10 Incompatible/important changes
11 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
13 - Adding disks with 'gnt-instance modify' now waits for the disks to sync per
14 default. Specify --no-wait-for-sync to override this behavior.
15 - The Ganeti python code now adheres to a private-module layout. In particular,
16 the module 'ganeti' is no longer in the python search path.
17 - On instance allocation, the iallocator now considers non-LVM storage
18 properly. In particular, actual file storage space information is used
19 when allocating space for a file/sharedfile instance.
20 - When disabling disk templates cluster-wide, the cluster now first
21 checks whether there are instances still using those templates.
22 - 'gnt-node list-storage' now also reports storage information about
23 file-based storage types.
28 - KVM hypervisors can now access RBD storage directly without having to
29 go through a block device.
30 - A new command 'gnt-cluster upgrade' was added that automates the upgrade
31 procedure between two Ganeti versions that are both 2.10 or higher.
32 - The move-instance command can now change disk templates when moving
33 instances, and does not require any node placement options to be
34 specified if the destination cluster has a default iallocator.
35 - Users can now change the soundhw and cpuid settings for XEN hypervisors.
36 - Hail and hbal now have the (optional) capability of accessing average CPU
37 load information through the monitoring deamon, and to use it to dynamically
38 adapt the allocation of instances.
39 - Hotplug support. Introduce new option '--hotplug' to ``gnt-instance modify``
40 so that disk and NIC modifications take effect without the need of actual
41 reboot. There are a couple of constrains currently for this feature:
43 - only KVM hypervisor (versions >= 1.0) supports it,
44 - one can not (yet) hotplug a disk using userspace access mode for RBD
45 - in case of a downgrade instances should suffer a reboot in order to
46 be migratable (due to core change of runtime files)
51 - A new test framework for logical units was introduced and the test
52 coverage for logical units was improved significantly.
53 - Opcodes are entirely generated from Haskell using the tool 'hs2py' and
54 the module 'src/Ganeti/OpCodes.hs'.
55 - Constants are also generated from Haskell using the tool
56 'hs2py-constants' and the module 'src/Ganeti/HsConstants.hs', with the
57 exception of socket related constants, which require changing the
58 cluster configuration file, and HVS related constants, because they
59 are part of a port of instance queries to Haskell. As a result, these
60 changes will be part of the next release of Ganeti.
65 The following new dependencies have been added/updated.
69 - The version requirements for ``python-mock`` have increased to at least
70 version 1.0.1. It is still used for testing only.
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
152 *(Released Tue, 15 Oct 2013)*
154 The third release candidate in the 2.9 series. Since 2.9.0 rc2:
156 - in implicit configuration upgrade, match ipolicy with enabled disk templates
157 - improved harep documentation (inherited from stable-2.8)
163 *(Released Wed, 9 Oct 2013)*
165 The second release candidate in the 2.9 series. Since 2.9.0 rc1:
167 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
168 at least one DRBD instance.
169 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
176 *(Released Tue, 1 Oct 2013)*
178 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
181 - update of the documentation, in particular installation instructions
182 - merging of LD_* constants into DT_* constants
183 - python style changes to be compatible with newer versions of pylint
189 *(Released Thu, 29 Aug 2013)*
191 This was the first beta release of the 2.9 series. All important changes
192 are listed in the latest 2.9 entry.
198 *(Released Thu, 17 Oct 2013)*
200 - Correctly start/stop luxid during gnt-cluster master-failover
201 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
202 - Fix path for the job queue serial file
203 - Improved harep man page
204 - Minor documentation improvements
210 *(Released Mon, 30 Sep 2013)*
212 Incompatible/important changes
213 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
215 - Instance policy can contain multiple instance specs, as described in
216 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
217 <design-partitioned>`. As a consequence, it's not possible to partially change
218 or override instance specs. Bounding specs (min and max) can be specified as a
219 whole using the new option ``--ipolicy-bounds-specs``, while standard
220 specs use the new option ``--ipolicy-std-specs``.
221 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
222 gnt-instance is a valid YAML object.
223 - hail now honors network restrictions when allocating nodes. This led to an
224 update of the IAllocator protocol. See the IAllocator documentation for
226 - confd now only answers static configuration request over the network. luxid
227 was extracted, listens on the local LUXI socket and responds to live queries.
228 This allows finer grained permissions if using separate users.
233 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
234 to always require authentication, ``--require-authentication``. It can
235 be specified in ``$sysconfdir/default/ganeti``.
236 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
237 be used to manage the disk templates to be used by instances in the cluster.
238 Initially, it will be set to a list that includes plain, drbd, if they were
239 enabled by specifying a volume group name, and file and sharedfile, if those
240 were enabled at configure time. Additionally, it will include all disk
241 templates that are currently used by instances. The order of disk templates
242 will be based on Ganeti's history of supporting them. In the future, the
243 first entry of the list will be used as a default disk template on instance
245 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
246 configuration back to the previous stable version.
247 - Disk templates in group ipolicy can be restored to the default value.
248 - Initial support for diskless instances and virtual clusters in QA.
249 - More QA and unit tests for instance policies.
250 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
251 describing why the opcode itself was executed.
252 - The monitoring daemon is now available. It allows users to query the cluster
253 for obtaining information about the status of the system. The daemon is only
254 responsible for providing the information over the network: the actual data
255 gathering is performed by data collectors (currently, only the DRBD status
256 collector is available).
257 - In order to help developers work on Ganeti, a new script
258 (``devel/build_chroot``) is provided, for building a chroot that contains all
259 the required development libraries and tools for compiling Ganeti on a Debian
261 - A new tool, ``harep``, for performing self-repair and recreation of instances
262 in Ganeti has been added.
263 - Split queries are enabled for tags, network, exports, cluster info, groups,
265 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
266 It prints the command line to set the current policies, to ease
268 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
269 devices for KVM virtio-net interfaces will get created with VNET_HDR
270 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
271 virtio-net interfaces, which prevents host kernel tainting and log
272 flooding, when dealing with broken or malicious virtio-net drivers.
273 It's set to true by default.
274 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
276 - Support 'viridian' parameter in Xen HVM
277 - Support DSA SSH keys in bootstrap
278 - To simplify the work of packaging frameworks that want to add the needed users
279 and groups in a split-user setup themselves, at build time three files in
280 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
281 the groups to be generated, the ``users`` file contains, one per line, the
282 users to be generated, optionally followed by their primary group, where
283 important. The ``groupmemberships`` file contains, one per line, additional
284 user-group membership relations that need to be established. The syntax of
285 these files will remain stable in all future versions.
290 The following new dependencies have been added:
293 - The ``curl`` library is not optional anymore for compiling the Haskell code.
294 - ``snap-server`` library (if monitoring is enabled).
297 - The minimum Python version needed to run Ganeti is now 2.6.
298 - ``yaml`` library (only for running the QA).
302 - Perform proper cleanup on termination of Haskell daemons
303 - Fix corner-case in handling of remaining retry time
309 *(Released Tue, 17 Sep 2013)*
311 - To simplify the work of packaging frameworks that want to add the needed users
312 and groups in a split-user setup themselves, at build time three files in
313 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
314 the groups to be generated, the ``users`` file contains, one per line, the
315 users to be generated, optionally followed by their primary group, where
316 important. The ``groupmemberships`` file contains, one per line, additional
317 user-group membership relations that need to be established. The syntax of
318 these files will remain stable in all future versions.
319 - Add a default to file-driver when unspecified over RAPI (Issue 571)
320 - Mark the DSA host pubkey as optional, and remove it during config downgrade
322 - Some documentation fixes
328 *(Released Tue, 27 Aug 2013)*
330 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
332 - Support 'viridian' parameter in Xen HVM (Issue 233)
333 - Include VCS version in ``gnt-cluster version``
334 - Support DSA SSH keys in bootstrap (Issue 338)
335 - Fix batch creation of instances
336 - Use FQDN to check master node status (Issue 551)
337 - Make the DRBD collector more failure-resilient
343 *(Released Fri, 2 Aug 2013)*
345 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
347 - Fix upgrading/downgrading from 2.7
348 - Increase maximum RAPI message size
349 - Documentation updates
350 - Split ``confd`` between ``luxid`` and ``confd``
351 - Merge 2.7 series up to the 2.7.1 release
352 - Allow the ``modify_etc_hosts`` option to be changed
353 - Add better debugging for ``luxid`` queries
354 - Expose bulk parameter for GetJobs in RAPI client
355 - Expose missing ``network`` fields in RAPI
356 - Add some ``cluster verify`` tests
357 - Some unittest fixes
358 - Fix a malfunction in ``hspace``'s tiered allocation
359 - Fix query compatibility between haskell and python implementations
360 - Add the ``vnet_hdr`` HV parameter for KVM
361 - Add ``--cleanup`` to instance failover
362 - Change the connected groups format in ``gnt-network info`` output; it
363 was previously displayed as a raw list by mistake. (Merged from 2.7)
369 *(Released Mon, 24 Jun 2013)*
371 This was the first beta release of the 2.8 series. All important changes
372 are listed in the latest 2.8 entry.
378 *(Released Thu, 26 Sep 2013)*
380 - Change the connected groups format in ``gnt-network info`` output; it
381 was previously displayed as a raw list by mistake
382 - Check disk template in right dict when copying
383 - Support multi-instance allocs without iallocator
384 - Fix some errors in the documentation
385 - Fix formatting of tuple in an error message
391 *(Released Thu, 25 Jul 2013)*
393 - Add logrotate functionality in daemon-util
394 - Add logrotate example file
395 - Add missing fields to network queries over rapi
396 - Fix network object timestamps
397 - Add support for querying network timestamps
398 - Fix a typo in the example crontab
399 - Fix a documentation typo
405 *(Released Thu, 04 Jul 2013)*
407 Incompatible/important changes
408 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
410 - Instance policies for disk size were documented to be on a per-disk
411 basis, but hail applied them to the sum of all disks. This has been
413 - ``hbal`` will now exit with status 0 if, during job execution over
414 LUXI, early exit has been requested and all jobs are successful;
415 before, exit status 1 was used, which cannot be differentiated from
417 - Compatibility with newer versions of rbd has been fixed
418 - ``gnt-instance batch-create`` has been changed to use the bulk create
419 opcode from Ganeti. This lead to incompatible changes in the format of
420 the JSON file. It's now not a custom dict anymore but a dict
421 compatible with the ``OpInstanceCreate`` opcode.
422 - Parent directories for file storage need to be listed in
423 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
424 write the file automatically based on old configuration values, but it
425 can not distribute it across all nodes and the file contents should be
426 verified. Use ``gnt-cluster copyfile
427 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
428 upgraded. The reason for requiring this list of paths now is that
429 before it would have been possible to inject new paths via RPC,
430 allowing files to be created in arbitrary locations. The RPC protocol
431 is protected using SSL/X.509 certificates, but as a design principle
432 Ganeti does not permit arbitrary paths to be passed.
433 - The parsing of the variants file for OSes (see
434 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
435 lines and comment lines (starting with ``#``) are ignored for better
437 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
438 longer available. ``gnt-node add`` now invokes a new tool on the
439 destination node, named ``prepare-node-join``, to configure the SSH
440 daemon. Paramiko is no longer necessary to configure nodes' SSH
441 daemons via ``gnt-node add``.
442 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
443 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
444 the flag is not reset after a master failover.
445 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
446 is recommended. 2.6 will be mandatory from the 2.8 series.
452 - New network management functionality to support automatic allocation
453 of IP addresses and managing of network parameters. See
454 :manpage:`gnt-network(8)` for more details.
455 - New external storage backend, to allow managing arbitrary storage
456 systems external to the cluster. See
457 :manpage:`ganeti-extstorage-interface(7)`.
458 - New ``exclusive-storage`` node parameter added, restricted to
459 nodegroup level. When it's set to true, physical disks are assigned in
460 an exclusive fashion to instances, as documented in :doc:`Partitioned
461 Ganeti <design-partitioned>`. Currently, only instances using the
462 ``plain`` disk template are supported.
463 - The KVM hypervisor has been updated with many new hypervisor
464 parameters, including a generic one for passing arbitrary command line
465 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
466 compatible up to qemu 1.4.
467 - A new tool, called ``mon-collector``, is the stand-alone executor of
468 the data collectors for a monitoring system. As of this version, it
469 just includes the DRBD data collector, that can be executed by calling
470 ``mon-collector`` using the ``drbd`` parameter. See
471 :manpage:`mon-collector(7)`.
472 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
473 for RAPI users. It allows granting permissions to query for
474 information to a specific user without giving
475 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
476 - A new tool named ``node-cleanup`` has been added. It cleans remains of
477 a cluster from a machine by stopping all daemons, removing
478 certificates and ssconf files. Unless the ``--no-backup`` option is
479 given, copies of the certificates are made.
480 - Instance creations now support the use of opportunistic locking,
481 potentially speeding up the (parallel) creation of multiple instances.
482 This feature is currently only available via the :doc:`RAPI
483 <rapi>` interface and when an instance allocator is used. If the
484 ``opportunistic_locking`` parameter is set the opcode will try to
485 acquire as many locks as possible, but will not wait for any locks
486 held by other opcodes. If not enough resources can be found to
487 allocate the instance, the temporary error code
488 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
489 retried thereafter, with or without opportunistic locking.
490 - New experimental linux-ha resource scripts.
491 - Restricted-commands support: ganeti can now be asked (via command line
492 or rapi) to perform commands on a node. These are passed via ganeti
493 RPC rather than ssh. This functionality is restricted to commands
494 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
495 reasons. The file is not copied automatically.
501 - Diskless instances are now externally mirrored (Issue 237). This for
502 now has only been tested in conjunction with explicit target nodes for
504 - Queries not needing locks or RPC access to the node can now be
505 performed by the confd daemon, making them independent from jobs, and
506 thus faster to execute. This is selectable at configure time.
507 - The functionality for allocating multiple instances at once has been
508 overhauled and is now also available through :doc:`RAPI <rapi>`.
510 There are no significant changes from version 2.7.0~rc3.
516 *(Released Tue, 25 Jun 2013)*
518 - Fix permissions on the confd query socket (Issue 477)
519 - Fix permissions on the job archive dir (Issue 498)
520 - Fix handling of an internal exception in replace-disks (Issue 472)
521 - Fix gnt-node info handling of shortened names (Issue 497)
522 - Fix gnt-instance grow-disk when wiping is enabled
523 - Documentation improvements, and support for newer pandoc
524 - Fix hspace honoring ipolicy for disks (Issue 484)
525 - Improve handling of the ``kvm_extra`` HV parameter
531 *(Released Fri, 24 May 2013)*
533 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
535 - Disks added through ``gnt-instance modify`` or created through
536 ``gnt-instance recreate-disks`` are wiped, if the
537 ``prealloc_wipe_disks`` flag is set.
538 - If wiping newly created disks fails, the disks are removed. Also,
539 partial failures in creating disks through ``gnt-instance modify``
540 triggers a cleanup of the partially-created disks.
541 - Removing the master IP address doesn't fail if the address has been
543 - Fix ownership of the OS log dir
544 - Workaround missing SO_PEERCRED constant (Issue 191)
550 *(Released Fri, 3 May 2013)*
552 This was the first release candidate of the 2.7 series. Since beta3:
554 - Fix kvm compatibility with qemu 1.4 (Issue 389)
555 - Documentation updates (admin guide, upgrade notes, install
556 instructions) (Issue 372)
557 - Fix gnt-group list nodes and instances count (Issue 436)
558 - Fix compilation without non-mandatory libraries (Issue 441)
559 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
560 - Make confd logging more verbose at INFO level (Issue 435)
561 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
562 - Fix failure path for instance storage type conversion (Issue 229)
563 - Update htools text backend documentation
564 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
565 - Disable inter-cluster instance move for file-based instances, because
566 it is dependant on instance export, which is not supported for
567 file-based instances. (Issue 414)
568 - Fix gnt-job crashes on non-ascii characters (Issue 427)
569 - Fix volume group checks on non-vm-capable nodes (Issue 432)
575 *(Released Mon, 22 Apr 2013)*
577 This was the third beta release of the 2.7 series. Since beta2:
579 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
580 - Fix data loss on wrong usage of ``gnt-instance move``
581 - Properly export errors in confd-based job queries
582 - Add ``users-setup`` tool
583 - Fix iallocator protocol to report 0 as a disk size for diskless
584 instances. This avoids hail breaking when a diskless instance is
586 - Fix job queue directory permission problem that made confd job queries
587 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
588 for access to archived jobs (Issue 406).
589 - Limit the sizes of networks supported by ``gnt-network`` to something
590 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
591 - Fix bugs in instance disk template conversion
592 - Fix GHC 7 compatibility
593 - Fix ``burnin`` install path (Issue 426).
594 - Allow very small disk grows (Issue 347).
595 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
596 that noded doesn't import masterd code (Issue 419).
597 - Make sure the default metavg at cluster init is the same as the vg, if
598 unspecified (Issue 358).
599 - Fix cleanup of partially created disks (part of Issue 416)
605 *(Released Tue, 2 Apr 2013)*
607 This was the second beta release of the 2.7 series. Since beta1:
609 - Networks no longer have a "type" slot, since this information was
610 unused in Ganeti: instead of it tags should be used.
611 - The rapi client now has a ``target_node`` option to MigrateInstance.
612 - Fix early exit return code for hbal (Issue 386).
613 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
614 - Fix ``rbd showmapped`` output parsing (Issue 312).
615 - Networks are now referenced indexed by UUID, rather than name. This
616 will require running cfgupgrade, from 2.7.0beta1, if networks are in
618 - The OS environment now includes network information.
619 - Deleting of a network is now disallowed if any instance nic is using
620 it, to prevent dangling references.
621 - External storage is now documented in man pages.
622 - The exclusive_storage flag can now only be set at nodegroup level.
623 - Hbal can now submit an explicit priority with its jobs.
624 - Many network related locking fixes.
625 - Bump up the required pylint version to 0.25.1.
626 - Fix the ``no_remember`` option in RAPI client.
627 - Many ipolicy related tests, qa, and fixes.
628 - Many documentation improvements and fixes.
629 - Fix building with ``--disable-file-storage``.
630 - Fix ``-q`` option in htools, which was broken if passed more than
632 - Some haskell/python interaction improvements and fixes.
633 - Fix iallocator in case of missing LVM storage.
634 - Fix confd config load in case of ``--no-lvm-storage``.
635 - The confd/query functionality is now mentioned in the security
642 *(Released Wed, 6 Feb 2013)*
644 This was the first beta release of the 2.7 series. All important changes
645 are listed in the latest 2.7 entry.
651 *(Released Fri, 21 Dec 2012)*
653 Important behaviour change: hbal won't rebalance anymore instances which
654 have the ``auto_balance`` attribute set to false. This was the intention
655 all along, but until now it only skipped those from the N+1 memory
656 reservation (DRBD-specific).
658 A significant number of bug fixes in this release:
660 - Fixed disk adoption interaction with ipolicy checks.
661 - Fixed networking issues when instances are started, stopped or
662 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
663 - Fixed the warning in cluster verify for shared storage instances not
665 - Fixed removal of storage directory on shared file storage (issue 262).
666 - Fixed validation of LVM volume group name in OpClusterSetParams
667 (``gnt-cluster modify``) (issue 285).
668 - Fixed runtime memory increases (``gnt-instance modify -m``).
669 - Fixed live migration under Xen's ``xl`` mode.
670 - Fixed ``gnt-instance console`` with ``xl``.
671 - Fixed building with newer Haskell compiler/libraries.
672 - Fixed PID file writing in Haskell daemons (confd); this prevents
673 restart issues if confd was launched manually (outside of
674 ``daemon-util``) while another copy of it was running
675 - Fixed a type error when doing live migrations with KVM (issue 297) and
676 the error messages for failing migrations have been improved.
677 - Fixed opcode validation for the out-of-band commands (``gnt-node
679 - Fixed a type error when unsetting OS hypervisor parameters (issue
680 311); now it's possible to unset all OS-specific hypervisor
682 - Fixed the ``dry-run`` mode for many operations: verification of
683 results was over-zealous but didn't take into account the ``dry-run``
684 operation, resulting in "wrong" failures.
685 - Fixed bash completion in ``gnt-job list`` when the job queue has
686 hundreds of entries; especially with older ``bash`` versions, this
687 results in significant CPU usage.
689 And lastly, a few other improvements have been made:
691 - Added option to force master-failover without voting (issue 282).
692 - Clarified error message on lock conflict (issue 287).
693 - Logging of newly submitted jobs has been improved (issue 290).
694 - Hostname checks have been made uniform between instance rename and
696 - The ``--submit`` option is now supported by ``gnt-debug delay``.
697 - Shutting down the master daemon by sending SIGTERM now stops it from
698 processing jobs waiting for locks; instead, those jobs will be started
699 once again after the master daemon is started the next time (issue
701 - Support for Xen's ``xl`` program has been improved (besides the fixes
703 - Reduced logging noise in the Haskell confd daemon (only show one log
704 entry for each config reload, instead of two).
705 - Several man page updates and typo fixes.
711 *(Released Fri, 12 Oct 2012)*
713 A small bugfix release. Among the bugs fixed:
715 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
716 made the command unusable.
717 - Commands that issue many jobs don't fail anymore just because some jobs
718 take so long that other jobs are archived.
719 - Failures during ``gnt-instance reinstall`` are reflected by the exit
721 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
723 - When ``always_failover`` is set, ``--allow-failover`` is not required
724 in migrate commands anymore.
725 - ``bash_completion`` works even if extglob is disabled.
726 - Fixed bug with locks that made failover for RDB-based instances fail.
727 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
728 a random node instead of one based on the allocator metric.
729 - Support for newer versions of pylint and pep8.
730 - Hail doesn't fail anymore when trying to add an instance of type
731 ``file``, ``sharedfile`` or ``rbd``.
732 - Added new Makefile target to rebuild the whole distribution, so that
733 all files are included.
739 *(Released Fri, 27 Jul 2012)*
742 .. attention:: The ``LUXI`` protocol has been made more consistent
743 regarding its handling of command arguments. This, however, leads to
744 incompatibility issues with previous versions. Please ensure that you
745 restart Ganeti daemons soon after the upgrade, otherwise most
746 ``LUXI`` calls (job submission, setting/resetting the drain flag,
747 pausing/resuming the watcher, cancelling and archiving jobs, querying
748 the cluster configuration) will fail.
757 The current ``admin_up`` field, which used to denote whether an instance
758 should be running or not, has been removed. Instead, ``admin_state`` is
759 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
761 The rational behind this is that an instance being “down” can have
764 - it could be down during a reboot
765 - it could be temporarily be down for a reinstall
766 - or it could be down because it is deprecated and kept just for its
769 The previous Boolean state was making it difficult to do capacity
770 calculations: should Ganeti reserve memory for a down instance? Now, the
771 tri-state field makes it clear:
773 - in ``up`` and ``down`` state, all resources are reserved for the
774 instance, and it can be at any time brought up if it is down
775 - in ``offline`` state, only disk space is reserved for it, but not
778 The field can have an extra use: since the transition between ``up`` and
779 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
780 transition between ``offline`` and ``down`` is done via ``gnt-instance
781 modify``, it is possible to given different rights to users. For
782 example, owners of an instance could be allowed to start/stop it, but
783 not transition it out of the offline state.
785 Instance policies and specs
786 +++++++++++++++++++++++++++
788 In previous Ganeti versions, an instance creation request was not
789 limited on the minimum size and on the maximum size just by the cluster
790 resources. As such, any policy could be implemented only in third-party
791 clients (RAPI clients, or shell wrappers over ``gnt-*``
792 tools). Furthermore, calculating cluster capacity via ``hspace`` again
793 required external input with regards to instance sizes.
795 In order to improve these workflows and to allow for example better
796 per-node group differentiation, we introduced instance specs, which
799 - minimum instance disk size, disk count, memory size, cpu count
800 - maximum values for the above metrics
801 - and “standard” values (used in ``hspace`` to calculate the standard
804 The minimum/maximum values can be also customised at node-group level,
805 for example allowing more powerful hardware to support bigger instance
808 Beside the instance specs, there are a few other settings belonging to
809 the instance policy framework. It is possible now to customise, per
810 cluster and node-group:
812 - the list of allowed disk templates
813 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
814 - the maximum ratio of instance to spindles (see below for more
815 information) for local storage
817 All these together should allow all tools that talk to Ganeti to know
818 what are the ranges of allowed values for instances and the
819 over-subscription that is allowed.
821 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
822 instance configuration, and the physical CPU configuration from the
823 node. For the spindle ratios however, we didn't track before these
824 values, so new parameters have been added:
826 - a new node parameter ``spindle_count``, defaults to 1, customisable at
827 node group or node level
828 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
830 Note that spindles in this context doesn't need to mean actual
831 mechanical hard-drives; it's just a relative number for both the node
832 I/O capacity and instance I/O consumption.
834 Instance migration behaviour
835 ++++++++++++++++++++++++++++
837 While live-migration is in general desirable over failover, it is
838 possible that for some workloads it is actually worse, due to the
839 variable time of the “suspend” phase during live migration.
841 To allow the tools to work consistently over such instances (without
842 having to hard-code instance names), a new backend parameter
843 ``always_failover`` has been added to control the migration/failover
844 behaviour. When set to True, all migration requests for an instance will
845 instead fall-back to failover.
847 Instance memory ballooning
848 ++++++++++++++++++++++++++
850 Initial support for memory ballooning has been added. The memory for an
851 instance is no longer fixed (backend parameter ``memory``), but instead
852 can vary between minimum and maximum values (backend parameters
853 ``minmem`` and ``maxmem``). Currently we only change an instance's
856 - live migrating or failing over and instance and the target node
857 doesn't have enough memory
858 - user requests changing the memory via ``gnt-instance modify
864 In order to control the use of specific CPUs by instance, support for
865 controlling CPU pinning has been added for the Xen, HVM and LXC
866 hypervisors. This is controlled by a new hypervisor parameter
867 ``cpu_mask``; details about possible values for this are in the
868 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
869 VCPU-to-CPU mapping) form will work well only when all nodes in your
870 cluster have the same amount of CPUs.
875 Another area in which Ganeti was not customisable were the parameters
876 used for storage configuration, e.g. how many stripes to use for LVM,
877 DRBD resync configuration, etc.
879 To improve this area, we've added disks parameters, which are
880 customisable at cluster and node group level, and which allow to
881 specify various parameters for disks (DRBD has the most parameters
882 currently), for example:
884 - DRBD resync algorithm and parameters (e.g. speed)
885 - the default VG for meta-data volumes for DRBD
886 - number of stripes for LVM (plain disk template)
889 These parameters can be modified via ``gnt-cluster modify -D …`` and
890 ``gnt-group modify -D …``, and are used at either instance creation (in
891 case of LVM stripes, for example) or at disk “activation” time
894 Rados block device support
895 ++++++++++++++++++++++++++
897 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
898 denoted by the ``rbd`` disk template type. This is considered
899 experimental, feedback is welcome. For details on configuring it, see
900 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
905 The existing master IP functionality works well only in simple setups (a
906 single network shared by all nodes); however, if nodes belong to
907 different networks, then the ``/32`` setup and lack of routing
908 information is not enough.
910 To allow the master IP to function well in more complex cases, the
911 system was reworked as follows:
913 - a master IP netmask setting has been added
914 - the master IP activation/turn-down code was moved from the node daemon
916 - whether to run the Ganeti-supplied master IP script or a user-supplied
917 on is a ``gnt-cluster init`` setting
919 Details about the location of the standard and custom setup scripts are
920 in the man page :manpage:`gnt-cluster(8)`; for information about the
921 setup script protocol, look at the Ganeti-supplied script.
926 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
929 It is now possible to use TLS-protected connections, and when renewing
930 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
931 it is now possible to specify spice or spice CA certificates. Also, it
932 is possible to configure a password for SPICE sessions via the
933 hypervisor parameter ``spice_password_file``.
935 There are also new parameters to control the compression and streaming
936 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
937 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
938 for the spice parameters.
940 Lastly, it is now possible to see the SPICE connection information via
941 ``gnt-instance console``.
946 A new tool (``tools/ovfconverter``) has been added that supports
947 conversion between Ganeti and the `Open Virtualization Format
948 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
951 This relies on the ``qemu-img`` tool to convert the disk formats, so the
952 actual compatibility with other virtualization solutions depends on it.
957 The configuration query daemon (``ganeti-confd``) is now optional, and
958 has been rewritten in Haskell; whether to use the daemon at all, use the
959 Python (default) or the Haskell version is selectable at configure time
960 via the ``--enable-confd`` parameter, which can take one of the
961 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
962 daemon will result in a smaller footprint; for larger systems, we
963 welcome feedback on the Haskell version which might become the default
966 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
967 daemon running. The Python version doesn't implement the new call.
970 User interface changes
971 ~~~~~~~~~~~~~~~~~~~~~~
973 We have replaced the ``--disks`` option of ``gnt-instance
974 replace-disks`` with a more flexible ``--disk`` option, which allows
975 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
976 disk size and mode can be changed upon recreation (via ``gnt-instance
977 recreate-disks``, which accepts the same ``--disk`` option).
979 As many people are used to a ``show`` command, we have added that as an
980 alias to ``info`` on all ``gnt-*`` commands.
982 The ``gnt-instance grow-disk`` command has a new mode in which it can
983 accept the target size of the disk, instead of the delta; this can be
984 more safe since two runs in absolute mode will be idempotent, and
985 sometimes it's also easier to specify the desired size directly.
987 Also the handling of instances with regard to offline secondaries has
988 been improved. Instance operations should not fail because one of it's
989 secondary nodes is offline, even though it's safe to proceed.
991 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
992 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
993 minors to instance name.
998 RAPI coverage has improved, with (for example) new resources for
999 recreate-disks, node power-cycle, etc.
1004 There is partial support for ``xl`` in the Xen hypervisor; feedback is
1007 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1008 whether to still support Python 2.4 or move to Python 2.6 as minimum
1011 Support for Fedora has been slightly improved; the provided example
1012 init.d script should work better on it and the INSTALL file should
1013 document the needed dependencies.
1018 The deprecated ``QueryLocks`` LUXI request has been removed. Use
1019 ``Query(what=QR_LOCK, ...)`` instead.
1021 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1022 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1023 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1024 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1025 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1027 RAPI client: ``CertificateError`` now derives from
1028 ``GanetiApiError``. This should make it more easy to handle Ganeti
1031 Deprecation warnings due to PyCrypto/paramiko import in
1032 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
1033 make sure to run an up-to-date paramiko version, if you use this tool.
1035 The QA scripts now depend on Python 2.5 or above (the main code base
1036 still works with Python 2.4).
1038 The configuration file (``config.data``) is now written without
1039 indentation for performance reasons; if you want to edit it, it can be
1040 re-formatted via ``tools/fmtjson``.
1042 A number of bugs has been fixed in the cluster merge tool.
1044 ``x509`` certification verification (used in import-export) has been
1045 changed to allow the same clock skew as permitted by the cluster
1046 verification. This will remove some rare but hard to diagnose errors in
1053 *(Released Thu, 19 Jul 2012)*
1055 Very few changes from rc4 to the final release, only bugfixes:
1057 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
1058 instance, fix CDROM booting for KVM instances)
1059 - fixed node group modification of node parameters
1060 - fixed issue in LUClusterVerifyGroup with multi-group clusters
1061 - fixed generation of bash completion to ensure a stable ordering
1068 *(Released Fri, 13 Jul 2012)*
1070 Third release candidate for 2.6. The following changes were done from
1073 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1074 - Fixed an inconsistency in the LUXI protocol with the provided
1075 arguments (NOT backwards compatible)
1076 - Fixed a bug with node groups ipolicy where ``min`` was greater than
1077 the cluster ``std`` value
1078 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1079 easier instance debugging on nodes (requires ``hconfd`` to work)
1085 *(Released Tue, 03 Jul 2012)*
1087 Second release candidate for 2.6. The following changes were done from
1090 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1092 - Fixed a RAPI regression on missing beparams/memory
1093 - Fixed redistribution of files on offline nodes
1094 - Added possibility to run activate-disks even though secondaries are
1095 offline. With this change it relaxes also the strictness on some other
1096 commands which use activate disks internally:
1097 * ``gnt-instance start|reboot|rename|backup|export``
1098 - Made it possible to remove safely an instance if its secondaries are
1100 - Made it possible to reinstall even though secondaries are offline
1106 *(Released Mon, 25 Jun 2012)*
1108 First release candidate for 2.6. The following changes were done from
1111 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1113 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1114 behaviour. It supports now the modify operation
1115 - ``hcheck`` implemented to analyze cluster health and possibility of
1116 improving health by rebalance
1117 - ``hbal`` has been improved in dealing with split instances
1123 *(Released Mon, 11 Jun 2012)*
1125 Second beta release of 2.6. The following changes were done from beta2
1128 - Fixed ``daemon-util`` with non-root user models
1129 - Fixed creation of plain instances with ``--no-wait-for-sync``
1130 - Fix wrong iv_names when running ``cfgupgrade``
1131 - Export more information in RAPI group queries
1132 - Fixed bug when changing instance network interfaces
1133 - Extended burnin to do NIC changes
1134 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1135 - Changed default for DRBD barriers
1136 - Fixed DRBD error reporting for syncer rate
1137 - Verify the options on disk parameters
1139 And of course various fixes to documentation and improved unittests and
1146 *(Released Wed, 23 May 2012)*
1148 First beta release of 2.6. The following changes were done from beta1 to
1151 - integrated patch for distributions without ``start-stop-daemon``
1152 - adapted example init.d script to work on Fedora
1153 - fixed log handling in Haskell daemons
1154 - adapted checks in the watcher for pycurl linked against libnss
1155 - add partial support for ``xl`` instead of ``xm`` for Xen
1156 - fixed a type issue in cluster verification
1157 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1160 Plus integrated fixes from the 2.5 branch:
1162 - fixed ``kvm-ifup`` to use ``/bin/bash``
1163 - fixed parallel build failures
1164 - KVM live migration when using a custom keymap
1170 *(Released Tue, 24 Jul 2012)*
1172 A small bugfix release, with no new features:
1174 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1175 different default shell (e.g. Debian, Ubuntu)
1176 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1177 243 and Debian bug #650664)
1178 - fixed compatibility with KVM versions that don't support multiple boot
1179 devices (fixes Issue 230 and Debian bug #624256)
1181 Additionally, a few fixes were done to the build system (fixed parallel
1182 build failures) and to the unittests (fixed race condition in test for
1183 FileID functions, and the default enable/disable mode for QA test is now
1190 *(Released Fri, 11 May 2012)*
1192 A small bugfix release.
1194 The main issues solved are on the topic of compatibility with newer LVM
1197 - fixed parsing of ``lv_attr`` field
1198 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1199 the ``--force`` flag is needed
1201 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1202 to accept kernel 3.x too (was hardcoded to 2.6.*).
1204 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1205 of hook results and that also made display of other errors suboptimal
1206 was fixed; the code behaves now like 2.4 and earlier.
1208 Another change in 2.5, the cleanup of the OS scripts environment, is too
1209 aggressive: it removed even the ``PATH`` variable, which requires the OS
1210 scripts to *always* need to export it. Since this is a bit too strict,
1211 we now export a minimal PATH, the same that we export for hooks.
1213 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1214 integrated into this release.
1216 Finally, a few other miscellaneous changes were done (no new features,
1217 just small improvements):
1219 - Fix ``gnt-group --help`` display
1220 - Fix hardcoded Xen kernel path
1221 - Fix grow-disk handling of invalid units
1222 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1223 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1229 *(Released Thu, 12 Apr 2012)*
1231 Incompatible/important changes and bugfixes
1232 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1234 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1235 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1236 to match the underlying LUXI interface.
1237 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1238 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1239 not maintain backwards-compatibility as the underlying operation
1240 changed in an incompatible way. The RAPI client can talk to old
1241 servers, but it needs to be told so as the return value changed.
1242 - When creating file-based instances via RAPI, the ``file_driver``
1243 parameter no longer defaults to ``loop`` and must be specified.
1244 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1246 - Support for the undocumented and deprecated RAPI instance creation
1247 request format version 0 has been dropped. Use version 1, supported
1248 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1249 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1250 documentation <install>`.
1251 - The "cluster-verify" hooks are now executed per group by the
1252 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1253 if you just run ``gnt-cluster verify``, which generates one opcode per
1255 - The environment as passed to the OS scripts is cleared, and thus no
1256 environment variables defined in the node daemon's environment will be
1257 inherited by the scripts.
1258 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1260 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1261 support operations involving multiple node groups.
1262 - Offline nodes are ignored when failing over an instance.
1263 - Support for KVM version 1.0, which changed the version reporting format
1265 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1267 - ``Makefile`` is now compatible with Automake 1.11.2
1268 - Includes all bugfixes made in the 2.4 series
1273 - The ganeti-htools project has been merged into the ganeti-core source
1274 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1275 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1276 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1277 GPT-style partition tables (requires `parted
1278 <http://www.gnu.org/s/parted/>`_).
1279 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1280 - Allowed adding tags on instance creation.
1281 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1282 - Allow instances to be started in a paused state, enabling the user to
1283 see the complete console output on boot using the console.
1284 - Added new hypervisor flag to control default reboot behaviour
1285 (``reboot_behavior``).
1286 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1287 - Improved out-of-band management support:
1289 - Added ``gnt-node health`` command reporting the health status of
1291 - Added ``gnt-node power`` command to manage power status of nodes.
1292 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1294 - Instance migration can fall back to failover if instance is not
1296 - Filters can be used when listing nodes, instances, groups and locks;
1297 see :manpage:`ganeti(7)` manpage.
1298 - Added post-execution status as variables to :doc:`hooks <hooks>`
1300 - Instance tags are exported/imported together with the instance.
1301 - When given an explicit job ID, ``gnt-job info`` will work for archived
1303 - Jobs can define dependencies on other jobs (not yet supported via
1304 RAPI or command line, but used by internal commands and usable via
1307 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1310 - Instance failover is now available as a RAPI resource
1311 (``/2/instances/[instance_name]/failover``).
1312 - ``gnt-instance info`` defaults to static information if primary node
1314 - Opcodes have a new ``comment`` attribute.
1315 - Added basic SPICE support to KVM hypervisor.
1316 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1318 Node group improvements
1319 ~~~~~~~~~~~~~~~~~~~~~~~
1321 - ``gnt-cluster verify`` has been modified to check groups separately,
1322 thereby improving performance.
1323 - Node group support has been added to ``gnt-cluster verify-disks``,
1324 which now operates per node group.
1325 - Watcher has been changed to work better with node groups.
1327 - One process and state file per node group.
1328 - Slow watcher in one group doesn't block other group's watcher.
1330 - Added new command, ``gnt-group evacuate``, to move all instances in a
1331 node group to other groups.
1332 - Added ``gnt-instance change-group`` to move an instance to another
1334 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1335 per-group operations.
1336 - Node groups can be tagged.
1337 - Some operations switch from an exclusive to a shared lock as soon as
1339 - Instance's primary and secondary nodes' groups are now available as
1340 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1341 and ``snodes.group.uuid``).
1346 - Numerous updates to documentation and manpages.
1348 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1350 - Some opcode/job results are now also documented, see :doc:`RAPI
1353 - A lockset's internal lock is now also visible in lock monitor.
1354 - Log messages from job queue workers now contain information about the
1355 opcode they're processing.
1356 - ``gnt-instance console`` no longer requires the instance lock.
1357 - A short delay when waiting for job changes reduces the number of LUXI
1358 requests significantly.
1359 - DRBD metadata volumes are overwritten with zeros during disk creation.
1360 - Out-of-band commands no longer acquire the cluster lock in exclusive
1362 - ``devel/upload`` now uses correct permissions for directories.
1368 *(Released Fri, 23 Mar 2012)*
1370 This was the sixth release candidate of the 2.5 series.
1376 *(Released Mon, 9 Jan 2012)*
1378 This was the fifth release candidate of the 2.5 series.
1384 *(Released Thu, 27 Oct 2011)*
1386 This was the fourth release candidate of the 2.5 series.
1392 *(Released Wed, 26 Oct 2011)*
1394 This was the third release candidate of the 2.5 series.
1400 *(Released Tue, 18 Oct 2011)*
1402 This was the second release candidate of the 2.5 series.
1408 *(Released Tue, 4 Oct 2011)*
1410 This was the first release candidate of the 2.5 series.
1416 *(Released Wed, 31 Aug 2011)*
1418 This was the third beta release of the 2.5 series.
1424 *(Released Mon, 22 Aug 2011)*
1426 This was the second beta release of the 2.5 series.
1432 *(Released Fri, 12 Aug 2011)*
1434 This was the first beta release of the 2.5 series.
1440 *(Released Thu, 27 Oct 2011)*
1442 - Fixed bug when parsing command line parameter values ending in
1444 - Fixed assertion error after unclean master shutdown
1445 - Disable HTTP client pool for RPC, significantly reducing memory usage
1447 - Fixed queue archive creation with wrong permissions
1453 *(Released Tue, 23 Aug 2011)*
1457 - Fixed documentation for importing with ``--src-dir`` option
1458 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1459 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1465 *(Released Fri, 5 Aug 2011)*
1467 Many bug-fixes and a few small features:
1469 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1470 issues when you tried to add an instance with two MAC addresses in one
1472 - KVM: fixed per-instance stored UID value
1473 - KVM: configure bridged NICs at migration start
1474 - KVM: Fix a bug where instance will not start with never KVM versions
1476 - Added OS search path to ``gnt-cluster info``
1477 - Fixed an issue with ``file_storage_dir`` where you were forced to
1478 provide an absolute path, but the documentation states it is a
1479 relative path, the documentation was right
1480 - Added a new parameter to instance stop/start called ``--no-remember``
1481 that will make the state change to not be remembered
1482 - Implemented ``no_remember`` at RAPI level
1483 - Improved the documentation
1484 - Node evacuation: don't call IAllocator if node is already empty
1485 - Fixed bug in DRBD8 replace disks on current nodes
1486 - Fixed bug in recreate-disks for DRBD instances
1487 - Moved assertion checking locks in ``gnt-instance replace-disks``
1488 causing it to abort with not owning the right locks for some situation
1489 - Job queue: Fixed potential race condition when cancelling queued jobs
1490 - Fixed off-by-one bug in job serial generation
1491 - ``gnt-node volumes``: Fix instance names
1492 - Fixed aliases in bash completion
1493 - Fixed a bug in reopening log files after being sent a SIGHUP
1494 - Added a flag to burnin to allow specifying VCPU count
1495 - Bugfixes to non-root Ganeti configuration
1501 *(Released Thu, 12 May 2011)*
1503 Many bug-fixes and a few new small features:
1505 - Fixed a bug related to log opening failures
1506 - Fixed a bug in instance listing with orphan instances
1507 - Fixed a bug which prevented resetting the cluster-level node parameter
1508 ``oob_program`` to the default
1509 - Many fixes related to the ``cluster-merge`` tool
1510 - Fixed a race condition in the lock monitor, which caused failures
1511 during (at least) creation of many instances in parallel
1512 - Improved output for gnt-job info
1513 - Removed the quiet flag on some ssh calls which prevented debugging
1515 - Improved the N+1 failure messages in cluster verify by actually
1516 showing the memory values (needed and available)
1517 - Increased lock attempt timeouts so that when executing long operations
1518 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1519 too early and thus prevent the use of the 'fair' mechanism
1520 - Changed instance query data (``gnt-instance info``) to not acquire
1521 locks unless needed, thus allowing its use on locked instance if only
1522 static information is asked for
1523 - Improved behaviour with filesystems that do not support rename on an
1525 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1526 kept locks on all nodes during the wipe, which is unneeded
1527 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1528 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1529 10GiB) which caused the wipe to fail right at the end in some cases
1530 - Fixed master IP activation when doing master failover with no-voting
1531 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1532 the master node itself
1533 - Fixed potential data-loss in under disk full conditions, where Ganeti
1534 wouldn't check correctly the return code and would consider
1535 partially-written files 'correct'
1536 - Fixed bug related to multiple VGs and DRBD disk replacing
1537 - Added new disk parameter ``metavg`` that allows placement of the meta
1538 device for DRBD in a different volume group
1539 - Fixed error handling in the node daemon when the system libc doesn't
1540 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1541 - Fixed lock release during replace-disks, which kept cluster-wide locks
1542 when doing disk replaces with an iallocator script
1543 - Added check for missing bridges in cluster verify
1544 - Handle EPIPE errors while writing to the terminal better, so that
1545 piping the output to e.g. ``less`` doesn't cause a backtrace
1546 - Fixed rare case where a ^C during Luxi calls could have been
1547 interpreted as server errors, instead of simply terminating
1548 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1550 - Added a few more parameters to the KVM hypervisor, allowing a second
1551 CDROM, custom disk type for CDROMs and a floppy image
1552 - Removed redundant message in instance rename when the name is given
1554 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1555 disks on new nodes, allowing recreation when the original instance
1556 nodes are completely gone
1557 - Added option when converting disk templates to DRBD to skip waiting
1558 for the resync, in order to make the instance available sooner
1559 - Added two new variables to the OS scripts environment (containing the
1561 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1562 to allow use of ``pvgrub`` as bootloader
1563 - Changed the instance memory modifications to only check out-of-memory
1564 conditions on memory increases, and turned the secondary node warnings
1565 into errors (they can still be overridden via ``--force``)
1566 - Fixed the handling of a corner case when the Python installation gets
1567 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1568 to execute a command that doesn't exist
1569 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1570 node of the instance returned failures during instance shutdown; this
1571 adds the option ``--ignore-consistency`` to gnt-instance move
1573 And as usual, various improvements to the error messages, documentation
1580 *(Released Wed, 09 Mar 2011)*
1582 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1583 the RAPI users file if run twice (even with ``--dry-run``).
1585 The release fixes that bug (nothing else changed).
1591 *(Released Mon, 07 Mar 2011)*
1593 Final 2.4.0 release. Just a few small fixes:
1595 - Fixed RAPI node evacuate
1596 - Fixed the kvm-ifup script
1597 - Fixed internal error handling for special job cases
1598 - Updated man page to specify the escaping feature for options
1604 *(Released Mon, 28 Feb 2011)*
1606 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1607 that this feature wiped the disks of the wrong instance, leading to loss
1612 - Fixed title of query field containing instance name
1613 - Expanded the glossary in the documentation
1614 - Fixed one unittest (internal issue)
1620 *(Released Mon, 21 Feb 2011)*
1622 A number of bug fixes plus just a couple functionality changes.
1624 On the user-visible side, the ``gnt-* list`` command output has changed
1625 with respect to "special" field states. The current rc1 style of display
1626 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1627 the default output mode special fields states are displayed as follows:
1629 - Offline resource: ``*``
1630 - Unavailable/not applicable: ``-``
1631 - Data missing (RPC failure): ``?``
1632 - Unknown field: ``??``
1634 Another user-visible change is the addition of ``--force-join`` to
1639 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1640 - Fixed regression in RAPI/instance reinstall where all parameters were
1641 required (instead of optional)
1642 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1643 - Fixed iallocator usage (offline nodes were not considered offline)
1644 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1645 - Fixed hypervisor and OS parameter validation with respect to
1646 non-vm_capable nodes
1647 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1649 - Fixed ``tools/listrunner`` with respect to agent-based usage
1655 *(Released Fri, 4 Feb 2011)*
1657 Many changes and fixes since the beta1 release. While there were some
1658 internal changes, the code has been mostly stabilised for the RC
1661 Note: the dumb allocator was removed in this release, as it was not kept
1662 up-to-date with the IAllocator protocol changes. It is recommended to
1663 use the ``hail`` command from the ganeti-htools package.
1665 Note: the 2.4 and up versions of Ganeti are not compatible with the
1666 0.2.x branch of ganeti-htools. You need to upgrade to
1667 ganeti-htools-0.3.0 (or later).
1669 Regressions fixed from 2.3
1670 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1672 - Fixed the ``gnt-cluster verify-disks`` command
1673 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1675 - Fixed disk adoption breakage
1676 - Fixed wrong headers in instance listing for field aliases
1681 - Fixed corner case in KVM handling of NICs
1682 - Fixed many cases of wrong handling of non-vm_capable nodes
1683 - Fixed a bug where a missing instance symlink was not possible to
1684 recreate with any ``gnt-*`` command (now ``gnt-instance
1685 activate-disks`` does it)
1686 - Fixed the volume group name as reported by ``gnt-cluster
1688 - Increased timeouts for the import-export code, hopefully leading to
1689 fewer aborts due network or instance timeouts
1690 - Fixed bug in ``gnt-node list-storage``
1691 - Fixed bug where not all daemons were started on cluster
1692 initialisation, but only at the first watcher run
1693 - Fixed many bugs in the OOB implementation
1694 - Fixed watcher behaviour in presence of instances with offline
1696 - Fixed instance list output for instances running on the wrong node
1697 - a few fixes to the cluster-merge tool, but it still cannot merge
1698 multi-node groups (currently it is not recommended to use this tool)
1704 - Improved network configuration for the KVM hypervisor
1705 - Added e1000 as a supported NIC for Xen-HVM
1706 - Improved the lvmstrap tool to also be able to use partitions, as
1707 opposed to full disks
1708 - Improved speed of disk wiping (the cluster parameter
1709 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1710 of instance creations
1711 - Added documentation for the OS parameters
1712 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1713 hypervisor is not responding
1714 - Added display of blacklisted and hidden OS information in
1715 ``gnt-cluster info``
1716 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1717 NIC and node parameters, which might create problems with currently
1718 invalid (but undetected) configuration files, but prevents validation
1719 failures when unrelated parameters are modified
1720 - Changed cluster initialisation to wait for the master daemon to become
1722 - Expanded the RAPI interface:
1724 - Added config redistribution resource
1725 - Added activation/deactivation of instance disks
1726 - Added export of console information
1728 - Implemented log file reopening on SIGHUP, which allows using
1729 logrotate(8) for the Ganeti log files
1730 - Added a basic OOB helper script as an example
1736 *(Released Fri, 14 Jan 2011)*
1741 - Fixed timezone issues when formatting timestamps
1742 - Added support for node groups, available via ``gnt-group`` and other
1744 - Added out-of-band framework and management, see :doc:`design
1745 document <design-oob>`
1746 - Removed support for roman numbers from ``gnt-node list`` and
1747 ``gnt-instance list``.
1748 - Allowed modification of master network interface via ``gnt-cluster
1749 modify --master-netdev``
1750 - Accept offline secondaries while shutting down instance disks
1751 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1752 - Added support for multiple LVM volume groups
1753 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1755 - Added commands to list available fields:
1757 - ``gnt-node list-fields``
1758 - ``gnt-group list-fields``
1759 - ``gnt-instance list-fields``
1761 - Updated documentation and man pages
1766 - Moved ``rapi_users`` file into separate directory, now named
1767 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1769 - Added new tool for running commands on many machines,
1770 ``tools/ganeti-listrunner``
1771 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1772 improving the ``gnt-instance console`` output
1773 - Allowed customisation of disk index separator at ``configure`` time
1774 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1775 - Added support for non-partitioned md disks in ``lvmstrap``
1776 - Added script to gracefully power off KVM instances
1777 - Split ``utils`` module into smaller parts
1778 - Changed query operations to return more detailed information, e.g.
1779 whether an information is unavailable due to an offline node. To use
1780 this new functionality, the LUXI call ``Query`` must be used. Field
1781 information is now stored by the master daemon and can be retrieved
1782 using ``QueryFields``. Instances, nodes and groups can also be queried
1783 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1784 exposed via RAPI). The following commands make use of this
1785 infrastructure change:
1787 - ``gnt-group list``
1788 - ``gnt-group list-fields``
1790 - ``gnt-node list-fields``
1791 - ``gnt-instance list``
1792 - ``gnt-instance list-fields``
1793 - ``gnt-debug locks``
1798 - New RAPI resources (see :doc:`rapi`):
1802 - ``/2/groups/[group_name]``
1803 - ``/2/groups/[group_name]/assign-nodes``
1804 - ``/2/groups/[group_name]/modify``
1805 - ``/2/groups/[group_name]/rename``
1806 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1810 - Implemented ``no_install`` for instance creation
1811 - Implemented OS parameters for instance reinstallation, allowing
1812 use of special settings on reinstallation (e.g. for preserving data)
1817 - Added IPv6 support in import/export
1818 - Pause DRBD synchronization while wiping disks on instance creation
1819 - Updated unittests and QA scripts
1820 - Improved network parameters passed to KVM
1821 - Converted man pages from docbook to reStructuredText
1827 *(Released Mon, 20 Dec 2010)*
1829 Released version 2.3.1~rc1 without any changes.
1835 *(Released Wed, 1 Dec 2010)*
1837 - impexpd: Disable OpenSSL compression in socat if possible (backport
1838 from master, commit e90739d625b, see :doc:`installation guide
1839 <install-quick>` for details)
1840 - Changed unittest coverage report to exclude test scripts
1841 - Added script to check version format
1847 *(Released Wed, 1 Dec 2010)*
1849 Released version 2.3.0~rc1 without any changes.
1855 *(Released Fri, 19 Nov 2010)*
1857 A number of bugfixes and documentation updates:
1859 - Update ganeti-os-interface documentation
1860 - Fixed a bug related to duplicate MACs or similar items which should be
1862 - Fix breakage in OS state modify
1863 - Reinstall instance: disallow offline secondaries (fixes bug related to
1864 OS changing but reinstall failing)
1865 - plus all the other fixes between 2.2.1 and 2.2.2
1871 *(Released Tue, 2 Nov 2010)*
1873 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1874 - Fixed master failover race with watcher
1875 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1877 - Accept previously stopped instance for export with instance removal
1878 - Simplify and extend the environment variables for instance OS scripts
1879 - Added new node flags, ``master_capable`` and ``vm_capable``
1880 - Added optional instance disk wiping prior during allocation. This is a
1881 cluster-wide option and can be set/modified using
1882 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1883 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1884 :doc:`install-quick`
1885 - Added a new watcher option (``--ignore-pause``)
1886 - Added option to ignore offline node on instance start/stop
1887 (``--ignore-offline``)
1888 - Allow overriding OS parameters with ``gnt-instance reinstall``
1889 - Added ability to change node's secondary IP address using ``gnt-node
1891 - Implemented privilege separation for all daemons except
1892 ``ganeti-noded``, see ``configure`` options
1893 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1895 - Implemented job priorities (see ``ganeti(7)`` manpage)
1896 - Ignore failures while shutting down instances during failover from
1898 - Exit daemon's bootstrap process only once daemon is ready
1899 - Export more information via ``LUInstanceQuery``/remote API
1900 - Improved documentation, QA and unittests
1901 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1902 restart if the file was created or changed
1903 - Added LUXI protocol version sent with each request and response,
1904 allowing detection of server/client mismatches
1905 - Moved the Python scripts among gnt-* and ganeti-* into modules
1906 - Moved all code related to setting up SSH to an external script,
1908 - Infrastructure changes for node group support in future versions
1914 *(Released Fri, 19 Nov 2010)*
1916 A few small bugs fixed, and some improvements to the build system:
1918 - Fix documentation regarding conversion to drbd
1919 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1921 - Fix error handling in node modify with multiple changes
1922 - Allow remote imports without checked names
1928 *(Released Tue, 19 Oct 2010)*
1930 - Disable SSL session ID cache in RPC client
1936 *(Released Thu, 14 Oct 2010)*
1938 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1939 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1941 - Fix problems with interaction between Curl and Python's HTTP server,
1942 resulting in increased speed in many RPC calls
1943 - Improve our release script to prevent breakage with older aclocal and
1950 *(Released Thu, 7 Oct 2010)*
1952 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1953 value retrieved from master
1954 - Added support for blacklisted or hidden OS definitions
1955 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1956 - Added support for -mem-path in KVM hypervisor abstraction layer
1957 - Allow overriding instance parameters in tool for inter-cluster
1958 instance moves (``tools/move-instance``)
1959 - Improved opcode summaries (e.g. in ``gnt-job list``)
1960 - Improve consistency of OS listing by sorting it
1961 - Documentation updates
1967 *(Released Fri, 8 Oct 2010)*
1969 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1975 *(Released Mon, 4 Oct 2010)*
1977 - Fixed regression in ``gnt-instance rename``
1983 *(Released Wed, 22 Sep 2010)*
1985 - Fixed OS_VARIANT variable for OS scripts
1986 - Fixed cluster tag operations via RAPI
1987 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1988 - Disabled RAPI CA checks in watcher
1994 *(Released Mon, 23 Aug 2010)*
1996 - Support DRBD versions of the format "a.b.c.d"
1998 - Re-introduce support for usage from multiple threads in RAPI client
1999 - Instance renames and modify via RAPI
2000 - Work around race condition between processing and archival in job
2002 - Mark opcodes following failed one as failed, too
2003 - Job field ``lock_status`` was removed due to difficulties making it
2004 work with the changed job queue in Ganeti 2.2; a better way to monitor
2005 locks is expected for a later 2.2.x release
2006 - Fixed dry-run behaviour with many commands
2007 - Support ``ssh-agent`` again when adding nodes
2008 - Many additional bugfixes
2014 *(Released Fri, 30 Jul 2010)*
2016 Important change: the internal RPC mechanism between Ganeti nodes has
2017 changed from using a home-grown http library (based on the Python base
2018 libraries) to use the PycURL library. This requires that PycURL is
2019 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2020 against GnuTLS by default. cURL's support for GnuTLS had known issues
2021 before cURL 7.21.0 and we recommend using the latest cURL release or
2022 linking against OpenSSL. Most other distributions already link PycURL
2023 and cURL against OpenSSL. The command::
2025 python -c 'import pycurl; print pycurl.version'
2027 can be used to determine the libraries PycURL and cURL are linked
2030 Other significant changes:
2032 - Rewrote much of the internals of the job queue, in order to achieve
2033 better parallelism; this decouples job query operations from the job
2034 processing, and it should allow much nicer behaviour of the master
2035 daemon under load, and it also has uncovered some long-standing bugs
2036 related to the job serialisation (now fixed)
2037 - Added a default iallocator setting to the cluster parameters,
2038 eliminating the need to always pass nodes or an iallocator for
2039 operations that require selection of new node(s)
2040 - Added experimental support for the LXC virtualization method
2041 - Added support for OS parameters, which allows the installation of
2042 instances to pass parameter to OS scripts in order to customise the
2044 - Added a hypervisor parameter controlling the migration type (live or
2045 non-live), since hypervisors have various levels of reliability; this
2046 has renamed the 'live' parameter to 'mode'
2047 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
2048 logical volumes, meaning that cluster verify will ignore them and not
2049 flag their presence as errors
2050 - The watcher will now reset the error count for failed instances after
2051 8 hours, thus allowing self-healing if the problem that caused the
2052 instances to be down/fail to start has cleared in the meantime
2053 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2054 check for, and warn, if the drbd module parameter ``usermode_helper``
2055 is not consistent with the cluster-wide setting; this is needed to
2056 make diagnose easier of failed drbd creations
2057 - Started adding base IPv6 support, but this is not yet
2058 enabled/available for use
2059 - Rename operations (cluster, instance) will now return the new name,
2060 which is especially useful if a short name was passed in
2061 - Added support for instance migration in RAPI
2062 - Added a tool to pre-configure nodes for the SSH setup, before joining
2063 them to the cluster; this will allow in the future a simplified model
2064 for node joining (but not yet fully enabled in 2.2); this needs the
2065 paramiko python library
2066 - Fixed handling of name-resolving errors
2067 - Fixed consistency of job results on the error path
2068 - Fixed master-failover race condition when executed multiple times in
2070 - Fixed many bugs related to the job queue (mostly introduced during the
2071 2.2 development cycle, so not all are impacting 2.1)
2072 - Fixed instance migration with missing disk symlinks
2073 - Fixed handling of unknown jobs in ``gnt-job archive``
2074 - And many other small fixes/improvements
2078 - Enhanced both the unittest and the QA coverage
2079 - Switched the opcode validation to a generic model, and extended the
2080 validation to all opcode parameters
2081 - Changed more parts of the code that write shell scripts to use the
2083 - Switched the master daemon to use the asyncore library for the Luxi
2090 *(Released Thu, 17 Jun 2010)*
2092 - Added tool (``move-instance``) and infrastructure to move instances
2093 between separate clusters (see :doc:`separate documentation
2094 <move-instance>` and :doc:`design document <design-2.2>`)
2095 - Added per-request RPC timeout
2096 - RAPI now requires a Content-Type header for requests with a body (e.g.
2097 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2098 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2099 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2101 - Implemented initial support for running Ganeti daemons as separate
2102 users, see configure-time flags ``--with-user-prefix`` and
2103 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2105 - Instances can be removed after export (``gnt-backup export
2106 --remove-instance``)
2107 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2108 key (instead of 1024 bit)
2109 - Added new cluster configuration file for cluster domain secret
2110 - Import/export now use SSL instead of SSH
2111 - Added support for showing estimated time when exporting an instance,
2112 see the ``ganeti-os-interface(7)`` manpage and look for
2119 *(Released Tue, 16 Nov 2010)*
2121 Some more bugfixes. Unless critical bugs occur, this will be the last
2124 - Fix case of MAC special-values
2125 - Fix mac checker regex
2126 - backend: Fix typo causing "out of range" error
2127 - Add missing --units in gnt-instance list man page
2133 *(Released Tue, 24 Aug 2010)*
2136 - Don't ignore secondary node silently on non-mirrored disk templates
2138 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2139 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2140 - Properly document the usb_mouse parameter
2141 - Fix path in ganeti-rapi(8) (issue 116)
2142 - Adjust error message when the ganeti user's .ssh directory is
2144 - Add same-node-check when changing the disk template to drbd
2150 *(Released Fri, 16 Jul 2010)*
2153 - Add an option to only select some reboot types during qa/burnin.
2154 (on some hypervisors consequent reboots are not supported)
2155 - Fix infrequent race condition in master failover. Sometimes the old
2156 master ip address would be still detected as up for a short time
2157 after it was removed, causing failover to fail.
2158 - Decrease mlockall warnings when the ctypes module is missing. On
2159 Python 2.4 we support running even if no ctypes module is installed,
2160 but we were too verbose about this issue.
2161 - Fix building on old distributions, on which man doesn't have a
2163 - Fix RAPI not to ignore the MAC address on instance creation
2164 - Implement the old instance creation format in the RAPI client.
2170 *(Released Thu, 01 Jul 2010)*
2172 A small bugfix release:
2173 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2174 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2175 a non-existing option
2176 - Fix instance create: the --force-variant option was ignored
2177 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2178 - Fix modify node storage with non-FQDN arguments
2179 - Fix RAPI client to authenticate under Python 2.6 when used
2180 for more than 5 requests needing authentication
2181 - Fix gnt-instance modify -t (storage) giving a wrong error message
2182 when converting a non-shutdown drbd instance to plain
2188 *(Released Fri, 18 Jun 2010)*
2190 A small bugfix release:
2192 - Fix live migration of KVM instances started with older Ganeti
2193 versions which had fewer hypervisor parameters
2194 - Fix gnt-instance grow-disk on down instances
2195 - Fix an error-reporting bug during instance migration
2196 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2197 silently ignoring broken ones
2198 - Fix an RPC error reporting bug affecting, for example, RAPI client
2200 - Fix bug triggered by different API version os-es on different nodes
2201 - Fix a bug in instance startup with custom hvparams: OS level
2202 parameters would fail to be applied.
2203 - Fix the RAPI client under Python 2.6 (but more work is needed to
2204 make it work completely well with OpenSSL)
2205 - Fix handling of errors when resolving names from DNS
2211 *(Released Thu, 3 Jun 2010)*
2213 A medium sized development cycle. Some new features, and some
2214 fixes/small improvements/cleanups.
2216 Significant features
2217 ~~~~~~~~~~~~~~~~~~~~
2219 The node deamon now tries to mlock itself into memory, unless the
2220 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2221 its logs, and falls back to console logging. This allows emergency
2222 features such as ``gnt-node powercycle`` to work even in the event of a
2223 broken node disk (tested offlining the disk hosting the node's
2224 filesystem and dropping its memory caches; don't try this at home)
2226 KVM: add vhost-net acceleration support. It can be tested with a new
2227 enough version of the kernel and of qemu-kvm.
2229 KVM: Add instance chrooting feature. If you use privilege dropping for
2230 your VMs you can also now force them to chroot to an empty directory,
2231 before starting the emulated guest.
2233 KVM: Add maximum migration bandwith and maximum downtime tweaking
2234 support (requires a new-enough version of qemu-kvm).
2236 Cluster verify will now warn if the master node doesn't have the master
2237 ip configured on it.
2239 Add a new (incompatible) instance creation request format to RAPI which
2240 supports all parameters (previously only a subset was supported, and it
2241 wasn't possible to extend the old format to accomodate all the new
2242 features. The old format is still supported, and a client can check for
2243 this feature, before using it, by checking for its presence in the
2244 ``features`` RAPI resource.
2246 Now with ancient latin support. Try it passing the ``--roman`` option to
2247 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2248 (requires the python-roman module to be installed, in order to work).
2253 As usual many internal code refactorings, documentation updates, and
2256 - Lots of improvements and cleanups to the experimental Remote API
2257 (RAPI) client library.
2258 - A new unit test suite for the core daemon libraries.
2259 - A fix to creating missing directories makes sure the umask is not
2260 applied anymore. This enforces the same directory permissions
2262 - Better handling terminating daemons with ctrl+c (used when running
2263 them in debugging mode).
2264 - Fix a race condition in live migrating a KVM instance, when stat()
2265 on the old proc status file returned EINVAL, which is an unexpected
2267 - Fixed manpage checking with newer man and utf-8 charachters. But now
2268 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2274 *(Released Fri, 7 May 2010)*
2276 Fix a bug which prevented untagged KVM instances from starting.
2282 *(Released Fri, 7 May 2010)*
2284 Another release with a long development cycle, during which many
2285 different features were added.
2287 Significant features
2288 ~~~~~~~~~~~~~~~~~~~~
2290 The KVM hypervisor now can run the individual instances as non-root, to
2291 reduce the impact of a VM being hijacked due to bugs in the
2292 hypervisor. It is possible to run all instances as a single (non-root)
2293 user, to manually specify a user for each instance, or to dynamically
2294 allocate a user out of a cluster-wide pool to each instance, with the
2295 guarantee that no two instances will run under the same user ID on any
2298 An experimental RAPI client library, that can be used standalone
2299 (without the other Ganeti libraries), is provided in the source tree as
2300 ``lib/rapi/client.py``. Note this client might change its interface in
2301 the future, as we iterate on its capabilities.
2303 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2304 replace the cluster's certificates and crypto keys. This might help in
2305 case they have been compromised, or have simply expired.
2307 A new disk option for instance creation has been added that allows one
2308 to "adopt" currently existing logical volumes, with data
2309 preservation. This should allow easier migration to Ganeti from
2310 unmanaged (or managed via other software) instances.
2312 Another disk improvement is the possibility to convert between redundant
2313 (DRBD) and plain (LVM) disk configuration for an instance. This should
2314 allow better scalability (starting with one node and growing the
2315 cluster, or shrinking a two-node cluster to one node).
2317 A new feature that could help with automated node failovers has been
2318 implemented: if a node sees itself as offline (by querying the master
2319 candidates), it will try to shutdown (hard) all instances and any active
2320 DRBD devices. This reduces the risk of duplicate instances if an
2321 external script automatically failovers the instances on such nodes. To
2322 enable this, the cluster parameter ``maintain_node_health`` should be
2323 enabled; in the future this option (per the name) will enable other
2324 automatic maintenance features.
2326 Instance export/import now will reuse the original instance
2327 specifications for all parameters; that means exporting an instance,
2328 deleting it and the importing it back should give an almost identical
2329 instance. Note that the default import behaviour has changed from
2330 before, where it created only one NIC; now it recreates the original
2333 Cluster verify has added a few new checks: SSL certificates validity,
2334 /etc/hosts consistency across the cluster, etc.
2339 As usual, many internal changes were done, documentation fixes,
2342 - Fixed cluster initialization with disabled cluster storage (regression
2343 introduced in 2.1.1)
2344 - File-based storage supports growing the disks
2345 - Fixed behaviour of node role changes
2346 - Fixed cluster verify for some corner cases, plus a general rewrite of
2347 cluster verify to allow future extension with more checks
2348 - Fixed log spamming by watcher and node daemon (regression introduced
2350 - Fixed possible validation issues when changing the list of enabled
2352 - Fixed cleanup of /etc/hosts during node removal
2353 - Fixed RAPI response for invalid methods
2354 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2355 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2356 from ide disks, etc.)
2357 - Allow OS changes without re-installation (to record a changed OS
2358 outside of Ganeti, or to allow OS renames)
2359 - Allow instance creation without OS installation (useful for example if
2360 the OS will be installed manually, or restored from a backup not in
2362 - Implemented option to make cluster ``copyfile`` use the replication
2364 - Added list of enabled hypervisors to ssconf (possibly useful for
2366 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2368 - A partial form of node re-IP is possible via node readd, which now
2369 allows changed node primary IP
2370 - Command line utilities now show an informational message if the job is
2372 - The logs of the master daemon now show the PID/UID/GID of the
2379 *(Released Fri, 12 Mar 2010)*
2381 During the 2.1.0 long release candidate cycle, a lot of improvements and
2382 changes have accumulated with were released later as 2.1.1.
2387 The node evacuate command (``gnt-node evacuate``) was significantly
2388 rewritten, and as such the IAllocator protocol was changed - a new
2389 request type has been added. This unfortunate change during a stable
2390 series is designed to improve performance of node evacuations; on
2391 clusters with more than about five nodes and which are well-balanced,
2392 evacuation should proceed in parallel for all instances of the node
2393 being evacuated. As such, any existing IAllocator scripts need to be
2394 updated, otherwise the above command will fail due to the unknown
2395 request. The provided "dumb" allocator has not been updated; but the
2396 ganeti-htools package supports the new protocol since version 0.2.4.
2398 Another important change is increased validation of node and instance
2399 names. This might create problems in special cases, if invalid host
2400 names are being used.
2402 Also, a new layer of hypervisor parameters has been added, that sits at
2403 OS level between the cluster defaults and the instance ones. This allows
2404 customisation of virtualization parameters depending on the installed
2405 OS. For example instances with OS 'X' may have a different KVM kernel
2406 (or any other parameter) than the cluster defaults. This is intended to
2407 help managing a multiple OSes on the same cluster, without manual
2408 modification of each instance's parameters.
2410 A tool for merging clusters, ``cluster-merge``, has been added in the
2411 tools sub-directory.
2416 - Improved the int/float conversions that should make the code more
2417 robust in face of errors from the node daemons
2418 - Fixed the remove node code in case of internal configuration errors
2419 - Fixed the node daemon behaviour in face of inconsistent queue
2420 directory (e.g. read-only file-system where we can't open the files
2422 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2423 now it either aborts cleanly or, if given the new "auto_promote"
2424 parameter, will automatically promote other nodes as needed
2425 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2426 completely prevent Ganeti from working
2427 - Fixed bug for instance export when not all disks were successfully
2429 - Fixed behaviour of node add when the new node is slow in starting up
2431 - Fixed handling of signals in the LUXI client, which should improve
2432 behaviour of command-line scripts
2433 - Added checks for invalid node/instance names in the configuration (now
2434 flagged during cluster verify)
2435 - Fixed watcher behaviour for disk activation errors
2436 - Fixed two potentially endless loops in http library, which led to the
2437 RAPI daemon hanging and consuming 100% CPU in some cases
2438 - Fixed bug in RAPI daemon related to hashed passwords
2439 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2441 - Enhanced compatibility with non-Debian OSes, but not using absolute
2442 path in some commands and allowing customisation of the ssh
2443 configuration directory
2444 - Fixed possible future issue with new Python versions by abiding to the
2445 proper use of ``__slots__`` attribute on classes
2446 - Added checks that should prevent directory traversal attacks
2447 - Many documentation fixes based on feedback from users
2452 - Added an "early_release" more for instance replace disks and node
2453 evacuate, where we release locks earlier and thus allow higher
2454 parallelism within the cluster
2455 - Added watcher hooks, intended to allow the watcher to restart other
2456 daemons (e.g. from the ganeti-nbma project), but they can be used of
2457 course for any other purpose
2458 - Added a compile-time disable for DRBD barriers, to increase
2459 performance if the administrator trusts the power supply or the
2460 storage system to not lose writes
2461 - Added the option of using syslog for logging instead of, or in
2462 addition to, Ganeti's own log files
2463 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2464 can indeed boot from a paravirtual NIC
2465 - Added a generic debug level for many operations; while this is not
2466 used widely yet, it allows one to pass the debug value all the way to
2468 - Enhanced the hooks environment for instance moves (failovers,
2469 migrations) where the primary/secondary nodes changed during the
2470 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2471 - Enhanced data validations for many user-supplied values; one important
2472 item is the restrictions imposed on instance and node names, which
2473 might reject some (invalid) host names
2474 - Add a configure-time option to disable file-based storage, if it's not
2475 needed; this allows greater security separation between the master
2476 node and the other nodes from the point of view of the inter-node RPC
2478 - Added user notification in interactive tools if job is waiting in the
2479 job queue or trying to acquire locks
2480 - Added log messages when a job is waiting for locks
2481 - Added filtering by node tags in instance operations which admit
2482 multiple instances (start, stop, reboot, reinstall)
2483 - Added a new tool for cluster mergers, ``cluster-merge``
2484 - Parameters from command line which are of the form ``a=b,c=d`` can now
2485 use backslash escapes to pass in values which contain commas,
2486 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2488 - For KVM, the instance name is the first parameter passed to KVM, so
2489 that it's more visible in the process list
2495 *(Released Tue, 2 Mar 2010)*
2497 Ganeti 2.1 brings many improvements with it. Major changes:
2499 - Added infrastructure to ease automated disk repairs
2500 - Added new daemon to export configuration data in a cheaper way than
2501 using the remote API
2502 - Instance NICs can now be routed instead of being associated with a
2504 - Improved job locking logic to reduce impact of jobs acquiring multiple
2505 locks waiting for other long-running jobs
2507 In-depth implementation details can be found in the Ganeti 2.1 design
2513 - Added chroot hypervisor
2514 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2516 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2517 ``bootloader_path`` and ``bootloader_args``)
2518 - Added the ``use_localtime`` option for the xen-hvm and kvm
2519 hypervisors, and the default value for this has changed to false (in
2520 2.0 xen-hvm always enabled it)
2521 - Added luxi call to submit multiple jobs in one go
2522 - Added cluster initialization option to not modify ``/etc/hosts``
2524 - Added network interface parameters
2525 - Added dry run mode to some LUs
2526 - Added RAPI resources:
2528 - ``/2/instances/[instance_name]/info``
2529 - ``/2/instances/[instance_name]/replace-disks``
2530 - ``/2/nodes/[node_name]/evacuate``
2531 - ``/2/nodes/[node_name]/migrate``
2532 - ``/2/nodes/[node_name]/role``
2533 - ``/2/nodes/[node_name]/storage``
2534 - ``/2/nodes/[node_name]/storage/modify``
2535 - ``/2/nodes/[node_name]/storage/repair``
2537 - Added OpCodes to evacuate or migrate all instances on a node
2538 - Added new command to list storage elements on nodes (``gnt-node
2539 list-storage``) and modify them (``gnt-node modify-storage``)
2540 - Added new ssconf files with master candidate IP address
2541 (``ssconf_master_candidates_ips``), node primary IP address
2542 (``ssconf_node_primary_ips``) and node secondary IP address
2543 (``ssconf_node_secondary_ips``)
2544 - Added ``ganeti-confd`` and a client library to query the Ganeti
2545 configuration via UDP
2546 - Added ability to run hooks after cluster initialization and before
2548 - Added automatic mode for disk replace (``gnt-instance replace-disks
2550 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2551 after catastrophic data-loss
2552 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2554 - Added ``gnt-instance move`` command to move instances
2555 - Added ``gnt-cluster watcher`` command to control watcher
2556 - Added ``gnt-node powercycle`` command to powercycle nodes
2557 - Added new job status field ``lock_status``
2558 - Added parseable error codes to cluster verification (``gnt-cluster
2559 verify --error-codes``) and made output less verbose (use
2560 ``--verbose`` to restore previous behaviour)
2561 - Added UUIDs to the main config entities (cluster, nodes, instances)
2562 - Added support for OS variants
2563 - Added support for hashed passwords in the Ganeti remote API users file
2565 - Added option to specify maximum timeout on instance shutdown
2566 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2567 - Added new helper script to start and stop Ganeti daemons
2568 (``daemon-util``), with the intent to reduce the work necessary to
2569 adjust Ganeti for non-Debian distributions and to start/stop daemons
2571 - Added more unittests
2572 - Fixed critical bug in ganeti-masterd startup
2573 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2574 now customisable at the cluster level for both the KVM and Xen
2575 hypervisors using the new ``migration_port`` parameter
2576 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2577 reinstalling an instance
2578 - Allowed ``@`` in tag names
2579 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2580 - Many documentation updates
2581 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2582 - ``gnt-instance reinstall`` can now reinstall multiple instances
2583 - Updated many command line parameters
2584 - Introduced new OS API version 15
2585 - No longer support a default hypervisor
2586 - Treat virtual LVs as inexistent
2587 - Improved job locking logic to reduce lock contention
2588 - Match instance and node names case insensitively
2589 - Reimplemented bash completion script to be more complete
2596 *(Released Thu, 4 Feb 2010)*
2598 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2599 - Fix a string formatting bug
2600 - Improve safety of the code in some error paths
2601 - Improve data validation in the master of values returned from nodes
2607 *(Released Thu, 17 Dec 2009)*
2609 - Fix security issue due to missing validation of iallocator names; this
2610 allows local and remote execution of arbitrary executables
2611 - Fix failure of gnt-node list during instance removal
2612 - Ship the RAPI documentation in the archive
2618 *(Released Wed, 30 Sep 2009)*
2620 - Fixed many wrong messages
2621 - Fixed a few bugs related to the locking library
2622 - Fixed MAC checking at instance creation time
2623 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2624 - Fixed a few issues related to signal handling in both daemons and
2626 - Fixed the example startup script provided
2627 - Fixed insserv dependencies in the example startup script (patch from
2629 - Fixed handling of drained nodes in the iallocator framework
2630 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2632 - Fixed error related to invalid job IDs in job polling
2633 - Fixed job/opcode persistence on unclean master shutdown
2634 - Fixed handling of partial job processing after unclean master
2636 - Fixed error reporting from LUs, previously all errors were converted
2637 into execution errors
2638 - Fixed error reporting from burnin
2639 - Decreased significantly the memory usage of the job queue
2640 - Optimised slightly multi-job submission
2641 - Optimised slightly opcode loading
2642 - Backported the multi-job submit framework from the development
2643 branch; multi-instance start and stop should be faster
2644 - Added script to clean archived jobs after 21 days; this will reduce
2645 the size of the queue directory
2646 - Added some extra checks in disk size tracking
2647 - Added an example ethers hook script
2648 - Added a cluster parameter that prevents Ganeti from modifying of
2650 - Added more node information to RAPI responses
2651 - Added a ``gnt-job watch`` command that allows following the ouput of a
2653 - Added a bind-address option to ganeti-rapi
2654 - Added more checks to the configuration verify
2655 - Enhanced the burnin script such that some operations can be retried
2657 - Converted instance reinstall to multi-instance model
2663 *(Released Fri, 7 Aug 2009)*
2665 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2666 to allow using the pre-2.0.2 behaviour in activation, if any existing
2667 instances have mismatched disk sizes in the configuration
2668 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2669 any configuration mismatches for disk sizes
2670 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2671 failover to work on two-node clusters
2672 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2674 - Fixed detection of OS script errors in ``gnt-backup export``
2675 - Fixed exit code of ``gnt-backup export``
2681 *(Released Fri, 17 Jul 2009)*
2683 - Added experimental support for stripped logical volumes; this should
2684 enhance performance but comes with a higher complexity in the block
2685 device handling; stripping is only enabled when passing
2686 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2687 affected even in the non-stripped mode
2688 - Improved resiliency against transient failures at the end of DRBD
2689 resyncs, and in general of DRBD resync checks
2690 - Fixed a couple of issues with exports and snapshot errors
2691 - Fixed a couple of issues in instance listing
2692 - Added display of the disk size in ``gnt-instance info``
2693 - Fixed checking for valid OSes in instance creation
2694 - Fixed handling of the "vcpus" parameter in instance listing and in
2695 general of invalid parameters
2696 - Fixed http server library, and thus RAPI, to handle invalid
2697 username/password combinations correctly; this means that now they
2698 report unauthorized for queries too, not only for modifications,
2699 allowing earlier detect of configuration problems
2700 - Added a new "role" node list field, equivalent to the master/master
2701 candidate/drained/offline flags combinations
2702 - Fixed cluster modify and changes of candidate pool size
2703 - Fixed cluster verify error messages for wrong files on regular nodes
2704 - Fixed a couple of issues with node demotion from master candidate role
2705 - Fixed node readd issues
2706 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2707 - Added a new ``--no-voting`` option for masterfailover to fix failover
2708 on two-nodes clusters when the former master node is unreachable
2709 - Added instance reinstall over RAPI
2715 *(Released Tue, 16 Jun 2009)*
2717 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2718 allow re-adding the start in single-user option (regression from 1.2)
2719 - the watcher writes the instance status to a file, to allow monitoring
2720 to report the instance status (from the master) based on cached
2721 results of the watcher's queries; while this can get stale if the
2722 watcher is being locked due to other work on the cluster, this is
2723 still an improvement
2724 - the watcher now also restarts the node daemon and the rapi daemon if
2726 - fixed the watcher to handle full and drained queue cases
2727 - hooks export more instance data in the environment, which helps if
2728 hook scripts need to take action based on the instance's properties
2729 (no longer need to query back into ganeti)
2730 - instance failovers when the instance is stopped do not check for free
2731 RAM, so that failing over a stopped instance is possible in low memory
2733 - rapi uses queries for tags instead of jobs (for less job traffic), and
2734 for cluster tags it won't talk to masterd at all but read them from
2736 - a couple of error handling fixes in RAPI
2737 - drbd handling: improved the error handling of inconsistent disks after
2738 resync to reduce the frequency of "there are some degraded disks for
2739 this instance" messages
2740 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2741 error handling path called a wrong function name)
2747 *(Released Wed, 27 May 2009)*
2749 - no changes from rc5
2755 *(Released Wed, 20 May 2009)*
2757 - fix a couple of bugs (validation, argument checks)
2758 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2759 - some small improvements to RAPI and IAllocator
2760 - make watcher automatically start the master daemon if down
2766 *(Released Mon, 27 Apr 2009)*
2768 - change the OS list to not require locks; this helps with big clusters
2769 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2770 volume group is broken
2771 - ``gnt-instance info``, without any arguments, doesn't run for all
2772 instances anymore; either pass ``--all`` or pass the desired
2773 instances; this helps against mistakes on big clusters where listing
2774 the information for all instances takes a long time
2775 - miscellaneous doc and man pages fixes
2781 *(Released Wed, 8 Apr 2009)*
2783 - Change the internal locking model of some ``gnt-node`` commands, in
2784 order to reduce contention (and blocking of master daemon) when
2785 batching many creation/reinstall jobs
2786 - Fixes to Xen soft reboot
2787 - No longer build documentation at build time, instead distribute it in
2788 the archive, in order to reduce the need for the whole docbook/rst
2795 *(Released Fri, 27 Mar 2009)*
2797 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2798 - Fix watcher startup sequence, improves the behaviour of busy clusters
2799 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2800 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2801 bind address and other places
2802 - Some documentation fixes and updates
2808 *(Released Mon, 2 Mar 2009)*
2810 - More documentation updates, now all docs should be more-or-less
2812 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2814 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2816 - Fix an issue related to $libdir/run/ganeti and cluster creation
2822 *(Released Thu, 19 Feb 2009)*
2824 - Xen PVM and KVM have switched the default value for the instance root
2825 disk to the first partition on the first drive, instead of the whole
2826 drive; this means that the OS installation scripts must be changed
2828 - Man pages have been updated
2829 - RAPI has been switched by default to HTTPS, and the exported functions
2830 should all work correctly
2831 - RAPI v1 has been removed
2832 - Many improvements to the KVM hypervisor
2833 - Block device errors are now better reported
2834 - Many other bugfixes and small improvements
2840 *(Released Mon, 26 Jan 2009)*
2842 - Version 2 is a general rewrite of the code and therefore the
2843 differences are too many to list, see the design document for 2.0 in
2844 the ``doc/`` subdirectory for more details
2845 - In this beta version there is not yet a migration path from 1.2 (there
2846 will be one in the final 2.0 release)
2847 - A few significant changes are:
2849 - all commands are executed by a daemon (``ganeti-masterd``) and the
2850 various ``gnt-*`` commands are just front-ends to it
2851 - all the commands are entered into, and executed from a job queue,
2852 see the ``gnt-job(8)`` manpage
2853 - the RAPI daemon supports read-write operations, secured by basic
2854 HTTP authentication on top of HTTPS
2855 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2856 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2857 to migrate to DRBD 8 first while still running Ganeti 1.2)
2858 - DRBD devices are using statically allocated minor numbers, which
2859 will be assigned to existing instances during the migration process
2860 - there is support for both Xen PVM and Xen HVM instances running on
2862 - KVM virtualization is supported too
2863 - file-based storage has been implemented, which means that it is
2864 possible to run the cluster without LVM and DRBD storage, for
2865 example using a shared filesystem exported from shared storage (and
2866 still have live migration)
2872 *(Released Tue, 13 Jan 2009)*
2874 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2875 - Reuse the old instance mac address by default on instance import, if
2876 the instance name is the same.
2877 - Handle situations in which the node info rpc returns incomplete
2879 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2880 - Improved version of batcher:
2882 - state file support
2883 - instance mac address support
2884 - support for HVM clusters/instances
2886 - Add an option to show the number of cpu sockets and nodes in
2888 - Support OSes that handle more than one version of the OS api (but do
2889 not change the current API in any other way)
2890 - Fix ``gnt-node migrate``
2891 - ``gnt-debug`` man page
2892 - Fixes various more typos and small issues
2893 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2899 *(Released Wed, 24 Sep 2008)*
2901 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2902 type of disk exported to fully virtualized instances.
2903 - provide access to the serial console of HVM instances
2904 - instance auto_balance flag, set by default. If turned off it will
2905 avoid warnings on cluster verify if there is not enough memory to fail
2906 over an instance. in the future it will prevent automatically failing
2907 it over when we will support that.
2908 - batcher tool for instance creation, see ``tools/README.batcher``
2909 - ``gnt-instance reinstall --select-os`` to interactively select a new
2910 operating system when reinstalling an instance.
2911 - when changing the memory amount on instance modify a check has been
2912 added that the instance will be able to start. also warnings are
2913 emitted if the instance will not be able to fail over, if auto_balance
2915 - documentation fixes
2916 - sync fields between ``gnt-instance list/modify/add/import``
2917 - fix a race condition in drbd when the sync speed was set after giving
2918 the device a remote peer.
2924 *(Released Tue, 22 Jul 2008)*
2926 - note: the allowed size and number of tags per object were reduced
2927 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2928 - fixed twisted 8.x compatibility
2929 - fixed ``gnt-instance replace-disks`` with iallocator
2930 - add TCP keepalives on twisted connections to detect restarted nodes
2931 - disk increase support, see ``gnt-instance grow-disk``
2932 - implement bulk node/instance query for RAPI
2933 - add tags in node/instance listing (optional)
2934 - experimental migration (and live migration) support, read the man page
2935 for ``gnt-instance migrate``
2936 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2937 has some small improvements in handling its state file
2943 *(Released Fri, 13 Jun 2008)*
2945 - Experimental readonly, REST-based remote API implementation;
2946 automatically started on master node, TCP port 5080, if enabled by
2947 ``--enable-rapi`` parameter to configure script.
2948 - Instance allocator support. Add and import instance accept a
2949 ``--iallocator`` parameter, and call that instance allocator to decide
2950 which node to use for the instance. The iallocator document describes
2951 what's expected from an allocator script.
2952 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2953 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2954 node is lost there is still enough memory to fail over the instances
2956 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2957 ``gnt-cluster verify``, to check for site-specific compliance. All the
2958 hooks will run, and their output, if any, will be displayed. Any
2959 failing hook will make the verification return an error value.
2960 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2961 primary and secondary interfaces
2962 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2963 that is still declared as part of the cluster and has failed.
2964 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2965 specifying output fields, that just adds the chosen fields to the
2967 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2968 export from the filesystem.
2969 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2970 have been added. Using them you can enable/disable acpi and pae
2971 support, and specify a path for a cd image to be exported to the
2972 instance. These parameters as the name suggest only work on HVM
2974 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2975 PAE support will be set to the previously hardcoded values, but the
2976 (previously hardcoded) path to the CDROM ISO image will be unset and
2977 if required, needs to be set manually with ``gnt-instance modify``
2979 - The address to which an instance's VNC console is bound is now
2980 selectable per-instance, rather than being cluster wide. Of course
2981 this only applies to instances controlled via VNC, so currently just
2982 applies to HVM clusters.
2988 *(Released Mon, 18 Feb 2008)*
2990 - more tweaks to the disk activation code (especially helpful for DRBD)
2991 - change the default ``gnt-instance list`` output format, now there is
2992 one combined status field (see the manpage for the exact values this
2994 - some more fixes for the mac export to hooks change
2995 - make Ganeti not break with DRBD 8.2.x (which changed the version
2996 format in ``/proc/drbd``) (issue 24)
2997 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2998 template, allowing migration from DRBD0.7+MD to DRBD8
3004 *(Released Wed, 30 Jan 2008)*
3006 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3008 - add command aliases infrastructure and a few aliases
3009 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
3010 man pages and the ``--help`` option of ``gnt-node
3011 list``/``gnt-instance list``
3012 - fix ``gnt-backup list`` with down nodes (issue 21)
3013 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3014 - fix the dist archive and add a check for including svn/git files in
3016 - some developer-related changes: improve the burnin and the QA suite,
3017 add an upload script for testing during development
3023 *(Released Wed, 16 Jan 2008)*
3025 - experimental HVM support, read the install document, section
3026 "Initializing the cluster"
3027 - allow for the PVM hypervisor per-instance kernel and initrd paths
3028 - add a new command ``gnt-cluster verify-disks`` which uses a new
3029 algorithm to improve the reconnection of the DRBD pairs if the device
3030 on the secondary node has gone away
3031 - make logical volume code auto-activate LVs at disk activation time
3032 - slightly improve the speed of activating disks
3033 - allow specification of the MAC address at instance creation time, and
3034 changing it later via ``gnt-instance modify``
3035 - fix handling of external commands that generate lots of output on
3037 - update documentation with regard to minimum version of DRBD8 supported
3043 *(Released Tue, 4 Dec 2007)*
3045 - Log the ``xm create`` output to the node daemon log on failure (to
3046 help diagnosing the error)
3047 - In debug mode, log all external commands output if failed to the logs
3048 - Change parsing of lvm commands to ignore stderr
3054 *(Released Wed, 28 Nov 2007)*
3056 - Another round of updates to the DRBD 8 code to deal with more failures
3057 in the replace secondary node operation
3058 - Some more logging of failures in disk operations (lvm, drbd)
3059 - A few documentation updates
3066 *(Released Tue, 13 Nov 2007)*
3068 - Change configuration file format from Python's Pickle to JSON.
3069 Upgrading is possible using the cfgupgrade utility.
3070 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3071 faster replace disks and is more stable (DRBD 8 has many improvements
3072 compared to DRBD 0.7)
3073 - Added command line tags support (see man pages for ``gnt-instance``,
3074 ``gnt-node``, ``gnt-cluster``)
3075 - Added instance rename support
3076 - Added multi-instance startup/shutdown
3077 - Added cluster rename support
3078 - Added ``gnt-node evacuate`` to simplify some node operations
3079 - Added instance reboot operation that can speedup reboot as compared to
3081 - Soften the requirement that hostnames are in FQDN format
3082 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3084 - Removed dependency on debian's patched fping that uses the
3085 non-standard ``-S`` option
3086 - Now the OS definitions are searched for in multiple, configurable
3087 paths (easier for distros to package)
3088 - Some changes to the hooks infrastructure (especially the new
3089 post-configuration update hook)
3090 - Other small bugfixes
3092 .. vim: set textwidth=72 syntax=rst :