*).
1247 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1248 of hook results and that also made display of other errors suboptimal
1249 was fixed; the code behaves now like 2.4 and earlier.
1251 Another change in 2.5, the cleanup of the OS scripts environment, is too
1252 aggressive: it removed even the ``PATH`` variable, which requires the OS
1253 scripts to *always* need to export it. Since this is a bit too strict,
1254 we now export a minimal PATH, the same that we export for hooks.
1256 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1257 integrated into this release.
1259 Finally, a few other miscellaneous changes were done (no new features,
1260 just small improvements):
1262 - Fix ``gnt-group --help`` display
1263 - Fix hardcoded Xen kernel path
1264 - Fix grow-disk handling of invalid units
1265 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1266 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1272 *(Released Thu, 12 Apr 2012)*
1274 Incompatible/important changes and bugfixes
1275 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1277 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1278 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1279 to match the underlying LUXI interface.
1280 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1281 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1282 not maintain backwards-compatibility as the underlying operation
1283 changed in an incompatible way. The RAPI client can talk to old
1284 servers, but it needs to be told so as the return value changed.
1285 - When creating file-based instances via RAPI, the ``file_driver``
1286 parameter no longer defaults to ``loop`` and must be specified.
1287 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1289 - Support for the undocumented and deprecated RAPI instance creation
1290 request format version 0 has been dropped. Use version 1, supported
1291 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1292 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1293 documentation <install>`.
1294 - The "cluster-verify" hooks are now executed per group by the
1295 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1296 if you just run ``gnt-cluster verify``, which generates one opcode per
1298 - The environment as passed to the OS scripts is cleared, and thus no
1299 environment variables defined in the node daemon's environment will be
1300 inherited by the scripts.
1301 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1303 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1304 support operations involving multiple node groups.
1305 - Offline nodes are ignored when failing over an instance.
1306 - Support for KVM version 1.0, which changed the version reporting format
1308 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1310 - ``Makefile`` is now compatible with Automake 1.11.2
1311 - Includes all bugfixes made in the 2.4 series
1316 - The ganeti-htools project has been merged into the ganeti-core source
1317 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1318 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1319 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1320 GPT-style partition tables (requires `parted
1321 <http://www.gnu.org/s/parted/>`_).
1322 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1323 - Allowed adding tags on instance creation.
1324 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1325 - Allow instances to be started in a paused state, enabling the user to
1326 see the complete console output on boot using the console.
1327 - Added new hypervisor flag to control default reboot behaviour
1328 (``reboot_behavior``).
1329 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1330 - Improved out-of-band management support:
1332 - Added ``gnt-node health`` command reporting the health status of
1334 - Added ``gnt-node power`` command to manage power status of nodes.
1335 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1337 - Instance migration can fall back to failover if instance is not
1339 - Filters can be used when listing nodes, instances, groups and locks;
1340 see :manpage:`ganeti(7)` manpage.
1341 - Added post-execution status as variables to :doc:`hooks <hooks>`
1343 - Instance tags are exported/imported together with the instance.
1344 - When given an explicit job ID, ``gnt-job info`` will work for archived
1346 - Jobs can define dependencies on other jobs (not yet supported via
1347 RAPI or command line, but used by internal commands and usable via
1350 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1353 - Instance failover is now available as a RAPI resource
1354 (``/2/instances/[instance_name]/failover``).
1355 - ``gnt-instance info`` defaults to static information if primary node
1357 - Opcodes have a new ``comment`` attribute.
1358 - Added basic SPICE support to KVM hypervisor.
1359 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1361 Node group improvements
1362 ~~~~~~~~~~~~~~~~~~~~~~~
1364 - ``gnt-cluster verify`` has been modified to check groups separately,
1365 thereby improving performance.
1366 - Node group support has been added to ``gnt-cluster verify-disks``,
1367 which now operates per node group.
1368 - Watcher has been changed to work better with node groups.
1370 - One process and state file per node group.
1371 - Slow watcher in one group doesn't block other group's watcher.
1373 - Added new command, ``gnt-group evacuate``, to move all instances in a
1374 node group to other groups.
1375 - Added ``gnt-instance change-group`` to move an instance to another
1377 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1378 per-group operations.
1379 - Node groups can be tagged.
1380 - Some operations switch from an exclusive to a shared lock as soon as
1382 - Instance's primary and secondary nodes' groups are now available as
1383 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1384 and ``snodes.group.uuid``).
1389 - Numerous updates to documentation and manpages.
1391 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1393 - Some opcode/job results are now also documented, see :doc:`RAPI
1396 - A lockset's internal lock is now also visible in lock monitor.
1397 - Log messages from job queue workers now contain information about the
1398 opcode they're processing.
1399 - ``gnt-instance console`` no longer requires the instance lock.
1400 - A short delay when waiting for job changes reduces the number of LUXI
1401 requests significantly.
1402 - DRBD metadata volumes are overwritten with zeros during disk creation.
1403 - Out-of-band commands no longer acquire the cluster lock in exclusive
1405 - ``devel/upload`` now uses correct permissions for directories.
1411 *(Released Fri, 23 Mar 2012)*
1413 This was the sixth release candidate of the 2.5 series.
1419 *(Released Mon, 9 Jan 2012)*
1421 This was the fifth release candidate of the 2.5 series.
1427 *(Released Thu, 27 Oct 2011)*
1429 This was the fourth release candidate of the 2.5 series.
1435 *(Released Wed, 26 Oct 2011)*
1437 This was the third release candidate of the 2.5 series.
1443 *(Released Tue, 18 Oct 2011)*
1445 This was the second release candidate of the 2.5 series.
1451 *(Released Tue, 4 Oct 2011)*
1453 This was the first release candidate of the 2.5 series.
1459 *(Released Wed, 31 Aug 2011)*
1461 This was the third beta release of the 2.5 series.
1467 *(Released Mon, 22 Aug 2011)*
1469 This was the second beta release of the 2.5 series.
1475 *(Released Fri, 12 Aug 2011)*
1477 This was the first beta release of the 2.5 series.
1483 *(Released Thu, 27 Oct 2011)*
1485 - Fixed bug when parsing command line parameter values ending in
1487 - Fixed assertion error after unclean master shutdown
1488 - Disable HTTP client pool for RPC, significantly reducing memory usage
1490 - Fixed queue archive creation with wrong permissions
1496 *(Released Tue, 23 Aug 2011)*
1500 - Fixed documentation for importing with ``--src-dir`` option
1501 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1502 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1508 *(Released Fri, 5 Aug 2011)*
1510 Many bug-fixes and a few small features:
1512 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1513 issues when you tried to add an instance with two MAC addresses in one
1515 - KVM: fixed per-instance stored UID value
1516 - KVM: configure bridged NICs at migration start
1517 - KVM: Fix a bug where instance will not start with never KVM versions
1519 - Added OS search path to ``gnt-cluster info``
1520 - Fixed an issue with ``file_storage_dir`` where you were forced to
1521 provide an absolute path, but the documentation states it is a
1522 relative path, the documentation was right
1523 - Added a new parameter to instance stop/start called ``--no-remember``
1524 that will make the state change to not be remembered
1525 - Implemented ``no_remember`` at RAPI level
1526 - Improved the documentation
1527 - Node evacuation: don't call IAllocator if node is already empty
1528 - Fixed bug in DRBD8 replace disks on current nodes
1529 - Fixed bug in recreate-disks for DRBD instances
1530 - Moved assertion checking locks in ``gnt-instance replace-disks``
1531 causing it to abort with not owning the right locks for some situation
1532 - Job queue: Fixed potential race condition when cancelling queued jobs
1533 - Fixed off-by-one bug in job serial generation
1534 - ``gnt-node volumes``: Fix instance names
1535 - Fixed aliases in bash completion
1536 - Fixed a bug in reopening log files after being sent a SIGHUP
1537 - Added a flag to burnin to allow specifying VCPU count
1538 - Bugfixes to non-root Ganeti configuration
1544 *(Released Thu, 12 May 2011)*
1546 Many bug-fixes and a few new small features:
1548 - Fixed a bug related to log opening failures
1549 - Fixed a bug in instance listing with orphan instances
1550 - Fixed a bug which prevented resetting the cluster-level node parameter
1551 ``oob_program`` to the default
1552 - Many fixes related to the ``cluster-merge`` tool
1553 - Fixed a race condition in the lock monitor, which caused failures
1554 during (at least) creation of many instances in parallel
1555 - Improved output for gnt-job info
1556 - Removed the quiet flag on some ssh calls which prevented debugging
1558 - Improved the N+1 failure messages in cluster verify by actually
1559 showing the memory values (needed and available)
1560 - Increased lock attempt timeouts so that when executing long operations
1561 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1562 too early and thus prevent the use of the 'fair' mechanism
1563 - Changed instance query data (``gnt-instance info``) to not acquire
1564 locks unless needed, thus allowing its use on locked instance if only
1565 static information is asked for
1566 - Improved behaviour with filesystems that do not support rename on an
1568 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1569 kept locks on all nodes during the wipe, which is unneeded
1570 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1571 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1572 10GiB) which caused the wipe to fail right at the end in some cases
1573 - Fixed master IP activation when doing master failover with no-voting
1574 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1575 the master node itself
1576 - Fixed potential data-loss in under disk full conditions, where Ganeti
1577 wouldn't check correctly the return code and would consider
1578 partially-written files 'correct'
1579 - Fixed bug related to multiple VGs and DRBD disk replacing
1580 - Added new disk parameter ``metavg`` that allows placement of the meta
1581 device for DRBD in a different volume group
1582 - Fixed error handling in the node daemon when the system libc doesn't
1583 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1584 - Fixed lock release during replace-disks, which kept cluster-wide locks
1585 when doing disk replaces with an iallocator script
1586 - Added check for missing bridges in cluster verify
1587 - Handle EPIPE errors while writing to the terminal better, so that
1588 piping the output to e.g. ``less`` doesn't cause a backtrace
1589 - Fixed rare case where a ^C during Luxi calls could have been
1590 interpreted as server errors, instead of simply terminating
1591 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1593 - Added a few more parameters to the KVM hypervisor, allowing a second
1594 CDROM, custom disk type for CDROMs and a floppy image
1595 - Removed redundant message in instance rename when the name is given
1597 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1598 disks on new nodes, allowing recreation when the original instance
1599 nodes are completely gone
1600 - Added option when converting disk templates to DRBD to skip waiting
1601 for the resync, in order to make the instance available sooner
1602 - Added two new variables to the OS scripts environment (containing the
1604 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1605 to allow use of ``pvgrub`` as bootloader
1606 - Changed the instance memory modifications to only check out-of-memory
1607 conditions on memory increases, and turned the secondary node warnings
1608 into errors (they can still be overridden via ``--force``)
1609 - Fixed the handling of a corner case when the Python installation gets
1610 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1611 to execute a command that doesn't exist
1612 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1613 node of the instance returned failures during instance shutdown; this
1614 adds the option ``--ignore-consistency`` to gnt-instance move
1616 And as usual, various improvements to the error messages, documentation
1623 *(Released Wed, 09 Mar 2011)*
1625 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1626 the RAPI users file if run twice (even with ``--dry-run``).
1628 The release fixes that bug (nothing else changed).
1634 *(Released Mon, 07 Mar 2011)*
1636 Final 2.4.0 release. Just a few small fixes:
1638 - Fixed RAPI node evacuate
1639 - Fixed the kvm-ifup script
1640 - Fixed internal error handling for special job cases
1641 - Updated man page to specify the escaping feature for options
1647 *(Released Mon, 28 Feb 2011)*
1649 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1650 that this feature wiped the disks of the wrong instance, leading to loss
1655 - Fixed title of query field containing instance name
1656 - Expanded the glossary in the documentation
1657 - Fixed one unittest (internal issue)
1663 *(Released Mon, 21 Feb 2011)*
1665 A number of bug fixes plus just a couple functionality changes.
1667 On the user-visible side, the ``gnt-* list`` command output has changed
1668 with respect to "special" field states. The current rc1 style of display
1669 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1670 the default output mode special fields states are displayed as follows:
1672 - Offline resource: ``*``
1673 - Unavailable/not applicable: ``-``
1674 - Data missing (RPC failure): ``?``
1675 - Unknown field: ``??``
1677 Another user-visible change is the addition of ``--force-join`` to
1682 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1683 - Fixed regression in RAPI/instance reinstall where all parameters were
1684 required (instead of optional)
1685 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1686 - Fixed iallocator usage (offline nodes were not considered offline)
1687 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1688 - Fixed hypervisor and OS parameter validation with respect to
1689 non-vm_capable nodes
1690 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1692 - Fixed ``tools/listrunner`` with respect to agent-based usage
1698 *(Released Fri, 4 Feb 2011)*
1700 Many changes and fixes since the beta1 release. While there were some
1701 internal changes, the code has been mostly stabilised for the RC
1704 Note: the dumb allocator was removed in this release, as it was not kept
1705 up-to-date with the IAllocator protocol changes. It is recommended to
1706 use the ``hail`` command from the ganeti-htools package.
1708 Note: the 2.4 and up versions of Ganeti are not compatible with the
1709 0.2.x branch of ganeti-htools. You need to upgrade to
1710 ganeti-htools-0.3.0 (or later).
1712 Regressions fixed from 2.3
1713 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1715 - Fixed the ``gnt-cluster verify-disks`` command
1716 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1718 - Fixed disk adoption breakage
1719 - Fixed wrong headers in instance listing for field aliases
1724 - Fixed corner case in KVM handling of NICs
1725 - Fixed many cases of wrong handling of non-vm_capable nodes
1726 - Fixed a bug where a missing instance symlink was not possible to
1727 recreate with any ``gnt-*`` command (now ``gnt-instance
1728 activate-disks`` does it)
1729 - Fixed the volume group name as reported by ``gnt-cluster
1731 - Increased timeouts for the import-export code, hopefully leading to
1732 fewer aborts due network or instance timeouts
1733 - Fixed bug in ``gnt-node list-storage``
1734 - Fixed bug where not all daemons were started on cluster
1735 initialisation, but only at the first watcher run
1736 - Fixed many bugs in the OOB implementation
1737 - Fixed watcher behaviour in presence of instances with offline
1739 - Fixed instance list output for instances running on the wrong node
1740 - a few fixes to the cluster-merge tool, but it still cannot merge
1741 multi-node groups (currently it is not recommended to use this tool)
1747 - Improved network configuration for the KVM hypervisor
1748 - Added e1000 as a supported NIC for Xen-HVM
1749 - Improved the lvmstrap tool to also be able to use partitions, as
1750 opposed to full disks
1751 - Improved speed of disk wiping (the cluster parameter
1752 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1753 of instance creations
1754 - Added documentation for the OS parameters
1755 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1756 hypervisor is not responding
1757 - Added display of blacklisted and hidden OS information in
1758 ``gnt-cluster info``
1759 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1760 NIC and node parameters, which might create problems with currently
1761 invalid (but undetected) configuration files, but prevents validation
1762 failures when unrelated parameters are modified
1763 - Changed cluster initialisation to wait for the master daemon to become
1765 - Expanded the RAPI interface:
1767 - Added config redistribution resource
1768 - Added activation/deactivation of instance disks
1769 - Added export of console information
1771 - Implemented log file reopening on SIGHUP, which allows using
1772 logrotate(8) for the Ganeti log files
1773 - Added a basic OOB helper script as an example
1779 *(Released Fri, 14 Jan 2011)*
1784 - Fixed timezone issues when formatting timestamps
1785 - Added support for node groups, available via ``gnt-group`` and other
1787 - Added out-of-band framework and management, see :doc:`design
1788 document <design-oob>`
1789 - Removed support for roman numbers from ``gnt-node list`` and
1790 ``gnt-instance list``.
1791 - Allowed modification of master network interface via ``gnt-cluster
1792 modify --master-netdev``
1793 - Accept offline secondaries while shutting down instance disks
1794 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1795 - Added support for multiple LVM volume groups
1796 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1798 - Added commands to list available fields:
1800 - ``gnt-node list-fields``
1801 - ``gnt-group list-fields``
1802 - ``gnt-instance list-fields``
1804 - Updated documentation and man pages
1809 - Moved ``rapi_users`` file into separate directory, now named
1810 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1812 - Added new tool for running commands on many machines,
1813 ``tools/ganeti-listrunner``
1814 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1815 improving the ``gnt-instance console`` output
1816 - Allowed customisation of disk index separator at ``configure`` time
1817 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1818 - Added support for non-partitioned md disks in ``lvmstrap``
1819 - Added script to gracefully power off KVM instances
1820 - Split ``utils`` module into smaller parts
1821 - Changed query operations to return more detailed information, e.g.
1822 whether an information is unavailable due to an offline node. To use
1823 this new functionality, the LUXI call ``Query`` must be used. Field
1824 information is now stored by the master daemon and can be retrieved
1825 using ``QueryFields``. Instances, nodes and groups can also be queried
1826 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1827 exposed via RAPI). The following commands make use of this
1828 infrastructure change:
1830 - ``gnt-group list``
1831 - ``gnt-group list-fields``
1833 - ``gnt-node list-fields``
1834 - ``gnt-instance list``
1835 - ``gnt-instance list-fields``
1836 - ``gnt-debug locks``
1841 - New RAPI resources (see :doc:`rapi`):
1845 - ``/2/groups/[group_name]``
1846 - ``/2/groups/[group_name]/assign-nodes``
1847 - ``/2/groups/[group_name]/modify``
1848 - ``/2/groups/[group_name]/rename``
1849 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1853 - Implemented ``no_install`` for instance creation
1854 - Implemented OS parameters for instance reinstallation, allowing
1855 use of special settings on reinstallation (e.g. for preserving data)
1860 - Added IPv6 support in import/export
1861 - Pause DRBD synchronization while wiping disks on instance creation
1862 - Updated unittests and QA scripts
1863 - Improved network parameters passed to KVM
1864 - Converted man pages from docbook to reStructuredText
1870 *(Released Mon, 20 Dec 2010)*
1872 Released version 2.3.1~rc1 without any changes.
1878 *(Released Wed, 1 Dec 2010)*
1880 - impexpd: Disable OpenSSL compression in socat if possible (backport
1881 from master, commit e90739d625b, see :doc:`installation guide
1882 <install-quick>` for details)
1883 - Changed unittest coverage report to exclude test scripts
1884 - Added script to check version format
1890 *(Released Wed, 1 Dec 2010)*
1892 Released version 2.3.0~rc1 without any changes.
1898 *(Released Fri, 19 Nov 2010)*
1900 A number of bugfixes and documentation updates:
1902 - Update ganeti-os-interface documentation
1903 - Fixed a bug related to duplicate MACs or similar items which should be
1905 - Fix breakage in OS state modify
1906 - Reinstall instance: disallow offline secondaries (fixes bug related to
1907 OS changing but reinstall failing)
1908 - plus all the other fixes between 2.2.1 and 2.2.2
1914 *(Released Tue, 2 Nov 2010)*
1916 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1917 - Fixed master failover race with watcher
1918 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1920 - Accept previously stopped instance for export with instance removal
1921 - Simplify and extend the environment variables for instance OS scripts
1922 - Added new node flags, ``master_capable`` and ``vm_capable``
1923 - Added optional instance disk wiping prior during allocation. This is a
1924 cluster-wide option and can be set/modified using
1925 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1926 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1927 :doc:`install-quick`
1928 - Added a new watcher option (``--ignore-pause``)
1929 - Added option to ignore offline node on instance start/stop
1930 (``--ignore-offline``)
1931 - Allow overriding OS parameters with ``gnt-instance reinstall``
1932 - Added ability to change node's secondary IP address using ``gnt-node
1934 - Implemented privilege separation for all daemons except
1935 ``ganeti-noded``, see ``configure`` options
1936 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1938 - Implemented job priorities (see ``ganeti(7)`` manpage)
1939 - Ignore failures while shutting down instances during failover from
1941 - Exit daemon's bootstrap process only once daemon is ready
1942 - Export more information via ``LUInstanceQuery``/remote API
1943 - Improved documentation, QA and unittests
1944 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1945 restart if the file was created or changed
1946 - Added LUXI protocol version sent with each request and response,
1947 allowing detection of server/client mismatches
1948 - Moved the Python scripts among gnt-* and ganeti-* into modules
1949 - Moved all code related to setting up SSH to an external script,
1951 - Infrastructure changes for node group support in future versions
1957 *(Released Fri, 19 Nov 2010)*
1959 A few small bugs fixed, and some improvements to the build system:
1961 - Fix documentation regarding conversion to drbd
1962 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1964 - Fix error handling in node modify with multiple changes
1965 - Allow remote imports without checked names
1971 *(Released Tue, 19 Oct 2010)*
1973 - Disable SSL session ID cache in RPC client
1979 *(Released Thu, 14 Oct 2010)*
1981 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1982 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1984 - Fix problems with interaction between Curl and Python's HTTP server,
1985 resulting in increased speed in many RPC calls
1986 - Improve our release script to prevent breakage with older aclocal and
1993 *(Released Thu, 7 Oct 2010)*
1995 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1996 value retrieved from master
1997 - Added support for blacklisted or hidden OS definitions
1998 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1999 - Added support for -mem-path in KVM hypervisor abstraction layer
2000 - Allow overriding instance parameters in tool for inter-cluster
2001 instance moves (``tools/move-instance``)
2002 - Improved opcode summaries (e.g. in ``gnt-job list``)
2003 - Improve consistency of OS listing by sorting it
2004 - Documentation updates
2010 *(Released Fri, 8 Oct 2010)*
2012 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
2018 *(Released Mon, 4 Oct 2010)*
2020 - Fixed regression in ``gnt-instance rename``
2026 *(Released Wed, 22 Sep 2010)*
2028 - Fixed OS_VARIANT variable for OS scripts
2029 - Fixed cluster tag operations via RAPI
2030 - Made ``setup-ssh`` exit with non-zero code if an error occurred
2031 - Disabled RAPI CA checks in watcher
2037 *(Released Mon, 23 Aug 2010)*
2039 - Support DRBD versions of the format "a.b.c.d"
2041 - Re-introduce support for usage from multiple threads in RAPI client
2042 - Instance renames and modify via RAPI
2043 - Work around race condition between processing and archival in job
2045 - Mark opcodes following failed one as failed, too
2046 - Job field ``lock_status`` was removed due to difficulties making it
2047 work with the changed job queue in Ganeti 2.2; a better way to monitor
2048 locks is expected for a later 2.2.x release
2049 - Fixed dry-run behaviour with many commands
2050 - Support ``ssh-agent`` again when adding nodes
2051 - Many additional bugfixes
2057 *(Released Fri, 30 Jul 2010)*
2059 Important change: the internal RPC mechanism between Ganeti nodes has
2060 changed from using a home-grown http library (based on the Python base
2061 libraries) to use the PycURL library. This requires that PycURL is
2062 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2063 against GnuTLS by default. cURL's support for GnuTLS had known issues
2064 before cURL 7.21.0 and we recommend using the latest cURL release or
2065 linking against OpenSSL. Most other distributions already link PycURL
2066 and cURL against OpenSSL. The command::
2068 python -c 'import pycurl; print pycurl.version'
2070 can be used to determine the libraries PycURL and cURL are linked
2073 Other significant changes:
2075 - Rewrote much of the internals of the job queue, in order to achieve
2076 better parallelism; this decouples job query operations from the job
2077 processing, and it should allow much nicer behaviour of the master
2078 daemon under load, and it also has uncovered some long-standing bugs
2079 related to the job serialisation (now fixed)
2080 - Added a default iallocator setting to the cluster parameters,
2081 eliminating the need to always pass nodes or an iallocator for
2082 operations that require selection of new node(s)
2083 - Added experimental support for the LXC virtualization method
2084 - Added support for OS parameters, which allows the installation of
2085 instances to pass parameter to OS scripts in order to customise the
2087 - Added a hypervisor parameter controlling the migration type (live or
2088 non-live), since hypervisors have various levels of reliability; this
2089 has renamed the 'live' parameter to 'mode'
2090 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
2091 logical volumes, meaning that cluster verify will ignore them and not
2092 flag their presence as errors
2093 - The watcher will now reset the error count for failed instances after
2094 8 hours, thus allowing self-healing if the problem that caused the
2095 instances to be down/fail to start has cleared in the meantime
2096 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2097 check for, and warn, if the drbd module parameter ``usermode_helper``
2098 is not consistent with the cluster-wide setting; this is needed to
2099 make diagnose easier of failed drbd creations
2100 - Started adding base IPv6 support, but this is not yet
2101 enabled/available for use
2102 - Rename operations (cluster, instance) will now return the new name,
2103 which is especially useful if a short name was passed in
2104 - Added support for instance migration in RAPI
2105 - Added a tool to pre-configure nodes for the SSH setup, before joining
2106 them to the cluster; this will allow in the future a simplified model
2107 for node joining (but not yet fully enabled in 2.2); this needs the
2108 paramiko python library
2109 - Fixed handling of name-resolving errors
2110 - Fixed consistency of job results on the error path
2111 - Fixed master-failover race condition when executed multiple times in
2113 - Fixed many bugs related to the job queue (mostly introduced during the
2114 2.2 development cycle, so not all are impacting 2.1)
2115 - Fixed instance migration with missing disk symlinks
2116 - Fixed handling of unknown jobs in ``gnt-job archive``
2117 - And many other small fixes/improvements
2121 - Enhanced both the unittest and the QA coverage
2122 - Switched the opcode validation to a generic model, and extended the
2123 validation to all opcode parameters
2124 - Changed more parts of the code that write shell scripts to use the
2126 - Switched the master daemon to use the asyncore library for the Luxi
2133 *(Released Thu, 17 Jun 2010)*
2135 - Added tool (``move-instance``) and infrastructure to move instances
2136 between separate clusters (see :doc:`separate documentation
2137 <move-instance>` and :doc:`design document <design-2.2>`)
2138 - Added per-request RPC timeout
2139 - RAPI now requires a Content-Type header for requests with a body (e.g.
2140 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2141 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2142 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2144 - Implemented initial support for running Ganeti daemons as separate
2145 users, see configure-time flags ``--with-user-prefix`` and
2146 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2148 - Instances can be removed after export (``gnt-backup export
2149 --remove-instance``)
2150 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2151 key (instead of 1024 bit)
2152 - Added new cluster configuration file for cluster domain secret
2153 - Import/export now use SSL instead of SSH
2154 - Added support for showing estimated time when exporting an instance,
2155 see the ``ganeti-os-interface(7)`` manpage and look for
2162 *(Released Tue, 16 Nov 2010)*
2164 Some more bugfixes. Unless critical bugs occur, this will be the last
2167 - Fix case of MAC special-values
2168 - Fix mac checker regex
2169 - backend: Fix typo causing "out of range" error
2170 - Add missing --units in gnt-instance list man page
2176 *(Released Tue, 24 Aug 2010)*
2179 - Don't ignore secondary node silently on non-mirrored disk templates
2181 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2182 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2183 - Properly document the usb_mouse parameter
2184 - Fix path in ganeti-rapi(8) (issue 116)
2185 - Adjust error message when the ganeti user's .ssh directory is
2187 - Add same-node-check when changing the disk template to drbd
2193 *(Released Fri, 16 Jul 2010)*
2196 - Add an option to only select some reboot types during qa/burnin.
2197 (on some hypervisors consequent reboots are not supported)
2198 - Fix infrequent race condition in master failover. Sometimes the old
2199 master ip address would be still detected as up for a short time
2200 after it was removed, causing failover to fail.
2201 - Decrease mlockall warnings when the ctypes module is missing. On
2202 Python 2.4 we support running even if no ctypes module is installed,
2203 but we were too verbose about this issue.
2204 - Fix building on old distributions, on which man doesn't have a
2206 - Fix RAPI not to ignore the MAC address on instance creation
2207 - Implement the old instance creation format in the RAPI client.
2213 *(Released Thu, 01 Jul 2010)*
2215 A small bugfix release:
2216 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2217 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2218 a non-existing option
2219 - Fix instance create: the --force-variant option was ignored
2220 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2221 - Fix modify node storage with non-FQDN arguments
2222 - Fix RAPI client to authenticate under Python 2.6 when used
2223 for more than 5 requests needing authentication
2224 - Fix gnt-instance modify -t (storage) giving a wrong error message
2225 when converting a non-shutdown drbd instance to plain
2231 *(Released Fri, 18 Jun 2010)*
2233 A small bugfix release:
2235 - Fix live migration of KVM instances started with older Ganeti
2236 versions which had fewer hypervisor parameters
2237 - Fix gnt-instance grow-disk on down instances
2238 - Fix an error-reporting bug during instance migration
2239 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2240 silently ignoring broken ones
2241 - Fix an RPC error reporting bug affecting, for example, RAPI client
2243 - Fix bug triggered by different API version os-es on different nodes
2244 - Fix a bug in instance startup with custom hvparams: OS level
2245 parameters would fail to be applied.
2246 - Fix the RAPI client under Python 2.6 (but more work is needed to
2247 make it work completely well with OpenSSL)
2248 - Fix handling of errors when resolving names from DNS
2254 *(Released Thu, 3 Jun 2010)*
2256 A medium sized development cycle. Some new features, and some
2257 fixes/small improvements/cleanups.
2259 Significant features
2260 ~~~~~~~~~~~~~~~~~~~~
2262 The node deamon now tries to mlock itself into memory, unless the
2263 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2264 its logs, and falls back to console logging. This allows emergency
2265 features such as ``gnt-node powercycle`` to work even in the event of a
2266 broken node disk (tested offlining the disk hosting the node's
2267 filesystem and dropping its memory caches; don't try this at home)
2269 KVM: add vhost-net acceleration support. It can be tested with a new
2270 enough version of the kernel and of qemu-kvm.
2272 KVM: Add instance chrooting feature. If you use privilege dropping for
2273 your VMs you can also now force them to chroot to an empty directory,
2274 before starting the emulated guest.
2276 KVM: Add maximum migration bandwith and maximum downtime tweaking
2277 support (requires a new-enough version of qemu-kvm).
2279 Cluster verify will now warn if the master node doesn't have the master
2280 ip configured on it.
2282 Add a new (incompatible) instance creation request format to RAPI which
2283 supports all parameters (previously only a subset was supported, and it
2284 wasn't possible to extend the old format to accomodate all the new
2285 features. The old format is still supported, and a client can check for
2286 this feature, before using it, by checking for its presence in the
2287 ``features`` RAPI resource.
2289 Now with ancient latin support. Try it passing the ``--roman`` option to
2290 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2291 (requires the python-roman module to be installed, in order to work).
2296 As usual many internal code refactorings, documentation updates, and
2299 - Lots of improvements and cleanups to the experimental Remote API
2300 (RAPI) client library.
2301 - A new unit test suite for the core daemon libraries.
2302 - A fix to creating missing directories makes sure the umask is not
2303 applied anymore. This enforces the same directory permissions
2305 - Better handling terminating daemons with ctrl+c (used when running
2306 them in debugging mode).
2307 - Fix a race condition in live migrating a KVM instance, when stat()
2308 on the old proc status file returned EINVAL, which is an unexpected
2310 - Fixed manpage checking with newer man and utf-8 charachters. But now
2311 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2317 *(Released Fri, 7 May 2010)*
2319 Fix a bug which prevented untagged KVM instances from starting.
2325 *(Released Fri, 7 May 2010)*
2327 Another release with a long development cycle, during which many
2328 different features were added.
2330 Significant features
2331 ~~~~~~~~~~~~~~~~~~~~
2333 The KVM hypervisor now can run the individual instances as non-root, to
2334 reduce the impact of a VM being hijacked due to bugs in the
2335 hypervisor. It is possible to run all instances as a single (non-root)
2336 user, to manually specify a user for each instance, or to dynamically
2337 allocate a user out of a cluster-wide pool to each instance, with the
2338 guarantee that no two instances will run under the same user ID on any
2341 An experimental RAPI client library, that can be used standalone
2342 (without the other Ganeti libraries), is provided in the source tree as
2343 ``lib/rapi/client.py``. Note this client might change its interface in
2344 the future, as we iterate on its capabilities.
2346 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2347 replace the cluster's certificates and crypto keys. This might help in
2348 case they have been compromised, or have simply expired.
2350 A new disk option for instance creation has been added that allows one
2351 to "adopt" currently existing logical volumes, with data
2352 preservation. This should allow easier migration to Ganeti from
2353 unmanaged (or managed via other software) instances.
2355 Another disk improvement is the possibility to convert between redundant
2356 (DRBD) and plain (LVM) disk configuration for an instance. This should
2357 allow better scalability (starting with one node and growing the
2358 cluster, or shrinking a two-node cluster to one node).
2360 A new feature that could help with automated node failovers has been
2361 implemented: if a node sees itself as offline (by querying the master
2362 candidates), it will try to shutdown (hard) all instances and any active
2363 DRBD devices. This reduces the risk of duplicate instances if an
2364 external script automatically failovers the instances on such nodes. To
2365 enable this, the cluster parameter ``maintain_node_health`` should be
2366 enabled; in the future this option (per the name) will enable other
2367 automatic maintenance features.
2369 Instance export/import now will reuse the original instance
2370 specifications for all parameters; that means exporting an instance,
2371 deleting it and the importing it back should give an almost identical
2372 instance. Note that the default import behaviour has changed from
2373 before, where it created only one NIC; now it recreates the original
2376 Cluster verify has added a few new checks: SSL certificates validity,
2377 /etc/hosts consistency across the cluster, etc.
2382 As usual, many internal changes were done, documentation fixes,
2385 - Fixed cluster initialization with disabled cluster storage (regression
2386 introduced in 2.1.1)
2387 - File-based storage supports growing the disks
2388 - Fixed behaviour of node role changes
2389 - Fixed cluster verify for some corner cases, plus a general rewrite of
2390 cluster verify to allow future extension with more checks
2391 - Fixed log spamming by watcher and node daemon (regression introduced
2393 - Fixed possible validation issues when changing the list of enabled
2395 - Fixed cleanup of /etc/hosts during node removal
2396 - Fixed RAPI response for invalid methods
2397 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2398 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2399 from ide disks, etc.)
2400 - Allow OS changes without re-installation (to record a changed OS
2401 outside of Ganeti, or to allow OS renames)
2402 - Allow instance creation without OS installation (useful for example if
2403 the OS will be installed manually, or restored from a backup not in
2405 - Implemented option to make cluster ``copyfile`` use the replication
2407 - Added list of enabled hypervisors to ssconf (possibly useful for
2409 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2411 - A partial form of node re-IP is possible via node readd, which now
2412 allows changed node primary IP
2413 - Command line utilities now show an informational message if the job is
2415 - The logs of the master daemon now show the PID/UID/GID of the
2422 *(Released Fri, 12 Mar 2010)*
2424 During the 2.1.0 long release candidate cycle, a lot of improvements and
2425 changes have accumulated with were released later as 2.1.1.
2430 The node evacuate command (``gnt-node evacuate``) was significantly
2431 rewritten, and as such the IAllocator protocol was changed - a new
2432 request type has been added. This unfortunate change during a stable
2433 series is designed to improve performance of node evacuations; on
2434 clusters with more than about five nodes and which are well-balanced,
2435 evacuation should proceed in parallel for all instances of the node
2436 being evacuated. As such, any existing IAllocator scripts need to be
2437 updated, otherwise the above command will fail due to the unknown
2438 request. The provided "dumb" allocator has not been updated; but the
2439 ganeti-htools package supports the new protocol since version 0.2.4.
2441 Another important change is increased validation of node and instance
2442 names. This might create problems in special cases, if invalid host
2443 names are being used.
2445 Also, a new layer of hypervisor parameters has been added, that sits at
2446 OS level between the cluster defaults and the instance ones. This allows
2447 customisation of virtualization parameters depending on the installed
2448 OS. For example instances with OS 'X' may have a different KVM kernel
2449 (or any other parameter) than the cluster defaults. This is intended to
2450 help managing a multiple OSes on the same cluster, without manual
2451 modification of each instance's parameters.
2453 A tool for merging clusters, ``cluster-merge``, has been added in the
2454 tools sub-directory.
2459 - Improved the int/float conversions that should make the code more
2460 robust in face of errors from the node daemons
2461 - Fixed the remove node code in case of internal configuration errors
2462 - Fixed the node daemon behaviour in face of inconsistent queue
2463 directory (e.g. read-only file-system where we can't open the files
2465 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2466 now it either aborts cleanly or, if given the new "auto_promote"
2467 parameter, will automatically promote other nodes as needed
2468 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2469 completely prevent Ganeti from working
2470 - Fixed bug for instance export when not all disks were successfully
2472 - Fixed behaviour of node add when the new node is slow in starting up
2474 - Fixed handling of signals in the LUXI client, which should improve
2475 behaviour of command-line scripts
2476 - Added checks for invalid node/instance names in the configuration (now
2477 flagged during cluster verify)
2478 - Fixed watcher behaviour for disk activation errors
2479 - Fixed two potentially endless loops in http library, which led to the
2480 RAPI daemon hanging and consuming 100% CPU in some cases
2481 - Fixed bug in RAPI daemon related to hashed passwords
2482 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2484 - Enhanced compatibility with non-Debian OSes, but not using absolute
2485 path in some commands and allowing customisation of the ssh
2486 configuration directory
2487 - Fixed possible future issue with new Python versions by abiding to the
2488 proper use of ``__slots__`` attribute on classes
2489 - Added checks that should prevent directory traversal attacks
2490 - Many documentation fixes based on feedback from users
2495 - Added an "early_release" more for instance replace disks and node
2496 evacuate, where we release locks earlier and thus allow higher
2497 parallelism within the cluster
2498 - Added watcher hooks, intended to allow the watcher to restart other
2499 daemons (e.g. from the ganeti-nbma project), but they can be used of
2500 course for any other purpose
2501 - Added a compile-time disable for DRBD barriers, to increase
2502 performance if the administrator trusts the power supply or the
2503 storage system to not lose writes
2504 - Added the option of using syslog for logging instead of, or in
2505 addition to, Ganeti's own log files
2506 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2507 can indeed boot from a paravirtual NIC
2508 - Added a generic debug level for many operations; while this is not
2509 used widely yet, it allows one to pass the debug value all the way to
2511 - Enhanced the hooks environment for instance moves (failovers,
2512 migrations) where the primary/secondary nodes changed during the
2513 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2514 - Enhanced data validations for many user-supplied values; one important
2515 item is the restrictions imposed on instance and node names, which
2516 might reject some (invalid) host names
2517 - Add a configure-time option to disable file-based storage, if it's not
2518 needed; this allows greater security separation between the master
2519 node and the other nodes from the point of view of the inter-node RPC
2521 - Added user notification in interactive tools if job is waiting in the
2522 job queue or trying to acquire locks
2523 - Added log messages when a job is waiting for locks
2524 - Added filtering by node tags in instance operations which admit
2525 multiple instances (start, stop, reboot, reinstall)
2526 - Added a new tool for cluster mergers, ``cluster-merge``
2527 - Parameters from command line which are of the form ``a=b,c=d`` can now
2528 use backslash escapes to pass in values which contain commas,
2529 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2531 - For KVM, the instance name is the first parameter passed to KVM, so
2532 that it's more visible in the process list
2538 *(Released Tue, 2 Mar 2010)*
2540 Ganeti 2.1 brings many improvements with it. Major changes:
2542 - Added infrastructure to ease automated disk repairs
2543 - Added new daemon to export configuration data in a cheaper way than
2544 using the remote API
2545 - Instance NICs can now be routed instead of being associated with a
2547 - Improved job locking logic to reduce impact of jobs acquiring multiple
2548 locks waiting for other long-running jobs
2550 In-depth implementation details can be found in the Ganeti 2.1 design
2556 - Added chroot hypervisor
2557 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2559 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2560 ``bootloader_path`` and ``bootloader_args``)
2561 - Added the ``use_localtime`` option for the xen-hvm and kvm
2562 hypervisors, and the default value for this has changed to false (in
2563 2.0 xen-hvm always enabled it)
2564 - Added luxi call to submit multiple jobs in one go
2565 - Added cluster initialization option to not modify ``/etc/hosts``
2567 - Added network interface parameters
2568 - Added dry run mode to some LUs
2569 - Added RAPI resources:
2571 - ``/2/instances/[instance_name]/info``
2572 - ``/2/instances/[instance_name]/replace-disks``
2573 - ``/2/nodes/[node_name]/evacuate``
2574 - ``/2/nodes/[node_name]/migrate``
2575 - ``/2/nodes/[node_name]/role``
2576 - ``/2/nodes/[node_name]/storage``
2577 - ``/2/nodes/[node_name]/storage/modify``
2578 - ``/2/nodes/[node_name]/storage/repair``
2580 - Added OpCodes to evacuate or migrate all instances on a node
2581 - Added new command to list storage elements on nodes (``gnt-node
2582 list-storage``) and modify them (``gnt-node modify-storage``)
2583 - Added new ssconf files with master candidate IP address
2584 (``ssconf_master_candidates_ips``), node primary IP address
2585 (``ssconf_node_primary_ips``) and node secondary IP address
2586 (``ssconf_node_secondary_ips``)
2587 - Added ``ganeti-confd`` and a client library to query the Ganeti
2588 configuration via UDP
2589 - Added ability to run hooks after cluster initialization and before
2591 - Added automatic mode for disk replace (``gnt-instance replace-disks
2593 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2594 after catastrophic data-loss
2595 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2597 - Added ``gnt-instance move`` command to move instances
2598 - Added ``gnt-cluster watcher`` command to control watcher
2599 - Added ``gnt-node powercycle`` command to powercycle nodes
2600 - Added new job status field ``lock_status``
2601 - Added parseable error codes to cluster verification (``gnt-cluster
2602 verify --error-codes``) and made output less verbose (use
2603 ``--verbose`` to restore previous behaviour)
2604 - Added UUIDs to the main config entities (cluster, nodes, instances)
2605 - Added support for OS variants
2606 - Added support for hashed passwords in the Ganeti remote API users file
2608 - Added option to specify maximum timeout on instance shutdown
2609 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2610 - Added new helper script to start and stop Ganeti daemons
2611 (``daemon-util``), with the intent to reduce the work necessary to
2612 adjust Ganeti for non-Debian distributions and to start/stop daemons
2614 - Added more unittests
2615 - Fixed critical bug in ganeti-masterd startup
2616 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2617 now customisable at the cluster level for both the KVM and Xen
2618 hypervisors using the new ``migration_port`` parameter
2619 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2620 reinstalling an instance
2621 - Allowed ``@`` in tag names
2622 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2623 - Many documentation updates
2624 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2625 - ``gnt-instance reinstall`` can now reinstall multiple instances
2626 - Updated many command line parameters
2627 - Introduced new OS API version 15
2628 - No longer support a default hypervisor
2629 - Treat virtual LVs as inexistent
2630 - Improved job locking logic to reduce lock contention
2631 - Match instance and node names case insensitively
2632 - Reimplemented bash completion script to be more complete
2639 *(Released Thu, 4 Feb 2010)*
2641 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2642 - Fix a string formatting bug
2643 - Improve safety of the code in some error paths
2644 - Improve data validation in the master of values returned from nodes
2650 *(Released Thu, 17 Dec 2009)*
2652 - Fix security issue due to missing validation of iallocator names; this
2653 allows local and remote execution of arbitrary executables
2654 - Fix failure of gnt-node list during instance removal
2655 - Ship the RAPI documentation in the archive
2661 *(Released Wed, 30 Sep 2009)*
2663 - Fixed many wrong messages
2664 - Fixed a few bugs related to the locking library
2665 - Fixed MAC checking at instance creation time
2666 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2667 - Fixed a few issues related to signal handling in both daemons and
2669 - Fixed the example startup script provided
2670 - Fixed insserv dependencies in the example startup script (patch from
2672 - Fixed handling of drained nodes in the iallocator framework
2673 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2675 - Fixed error related to invalid job IDs in job polling
2676 - Fixed job/opcode persistence on unclean master shutdown
2677 - Fixed handling of partial job processing after unclean master
2679 - Fixed error reporting from LUs, previously all errors were converted
2680 into execution errors
2681 - Fixed error reporting from burnin
2682 - Decreased significantly the memory usage of the job queue
2683 - Optimised slightly multi-job submission
2684 - Optimised slightly opcode loading
2685 - Backported the multi-job submit framework from the development
2686 branch; multi-instance start and stop should be faster
2687 - Added script to clean archived jobs after 21 days; this will reduce
2688 the size of the queue directory
2689 - Added some extra checks in disk size tracking
2690 - Added an example ethers hook script
2691 - Added a cluster parameter that prevents Ganeti from modifying of
2693 - Added more node information to RAPI responses
2694 - Added a ``gnt-job watch`` command that allows following the ouput of a
2696 - Added a bind-address option to ganeti-rapi
2697 - Added more checks to the configuration verify
2698 - Enhanced the burnin script such that some operations can be retried
2700 - Converted instance reinstall to multi-instance model
2706 *(Released Fri, 7 Aug 2009)*
2708 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2709 to allow using the pre-2.0.2 behaviour in activation, if any existing
2710 instances have mismatched disk sizes in the configuration
2711 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2712 any configuration mismatches for disk sizes
2713 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2714 failover to work on two-node clusters
2715 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2717 - Fixed detection of OS script errors in ``gnt-backup export``
2718 - Fixed exit code of ``gnt-backup export``
2724 *(Released Fri, 17 Jul 2009)*
2726 - Added experimental support for stripped logical volumes; this should
2727 enhance performance but comes with a higher complexity in the block
2728 device handling; stripping is only enabled when passing
2729 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2730 affected even in the non-stripped mode
2731 - Improved resiliency against transient failures at the end of DRBD
2732 resyncs, and in general of DRBD resync checks
2733 - Fixed a couple of issues with exports and snapshot errors
2734 - Fixed a couple of issues in instance listing
2735 - Added display of the disk size in ``gnt-instance info``
2736 - Fixed checking for valid OSes in instance creation
2737 - Fixed handling of the "vcpus" parameter in instance listing and in
2738 general of invalid parameters
2739 - Fixed http server library, and thus RAPI, to handle invalid
2740 username/password combinations correctly; this means that now they
2741 report unauthorized for queries too, not only for modifications,
2742 allowing earlier detect of configuration problems
2743 - Added a new "role" node list field, equivalent to the master/master
2744 candidate/drained/offline flags combinations
2745 - Fixed cluster modify and changes of candidate pool size
2746 - Fixed cluster verify error messages for wrong files on regular nodes
2747 - Fixed a couple of issues with node demotion from master candidate role
2748 - Fixed node readd issues
2749 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2750 - Added a new ``--no-voting`` option for masterfailover to fix failover
2751 on two-nodes clusters when the former master node is unreachable
2752 - Added instance reinstall over RAPI
2758 *(Released Tue, 16 Jun 2009)*
2760 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2761 allow re-adding the start in single-user option (regression from 1.2)
2762 - the watcher writes the instance status to a file, to allow monitoring
2763 to report the instance status (from the master) based on cached
2764 results of the watcher's queries; while this can get stale if the
2765 watcher is being locked due to other work on the cluster, this is
2766 still an improvement
2767 - the watcher now also restarts the node daemon and the rapi daemon if
2769 - fixed the watcher to handle full and drained queue cases
2770 - hooks export more instance data in the environment, which helps if
2771 hook scripts need to take action based on the instance's properties
2772 (no longer need to query back into ganeti)
2773 - instance failovers when the instance is stopped do not check for free
2774 RAM, so that failing over a stopped instance is possible in low memory
2776 - rapi uses queries for tags instead of jobs (for less job traffic), and
2777 for cluster tags it won't talk to masterd at all but read them from
2779 - a couple of error handling fixes in RAPI
2780 - drbd handling: improved the error handling of inconsistent disks after
2781 resync to reduce the frequency of "there are some degraded disks for
2782 this instance" messages
2783 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2784 error handling path called a wrong function name)
2790 *(Released Wed, 27 May 2009)*
2792 - no changes from rc5
2798 *(Released Wed, 20 May 2009)*
2800 - fix a couple of bugs (validation, argument checks)
2801 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2802 - some small improvements to RAPI and IAllocator
2803 - make watcher automatically start the master daemon if down
2809 *(Released Mon, 27 Apr 2009)*
2811 - change the OS list to not require locks; this helps with big clusters
2812 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2813 volume group is broken
2814 - ``gnt-instance info``, without any arguments, doesn't run for all
2815 instances anymore; either pass ``--all`` or pass the desired
2816 instances; this helps against mistakes on big clusters where listing
2817 the information for all instances takes a long time
2818 - miscellaneous doc and man pages fixes
2824 *(Released Wed, 8 Apr 2009)*
2826 - Change the internal locking model of some ``gnt-node`` commands, in
2827 order to reduce contention (and blocking of master daemon) when
2828 batching many creation/reinstall jobs
2829 - Fixes to Xen soft reboot
2830 - No longer build documentation at build time, instead distribute it in
2831 the archive, in order to reduce the need for the whole docbook/rst
2838 *(Released Fri, 27 Mar 2009)*
2840 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2841 - Fix watcher startup sequence, improves the behaviour of busy clusters
2842 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2843 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2844 bind address and other places
2845 - Some documentation fixes and updates
2851 *(Released Mon, 2 Mar 2009)*
2853 - More documentation updates, now all docs should be more-or-less
2855 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2857 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2859 - Fix an issue related to $libdir/run/ganeti and cluster creation
2865 *(Released Thu, 19 Feb 2009)*
2867 - Xen PVM and KVM have switched the default value for the instance root
2868 disk to the first partition on the first drive, instead of the whole
2869 drive; this means that the OS installation scripts must be changed
2871 - Man pages have been updated
2872 - RAPI has been switched by default to HTTPS, and the exported functions
2873 should all work correctly
2874 - RAPI v1 has been removed
2875 - Many improvements to the KVM hypervisor
2876 - Block device errors are now better reported
2877 - Many other bugfixes and small improvements
2883 *(Released Mon, 26 Jan 2009)*
2885 - Version 2 is a general rewrite of the code and therefore the
2886 differences are too many to list, see the design document for 2.0 in
2887 the ``doc/`` subdirectory for more details
2888 - In this beta version there is not yet a migration path from 1.2 (there
2889 will be one in the final 2.0 release)
2890 - A few significant changes are:
2892 - all commands are executed by a daemon (``ganeti-masterd``) and the
2893 various ``gnt-*`` commands are just front-ends to it
2894 - all the commands are entered into, and executed from a job queue,
2895 see the ``gnt-job(8)`` manpage
2896 - the RAPI daemon supports read-write operations, secured by basic
2897 HTTP authentication on top of HTTPS
2898 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2899 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2900 to migrate to DRBD 8 first while still running Ganeti 1.2)
2901 - DRBD devices are using statically allocated minor numbers, which
2902 will be assigned to existing instances during the migration process
2903 - there is support for both Xen PVM and Xen HVM instances running on
2905 - KVM virtualization is supported too
2906 - file-based storage has been implemented, which means that it is
2907 possible to run the cluster without LVM and DRBD storage, for
2908 example using a shared filesystem exported from shared storage (and
2909 still have live migration)
2915 *(Released Tue, 13 Jan 2009)*
2917 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2918 - Reuse the old instance mac address by default on instance import, if
2919 the instance name is the same.
2920 - Handle situations in which the node info rpc returns incomplete
2922 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2923 - Improved version of batcher:
2925 - state file support
2926 - instance mac address support
2927 - support for HVM clusters/instances
2929 - Add an option to show the number of cpu sockets and nodes in
2931 - Support OSes that handle more than one version of the OS api (but do
2932 not change the current API in any other way)
2933 - Fix ``gnt-node migrate``
2934 - ``gnt-debug`` man page
2935 - Fixes various more typos and small issues
2936 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2942 *(Released Wed, 24 Sep 2008)*
2944 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2945 type of disk exported to fully virtualized instances.
2946 - provide access to the serial console of HVM instances
2947 - instance auto_balance flag, set by default. If turned off it will
2948 avoid warnings on cluster verify if there is not enough memory to fail
2949 over an instance. in the future it will prevent automatically failing
2950 it over when we will support that.
2951 - batcher tool for instance creation, see ``tools/README.batcher``
2952 - ``gnt-instance reinstall --select-os`` to interactively select a new
2953 operating system when reinstalling an instance.
2954 - when changing the memory amount on instance modify a check has been
2955 added that the instance will be able to start. also warnings are
2956 emitted if the instance will not be able to fail over, if auto_balance
2958 - documentation fixes
2959 - sync fields between ``gnt-instance list/modify/add/import``
2960 - fix a race condition in drbd when the sync speed was set after giving
2961 the device a remote peer.
2967 *(Released Tue, 22 Jul 2008)*
2969 - note: the allowed size and number of tags per object were reduced
2970 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2971 - fixed twisted 8.x compatibility
2972 - fixed ``gnt-instance replace-disks`` with iallocator
2973 - add TCP keepalives on twisted connections to detect restarted nodes
2974 - disk increase support, see ``gnt-instance grow-disk``
2975 - implement bulk node/instance query for RAPI
2976 - add tags in node/instance listing (optional)
2977 - experimental migration (and live migration) support, read the man page
2978 for ``gnt-instance migrate``
2979 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2980 has some small improvements in handling its state file
2986 *(Released Fri, 13 Jun 2008)*
2988 - Experimental readonly, REST-based remote API implementation;
2989 automatically started on master node, TCP port 5080, if enabled by
2990 ``--enable-rapi`` parameter to configure script.
2991 - Instance allocator support. Add and import instance accept a
2992 ``--iallocator`` parameter, and call that instance allocator to decide
2993 which node to use for the instance. The iallocator document describes
2994 what's expected from an allocator script.
2995 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2996 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2997 node is lost there is still enough memory to fail over the instances
2999 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
3000 ``gnt-cluster verify``, to check for site-specific compliance. All the
3001 hooks will run, and their output, if any, will be displayed. Any
3002 failing hook will make the verification return an error value.
3003 - ``gnt-cluster verify`` now checks that its peers are reachable on the
3004 primary and secondary interfaces
3005 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
3006 that is still declared as part of the cluster and has failed.
3007 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
3008 specifying output fields, that just adds the chosen fields to the
3010 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
3011 export from the filesystem.
3012 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
3013 have been added. Using them you can enable/disable acpi and pae
3014 support, and specify a path for a cd image to be exported to the
3015 instance. These parameters as the name suggest only work on HVM
3017 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
3018 PAE support will be set to the previously hardcoded values, but the
3019 (previously hardcoded) path to the CDROM ISO image will be unset and
3020 if required, needs to be set manually with ``gnt-instance modify``
3022 - The address to which an instance's VNC console is bound is now
3023 selectable per-instance, rather than being cluster wide. Of course
3024 this only applies to instances controlled via VNC, so currently just
3025 applies to HVM clusters.
3031 *(Released Mon, 18 Feb 2008)*
3033 - more tweaks to the disk activation code (especially helpful for DRBD)
3034 - change the default ``gnt-instance list`` output format, now there is
3035 one combined status field (see the manpage for the exact values this
3037 - some more fixes for the mac export to hooks change
3038 - make Ganeti not break with DRBD 8.2.x (which changed the version
3039 format in ``/proc/drbd``) (issue 24)
3040 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3041 template, allowing migration from DRBD0.7+MD to DRBD8
3047 *(Released Wed, 30 Jan 2008)*
3049 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3051 - add command aliases infrastructure and a few aliases
3052 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
3053 man pages and the ``--help`` option of ``gnt-node
3054 list``/``gnt-instance list``
3055 - fix ``gnt-backup list`` with down nodes (issue 21)
3056 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3057 - fix the dist archive and add a check for including svn/git files in
3059 - some developer-related changes: improve the burnin and the QA suite,
3060 add an upload script for testing during development
3066 *(Released Wed, 16 Jan 2008)*
3068 - experimental HVM support, read the install document, section
3069 "Initializing the cluster"
3070 - allow for the PVM hypervisor per-instance kernel and initrd paths
3071 - add a new command ``gnt-cluster verify-disks`` which uses a new
3072 algorithm to improve the reconnection of the DRBD pairs if the device
3073 on the secondary node has gone away
3074 - make logical volume code auto-activate LVs at disk activation time
3075 - slightly improve the speed of activating disks
3076 - allow specification of the MAC address at instance creation time, and
3077 changing it later via ``gnt-instance modify``
3078 - fix handling of external commands that generate lots of output on
3080 - update documentation with regard to minimum version of DRBD8 supported
3086 *(Released Tue, 4 Dec 2007)*
3088 - Log the ``xm create`` output to the node daemon log on failure (to
3089 help diagnosing the error)
3090 - In debug mode, log all external commands output if failed to the logs
3091 - Change parsing of lvm commands to ignore stderr
3097 *(Released Wed, 28 Nov 2007)*
3099 - Another round of updates to the DRBD 8 code to deal with more failures
3100 in the replace secondary node operation
3101 - Some more logging of failures in disk operations (lvm, drbd)
3102 - A few documentation updates
3109 *(Released Tue, 13 Nov 2007)*
3111 - Change configuration file format from Python's Pickle to JSON.
3112 Upgrading is possible using the cfgupgrade utility.
3113 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3114 faster replace disks and is more stable (DRBD 8 has many improvements
3115 compared to DRBD 0.7)
3116 - Added command line tags support (see man pages for ``gnt-instance``,
3117 ``gnt-node``, ``gnt-cluster``)
3118 - Added instance rename support
3119 - Added multi-instance startup/shutdown
3120 - Added cluster rename support
3121 - Added ``gnt-node evacuate`` to simplify some node operations
3122 - Added instance reboot operation that can speedup reboot as compared to
3124 - Soften the requirement that hostnames are in FQDN format
3125 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3127 - Removed dependency on debian's patched fping that uses the
3128 non-standard ``-S`` option
3129 - Now the OS definitions are searched for in multiple, configurable
3130 paths (easier for distros to package)
3131 - Some changes to the hooks infrastructure (especially the new
3132 post-configuration update hook)
3133 - Other small bugfixes
3135 .. vim: set textwidth=72 syntax=rst :