8 *(Released Mon, 27 Jan 2014)*
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)
13 - gnt-instance now honors the KVM path correctly (Issue 691)
15 Inherited from the 2.8 branch:
17 - Change the list separator for the usb_devices parameter from comma to space.
18 Commas could not work because they are already the hypervisor option
20 - Add support for blktap2 file-driver (Issue 638)
21 - Add network tag definitions to the haskell codebase (Issue 641)
22 - Fix RAPI network tag handling
23 - Add the network tags to the tags searched by gnt-cluster search-tags
24 - Fix caching bug preventing jobs from being cancelled
25 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
31 *(Released Fri, 13 Dec 2013)*
33 - use custom KVM path if set for version checking
34 - SingleNotifyPipeCondition: don't share pollers
36 Inherited from the 2.8 branch:
38 - Fixed Luxi daemon socket permissions after master-failover
39 - Improve IP version detection code directly checking for colons rather than
40 passing the family from the cluster object
41 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
42 opportunistically anymore (Issue 622)
43 - Allow link local IPv6 gateways (Issue 624)
44 - Fix error printing (Issue 616)
45 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
46 disk modifications, keep the old one. If name=none then set disk name to
48 - Update build_chroot script to work with the latest hackage packages
49 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
50 - Fix evacuation out of drained node (Issue 615)
51 - Add default file_driver if missing (Issue 571)
52 - Fix job error message after unclean master shutdown (Issue 618)
53 - Lock group(s) when creating instances (Issue 621)
54 - SetDiskID() before accepting an instance (Issue 633)
55 - Allow the ext template disks to receive arbitrary parameters, both at creation
56 time and while being modified
57 - Xen handle domain shutdown (future proofing cherry-pick)
58 - Refactor reading live data in htools (future proofing cherry-pick)
64 *(Released Wed, 13 Nov 2013)*
66 - fix bug, that kept nodes offline when readding
67 - when verifying DRBD versions, ignore unavailable nodes
68 - fix bug that made the console unavailable on kvm in split-user
70 - DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
76 *(Released Tue, 5 Nov 2013)*
78 Incompatible/important changes
79 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
81 - hroller now also plans for capacity to move non-redundant instances off
82 any node to be rebooted; the old behavior of completely ignoring any
83 non-redundant instances can be restored by adding the --ignore-non-redundant
85 - The cluster option '--no-lvm-storage' was removed in favor of the new option
86 '--enabled-disk-templates'.
87 - On instance creation, disk templates no longer need to be specified
88 with '-t'. The default disk template will be taken from the list of
89 enabled disk templates.
90 - The monitoring daemon is now running as root, in order to be able to collect
91 information only available to root (such as the state of Xen instances).
92 - The ConfD client is now IPv6 compatible.
93 - File and shared file storage is no longer dis/enabled at configure time,
94 but using the option '--enabled-disk-templates' at cluster initialization and
96 - The default directories for file and shared file storage are not anymore
97 specified at configure time, but taken from the cluster's configuration.
98 They can be set at cluster initialization and modification with
99 '--file-storage-dir' and '--shared-file-storage-dir'.
100 - Cluster verification now includes stricter checks regarding the
101 default file and shared file storage directories. It now checks that
102 the directories are explicitely allowed in the 'file-storage-paths' file and
103 that the directories exist on all nodes.
104 - The list of allowed disk templates in the instance policy and the list
105 of cluster-wide enabled disk templates is now checked for consistency
106 on cluster or group modification. On cluster initialization, the ipolicy
107 disk templates are ensured to be a subset of the cluster-wide enabled
113 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
114 the correct command syntax. It is possible to use different DRBD versions
115 on different nodes as long as they are compatible to each other. This
116 enables rolling upgrades of DRBD with no downtime. As permanent operation
117 of different DRBD versions within a node group is discouraged,
118 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
119 - New "inst-status-xen" data collector for the monitoring daemon, providing
120 information about the state of the xen instances on the nodes.
121 - New "lv" data collector for the monitoring daemon, collecting data about the
122 logical volumes on the nodes, and pairing them with the name of the instances
124 - New "diskstats" data collector, collecting the data from /proc/diskstats and
125 presenting them over the monitoring daemon interface.
126 - The ConfD client is now IPv6 compatible.
130 The following new dependencies have been added.
134 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
135 for the unit tests (and only used for testing).
139 - ``hslogger`` (http://software.complete.org/hslogger) is now always
140 required, even if confd is not enabled.
145 - Correctly start/stop luxid during gnt-cluster master-failover (inherited
147 - Improved error messsages (inherited from stable-2.8)
153 *(Released Tue, 15 Oct 2013)*
155 The third release candidate in the 2.9 series. Since 2.9.0 rc2:
157 - in implicit configuration upgrade, match ipolicy with enabled disk templates
158 - improved harep documentation (inherited from stable-2.8)
164 *(Released Wed, 9 Oct 2013)*
166 The second release candidate in the 2.9 series. Since 2.9.0 rc1:
168 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
169 at least one DRBD instance.
170 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
177 *(Released Tue, 1 Oct 2013)*
179 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
182 - update of the documentation, in particular installation instructions
183 - merging of LD_* constants into DT_* constants
184 - python style changes to be compatible with newer versions of pylint
190 *(Released Thu, 29 Aug 2013)*
192 This was the first beta release of the 2.9 series. All important changes
193 are listed in the latest 2.9 entry.
199 *(Released Thu, 23 Jan 2014)*
201 - Change the list separator for the usb_devices parameter from comma to space.
202 Commas could not work because they are already the hypervisor option
203 separator (Issue 649)
204 - Add support for blktap2 file-driver (Issue 638)
205 - Add network tag definitions to the haskell codebase (Issue 641)
206 - Fix RAPI network tag handling
207 - Add the network tags to the tags searched by gnt-cluster search-tags
208 - Fix caching bug preventing jobs from being cancelled
209 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
215 *(Released Thu, 12 Dec 2013)*
217 - Fixed Luxi daemon socket permissions after master-failover
218 - Improve IP version detection code directly checking for colons rather than
219 passing the family from the cluster object
220 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
221 opportunistically anymore (Issue 622)
222 - Allow link local IPv6 gateways (Issue 624)
223 - Fix error printing (Issue 616)
224 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
225 disk modifications, keep the old one. If name=none then set disk name to
227 - Update build_chroot script to work with the latest hackage packages
228 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
229 - Fix evacuation out of drained node (Issue 615)
230 - Add default file_driver if missing (Issue 571)
231 - Fix job error message after unclean master shutdown (Issue 618)
232 - Lock group(s) when creating instances (Issue 621)
233 - SetDiskID() before accepting an instance (Issue 633)
234 - Allow the ext template disks to receive arbitrary parameters, both at creation
235 time and while being modified
236 - Xen handle domain shutdown (future proofing cherry-pick)
237 - Refactor reading live data in htools (future proofing cherry-pick)
243 *(Released Thu, 07 Nov 2013)*
245 - DRBD: ensure peers are UpToDate for dual-primary
246 - Improve error message for replace-disks
247 - More dependency checks at configure time
248 - Placate warnings on ganeti.outils_unittest.py
254 *(Released Thu, 17 Oct 2013)*
256 - Correctly start/stop luxid during gnt-cluster master-failover
257 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
258 - Fix path for the job queue serial file
259 - Improved harep man page
260 - Minor documentation improvements
266 *(Released Mon, 30 Sep 2013)*
268 Incompatible/important changes
269 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
271 - Instance policy can contain multiple instance specs, as described in
272 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
273 <design-partitioned>`. As a consequence, it's not possible to partially change
274 or override instance specs. Bounding specs (min and max) can be specified as a
275 whole using the new option ``--ipolicy-bounds-specs``, while standard
276 specs use the new option ``--ipolicy-std-specs``.
277 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
278 gnt-instance is a valid YAML object.
279 - hail now honors network restrictions when allocating nodes. This led to an
280 update of the IAllocator protocol. See the IAllocator documentation for
282 - confd now only answers static configuration request over the network. luxid
283 was extracted, listens on the local LUXI socket and responds to live queries.
284 This allows finer grained permissions if using separate users.
289 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
290 to always require authentication, ``--require-authentication``. It can
291 be specified in ``$sysconfdir/default/ganeti``.
292 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
293 be used to manage the disk templates to be used by instances in the cluster.
294 Initially, it will be set to a list that includes plain, drbd, if they were
295 enabled by specifying a volume group name, and file and sharedfile, if those
296 were enabled at configure time. Additionally, it will include all disk
297 templates that are currently used by instances. The order of disk templates
298 will be based on Ganeti's history of supporting them. In the future, the
299 first entry of the list will be used as a default disk template on instance
301 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
302 configuration back to the previous stable version.
303 - Disk templates in group ipolicy can be restored to the default value.
304 - Initial support for diskless instances and virtual clusters in QA.
305 - More QA and unit tests for instance policies.
306 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
307 describing why the opcode itself was executed.
308 - The monitoring daemon is now available. It allows users to query the cluster
309 for obtaining information about the status of the system. The daemon is only
310 responsible for providing the information over the network: the actual data
311 gathering is performed by data collectors (currently, only the DRBD status
312 collector is available).
313 - In order to help developers work on Ganeti, a new script
314 (``devel/build_chroot``) is provided, for building a chroot that contains all
315 the required development libraries and tools for compiling Ganeti on a Debian
317 - A new tool, ``harep``, for performing self-repair and recreation of instances
318 in Ganeti has been added.
319 - Split queries are enabled for tags, network, exports, cluster info, groups,
321 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
322 It prints the command line to set the current policies, to ease
324 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
325 devices for KVM virtio-net interfaces will get created with VNET_HDR
326 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
327 virtio-net interfaces, which prevents host kernel tainting and log
328 flooding, when dealing with broken or malicious virtio-net drivers.
329 It's set to true by default.
330 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
332 - Support 'viridian' parameter in Xen HVM
333 - Support DSA SSH keys in bootstrap
334 - To simplify the work of packaging frameworks that want to add the needed users
335 and groups in a split-user setup themselves, at build time three files in
336 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
337 the groups to be generated, the ``users`` file contains, one per line, the
338 users to be generated, optionally followed by their primary group, where
339 important. The ``groupmemberships`` file contains, one per line, additional
340 user-group membership relations that need to be established. The syntax of
341 these files will remain stable in all future versions.
346 The following new dependencies have been added:
349 - The ``curl`` library is not optional anymore for compiling the Haskell code.
350 - ``snap-server`` library (if monitoring is enabled).
353 - The minimum Python version needed to run Ganeti is now 2.6.
354 - ``yaml`` library (only for running the QA).
358 - Perform proper cleanup on termination of Haskell daemons
359 - Fix corner-case in handling of remaining retry time
365 *(Released Tue, 17 Sep 2013)*
367 - To simplify the work of packaging frameworks that want to add the needed users
368 and groups in a split-user setup themselves, at build time three files in
369 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
370 the groups to be generated, the ``users`` file contains, one per line, the
371 users to be generated, optionally followed by their primary group, where
372 important. The ``groupmemberships`` file contains, one per line, additional
373 user-group membership relations that need to be established. The syntax of
374 these files will remain stable in all future versions.
375 - Add a default to file-driver when unspecified over RAPI (Issue 571)
376 - Mark the DSA host pubkey as optional, and remove it during config downgrade
378 - Some documentation fixes
384 *(Released Tue, 27 Aug 2013)*
386 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
388 - Support 'viridian' parameter in Xen HVM (Issue 233)
389 - Include VCS version in ``gnt-cluster version``
390 - Support DSA SSH keys in bootstrap (Issue 338)
391 - Fix batch creation of instances
392 - Use FQDN to check master node status (Issue 551)
393 - Make the DRBD collector more failure-resilient
399 *(Released Fri, 2 Aug 2013)*
401 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
403 - Fix upgrading/downgrading from 2.7
404 - Increase maximum RAPI message size
405 - Documentation updates
406 - Split ``confd`` between ``luxid`` and ``confd``
407 - Merge 2.7 series up to the 2.7.1 release
408 - Allow the ``modify_etc_hosts`` option to be changed
409 - Add better debugging for ``luxid`` queries
410 - Expose bulk parameter for GetJobs in RAPI client
411 - Expose missing ``network`` fields in RAPI
412 - Add some ``cluster verify`` tests
413 - Some unittest fixes
414 - Fix a malfunction in ``hspace``'s tiered allocation
415 - Fix query compatibility between haskell and python implementations
416 - Add the ``vnet_hdr`` HV parameter for KVM
417 - Add ``--cleanup`` to instance failover
418 - Change the connected groups format in ``gnt-network info`` output; it
419 was previously displayed as a raw list by mistake. (Merged from 2.7)
425 *(Released Mon, 24 Jun 2013)*
427 This was the first beta release of the 2.8 series. All important changes
428 are listed in the latest 2.8 entry.
434 *(Released Thu, 26 Sep 2013)*
436 - Change the connected groups format in ``gnt-network info`` output; it
437 was previously displayed as a raw list by mistake
438 - Check disk template in right dict when copying
439 - Support multi-instance allocs without iallocator
440 - Fix some errors in the documentation
441 - Fix formatting of tuple in an error message
447 *(Released Thu, 25 Jul 2013)*
449 - Add logrotate functionality in daemon-util
450 - Add logrotate example file
451 - Add missing fields to network queries over rapi
452 - Fix network object timestamps
453 - Add support for querying network timestamps
454 - Fix a typo in the example crontab
455 - Fix a documentation typo
461 *(Released Thu, 04 Jul 2013)*
463 Incompatible/important changes
464 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
466 - Instance policies for disk size were documented to be on a per-disk
467 basis, but hail applied them to the sum of all disks. This has been
469 - ``hbal`` will now exit with status 0 if, during job execution over
470 LUXI, early exit has been requested and all jobs are successful;
471 before, exit status 1 was used, which cannot be differentiated from
473 - Compatibility with newer versions of rbd has been fixed
474 - ``gnt-instance batch-create`` has been changed to use the bulk create
475 opcode from Ganeti. This lead to incompatible changes in the format of
476 the JSON file. It's now not a custom dict anymore but a dict
477 compatible with the ``OpInstanceCreate`` opcode.
478 - Parent directories for file storage need to be listed in
479 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
480 write the file automatically based on old configuration values, but it
481 can not distribute it across all nodes and the file contents should be
482 verified. Use ``gnt-cluster copyfile
483 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
484 upgraded. The reason for requiring this list of paths now is that
485 before it would have been possible to inject new paths via RPC,
486 allowing files to be created in arbitrary locations. The RPC protocol
487 is protected using SSL/X.509 certificates, but as a design principle
488 Ganeti does not permit arbitrary paths to be passed.
489 - The parsing of the variants file for OSes (see
490 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
491 lines and comment lines (starting with ``#``) are ignored for better
493 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
494 longer available. ``gnt-node add`` now invokes a new tool on the
495 destination node, named ``prepare-node-join``, to configure the SSH
496 daemon. Paramiko is no longer necessary to configure nodes' SSH
497 daemons via ``gnt-node add``.
498 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
499 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
500 the flag is not reset after a master failover.
501 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
502 is recommended. 2.6 will be mandatory from the 2.8 series.
508 - New network management functionality to support automatic allocation
509 of IP addresses and managing of network parameters. See
510 :manpage:`gnt-network(8)` for more details.
511 - New external storage backend, to allow managing arbitrary storage
512 systems external to the cluster. See
513 :manpage:`ganeti-extstorage-interface(7)`.
514 - New ``exclusive-storage`` node parameter added, restricted to
515 nodegroup level. When it's set to true, physical disks are assigned in
516 an exclusive fashion to instances, as documented in :doc:`Partitioned
517 Ganeti <design-partitioned>`. Currently, only instances using the
518 ``plain`` disk template are supported.
519 - The KVM hypervisor has been updated with many new hypervisor
520 parameters, including a generic one for passing arbitrary command line
521 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
522 compatible up to qemu 1.4.
523 - A new tool, called ``mon-collector``, is the stand-alone executor of
524 the data collectors for a monitoring system. As of this version, it
525 just includes the DRBD data collector, that can be executed by calling
526 ``mon-collector`` using the ``drbd`` parameter. See
527 :manpage:`mon-collector(7)`.
528 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
529 for RAPI users. It allows granting permissions to query for
530 information to a specific user without giving
531 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
532 - A new tool named ``node-cleanup`` has been added. It cleans remains of
533 a cluster from a machine by stopping all daemons, removing
534 certificates and ssconf files. Unless the ``--no-backup`` option is
535 given, copies of the certificates are made.
536 - Instance creations now support the use of opportunistic locking,
537 potentially speeding up the (parallel) creation of multiple instances.
538 This feature is currently only available via the :doc:`RAPI
539 <rapi>` interface and when an instance allocator is used. If the
540 ``opportunistic_locking`` parameter is set the opcode will try to
541 acquire as many locks as possible, but will not wait for any locks
542 held by other opcodes. If not enough resources can be found to
543 allocate the instance, the temporary error code
544 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
545 retried thereafter, with or without opportunistic locking.
546 - New experimental linux-ha resource scripts.
547 - Restricted-commands support: ganeti can now be asked (via command line
548 or rapi) to perform commands on a node. These are passed via ganeti
549 RPC rather than ssh. This functionality is restricted to commands
550 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
551 reasons. The file is not copied automatically.
557 - Diskless instances are now externally mirrored (Issue 237). This for
558 now has only been tested in conjunction with explicit target nodes for
560 - Queries not needing locks or RPC access to the node can now be
561 performed by the confd daemon, making them independent from jobs, and
562 thus faster to execute. This is selectable at configure time.
563 - The functionality for allocating multiple instances at once has been
564 overhauled and is now also available through :doc:`RAPI <rapi>`.
566 There are no significant changes from version 2.7.0~rc3.
572 *(Released Tue, 25 Jun 2013)*
574 - Fix permissions on the confd query socket (Issue 477)
575 - Fix permissions on the job archive dir (Issue 498)
576 - Fix handling of an internal exception in replace-disks (Issue 472)
577 - Fix gnt-node info handling of shortened names (Issue 497)
578 - Fix gnt-instance grow-disk when wiping is enabled
579 - Documentation improvements, and support for newer pandoc
580 - Fix hspace honoring ipolicy for disks (Issue 484)
581 - Improve handling of the ``kvm_extra`` HV parameter
587 *(Released Fri, 24 May 2013)*
589 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
591 - Disks added through ``gnt-instance modify`` or created through
592 ``gnt-instance recreate-disks`` are wiped, if the
593 ``prealloc_wipe_disks`` flag is set.
594 - If wiping newly created disks fails, the disks are removed. Also,
595 partial failures in creating disks through ``gnt-instance modify``
596 triggers a cleanup of the partially-created disks.
597 - Removing the master IP address doesn't fail if the address has been
599 - Fix ownership of the OS log dir
600 - Workaround missing SO_PEERCRED constant (Issue 191)
606 *(Released Fri, 3 May 2013)*
608 This was the first release candidate of the 2.7 series. Since beta3:
610 - Fix kvm compatibility with qemu 1.4 (Issue 389)
611 - Documentation updates (admin guide, upgrade notes, install
612 instructions) (Issue 372)
613 - Fix gnt-group list nodes and instances count (Issue 436)
614 - Fix compilation without non-mandatory libraries (Issue 441)
615 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
616 - Make confd logging more verbose at INFO level (Issue 435)
617 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
618 - Fix failure path for instance storage type conversion (Issue 229)
619 - Update htools text backend documentation
620 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
621 - Disable inter-cluster instance move for file-based instances, because
622 it is dependant on instance export, which is not supported for
623 file-based instances. (Issue 414)
624 - Fix gnt-job crashes on non-ascii characters (Issue 427)
625 - Fix volume group checks on non-vm-capable nodes (Issue 432)
631 *(Released Mon, 22 Apr 2013)*
633 This was the third beta release of the 2.7 series. Since beta2:
635 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
636 - Fix data loss on wrong usage of ``gnt-instance move``
637 - Properly export errors in confd-based job queries
638 - Add ``users-setup`` tool
639 - Fix iallocator protocol to report 0 as a disk size for diskless
640 instances. This avoids hail breaking when a diskless instance is
642 - Fix job queue directory permission problem that made confd job queries
643 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
644 for access to archived jobs (Issue 406).
645 - Limit the sizes of networks supported by ``gnt-network`` to something
646 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
647 - Fix bugs in instance disk template conversion
648 - Fix GHC 7 compatibility
649 - Fix ``burnin`` install path (Issue 426).
650 - Allow very small disk grows (Issue 347).
651 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
652 that noded doesn't import masterd code (Issue 419).
653 - Make sure the default metavg at cluster init is the same as the vg, if
654 unspecified (Issue 358).
655 - Fix cleanup of partially created disks (part of Issue 416)
661 *(Released Tue, 2 Apr 2013)*
663 This was the second beta release of the 2.7 series. Since beta1:
665 - Networks no longer have a "type" slot, since this information was
666 unused in Ganeti: instead of it tags should be used.
667 - The rapi client now has a ``target_node`` option to MigrateInstance.
668 - Fix early exit return code for hbal (Issue 386).
669 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
670 - Fix ``rbd showmapped`` output parsing (Issue 312).
671 - Networks are now referenced indexed by UUID, rather than name. This
672 will require running cfgupgrade, from 2.7.0beta1, if networks are in
674 - The OS environment now includes network information.
675 - Deleting of a network is now disallowed if any instance nic is using
676 it, to prevent dangling references.
677 - External storage is now documented in man pages.
678 - The exclusive_storage flag can now only be set at nodegroup level.
679 - Hbal can now submit an explicit priority with its jobs.
680 - Many network related locking fixes.
681 - Bump up the required pylint version to 0.25.1.
682 - Fix the ``no_remember`` option in RAPI client.
683 - Many ipolicy related tests, qa, and fixes.
684 - Many documentation improvements and fixes.
685 - Fix building with ``--disable-file-storage``.
686 - Fix ``-q`` option in htools, which was broken if passed more than
688 - Some haskell/python interaction improvements and fixes.
689 - Fix iallocator in case of missing LVM storage.
690 - Fix confd config load in case of ``--no-lvm-storage``.
691 - The confd/query functionality is now mentioned in the security
698 *(Released Wed, 6 Feb 2013)*
700 This was the first beta release of the 2.7 series. All important changes
701 are listed in the latest 2.7 entry.
707 *(Released Fri, 21 Dec 2012)*
709 Important behaviour change: hbal won't rebalance anymore instances which
710 have the ``auto_balance`` attribute set to false. This was the intention
711 all along, but until now it only skipped those from the N+1 memory
712 reservation (DRBD-specific).
714 A significant number of bug fixes in this release:
716 - Fixed disk adoption interaction with ipolicy checks.
717 - Fixed networking issues when instances are started, stopped or
718 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
719 - Fixed the warning in cluster verify for shared storage instances not
721 - Fixed removal of storage directory on shared file storage (issue 262).
722 - Fixed validation of LVM volume group name in OpClusterSetParams
723 (``gnt-cluster modify``) (issue 285).
724 - Fixed runtime memory increases (``gnt-instance modify -m``).
725 - Fixed live migration under Xen's ``xl`` mode.
726 - Fixed ``gnt-instance console`` with ``xl``.
727 - Fixed building with newer Haskell compiler/libraries.
728 - Fixed PID file writing in Haskell daemons (confd); this prevents
729 restart issues if confd was launched manually (outside of
730 ``daemon-util``) while another copy of it was running
731 - Fixed a type error when doing live migrations with KVM (issue 297) and
732 the error messages for failing migrations have been improved.
733 - Fixed opcode validation for the out-of-band commands (``gnt-node
735 - Fixed a type error when unsetting OS hypervisor parameters (issue
736 311); now it's possible to unset all OS-specific hypervisor
738 - Fixed the ``dry-run`` mode for many operations: verification of
739 results was over-zealous but didn't take into account the ``dry-run``
740 operation, resulting in "wrong" failures.
741 - Fixed bash completion in ``gnt-job list`` when the job queue has
742 hundreds of entries; especially with older ``bash`` versions, this
743 results in significant CPU usage.
745 And lastly, a few other improvements have been made:
747 - Added option to force master-failover without voting (issue 282).
748 - Clarified error message on lock conflict (issue 287).
749 - Logging of newly submitted jobs has been improved (issue 290).
750 - Hostname checks have been made uniform between instance rename and
752 - The ``--submit`` option is now supported by ``gnt-debug delay``.
753 - Shutting down the master daemon by sending SIGTERM now stops it from
754 processing jobs waiting for locks; instead, those jobs will be started
755 once again after the master daemon is started the next time (issue
757 - Support for Xen's ``xl`` program has been improved (besides the fixes
759 - Reduced logging noise in the Haskell confd daemon (only show one log
760 entry for each config reload, instead of two).
761 - Several man page updates and typo fixes.
767 *(Released Fri, 12 Oct 2012)*
769 A small bugfix release. Among the bugs fixed:
771 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
772 made the command unusable.
773 - Commands that issue many jobs don't fail anymore just because some jobs
774 take so long that other jobs are archived.
775 - Failures during ``gnt-instance reinstall`` are reflected by the exit
777 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
779 - When ``always_failover`` is set, ``--allow-failover`` is not required
780 in migrate commands anymore.
781 - ``bash_completion`` works even if extglob is disabled.
782 - Fixed bug with locks that made failover for RDB-based instances fail.
783 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
784 a random node instead of one based on the allocator metric.
785 - Support for newer versions of pylint and pep8.
786 - Hail doesn't fail anymore when trying to add an instance of type
787 ``file``, ``sharedfile`` or ``rbd``.
788 - Added new Makefile target to rebuild the whole distribution, so that
789 all files are included.
795 *(Released Fri, 27 Jul 2012)*
798 .. attention:: The ``LUXI`` protocol has been made more consistent
799 regarding its handling of command arguments. This, however, leads to
800 incompatibility issues with previous versions. Please ensure that you
801 restart Ganeti daemons soon after the upgrade, otherwise most
802 ``LUXI`` calls (job submission, setting/resetting the drain flag,
803 pausing/resuming the watcher, cancelling and archiving jobs, querying
804 the cluster configuration) will fail.
813 The current ``admin_up`` field, which used to denote whether an instance
814 should be running or not, has been removed. Instead, ``admin_state`` is
815 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
817 The rational behind this is that an instance being “down” can have
820 - it could be down during a reboot
821 - it could be temporarily be down for a reinstall
822 - or it could be down because it is deprecated and kept just for its
825 The previous Boolean state was making it difficult to do capacity
826 calculations: should Ganeti reserve memory for a down instance? Now, the
827 tri-state field makes it clear:
829 - in ``up`` and ``down`` state, all resources are reserved for the
830 instance, and it can be at any time brought up if it is down
831 - in ``offline`` state, only disk space is reserved for it, but not
834 The field can have an extra use: since the transition between ``up`` and
835 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
836 transition between ``offline`` and ``down`` is done via ``gnt-instance
837 modify``, it is possible to given different rights to users. For
838 example, owners of an instance could be allowed to start/stop it, but
839 not transition it out of the offline state.
841 Instance policies and specs
842 +++++++++++++++++++++++++++
844 In previous Ganeti versions, an instance creation request was not
845 limited on the minimum size and on the maximum size just by the cluster
846 resources. As such, any policy could be implemented only in third-party
847 clients (RAPI clients, or shell wrappers over ``gnt-*``
848 tools). Furthermore, calculating cluster capacity via ``hspace`` again
849 required external input with regards to instance sizes.
851 In order to improve these workflows and to allow for example better
852 per-node group differentiation, we introduced instance specs, which
855 - minimum instance disk size, disk count, memory size, cpu count
856 - maximum values for the above metrics
857 - and “standard” values (used in ``hspace`` to calculate the standard
860 The minimum/maximum values can be also customised at node-group level,
861 for example allowing more powerful hardware to support bigger instance
864 Beside the instance specs, there are a few other settings belonging to
865 the instance policy framework. It is possible now to customise, per
866 cluster and node-group:
868 - the list of allowed disk templates
869 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
870 - the maximum ratio of instance to spindles (see below for more
871 information) for local storage
873 All these together should allow all tools that talk to Ganeti to know
874 what are the ranges of allowed values for instances and the
875 over-subscription that is allowed.
877 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
878 instance configuration, and the physical CPU configuration from the
879 node. For the spindle ratios however, we didn't track before these
880 values, so new parameters have been added:
882 - a new node parameter ``spindle_count``, defaults to 1, customisable at
883 node group or node level
884 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
886 Note that spindles in this context doesn't need to mean actual
887 mechanical hard-drives; it's just a relative number for both the node
888 I/O capacity and instance I/O consumption.
890 Instance migration behaviour
891 ++++++++++++++++++++++++++++
893 While live-migration is in general desirable over failover, it is
894 possible that for some workloads it is actually worse, due to the
895 variable time of the “suspend” phase during live migration.
897 To allow the tools to work consistently over such instances (without
898 having to hard-code instance names), a new backend parameter
899 ``always_failover`` has been added to control the migration/failover
900 behaviour. When set to True, all migration requests for an instance will
901 instead fall-back to failover.
903 Instance memory ballooning
904 ++++++++++++++++++++++++++
906 Initial support for memory ballooning has been added. The memory for an
907 instance is no longer fixed (backend parameter ``memory``), but instead
908 can vary between minimum and maximum values (backend parameters
909 ``minmem`` and ``maxmem``). Currently we only change an instance's
912 - live migrating or failing over and instance and the target node
913 doesn't have enough memory
914 - user requests changing the memory via ``gnt-instance modify
920 In order to control the use of specific CPUs by instance, support for
921 controlling CPU pinning has been added for the Xen, HVM and LXC
922 hypervisors. This is controlled by a new hypervisor parameter
923 ``cpu_mask``; details about possible values for this are in the
924 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
925 VCPU-to-CPU mapping) form will work well only when all nodes in your
926 cluster have the same amount of CPUs.
931 Another area in which Ganeti was not customisable were the parameters
932 used for storage configuration, e.g. how many stripes to use for LVM,
933 DRBD resync configuration, etc.
935 To improve this area, we've added disks parameters, which are
936 customisable at cluster and node group level, and which allow to
937 specify various parameters for disks (DRBD has the most parameters
938 currently), for example:
940 - DRBD resync algorithm and parameters (e.g. speed)
941 - the default VG for meta-data volumes for DRBD
942 - number of stripes for LVM (plain disk template)
945 These parameters can be modified via ``gnt-cluster modify -D …`` and
946 ``gnt-group modify -D …``, and are used at either instance creation (in
947 case of LVM stripes, for example) or at disk “activation” time
950 Rados block device support
951 ++++++++++++++++++++++++++
953 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
954 denoted by the ``rbd`` disk template type. This is considered
955 experimental, feedback is welcome. For details on configuring it, see
956 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
961 The existing master IP functionality works well only in simple setups (a
962 single network shared by all nodes); however, if nodes belong to
963 different networks, then the ``/32`` setup and lack of routing
964 information is not enough.
966 To allow the master IP to function well in more complex cases, the
967 system was reworked as follows:
969 - a master IP netmask setting has been added
970 - the master IP activation/turn-down code was moved from the node daemon
972 - whether to run the Ganeti-supplied master IP script or a user-supplied
973 on is a ``gnt-cluster init`` setting
975 Details about the location of the standard and custom setup scripts are
976 in the man page :manpage:`gnt-cluster(8)`; for information about the
977 setup script protocol, look at the Ganeti-supplied script.
982 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
985 It is now possible to use TLS-protected connections, and when renewing
986 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
987 it is now possible to specify spice or spice CA certificates. Also, it
988 is possible to configure a password for SPICE sessions via the
989 hypervisor parameter ``spice_password_file``.
991 There are also new parameters to control the compression and streaming
992 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
993 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
994 for the spice parameters.
996 Lastly, it is now possible to see the SPICE connection information via
997 ``gnt-instance console``.
1002 A new tool (``tools/ovfconverter``) has been added that supports
1003 conversion between Ganeti and the `Open Virtualization Format
1004 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
1007 This relies on the ``qemu-img`` tool to convert the disk formats, so the
1008 actual compatibility with other virtualization solutions depends on it.
1010 Confd daemon changes
1011 ++++++++++++++++++++
1013 The configuration query daemon (``ganeti-confd``) is now optional, and
1014 has been rewritten in Haskell; whether to use the daemon at all, use the
1015 Python (default) or the Haskell version is selectable at configure time
1016 via the ``--enable-confd`` parameter, which can take one of the
1017 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
1018 daemon will result in a smaller footprint; for larger systems, we
1019 welcome feedback on the Haskell version which might become the default
1022 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
1023 daemon running. The Python version doesn't implement the new call.
1026 User interface changes
1027 ~~~~~~~~~~~~~~~~~~~~~~
1029 We have replaced the ``--disks`` option of ``gnt-instance
1030 replace-disks`` with a more flexible ``--disk`` option, which allows
1031 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
1032 disk size and mode can be changed upon recreation (via ``gnt-instance
1033 recreate-disks``, which accepts the same ``--disk`` option).
1035 As many people are used to a ``show`` command, we have added that as an
1036 alias to ``info`` on all ``gnt-*`` commands.
1038 The ``gnt-instance grow-disk`` command has a new mode in which it can
1039 accept the target size of the disk, instead of the delta; this can be
1040 more safe since two runs in absolute mode will be idempotent, and
1041 sometimes it's also easier to specify the desired size directly.
1043 Also the handling of instances with regard to offline secondaries has
1044 been improved. Instance operations should not fail because one of it's
1045 secondary nodes is offline, even though it's safe to proceed.
1047 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
1048 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1049 minors to instance name.
1054 RAPI coverage has improved, with (for example) new resources for
1055 recreate-disks, node power-cycle, etc.
1060 There is partial support for ``xl`` in the Xen hypervisor; feedback is
1063 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1064 whether to still support Python 2.4 or move to Python 2.6 as minimum
1067 Support for Fedora has been slightly improved; the provided example
1068 init.d script should work better on it and the INSTALL file should
1069 document the needed dependencies.
1074 The deprecated ``QueryLocks`` LUXI request has been removed. Use
1075 ``Query(what=QR_LOCK, ...)`` instead.
1077 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1078 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1079 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1080 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1081 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1083 RAPI client: ``CertificateError`` now derives from
1084 ``GanetiApiError``. This should make it more easy to handle Ganeti
1087 Deprecation warnings due to PyCrypto/paramiko import in
1088 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
1089 make sure to run an up-to-date paramiko version, if you use this tool.
1091 The QA scripts now depend on Python 2.5 or above (the main code base
1092 still works with Python 2.4).
1094 The configuration file (``config.data``) is now written without
1095 indentation for performance reasons; if you want to edit it, it can be
1096 re-formatted via ``tools/fmtjson``.
1098 A number of bugs has been fixed in the cluster merge tool.
1100 ``x509`` certification verification (used in import-export) has been
1101 changed to allow the same clock skew as permitted by the cluster
1102 verification. This will remove some rare but hard to diagnose errors in
1109 *(Released Thu, 19 Jul 2012)*
1111 Very few changes from rc4 to the final release, only bugfixes:
1113 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
1114 instance, fix CDROM booting for KVM instances)
1115 - fixed node group modification of node parameters
1116 - fixed issue in LUClusterVerifyGroup with multi-group clusters
1117 - fixed generation of bash completion to ensure a stable ordering
1124 *(Released Fri, 13 Jul 2012)*
1126 Third release candidate for 2.6. The following changes were done from
1129 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1130 - Fixed an inconsistency in the LUXI protocol with the provided
1131 arguments (NOT backwards compatible)
1132 - Fixed a bug with node groups ipolicy where ``min`` was greater than
1133 the cluster ``std`` value
1134 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1135 easier instance debugging on nodes (requires ``hconfd`` to work)
1141 *(Released Tue, 03 Jul 2012)*
1143 Second release candidate for 2.6. The following changes were done from
1146 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1148 - Fixed a RAPI regression on missing beparams/memory
1149 - Fixed redistribution of files on offline nodes
1150 - Added possibility to run activate-disks even though secondaries are
1151 offline. With this change it relaxes also the strictness on some other
1152 commands which use activate disks internally:
1153 * ``gnt-instance start|reboot|rename|backup|export``
1154 - Made it possible to remove safely an instance if its secondaries are
1156 - Made it possible to reinstall even though secondaries are offline
1162 *(Released Mon, 25 Jun 2012)*
1164 First release candidate for 2.6. The following changes were done from
1167 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1169 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1170 behaviour. It supports now the modify operation
1171 - ``hcheck`` implemented to analyze cluster health and possibility of
1172 improving health by rebalance
1173 - ``hbal`` has been improved in dealing with split instances
1179 *(Released Mon, 11 Jun 2012)*
1181 Second beta release of 2.6. The following changes were done from beta2
1184 - Fixed ``daemon-util`` with non-root user models
1185 - Fixed creation of plain instances with ``--no-wait-for-sync``
1186 - Fix wrong iv_names when running ``cfgupgrade``
1187 - Export more information in RAPI group queries
1188 - Fixed bug when changing instance network interfaces
1189 - Extended burnin to do NIC changes
1190 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1191 - Changed default for DRBD barriers
1192 - Fixed DRBD error reporting for syncer rate
1193 - Verify the options on disk parameters
1195 And of course various fixes to documentation and improved unittests and
1202 *(Released Wed, 23 May 2012)*
1204 First beta release of 2.6. The following changes were done from beta1 to
1207 - integrated patch for distributions without ``start-stop-daemon``
1208 - adapted example init.d script to work on Fedora
1209 - fixed log handling in Haskell daemons
1210 - adapted checks in the watcher for pycurl linked against libnss
1211 - add partial support for ``xl`` instead of ``xm`` for Xen
1212 - fixed a type issue in cluster verification
1213 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1216 Plus integrated fixes from the 2.5 branch:
1218 - fixed ``kvm-ifup`` to use ``/bin/bash``
1219 - fixed parallel build failures
1220 - KVM live migration when using a custom keymap
1226 *(Released Tue, 24 Jul 2012)*
1228 A small bugfix release, with no new features:
1230 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1231 different default shell (e.g. Debian, Ubuntu)
1232 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1233 243 and Debian bug #650664)
1234 - fixed compatibility with KVM versions that don't support multiple boot
1235 devices (fixes Issue 230 and Debian bug #624256)
1237 Additionally, a few fixes were done to the build system (fixed parallel
1238 build failures) and to the unittests (fixed race condition in test for
1239 FileID functions, and the default enable/disable mode for QA test is now
1246 *(Released Fri, 11 May 2012)*
1248 A small bugfix release.
1250 The main issues solved are on the topic of compatibility with newer LVM
1253 - fixed parsing of ``lv_attr`` field
1254 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1255 the ``--force`` flag is needed
1257 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1258 to accept kernel 3.x too (was hardcoded to 2.6.*).
1260 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1261 of hook results and that also made display of other errors suboptimal
1262 was fixed; the code behaves now like 2.4 and earlier.
1264 Another change in 2.5, the cleanup of the OS scripts environment, is too
1265 aggressive: it removed even the ``PATH`` variable, which requires the OS
1266 scripts to *always* need to export it. Since this is a bit too strict,
1267 we now export a minimal PATH, the same that we export for hooks.
1269 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1270 integrated into this release.
1272 Finally, a few other miscellaneous changes were done (no new features,
1273 just small improvements):
1275 - Fix ``gnt-group --help`` display
1276 - Fix hardcoded Xen kernel path
1277 - Fix grow-disk handling of invalid units
1278 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1279 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1285 *(Released Thu, 12 Apr 2012)*
1287 Incompatible/important changes and bugfixes
1288 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1290 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1291 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1292 to match the underlying LUXI interface.
1293 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1294 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1295 not maintain backwards-compatibility as the underlying operation
1296 changed in an incompatible way. The RAPI client can talk to old
1297 servers, but it needs to be told so as the return value changed.
1298 - When creating file-based instances via RAPI, the ``file_driver``
1299 parameter no longer defaults to ``loop`` and must be specified.
1300 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1302 - Support for the undocumented and deprecated RAPI instance creation
1303 request format version 0 has been dropped. Use version 1, supported
1304 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1305 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1306 documentation <install>`.
1307 - The "cluster-verify" hooks are now executed per group by the
1308 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1309 if you just run ``gnt-cluster verify``, which generates one opcode per
1311 - The environment as passed to the OS scripts is cleared, and thus no
1312 environment variables defined in the node daemon's environment will be
1313 inherited by the scripts.
1314 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1316 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1317 support operations involving multiple node groups.
1318 - Offline nodes are ignored when failing over an instance.
1319 - Support for KVM version 1.0, which changed the version reporting format
1321 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1323 - ``Makefile`` is now compatible with Automake 1.11.2
1324 - Includes all bugfixes made in the 2.4 series
1329 - The ganeti-htools project has been merged into the ganeti-core source
1330 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1331 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1332 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1333 GPT-style partition tables (requires `parted
1334 <http://www.gnu.org/s/parted/>`_).
1335 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1336 - Allowed adding tags on instance creation.
1337 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1338 - Allow instances to be started in a paused state, enabling the user to
1339 see the complete console output on boot using the console.
1340 - Added new hypervisor flag to control default reboot behaviour
1341 (``reboot_behavior``).
1342 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1343 - Improved out-of-band management support:
1345 - Added ``gnt-node health`` command reporting the health status of
1347 - Added ``gnt-node power`` command to manage power status of nodes.
1348 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1350 - Instance migration can fall back to failover if instance is not
1352 - Filters can be used when listing nodes, instances, groups and locks;
1353 see :manpage:`ganeti(7)` manpage.
1354 - Added post-execution status as variables to :doc:`hooks <hooks>`
1356 - Instance tags are exported/imported together with the instance.
1357 - When given an explicit job ID, ``gnt-job info`` will work for archived
1359 - Jobs can define dependencies on other jobs (not yet supported via
1360 RAPI or command line, but used by internal commands and usable via
1363 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1366 - Instance failover is now available as a RAPI resource
1367 (``/2/instances/[instance_name]/failover``).
1368 - ``gnt-instance info`` defaults to static information if primary node
1370 - Opcodes have a new ``comment`` attribute.
1371 - Added basic SPICE support to KVM hypervisor.
1372 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1374 Node group improvements
1375 ~~~~~~~~~~~~~~~~~~~~~~~
1377 - ``gnt-cluster verify`` has been modified to check groups separately,
1378 thereby improving performance.
1379 - Node group support has been added to ``gnt-cluster verify-disks``,
1380 which now operates per node group.
1381 - Watcher has been changed to work better with node groups.
1383 - One process and state file per node group.
1384 - Slow watcher in one group doesn't block other group's watcher.
1386 - Added new command, ``gnt-group evacuate``, to move all instances in a
1387 node group to other groups.
1388 - Added ``gnt-instance change-group`` to move an instance to another
1390 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1391 per-group operations.
1392 - Node groups can be tagged.
1393 - Some operations switch from an exclusive to a shared lock as soon as
1395 - Instance's primary and secondary nodes' groups are now available as
1396 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1397 and ``snodes.group.uuid``).
1402 - Numerous updates to documentation and manpages.
1404 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1406 - Some opcode/job results are now also documented, see :doc:`RAPI
1409 - A lockset's internal lock is now also visible in lock monitor.
1410 - Log messages from job queue workers now contain information about the
1411 opcode they're processing.
1412 - ``gnt-instance console`` no longer requires the instance lock.
1413 - A short delay when waiting for job changes reduces the number of LUXI
1414 requests significantly.
1415 - DRBD metadata volumes are overwritten with zeros during disk creation.
1416 - Out-of-band commands no longer acquire the cluster lock in exclusive
1418 - ``devel/upload`` now uses correct permissions for directories.
1424 *(Released Fri, 23 Mar 2012)*
1426 This was the sixth release candidate of the 2.5 series.
1432 *(Released Mon, 9 Jan 2012)*
1434 This was the fifth release candidate of the 2.5 series.
1440 *(Released Thu, 27 Oct 2011)*
1442 This was the fourth release candidate of the 2.5 series.
1448 *(Released Wed, 26 Oct 2011)*
1450 This was the third release candidate of the 2.5 series.
1456 *(Released Tue, 18 Oct 2011)*
1458 This was the second release candidate of the 2.5 series.
1464 *(Released Tue, 4 Oct 2011)*
1466 This was the first release candidate of the 2.5 series.
1472 *(Released Wed, 31 Aug 2011)*
1474 This was the third beta release of the 2.5 series.
1480 *(Released Mon, 22 Aug 2011)*
1482 This was the second beta release of the 2.5 series.
1488 *(Released Fri, 12 Aug 2011)*
1490 This was the first beta release of the 2.5 series.
1496 *(Released Thu, 27 Oct 2011)*
1498 - Fixed bug when parsing command line parameter values ending in
1500 - Fixed assertion error after unclean master shutdown
1501 - Disable HTTP client pool for RPC, significantly reducing memory usage
1503 - Fixed queue archive creation with wrong permissions
1509 *(Released Tue, 23 Aug 2011)*
1513 - Fixed documentation for importing with ``--src-dir`` option
1514 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1515 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1521 *(Released Fri, 5 Aug 2011)*
1523 Many bug-fixes and a few small features:
1525 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1526 issues when you tried to add an instance with two MAC addresses in one
1528 - KVM: fixed per-instance stored UID value
1529 - KVM: configure bridged NICs at migration start
1530 - KVM: Fix a bug where instance will not start with never KVM versions
1532 - Added OS search path to ``gnt-cluster info``
1533 - Fixed an issue with ``file_storage_dir`` where you were forced to
1534 provide an absolute path, but the documentation states it is a
1535 relative path, the documentation was right
1536 - Added a new parameter to instance stop/start called ``--no-remember``
1537 that will make the state change to not be remembered
1538 - Implemented ``no_remember`` at RAPI level
1539 - Improved the documentation
1540 - Node evacuation: don't call IAllocator if node is already empty
1541 - Fixed bug in DRBD8 replace disks on current nodes
1542 - Fixed bug in recreate-disks for DRBD instances
1543 - Moved assertion checking locks in ``gnt-instance replace-disks``
1544 causing it to abort with not owning the right locks for some situation
1545 - Job queue: Fixed potential race condition when cancelling queued jobs
1546 - Fixed off-by-one bug in job serial generation
1547 - ``gnt-node volumes``: Fix instance names
1548 - Fixed aliases in bash completion
1549 - Fixed a bug in reopening log files after being sent a SIGHUP
1550 - Added a flag to burnin to allow specifying VCPU count
1551 - Bugfixes to non-root Ganeti configuration
1557 *(Released Thu, 12 May 2011)*
1559 Many bug-fixes and a few new small features:
1561 - Fixed a bug related to log opening failures
1562 - Fixed a bug in instance listing with orphan instances
1563 - Fixed a bug which prevented resetting the cluster-level node parameter
1564 ``oob_program`` to the default
1565 - Many fixes related to the ``cluster-merge`` tool
1566 - Fixed a race condition in the lock monitor, which caused failures
1567 during (at least) creation of many instances in parallel
1568 - Improved output for gnt-job info
1569 - Removed the quiet flag on some ssh calls which prevented debugging
1571 - Improved the N+1 failure messages in cluster verify by actually
1572 showing the memory values (needed and available)
1573 - Increased lock attempt timeouts so that when executing long operations
1574 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1575 too early and thus prevent the use of the 'fair' mechanism
1576 - Changed instance query data (``gnt-instance info``) to not acquire
1577 locks unless needed, thus allowing its use on locked instance if only
1578 static information is asked for
1579 - Improved behaviour with filesystems that do not support rename on an
1581 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1582 kept locks on all nodes during the wipe, which is unneeded
1583 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1584 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1585 10GiB) which caused the wipe to fail right at the end in some cases
1586 - Fixed master IP activation when doing master failover with no-voting
1587 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1588 the master node itself
1589 - Fixed potential data-loss in under disk full conditions, where Ganeti
1590 wouldn't check correctly the return code and would consider
1591 partially-written files 'correct'
1592 - Fixed bug related to multiple VGs and DRBD disk replacing
1593 - Added new disk parameter ``metavg`` that allows placement of the meta
1594 device for DRBD in a different volume group
1595 - Fixed error handling in the node daemon when the system libc doesn't
1596 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1597 - Fixed lock release during replace-disks, which kept cluster-wide locks
1598 when doing disk replaces with an iallocator script
1599 - Added check for missing bridges in cluster verify
1600 - Handle EPIPE errors while writing to the terminal better, so that
1601 piping the output to e.g. ``less`` doesn't cause a backtrace
1602 - Fixed rare case where a ^C during Luxi calls could have been
1603 interpreted as server errors, instead of simply terminating
1604 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1606 - Added a few more parameters to the KVM hypervisor, allowing a second
1607 CDROM, custom disk type for CDROMs and a floppy image
1608 - Removed redundant message in instance rename when the name is given
1610 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1611 disks on new nodes, allowing recreation when the original instance
1612 nodes are completely gone
1613 - Added option when converting disk templates to DRBD to skip waiting
1614 for the resync, in order to make the instance available sooner
1615 - Added two new variables to the OS scripts environment (containing the
1617 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1618 to allow use of ``pvgrub`` as bootloader
1619 - Changed the instance memory modifications to only check out-of-memory
1620 conditions on memory increases, and turned the secondary node warnings
1621 into errors (they can still be overridden via ``--force``)
1622 - Fixed the handling of a corner case when the Python installation gets
1623 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1624 to execute a command that doesn't exist
1625 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1626 node of the instance returned failures during instance shutdown; this
1627 adds the option ``--ignore-consistency`` to gnt-instance move
1629 And as usual, various improvements to the error messages, documentation
1636 *(Released Wed, 09 Mar 2011)*
1638 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1639 the RAPI users file if run twice (even with ``--dry-run``).
1641 The release fixes that bug (nothing else changed).
1647 *(Released Mon, 07 Mar 2011)*
1649 Final 2.4.0 release. Just a few small fixes:
1651 - Fixed RAPI node evacuate
1652 - Fixed the kvm-ifup script
1653 - Fixed internal error handling for special job cases
1654 - Updated man page to specify the escaping feature for options
1660 *(Released Mon, 28 Feb 2011)*
1662 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1663 that this feature wiped the disks of the wrong instance, leading to loss
1668 - Fixed title of query field containing instance name
1669 - Expanded the glossary in the documentation
1670 - Fixed one unittest (internal issue)
1676 *(Released Mon, 21 Feb 2011)*
1678 A number of bug fixes plus just a couple functionality changes.
1680 On the user-visible side, the ``gnt-* list`` command output has changed
1681 with respect to "special" field states. The current rc1 style of display
1682 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1683 the default output mode special fields states are displayed as follows:
1685 - Offline resource: ``*``
1686 - Unavailable/not applicable: ``-``
1687 - Data missing (RPC failure): ``?``
1688 - Unknown field: ``??``
1690 Another user-visible change is the addition of ``--force-join`` to
1695 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1696 - Fixed regression in RAPI/instance reinstall where all parameters were
1697 required (instead of optional)
1698 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1699 - Fixed iallocator usage (offline nodes were not considered offline)
1700 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1701 - Fixed hypervisor and OS parameter validation with respect to
1702 non-vm_capable nodes
1703 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1705 - Fixed ``tools/listrunner`` with respect to agent-based usage
1711 *(Released Fri, 4 Feb 2011)*
1713 Many changes and fixes since the beta1 release. While there were some
1714 internal changes, the code has been mostly stabilised for the RC
1717 Note: the dumb allocator was removed in this release, as it was not kept
1718 up-to-date with the IAllocator protocol changes. It is recommended to
1719 use the ``hail`` command from the ganeti-htools package.
1721 Note: the 2.4 and up versions of Ganeti are not compatible with the
1722 0.2.x branch of ganeti-htools. You need to upgrade to
1723 ganeti-htools-0.3.0 (or later).
1725 Regressions fixed from 2.3
1726 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1728 - Fixed the ``gnt-cluster verify-disks`` command
1729 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1731 - Fixed disk adoption breakage
1732 - Fixed wrong headers in instance listing for field aliases
1737 - Fixed corner case in KVM handling of NICs
1738 - Fixed many cases of wrong handling of non-vm_capable nodes
1739 - Fixed a bug where a missing instance symlink was not possible to
1740 recreate with any ``gnt-*`` command (now ``gnt-instance
1741 activate-disks`` does it)
1742 - Fixed the volume group name as reported by ``gnt-cluster
1744 - Increased timeouts for the import-export code, hopefully leading to
1745 fewer aborts due network or instance timeouts
1746 - Fixed bug in ``gnt-node list-storage``
1747 - Fixed bug where not all daemons were started on cluster
1748 initialisation, but only at the first watcher run
1749 - Fixed many bugs in the OOB implementation
1750 - Fixed watcher behaviour in presence of instances with offline
1752 - Fixed instance list output for instances running on the wrong node
1753 - a few fixes to the cluster-merge tool, but it still cannot merge
1754 multi-node groups (currently it is not recommended to use this tool)
1760 - Improved network configuration for the KVM hypervisor
1761 - Added e1000 as a supported NIC for Xen-HVM
1762 - Improved the lvmstrap tool to also be able to use partitions, as
1763 opposed to full disks
1764 - Improved speed of disk wiping (the cluster parameter
1765 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1766 of instance creations
1767 - Added documentation for the OS parameters
1768 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1769 hypervisor is not responding
1770 - Added display of blacklisted and hidden OS information in
1771 ``gnt-cluster info``
1772 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1773 NIC and node parameters, which might create problems with currently
1774 invalid (but undetected) configuration files, but prevents validation
1775 failures when unrelated parameters are modified
1776 - Changed cluster initialisation to wait for the master daemon to become
1778 - Expanded the RAPI interface:
1780 - Added config redistribution resource
1781 - Added activation/deactivation of instance disks
1782 - Added export of console information
1784 - Implemented log file reopening on SIGHUP, which allows using
1785 logrotate(8) for the Ganeti log files
1786 - Added a basic OOB helper script as an example
1792 *(Released Fri, 14 Jan 2011)*
1797 - Fixed timezone issues when formatting timestamps
1798 - Added support for node groups, available via ``gnt-group`` and other
1800 - Added out-of-band framework and management, see :doc:`design
1801 document <design-oob>`
1802 - Removed support for roman numbers from ``gnt-node list`` and
1803 ``gnt-instance list``.
1804 - Allowed modification of master network interface via ``gnt-cluster
1805 modify --master-netdev``
1806 - Accept offline secondaries while shutting down instance disks
1807 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1808 - Added support for multiple LVM volume groups
1809 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1811 - Added commands to list available fields:
1813 - ``gnt-node list-fields``
1814 - ``gnt-group list-fields``
1815 - ``gnt-instance list-fields``
1817 - Updated documentation and man pages
1822 - Moved ``rapi_users`` file into separate directory, now named
1823 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1825 - Added new tool for running commands on many machines,
1826 ``tools/ganeti-listrunner``
1827 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1828 improving the ``gnt-instance console`` output
1829 - Allowed customisation of disk index separator at ``configure`` time
1830 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1831 - Added support for non-partitioned md disks in ``lvmstrap``
1832 - Added script to gracefully power off KVM instances
1833 - Split ``utils`` module into smaller parts
1834 - Changed query operations to return more detailed information, e.g.
1835 whether an information is unavailable due to an offline node. To use
1836 this new functionality, the LUXI call ``Query`` must be used. Field
1837 information is now stored by the master daemon and can be retrieved
1838 using ``QueryFields``. Instances, nodes and groups can also be queried
1839 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1840 exposed via RAPI). The following commands make use of this
1841 infrastructure change:
1843 - ``gnt-group list``
1844 - ``gnt-group list-fields``
1846 - ``gnt-node list-fields``
1847 - ``gnt-instance list``
1848 - ``gnt-instance list-fields``
1849 - ``gnt-debug locks``
1854 - New RAPI resources (see :doc:`rapi`):
1858 - ``/2/groups/[group_name]``
1859 - ``/2/groups/[group_name]/assign-nodes``
1860 - ``/2/groups/[group_name]/modify``
1861 - ``/2/groups/[group_name]/rename``
1862 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1866 - Implemented ``no_install`` for instance creation
1867 - Implemented OS parameters for instance reinstallation, allowing
1868 use of special settings on reinstallation (e.g. for preserving data)
1873 - Added IPv6 support in import/export
1874 - Pause DRBD synchronization while wiping disks on instance creation
1875 - Updated unittests and QA scripts
1876 - Improved network parameters passed to KVM
1877 - Converted man pages from docbook to reStructuredText
1883 *(Released Mon, 20 Dec 2010)*
1885 Released version 2.3.1~rc1 without any changes.
1891 *(Released Wed, 1 Dec 2010)*
1893 - impexpd: Disable OpenSSL compression in socat if possible (backport
1894 from master, commit e90739d625b, see :doc:`installation guide
1895 <install-quick>` for details)
1896 - Changed unittest coverage report to exclude test scripts
1897 - Added script to check version format
1903 *(Released Wed, 1 Dec 2010)*
1905 Released version 2.3.0~rc1 without any changes.
1911 *(Released Fri, 19 Nov 2010)*
1913 A number of bugfixes and documentation updates:
1915 - Update ganeti-os-interface documentation
1916 - Fixed a bug related to duplicate MACs or similar items which should be
1918 - Fix breakage in OS state modify
1919 - Reinstall instance: disallow offline secondaries (fixes bug related to
1920 OS changing but reinstall failing)
1921 - plus all the other fixes between 2.2.1 and 2.2.2
1927 *(Released Tue, 2 Nov 2010)*
1929 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1930 - Fixed master failover race with watcher
1931 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1933 - Accept previously stopped instance for export with instance removal
1934 - Simplify and extend the environment variables for instance OS scripts
1935 - Added new node flags, ``master_capable`` and ``vm_capable``
1936 - Added optional instance disk wiping prior during allocation. This is a
1937 cluster-wide option and can be set/modified using
1938 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1939 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1940 :doc:`install-quick`
1941 - Added a new watcher option (``--ignore-pause``)
1942 - Added option to ignore offline node on instance start/stop
1943 (``--ignore-offline``)
1944 - Allow overriding OS parameters with ``gnt-instance reinstall``
1945 - Added ability to change node's secondary IP address using ``gnt-node
1947 - Implemented privilege separation for all daemons except
1948 ``ganeti-noded``, see ``configure`` options
1949 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1951 - Implemented job priorities (see ``ganeti(7)`` manpage)
1952 - Ignore failures while shutting down instances during failover from
1954 - Exit daemon's bootstrap process only once daemon is ready
1955 - Export more information via ``LUInstanceQuery``/remote API
1956 - Improved documentation, QA and unittests
1957 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1958 restart if the file was created or changed
1959 - Added LUXI protocol version sent with each request and response,
1960 allowing detection of server/client mismatches
1961 - Moved the Python scripts among gnt-* and ganeti-* into modules
1962 - Moved all code related to setting up SSH to an external script,
1964 - Infrastructure changes for node group support in future versions
1970 *(Released Fri, 19 Nov 2010)*
1972 A few small bugs fixed, and some improvements to the build system:
1974 - Fix documentation regarding conversion to drbd
1975 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1977 - Fix error handling in node modify with multiple changes
1978 - Allow remote imports without checked names
1984 *(Released Tue, 19 Oct 2010)*
1986 - Disable SSL session ID cache in RPC client
1992 *(Released Thu, 14 Oct 2010)*
1994 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1995 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1997 - Fix problems with interaction between Curl and Python's HTTP server,
1998 resulting in increased speed in many RPC calls
1999 - Improve our release script to prevent breakage with older aclocal and
2006 *(Released Thu, 7 Oct 2010)*
2008 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
2009 value retrieved from master
2010 - Added support for blacklisted or hidden OS definitions
2011 - Added simple lock monitor (accessible via (``gnt-debug locks``)
2012 - Added support for -mem-path in KVM hypervisor abstraction layer
2013 - Allow overriding instance parameters in tool for inter-cluster
2014 instance moves (``tools/move-instance``)
2015 - Improved opcode summaries (e.g. in ``gnt-job list``)
2016 - Improve consistency of OS listing by sorting it
2017 - Documentation updates
2023 *(Released Fri, 8 Oct 2010)*
2025 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
2031 *(Released Mon, 4 Oct 2010)*
2033 - Fixed regression in ``gnt-instance rename``
2039 *(Released Wed, 22 Sep 2010)*
2041 - Fixed OS_VARIANT variable for OS scripts
2042 - Fixed cluster tag operations via RAPI
2043 - Made ``setup-ssh`` exit with non-zero code if an error occurred
2044 - Disabled RAPI CA checks in watcher
2050 *(Released Mon, 23 Aug 2010)*
2052 - Support DRBD versions of the format "a.b.c.d"
2054 - Re-introduce support for usage from multiple threads in RAPI client
2055 - Instance renames and modify via RAPI
2056 - Work around race condition between processing and archival in job
2058 - Mark opcodes following failed one as failed, too
2059 - Job field ``lock_status`` was removed due to difficulties making it
2060 work with the changed job queue in Ganeti 2.2; a better way to monitor
2061 locks is expected for a later 2.2.x release
2062 - Fixed dry-run behaviour with many commands
2063 - Support ``ssh-agent`` again when adding nodes
2064 - Many additional bugfixes
2070 *(Released Fri, 30 Jul 2010)*
2072 Important change: the internal RPC mechanism between Ganeti nodes has
2073 changed from using a home-grown http library (based on the Python base
2074 libraries) to use the PycURL library. This requires that PycURL is
2075 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2076 against GnuTLS by default. cURL's support for GnuTLS had known issues
2077 before cURL 7.21.0 and we recommend using the latest cURL release or
2078 linking against OpenSSL. Most other distributions already link PycURL
2079 and cURL against OpenSSL. The command::
2081 python -c 'import pycurl; print pycurl.version'
2083 can be used to determine the libraries PycURL and cURL are linked
2086 Other significant changes:
2088 - Rewrote much of the internals of the job queue, in order to achieve
2089 better parallelism; this decouples job query operations from the job
2090 processing, and it should allow much nicer behaviour of the master
2091 daemon under load, and it also has uncovered some long-standing bugs
2092 related to the job serialisation (now fixed)
2093 - Added a default iallocator setting to the cluster parameters,
2094 eliminating the need to always pass nodes or an iallocator for
2095 operations that require selection of new node(s)
2096 - Added experimental support for the LXC virtualization method
2097 - Added support for OS parameters, which allows the installation of
2098 instances to pass parameter to OS scripts in order to customise the
2100 - Added a hypervisor parameter controlling the migration type (live or
2101 non-live), since hypervisors have various levels of reliability; this
2102 has renamed the 'live' parameter to 'mode'
2103 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
2104 logical volumes, meaning that cluster verify will ignore them and not
2105 flag their presence as errors
2106 - The watcher will now reset the error count for failed instances after
2107 8 hours, thus allowing self-healing if the problem that caused the
2108 instances to be down/fail to start has cleared in the meantime
2109 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2110 check for, and warn, if the drbd module parameter ``usermode_helper``
2111 is not consistent with the cluster-wide setting; this is needed to
2112 make diagnose easier of failed drbd creations
2113 - Started adding base IPv6 support, but this is not yet
2114 enabled/available for use
2115 - Rename operations (cluster, instance) will now return the new name,
2116 which is especially useful if a short name was passed in
2117 - Added support for instance migration in RAPI
2118 - Added a tool to pre-configure nodes for the SSH setup, before joining
2119 them to the cluster; this will allow in the future a simplified model
2120 for node joining (but not yet fully enabled in 2.2); this needs the
2121 paramiko python library
2122 - Fixed handling of name-resolving errors
2123 - Fixed consistency of job results on the error path
2124 - Fixed master-failover race condition when executed multiple times in
2126 - Fixed many bugs related to the job queue (mostly introduced during the
2127 2.2 development cycle, so not all are impacting 2.1)
2128 - Fixed instance migration with missing disk symlinks
2129 - Fixed handling of unknown jobs in ``gnt-job archive``
2130 - And many other small fixes/improvements
2134 - Enhanced both the unittest and the QA coverage
2135 - Switched the opcode validation to a generic model, and extended the
2136 validation to all opcode parameters
2137 - Changed more parts of the code that write shell scripts to use the
2139 - Switched the master daemon to use the asyncore library for the Luxi
2146 *(Released Thu, 17 Jun 2010)*
2148 - Added tool (``move-instance``) and infrastructure to move instances
2149 between separate clusters (see :doc:`separate documentation
2150 <move-instance>` and :doc:`design document <design-2.2>`)
2151 - Added per-request RPC timeout
2152 - RAPI now requires a Content-Type header for requests with a body (e.g.
2153 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2154 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2155 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2157 - Implemented initial support for running Ganeti daemons as separate
2158 users, see configure-time flags ``--with-user-prefix`` and
2159 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2161 - Instances can be removed after export (``gnt-backup export
2162 --remove-instance``)
2163 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2164 key (instead of 1024 bit)
2165 - Added new cluster configuration file for cluster domain secret
2166 - Import/export now use SSL instead of SSH
2167 - Added support for showing estimated time when exporting an instance,
2168 see the ``ganeti-os-interface(7)`` manpage and look for
2175 *(Released Tue, 16 Nov 2010)*
2177 Some more bugfixes. Unless critical bugs occur, this will be the last
2180 - Fix case of MAC special-values
2181 - Fix mac checker regex
2182 - backend: Fix typo causing "out of range" error
2183 - Add missing --units in gnt-instance list man page
2189 *(Released Tue, 24 Aug 2010)*
2192 - Don't ignore secondary node silently on non-mirrored disk templates
2194 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2195 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2196 - Properly document the usb_mouse parameter
2197 - Fix path in ganeti-rapi(8) (issue 116)
2198 - Adjust error message when the ganeti user's .ssh directory is
2200 - Add same-node-check when changing the disk template to drbd
2206 *(Released Fri, 16 Jul 2010)*
2209 - Add an option to only select some reboot types during qa/burnin.
2210 (on some hypervisors consequent reboots are not supported)
2211 - Fix infrequent race condition in master failover. Sometimes the old
2212 master ip address would be still detected as up for a short time
2213 after it was removed, causing failover to fail.
2214 - Decrease mlockall warnings when the ctypes module is missing. On
2215 Python 2.4 we support running even if no ctypes module is installed,
2216 but we were too verbose about this issue.
2217 - Fix building on old distributions, on which man doesn't have a
2219 - Fix RAPI not to ignore the MAC address on instance creation
2220 - Implement the old instance creation format in the RAPI client.
2226 *(Released Thu, 01 Jul 2010)*
2228 A small bugfix release:
2229 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2230 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2231 a non-existing option
2232 - Fix instance create: the --force-variant option was ignored
2233 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2234 - Fix modify node storage with non-FQDN arguments
2235 - Fix RAPI client to authenticate under Python 2.6 when used
2236 for more than 5 requests needing authentication
2237 - Fix gnt-instance modify -t (storage) giving a wrong error message
2238 when converting a non-shutdown drbd instance to plain
2244 *(Released Fri, 18 Jun 2010)*
2246 A small bugfix release:
2248 - Fix live migration of KVM instances started with older Ganeti
2249 versions which had fewer hypervisor parameters
2250 - Fix gnt-instance grow-disk on down instances
2251 - Fix an error-reporting bug during instance migration
2252 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2253 silently ignoring broken ones
2254 - Fix an RPC error reporting bug affecting, for example, RAPI client
2256 - Fix bug triggered by different API version os-es on different nodes
2257 - Fix a bug in instance startup with custom hvparams: OS level
2258 parameters would fail to be applied.
2259 - Fix the RAPI client under Python 2.6 (but more work is needed to
2260 make it work completely well with OpenSSL)
2261 - Fix handling of errors when resolving names from DNS
2267 *(Released Thu, 3 Jun 2010)*
2269 A medium sized development cycle. Some new features, and some
2270 fixes/small improvements/cleanups.
2272 Significant features
2273 ~~~~~~~~~~~~~~~~~~~~
2275 The node deamon now tries to mlock itself into memory, unless the
2276 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2277 its logs, and falls back to console logging. This allows emergency
2278 features such as ``gnt-node powercycle`` to work even in the event of a
2279 broken node disk (tested offlining the disk hosting the node's
2280 filesystem and dropping its memory caches; don't try this at home)
2282 KVM: add vhost-net acceleration support. It can be tested with a new
2283 enough version of the kernel and of qemu-kvm.
2285 KVM: Add instance chrooting feature. If you use privilege dropping for
2286 your VMs you can also now force them to chroot to an empty directory,
2287 before starting the emulated guest.
2289 KVM: Add maximum migration bandwith and maximum downtime tweaking
2290 support (requires a new-enough version of qemu-kvm).
2292 Cluster verify will now warn if the master node doesn't have the master
2293 ip configured on it.
2295 Add a new (incompatible) instance creation request format to RAPI which
2296 supports all parameters (previously only a subset was supported, and it
2297 wasn't possible to extend the old format to accomodate all the new
2298 features. The old format is still supported, and a client can check for
2299 this feature, before using it, by checking for its presence in the
2300 ``features`` RAPI resource.
2302 Now with ancient latin support. Try it passing the ``--roman`` option to
2303 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2304 (requires the python-roman module to be installed, in order to work).
2309 As usual many internal code refactorings, documentation updates, and
2312 - Lots of improvements and cleanups to the experimental Remote API
2313 (RAPI) client library.
2314 - A new unit test suite for the core daemon libraries.
2315 - A fix to creating missing directories makes sure the umask is not
2316 applied anymore. This enforces the same directory permissions
2318 - Better handling terminating daemons with ctrl+c (used when running
2319 them in debugging mode).
2320 - Fix a race condition in live migrating a KVM instance, when stat()
2321 on the old proc status file returned EINVAL, which is an unexpected
2323 - Fixed manpage checking with newer man and utf-8 charachters. But now
2324 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2330 *(Released Fri, 7 May 2010)*
2332 Fix a bug which prevented untagged KVM instances from starting.
2338 *(Released Fri, 7 May 2010)*
2340 Another release with a long development cycle, during which many
2341 different features were added.
2343 Significant features
2344 ~~~~~~~~~~~~~~~~~~~~
2346 The KVM hypervisor now can run the individual instances as non-root, to
2347 reduce the impact of a VM being hijacked due to bugs in the
2348 hypervisor. It is possible to run all instances as a single (non-root)
2349 user, to manually specify a user for each instance, or to dynamically
2350 allocate a user out of a cluster-wide pool to each instance, with the
2351 guarantee that no two instances will run under the same user ID on any
2354 An experimental RAPI client library, that can be used standalone
2355 (without the other Ganeti libraries), is provided in the source tree as
2356 ``lib/rapi/client.py``. Note this client might change its interface in
2357 the future, as we iterate on its capabilities.
2359 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2360 replace the cluster's certificates and crypto keys. This might help in
2361 case they have been compromised, or have simply expired.
2363 A new disk option for instance creation has been added that allows one
2364 to "adopt" currently existing logical volumes, with data
2365 preservation. This should allow easier migration to Ganeti from
2366 unmanaged (or managed via other software) instances.
2368 Another disk improvement is the possibility to convert between redundant
2369 (DRBD) and plain (LVM) disk configuration for an instance. This should
2370 allow better scalability (starting with one node and growing the
2371 cluster, or shrinking a two-node cluster to one node).
2373 A new feature that could help with automated node failovers has been
2374 implemented: if a node sees itself as offline (by querying the master
2375 candidates), it will try to shutdown (hard) all instances and any active
2376 DRBD devices. This reduces the risk of duplicate instances if an
2377 external script automatically failovers the instances on such nodes. To
2378 enable this, the cluster parameter ``maintain_node_health`` should be
2379 enabled; in the future this option (per the name) will enable other
2380 automatic maintenance features.
2382 Instance export/import now will reuse the original instance
2383 specifications for all parameters; that means exporting an instance,
2384 deleting it and the importing it back should give an almost identical
2385 instance. Note that the default import behaviour has changed from
2386 before, where it created only one NIC; now it recreates the original
2389 Cluster verify has added a few new checks: SSL certificates validity,
2390 /etc/hosts consistency across the cluster, etc.
2395 As usual, many internal changes were done, documentation fixes,
2398 - Fixed cluster initialization with disabled cluster storage (regression
2399 introduced in 2.1.1)
2400 - File-based storage supports growing the disks
2401 - Fixed behaviour of node role changes
2402 - Fixed cluster verify for some corner cases, plus a general rewrite of
2403 cluster verify to allow future extension with more checks
2404 - Fixed log spamming by watcher and node daemon (regression introduced
2406 - Fixed possible validation issues when changing the list of enabled
2408 - Fixed cleanup of /etc/hosts during node removal
2409 - Fixed RAPI response for invalid methods
2410 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2411 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2412 from ide disks, etc.)
2413 - Allow OS changes without re-installation (to record a changed OS
2414 outside of Ganeti, or to allow OS renames)
2415 - Allow instance creation without OS installation (useful for example if
2416 the OS will be installed manually, or restored from a backup not in
2418 - Implemented option to make cluster ``copyfile`` use the replication
2420 - Added list of enabled hypervisors to ssconf (possibly useful for
2422 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2424 - A partial form of node re-IP is possible via node readd, which now
2425 allows changed node primary IP
2426 - Command line utilities now show an informational message if the job is
2428 - The logs of the master daemon now show the PID/UID/GID of the
2435 *(Released Fri, 12 Mar 2010)*
2437 During the 2.1.0 long release candidate cycle, a lot of improvements and
2438 changes have accumulated with were released later as 2.1.1.
2443 The node evacuate command (``gnt-node evacuate``) was significantly
2444 rewritten, and as such the IAllocator protocol was changed - a new
2445 request type has been added. This unfortunate change during a stable
2446 series is designed to improve performance of node evacuations; on
2447 clusters with more than about five nodes and which are well-balanced,
2448 evacuation should proceed in parallel for all instances of the node
2449 being evacuated. As such, any existing IAllocator scripts need to be
2450 updated, otherwise the above command will fail due to the unknown
2451 request. The provided "dumb" allocator has not been updated; but the
2452 ganeti-htools package supports the new protocol since version 0.2.4.
2454 Another important change is increased validation of node and instance
2455 names. This might create problems in special cases, if invalid host
2456 names are being used.
2458 Also, a new layer of hypervisor parameters has been added, that sits at
2459 OS level between the cluster defaults and the instance ones. This allows
2460 customisation of virtualization parameters depending on the installed
2461 OS. For example instances with OS 'X' may have a different KVM kernel
2462 (or any other parameter) than the cluster defaults. This is intended to
2463 help managing a multiple OSes on the same cluster, without manual
2464 modification of each instance's parameters.
2466 A tool for merging clusters, ``cluster-merge``, has been added in the
2467 tools sub-directory.
2472 - Improved the int/float conversions that should make the code more
2473 robust in face of errors from the node daemons
2474 - Fixed the remove node code in case of internal configuration errors
2475 - Fixed the node daemon behaviour in face of inconsistent queue
2476 directory (e.g. read-only file-system where we can't open the files
2478 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2479 now it either aborts cleanly or, if given the new "auto_promote"
2480 parameter, will automatically promote other nodes as needed
2481 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2482 completely prevent Ganeti from working
2483 - Fixed bug for instance export when not all disks were successfully
2485 - Fixed behaviour of node add when the new node is slow in starting up
2487 - Fixed handling of signals in the LUXI client, which should improve
2488 behaviour of command-line scripts
2489 - Added checks for invalid node/instance names in the configuration (now
2490 flagged during cluster verify)
2491 - Fixed watcher behaviour for disk activation errors
2492 - Fixed two potentially endless loops in http library, which led to the
2493 RAPI daemon hanging and consuming 100% CPU in some cases
2494 - Fixed bug in RAPI daemon related to hashed passwords
2495 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2497 - Enhanced compatibility with non-Debian OSes, but not using absolute
2498 path in some commands and allowing customisation of the ssh
2499 configuration directory
2500 - Fixed possible future issue with new Python versions by abiding to the
2501 proper use of ``__slots__`` attribute on classes
2502 - Added checks that should prevent directory traversal attacks
2503 - Many documentation fixes based on feedback from users
2508 - Added an "early_release" more for instance replace disks and node
2509 evacuate, where we release locks earlier and thus allow higher
2510 parallelism within the cluster
2511 - Added watcher hooks, intended to allow the watcher to restart other
2512 daemons (e.g. from the ganeti-nbma project), but they can be used of
2513 course for any other purpose
2514 - Added a compile-time disable for DRBD barriers, to increase
2515 performance if the administrator trusts the power supply or the
2516 storage system to not lose writes
2517 - Added the option of using syslog for logging instead of, or in
2518 addition to, Ganeti's own log files
2519 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2520 can indeed boot from a paravirtual NIC
2521 - Added a generic debug level for many operations; while this is not
2522 used widely yet, it allows one to pass the debug value all the way to
2524 - Enhanced the hooks environment for instance moves (failovers,
2525 migrations) where the primary/secondary nodes changed during the
2526 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2527 - Enhanced data validations for many user-supplied values; one important
2528 item is the restrictions imposed on instance and node names, which
2529 might reject some (invalid) host names
2530 - Add a configure-time option to disable file-based storage, if it's not
2531 needed; this allows greater security separation between the master
2532 node and the other nodes from the point of view of the inter-node RPC
2534 - Added user notification in interactive tools if job is waiting in the
2535 job queue or trying to acquire locks
2536 - Added log messages when a job is waiting for locks
2537 - Added filtering by node tags in instance operations which admit
2538 multiple instances (start, stop, reboot, reinstall)
2539 - Added a new tool for cluster mergers, ``cluster-merge``
2540 - Parameters from command line which are of the form ``a=b,c=d`` can now
2541 use backslash escapes to pass in values which contain commas,
2542 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2544 - For KVM, the instance name is the first parameter passed to KVM, so
2545 that it's more visible in the process list
2551 *(Released Tue, 2 Mar 2010)*
2553 Ganeti 2.1 brings many improvements with it. Major changes:
2555 - Added infrastructure to ease automated disk repairs
2556 - Added new daemon to export configuration data in a cheaper way than
2557 using the remote API
2558 - Instance NICs can now be routed instead of being associated with a
2560 - Improved job locking logic to reduce impact of jobs acquiring multiple
2561 locks waiting for other long-running jobs
2563 In-depth implementation details can be found in the Ganeti 2.1 design
2569 - Added chroot hypervisor
2570 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2572 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2573 ``bootloader_path`` and ``bootloader_args``)
2574 - Added the ``use_localtime`` option for the xen-hvm and kvm
2575 hypervisors, and the default value for this has changed to false (in
2576 2.0 xen-hvm always enabled it)
2577 - Added luxi call to submit multiple jobs in one go
2578 - Added cluster initialization option to not modify ``/etc/hosts``
2580 - Added network interface parameters
2581 - Added dry run mode to some LUs
2582 - Added RAPI resources:
2584 - ``/2/instances/[instance_name]/info``
2585 - ``/2/instances/[instance_name]/replace-disks``
2586 - ``/2/nodes/[node_name]/evacuate``
2587 - ``/2/nodes/[node_name]/migrate``
2588 - ``/2/nodes/[node_name]/role``
2589 - ``/2/nodes/[node_name]/storage``
2590 - ``/2/nodes/[node_name]/storage/modify``
2591 - ``/2/nodes/[node_name]/storage/repair``
2593 - Added OpCodes to evacuate or migrate all instances on a node
2594 - Added new command to list storage elements on nodes (``gnt-node
2595 list-storage``) and modify them (``gnt-node modify-storage``)
2596 - Added new ssconf files with master candidate IP address
2597 (``ssconf_master_candidates_ips``), node primary IP address
2598 (``ssconf_node_primary_ips``) and node secondary IP address
2599 (``ssconf_node_secondary_ips``)
2600 - Added ``ganeti-confd`` and a client library to query the Ganeti
2601 configuration via UDP
2602 - Added ability to run hooks after cluster initialization and before
2604 - Added automatic mode for disk replace (``gnt-instance replace-disks
2606 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2607 after catastrophic data-loss
2608 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2610 - Added ``gnt-instance move`` command to move instances
2611 - Added ``gnt-cluster watcher`` command to control watcher
2612 - Added ``gnt-node powercycle`` command to powercycle nodes
2613 - Added new job status field ``lock_status``
2614 - Added parseable error codes to cluster verification (``gnt-cluster
2615 verify --error-codes``) and made output less verbose (use
2616 ``--verbose`` to restore previous behaviour)
2617 - Added UUIDs to the main config entities (cluster, nodes, instances)
2618 - Added support for OS variants
2619 - Added support for hashed passwords in the Ganeti remote API users file
2621 - Added option to specify maximum timeout on instance shutdown
2622 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2623 - Added new helper script to start and stop Ganeti daemons
2624 (``daemon-util``), with the intent to reduce the work necessary to
2625 adjust Ganeti for non-Debian distributions and to start/stop daemons
2627 - Added more unittests
2628 - Fixed critical bug in ganeti-masterd startup
2629 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2630 now customisable at the cluster level for both the KVM and Xen
2631 hypervisors using the new ``migration_port`` parameter
2632 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2633 reinstalling an instance
2634 - Allowed ``@`` in tag names
2635 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2636 - Many documentation updates
2637 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2638 - ``gnt-instance reinstall`` can now reinstall multiple instances
2639 - Updated many command line parameters
2640 - Introduced new OS API version 15
2641 - No longer support a default hypervisor
2642 - Treat virtual LVs as inexistent
2643 - Improved job locking logic to reduce lock contention
2644 - Match instance and node names case insensitively
2645 - Reimplemented bash completion script to be more complete
2652 *(Released Thu, 4 Feb 2010)*
2654 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2655 - Fix a string formatting bug
2656 - Improve safety of the code in some error paths
2657 - Improve data validation in the master of values returned from nodes
2663 *(Released Thu, 17 Dec 2009)*
2665 - Fix security issue due to missing validation of iallocator names; this
2666 allows local and remote execution of arbitrary executables
2667 - Fix failure of gnt-node list during instance removal
2668 - Ship the RAPI documentation in the archive
2674 *(Released Wed, 30 Sep 2009)*
2676 - Fixed many wrong messages
2677 - Fixed a few bugs related to the locking library
2678 - Fixed MAC checking at instance creation time
2679 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2680 - Fixed a few issues related to signal handling in both daemons and
2682 - Fixed the example startup script provided
2683 - Fixed insserv dependencies in the example startup script (patch from
2685 - Fixed handling of drained nodes in the iallocator framework
2686 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2688 - Fixed error related to invalid job IDs in job polling
2689 - Fixed job/opcode persistence on unclean master shutdown
2690 - Fixed handling of partial job processing after unclean master
2692 - Fixed error reporting from LUs, previously all errors were converted
2693 into execution errors
2694 - Fixed error reporting from burnin
2695 - Decreased significantly the memory usage of the job queue
2696 - Optimised slightly multi-job submission
2697 - Optimised slightly opcode loading
2698 - Backported the multi-job submit framework from the development
2699 branch; multi-instance start and stop should be faster
2700 - Added script to clean archived jobs after 21 days; this will reduce
2701 the size of the queue directory
2702 - Added some extra checks in disk size tracking
2703 - Added an example ethers hook script
2704 - Added a cluster parameter that prevents Ganeti from modifying of
2706 - Added more node information to RAPI responses
2707 - Added a ``gnt-job watch`` command that allows following the ouput of a
2709 - Added a bind-address option to ganeti-rapi
2710 - Added more checks to the configuration verify
2711 - Enhanced the burnin script such that some operations can be retried
2713 - Converted instance reinstall to multi-instance model
2719 *(Released Fri, 7 Aug 2009)*
2721 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2722 to allow using the pre-2.0.2 behaviour in activation, if any existing
2723 instances have mismatched disk sizes in the configuration
2724 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2725 any configuration mismatches for disk sizes
2726 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2727 failover to work on two-node clusters
2728 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2730 - Fixed detection of OS script errors in ``gnt-backup export``
2731 - Fixed exit code of ``gnt-backup export``
2737 *(Released Fri, 17 Jul 2009)*
2739 - Added experimental support for stripped logical volumes; this should
2740 enhance performance but comes with a higher complexity in the block
2741 device handling; stripping is only enabled when passing
2742 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2743 affected even in the non-stripped mode
2744 - Improved resiliency against transient failures at the end of DRBD
2745 resyncs, and in general of DRBD resync checks
2746 - Fixed a couple of issues with exports and snapshot errors
2747 - Fixed a couple of issues in instance listing
2748 - Added display of the disk size in ``gnt-instance info``
2749 - Fixed checking for valid OSes in instance creation
2750 - Fixed handling of the "vcpus" parameter in instance listing and in
2751 general of invalid parameters
2752 - Fixed http server library, and thus RAPI, to handle invalid
2753 username/password combinations correctly; this means that now they
2754 report unauthorized for queries too, not only for modifications,
2755 allowing earlier detect of configuration problems
2756 - Added a new "role" node list field, equivalent to the master/master
2757 candidate/drained/offline flags combinations
2758 - Fixed cluster modify and changes of candidate pool size
2759 - Fixed cluster verify error messages for wrong files on regular nodes
2760 - Fixed a couple of issues with node demotion from master candidate role
2761 - Fixed node readd issues
2762 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2763 - Added a new ``--no-voting`` option for masterfailover to fix failover
2764 on two-nodes clusters when the former master node is unreachable
2765 - Added instance reinstall over RAPI
2771 *(Released Tue, 16 Jun 2009)*
2773 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2774 allow re-adding the start in single-user option (regression from 1.2)
2775 - the watcher writes the instance status to a file, to allow monitoring
2776 to report the instance status (from the master) based on cached
2777 results of the watcher's queries; while this can get stale if the
2778 watcher is being locked due to other work on the cluster, this is
2779 still an improvement
2780 - the watcher now also restarts the node daemon and the rapi daemon if
2782 - fixed the watcher to handle full and drained queue cases
2783 - hooks export more instance data in the environment, which helps if
2784 hook scripts need to take action based on the instance's properties
2785 (no longer need to query back into ganeti)
2786 - instance failovers when the instance is stopped do not check for free
2787 RAM, so that failing over a stopped instance is possible in low memory
2789 - rapi uses queries for tags instead of jobs (for less job traffic), and
2790 for cluster tags it won't talk to masterd at all but read them from
2792 - a couple of error handling fixes in RAPI
2793 - drbd handling: improved the error handling of inconsistent disks after
2794 resync to reduce the frequency of "there are some degraded disks for
2795 this instance" messages
2796 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2797 error handling path called a wrong function name)
2803 *(Released Wed, 27 May 2009)*
2805 - no changes from rc5
2811 *(Released Wed, 20 May 2009)*
2813 - fix a couple of bugs (validation, argument checks)
2814 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2815 - some small improvements to RAPI and IAllocator
2816 - make watcher automatically start the master daemon if down
2822 *(Released Mon, 27 Apr 2009)*
2824 - change the OS list to not require locks; this helps with big clusters
2825 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2826 volume group is broken
2827 - ``gnt-instance info``, without any arguments, doesn't run for all
2828 instances anymore; either pass ``--all`` or pass the desired
2829 instances; this helps against mistakes on big clusters where listing
2830 the information for all instances takes a long time
2831 - miscellaneous doc and man pages fixes
2837 *(Released Wed, 8 Apr 2009)*
2839 - Change the internal locking model of some ``gnt-node`` commands, in
2840 order to reduce contention (and blocking of master daemon) when
2841 batching many creation/reinstall jobs
2842 - Fixes to Xen soft reboot
2843 - No longer build documentation at build time, instead distribute it in
2844 the archive, in order to reduce the need for the whole docbook/rst
2851 *(Released Fri, 27 Mar 2009)*
2853 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2854 - Fix watcher startup sequence, improves the behaviour of busy clusters
2855 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2856 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2857 bind address and other places
2858 - Some documentation fixes and updates
2864 *(Released Mon, 2 Mar 2009)*
2866 - More documentation updates, now all docs should be more-or-less
2868 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2870 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2872 - Fix an issue related to $libdir/run/ganeti and cluster creation
2878 *(Released Thu, 19 Feb 2009)*
2880 - Xen PVM and KVM have switched the default value for the instance root
2881 disk to the first partition on the first drive, instead of the whole
2882 drive; this means that the OS installation scripts must be changed
2884 - Man pages have been updated
2885 - RAPI has been switched by default to HTTPS, and the exported functions
2886 should all work correctly
2887 - RAPI v1 has been removed
2888 - Many improvements to the KVM hypervisor
2889 - Block device errors are now better reported
2890 - Many other bugfixes and small improvements
2896 *(Released Mon, 26 Jan 2009)*
2898 - Version 2 is a general rewrite of the code and therefore the
2899 differences are too many to list, see the design document for 2.0 in
2900 the ``doc/`` subdirectory for more details
2901 - In this beta version there is not yet a migration path from 1.2 (there
2902 will be one in the final 2.0 release)
2903 - A few significant changes are:
2905 - all commands are executed by a daemon (``ganeti-masterd``) and the
2906 various ``gnt-*`` commands are just front-ends to it
2907 - all the commands are entered into, and executed from a job queue,
2908 see the ``gnt-job(8)`` manpage
2909 - the RAPI daemon supports read-write operations, secured by basic
2910 HTTP authentication on top of HTTPS
2911 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2912 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2913 to migrate to DRBD 8 first while still running Ganeti 1.2)
2914 - DRBD devices are using statically allocated minor numbers, which
2915 will be assigned to existing instances during the migration process
2916 - there is support for both Xen PVM and Xen HVM instances running on
2918 - KVM virtualization is supported too
2919 - file-based storage has been implemented, which means that it is
2920 possible to run the cluster without LVM and DRBD storage, for
2921 example using a shared filesystem exported from shared storage (and
2922 still have live migration)
2928 *(Released Tue, 13 Jan 2009)*
2930 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2931 - Reuse the old instance mac address by default on instance import, if
2932 the instance name is the same.
2933 - Handle situations in which the node info rpc returns incomplete
2935 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2936 - Improved version of batcher:
2938 - state file support
2939 - instance mac address support
2940 - support for HVM clusters/instances
2942 - Add an option to show the number of cpu sockets and nodes in
2944 - Support OSes that handle more than one version of the OS api (but do
2945 not change the current API in any other way)
2946 - Fix ``gnt-node migrate``
2947 - ``gnt-debug`` man page
2948 - Fixes various more typos and small issues
2949 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2955 *(Released Wed, 24 Sep 2008)*
2957 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2958 type of disk exported to fully virtualized instances.
2959 - provide access to the serial console of HVM instances
2960 - instance auto_balance flag, set by default. If turned off it will
2961 avoid warnings on cluster verify if there is not enough memory to fail
2962 over an instance. in the future it will prevent automatically failing
2963 it over when we will support that.
2964 - batcher tool for instance creation, see ``tools/README.batcher``
2965 - ``gnt-instance reinstall --select-os`` to interactively select a new
2966 operating system when reinstalling an instance.
2967 - when changing the memory amount on instance modify a check has been
2968 added that the instance will be able to start. also warnings are
2969 emitted if the instance will not be able to fail over, if auto_balance
2971 - documentation fixes
2972 - sync fields between ``gnt-instance list/modify/add/import``
2973 - fix a race condition in drbd when the sync speed was set after giving
2974 the device a remote peer.
2980 *(Released Tue, 22 Jul 2008)*
2982 - note: the allowed size and number of tags per object were reduced
2983 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2984 - fixed twisted 8.x compatibility
2985 - fixed ``gnt-instance replace-disks`` with iallocator
2986 - add TCP keepalives on twisted connections to detect restarted nodes
2987 - disk increase support, see ``gnt-instance grow-disk``
2988 - implement bulk node/instance query for RAPI
2989 - add tags in node/instance listing (optional)
2990 - experimental migration (and live migration) support, read the man page
2991 for ``gnt-instance migrate``
2992 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2993 has some small improvements in handling its state file
2999 *(Released Fri, 13 Jun 2008)*
3001 - Experimental readonly, REST-based remote API implementation;
3002 automatically started on master node, TCP port 5080, if enabled by
3003 ``--enable-rapi`` parameter to configure script.
3004 - Instance allocator support. Add and import instance accept a
3005 ``--iallocator`` parameter, and call that instance allocator to decide
3006 which node to use for the instance. The iallocator document describes
3007 what's expected from an allocator script.
3008 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
3009 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
3010 node is lost there is still enough memory to fail over the instances
3012 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
3013 ``gnt-cluster verify``, to check for site-specific compliance. All the
3014 hooks will run, and their output, if any, will be displayed. Any
3015 failing hook will make the verification return an error value.
3016 - ``gnt-cluster verify`` now checks that its peers are reachable on the
3017 primary and secondary interfaces
3018 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
3019 that is still declared as part of the cluster and has failed.
3020 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
3021 specifying output fields, that just adds the chosen fields to the
3023 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
3024 export from the filesystem.
3025 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
3026 have been added. Using them you can enable/disable acpi and pae
3027 support, and specify a path for a cd image to be exported to the
3028 instance. These parameters as the name suggest only work on HVM
3030 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
3031 PAE support will be set to the previously hardcoded values, but the
3032 (previously hardcoded) path to the CDROM ISO image will be unset and
3033 if required, needs to be set manually with ``gnt-instance modify``
3035 - The address to which an instance's VNC console is bound is now
3036 selectable per-instance, rather than being cluster wide. Of course
3037 this only applies to instances controlled via VNC, so currently just
3038 applies to HVM clusters.
3044 *(Released Mon, 18 Feb 2008)*
3046 - more tweaks to the disk activation code (especially helpful for DRBD)
3047 - change the default ``gnt-instance list`` output format, now there is
3048 one combined status field (see the manpage for the exact values this
3050 - some more fixes for the mac export to hooks change
3051 - make Ganeti not break with DRBD 8.2.x (which changed the version
3052 format in ``/proc/drbd``) (issue 24)
3053 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3054 template, allowing migration from DRBD0.7+MD to DRBD8
3060 *(Released Wed, 30 Jan 2008)*
3062 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3064 - add command aliases infrastructure and a few aliases
3065 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
3066 man pages and the ``--help`` option of ``gnt-node
3067 list``/``gnt-instance list``
3068 - fix ``gnt-backup list`` with down nodes (issue 21)
3069 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3070 - fix the dist archive and add a check for including svn/git files in
3072 - some developer-related changes: improve the burnin and the QA suite,
3073 add an upload script for testing during development
3079 *(Released Wed, 16 Jan 2008)*
3081 - experimental HVM support, read the install document, section
3082 "Initializing the cluster"
3083 - allow for the PVM hypervisor per-instance kernel and initrd paths
3084 - add a new command ``gnt-cluster verify-disks`` which uses a new
3085 algorithm to improve the reconnection of the DRBD pairs if the device
3086 on the secondary node has gone away
3087 - make logical volume code auto-activate LVs at disk activation time
3088 - slightly improve the speed of activating disks
3089 - allow specification of the MAC address at instance creation time, and
3090 changing it later via ``gnt-instance modify``
3091 - fix handling of external commands that generate lots of output on
3093 - update documentation with regard to minimum version of DRBD8 supported
3099 *(Released Tue, 4 Dec 2007)*
3101 - Log the ``xm create`` output to the node daemon log on failure (to
3102 help diagnosing the error)
3103 - In debug mode, log all external commands output if failed to the logs
3104 - Change parsing of lvm commands to ignore stderr
3110 *(Released Wed, 28 Nov 2007)*
3112 - Another round of updates to the DRBD 8 code to deal with more failures
3113 in the replace secondary node operation
3114 - Some more logging of failures in disk operations (lvm, drbd)
3115 - A few documentation updates
3122 *(Released Tue, 13 Nov 2007)*
3124 - Change configuration file format from Python's Pickle to JSON.
3125 Upgrading is possible using the cfgupgrade utility.
3126 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3127 faster replace disks and is more stable (DRBD 8 has many improvements
3128 compared to DRBD 0.7)
3129 - Added command line tags support (see man pages for ``gnt-instance``,
3130 ``gnt-node``, ``gnt-cluster``)
3131 - Added instance rename support
3132 - Added multi-instance startup/shutdown
3133 - Added cluster rename support
3134 - Added ``gnt-node evacuate`` to simplify some node operations
3135 - Added instance reboot operation that can speedup reboot as compared to
3137 - Soften the requirement that hostnames are in FQDN format
3138 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3140 - Removed dependency on debian's patched fping that uses the
3141 non-standard ``-S`` option
3142 - Now the OS definitions are searched for in multiple, configurable
3143 paths (easier for distros to package)
3144 - Some changes to the hooks infrastructure (especially the new
3145 post-configuration update hook)
3146 - Other small bugfixes
3148 .. vim: set textwidth=72 syntax=rst :