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.
27 - KVM hypervisors can now access RBD storage directly without having to
28 go through a block device.
29 - A new command 'gnt-cluster upgrade' was added that automates the upgrade
30 procedure between two Ganeti versions that are both 2.10 or higher.
36 *(Released Wed, 9 Oct 2013)*
38 Incompatible/important changes
39 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
41 - hroller now also plans for capacity to move non-redundant instances off
42 any node to be rebooted; the old behavior of completely ignoring any
43 non-redundant instances can be restored by adding the --ignore-non-redundant
45 - The cluster option '--no-lvm-storage' was removed in favor of the new option
46 '--enabled-disk-templates'.
47 - On instance creation, disk templates no longer need to be specified
48 with '-t'. The default disk template will be taken from the list of
49 enabled disk templates.
50 - The monitoring daemon is now running as root, in order to be able to collect
51 information only available to root (such as the state of Xen instances).
52 - The ConfD client is now IPv6 compatible.
53 - File and shared file storage is no longer dis/enabled at configure time,
54 but using the option '--enabled-disk-templates' at cluster initialization and
56 - The default directories for file and shared file storage are not anymore
57 specified at configure time, but taken from the cluster's configuration.
58 They can be set at cluster initialization and modification with
59 '--file-storage-dir' and '--shared-file-storage-dir'.
60 - Cluster verification now includes stricter checks regarding the
61 default file and shared file storage directories. It now checks that
62 the directories are explicitely allowed in the 'file-storage-paths' file and
63 that the directories exist on all nodes.
64 - The list of allowed disk templates in the instance policy and the list
65 of cluster-wide enabled disk templates is now checked for consistency
66 on cluster or group modification. On cluster initialization, the ipolicy
67 disk templates are ensured to be a subset of the cluster-wide enabled
73 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
74 the correct command syntax. It is possible to use different DRBD versions
75 on different nodes as long as they are compatible to each other. This
76 enables rolling upgrades of DRBD with no downtime. As permanent operation
77 of different DRBD versions within a node group is discouraged,
78 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
79 - New "inst-status-xen" data collector for the monitoring daemon, providing
80 information about the state of the xen instances on the nodes.
81 - New "lv" data collector for the monitoring daemon, collecting data about the
82 logical volumes on the nodes, and pairing them with the name of the instances
84 - New "diskstats" data collector, collecting the data from /proc/diskstats and
85 presenting them over the monitoring daemon interface.
86 - The ConfD client is now IPv6 compatible.
90 The following new dependencies have been added.
94 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
95 for the unit tests (and only used for testing).
99 - ``hslogger`` (http://software.complete.org/hslogger) is now always
100 required, even if confd is not enabled.
105 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
106 at least one DRBD instance.
107 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
114 *(Released Tue, 1 Oct 2013)*
116 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
119 - update of the documentation, in particular installation instructions
120 - merging of LD_* constants into DT_* constants
121 - python style changes to be compatible with newer versions of pylint
127 *(Released Thu, 29 Aug 2013)*
129 This was the first beta release of the 2.9 series. All important changes
130 are listed in the latest 2.9 entry.
136 *(Released Mon, 30 Sep 2013)*
138 Incompatible/important changes
139 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
141 - Instance policy can contain multiple instance specs, as described in
142 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
143 <design-partitioned>`. As a consequence, it's not possible to partially change
144 or override instance specs. Bounding specs (min and max) can be specified as a
145 whole using the new option ``--ipolicy-bounds-specs``, while standard
146 specs use the new option ``--ipolicy-std-specs``.
147 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
148 gnt-instance is a valid YAML object.
149 - hail now honors network restrictions when allocating nodes. This led to an
150 update of the IAllocator protocol. See the IAllocator documentation for
152 - confd now only answers static configuration request over the network. luxid
153 was extracted, listens on the local LUXI socket and responds to live queries.
154 This allows finer grained permissions if using separate users.
159 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
160 to always require authentication, ``--require-authentication``. It can
161 be specified in ``$sysconfdir/default/ganeti``.
162 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
163 be used to manage the disk templates to be used by instances in the cluster.
164 Initially, it will be set to a list that includes plain, drbd, if they were
165 enabled by specifying a volume group name, and file and sharedfile, if those
166 were enabled at configure time. Additionally, it will include all disk
167 templates that are currently used by instances. The order of disk templates
168 will be based on Ganeti's history of supporting them. In the future, the
169 first entry of the list will be used as a default disk template on instance
171 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
172 configuration back to the previous stable version.
173 - Disk templates in group ipolicy can be restored to the default value.
174 - Initial support for diskless instances and virtual clusters in QA.
175 - More QA and unit tests for instance policies.
176 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
177 describing why the opcode itself was executed.
178 - The monitoring daemon is now available. It allows users to query the cluster
179 for obtaining information about the status of the system. The daemon is only
180 responsible for providing the information over the network: the actual data
181 gathering is performed by data collectors (currently, only the DRBD status
182 collector is available).
183 - In order to help developers work on Ganeti, a new script
184 (``devel/build_chroot``) is provided, for building a chroot that contains all
185 the required development libraries and tools for compiling Ganeti on a Debian
187 - A new tool, ``harep``, for performing self-repair and recreation of instances
188 in Ganeti has been added.
189 - Split queries are enabled for tags, network, exports, cluster info, groups,
191 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
192 It prints the command line to set the current policies, to ease
194 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
195 devices for KVM virtio-net interfaces will get created with VNET_HDR
196 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
197 virtio-net interfaces, which prevents host kernel tainting and log
198 flooding, when dealing with broken or malicious virtio-net drivers.
199 It's set to true by default.
200 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
202 - Support 'viridian' parameter in Xen HVM
203 - Support DSA SSH keys in bootstrap
204 - To simplify the work of packaging frameworks that want to add the needed users
205 and groups in a split-user setup themselves, at build time three files in
206 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
207 the groups to be generated, the ``users`` file contains, one per line, the
208 users to be generated, optionally followed by their primary group, where
209 important. The ``groupmemberships`` file contains, one per line, additional
210 user-group membership relations that need to be established. The syntax of
211 these files will remain stable in all future versions.
216 The following new dependencies have been added:
219 - The ``curl`` library is not optional anymore for compiling the Haskell code.
220 - ``snap-server`` library (if monitoring is enabled).
223 - The minimum Python version needed to run Ganeti is now 2.6.
224 - ``yaml`` library (only for running the QA).
228 - Perform proper cleanup on termination of Haskell daemons
229 - Fix corner-case in handling of remaining retry time
235 *(Released Tue, 17 Sep 2013)*
237 - To simplify the work of packaging frameworks that want to add the needed users
238 and groups in a split-user setup themselves, at build time three files in
239 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
240 the groups to be generated, the ``users`` file contains, one per line, the
241 users to be generated, optionally followed by their primary group, where
242 important. The ``groupmemberships`` file contains, one per line, additional
243 user-group membership relations that need to be established. The syntax of
244 these files will remain stable in all future versions.
245 - Add a default to file-driver when unspecified over RAPI (Issue 571)
246 - Mark the DSA host pubkey as optional, and remove it during config downgrade
248 - Some documentation fixes
254 *(Released Tue, 27 Aug 2013)*
256 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
258 - Support 'viridian' parameter in Xen HVM (Issue 233)
259 - Include VCS version in ``gnt-cluster version``
260 - Support DSA SSH keys in bootstrap (Issue 338)
261 - Fix batch creation of instances
262 - Use FQDN to check master node status (Issue 551)
263 - Make the DRBD collector more failure-resilient
269 *(Released Fri, 2 Aug 2013)*
271 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
273 - Fix upgrading/downgrading from 2.7
274 - Increase maximum RAPI message size
275 - Documentation updates
276 - Split ``confd`` between ``luxid`` and ``confd``
277 - Merge 2.7 series up to the 2.7.1 release
278 - Allow the ``modify_etc_hosts`` option to be changed
279 - Add better debugging for ``luxid`` queries
280 - Expose bulk parameter for GetJobs in RAPI client
281 - Expose missing ``network`` fields in RAPI
282 - Add some ``cluster verify`` tests
283 - Some unittest fixes
284 - Fix a malfunction in ``hspace``'s tiered allocation
285 - Fix query compatibility between haskell and python implementations
286 - Add the ``vnet_hdr`` HV parameter for KVM
287 - Add ``--cleanup`` to instance failover
288 - Change the connected groups format in ``gnt-network info`` output; it
289 was previously displayed as a raw list by mistake. (Merged from 2.7)
295 *(Released Mon, 24 Jun 2013)*
297 This was the first beta release of the 2.8 series. All important changes
298 are listed in the latest 2.8 entry.
304 *(Released Thu, 26 Sep 2013)*
306 - Change the connected groups format in ``gnt-network info`` output; it
307 was previously displayed as a raw list by mistake
308 - Check disk template in right dict when copying
309 - Support multi-instance allocs without iallocator
310 - Fix some errors in the documentation
311 - Fix formatting of tuple in an error message
317 *(Released Thu, 25 Jul 2013)*
319 - Add logrotate functionality in daemon-util
320 - Add logrotate example file
321 - Add missing fields to network queries over rapi
322 - Fix network object timestamps
323 - Add support for querying network timestamps
324 - Fix a typo in the example crontab
325 - Fix a documentation typo
331 *(Released Thu, 04 Jul 2013)*
333 Incompatible/important changes
334 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
336 - Instance policies for disk size were documented to be on a per-disk
337 basis, but hail applied them to the sum of all disks. This has been
339 - ``hbal`` will now exit with status 0 if, during job execution over
340 LUXI, early exit has been requested and all jobs are successful;
341 before, exit status 1 was used, which cannot be differentiated from
343 - Compatibility with newer versions of rbd has been fixed
344 - ``gnt-instance batch-create`` has been changed to use the bulk create
345 opcode from Ganeti. This lead to incompatible changes in the format of
346 the JSON file. It's now not a custom dict anymore but a dict
347 compatible with the ``OpInstanceCreate`` opcode.
348 - Parent directories for file storage need to be listed in
349 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
350 write the file automatically based on old configuration values, but it
351 can not distribute it across all nodes and the file contents should be
352 verified. Use ``gnt-cluster copyfile
353 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
354 upgraded. The reason for requiring this list of paths now is that
355 before it would have been possible to inject new paths via RPC,
356 allowing files to be created in arbitrary locations. The RPC protocol
357 is protected using SSL/X.509 certificates, but as a design principle
358 Ganeti does not permit arbitrary paths to be passed.
359 - The parsing of the variants file for OSes (see
360 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
361 lines and comment lines (starting with ``#``) are ignored for better
363 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
364 longer available. ``gnt-node add`` now invokes a new tool on the
365 destination node, named ``prepare-node-join``, to configure the SSH
366 daemon. Paramiko is no longer necessary to configure nodes' SSH
367 daemons via ``gnt-node add``.
368 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
369 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
370 the flag is not reset after a master failover.
371 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
372 is recommended. 2.6 will be mandatory from the 2.8 series.
378 - New network management functionality to support automatic allocation
379 of IP addresses and managing of network parameters. See
380 :manpage:`gnt-network(8)` for more details.
381 - New external storage backend, to allow managing arbitrary storage
382 systems external to the cluster. See
383 :manpage:`ganeti-extstorage-interface(7)`.
384 - New ``exclusive-storage`` node parameter added, restricted to
385 nodegroup level. When it's set to true, physical disks are assigned in
386 an exclusive fashion to instances, as documented in :doc:`Partitioned
387 Ganeti <design-partitioned>`. Currently, only instances using the
388 ``plain`` disk template are supported.
389 - The KVM hypervisor has been updated with many new hypervisor
390 parameters, including a generic one for passing arbitrary command line
391 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
392 compatible up to qemu 1.4.
393 - A new tool, called ``mon-collector``, is the stand-alone executor of
394 the data collectors for a monitoring system. As of this version, it
395 just includes the DRBD data collector, that can be executed by calling
396 ``mon-collector`` using the ``drbd`` parameter. See
397 :manpage:`mon-collector(7)`.
398 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
399 for RAPI users. It allows granting permissions to query for
400 information to a specific user without giving
401 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
402 - A new tool named ``node-cleanup`` has been added. It cleans remains of
403 a cluster from a machine by stopping all daemons, removing
404 certificates and ssconf files. Unless the ``--no-backup`` option is
405 given, copies of the certificates are made.
406 - Instance creations now support the use of opportunistic locking,
407 potentially speeding up the (parallel) creation of multiple instances.
408 This feature is currently only available via the :doc:`RAPI
409 <rapi>` interface and when an instance allocator is used. If the
410 ``opportunistic_locking`` parameter is set the opcode will try to
411 acquire as many locks as possible, but will not wait for any locks
412 held by other opcodes. If not enough resources can be found to
413 allocate the instance, the temporary error code
414 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
415 retried thereafter, with or without opportunistic locking.
416 - New experimental linux-ha resource scripts.
417 - Restricted-commands support: ganeti can now be asked (via command line
418 or rapi) to perform commands on a node. These are passed via ganeti
419 RPC rather than ssh. This functionality is restricted to commands
420 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
421 reasons. The file is not copied automatically.
427 - Diskless instances are now externally mirrored (Issue 237). This for
428 now has only been tested in conjunction with explicit target nodes for
430 - Queries not needing locks or RPC access to the node can now be
431 performed by the confd daemon, making them independent from jobs, and
432 thus faster to execute. This is selectable at configure time.
433 - The functionality for allocating multiple instances at once has been
434 overhauled and is now also available through :doc:`RAPI <rapi>`.
436 There are no significant changes from version 2.7.0~rc3.
442 *(Released Tue, 25 Jun 2013)*
444 - Fix permissions on the confd query socket (Issue 477)
445 - Fix permissions on the job archive dir (Issue 498)
446 - Fix handling of an internal exception in replace-disks (Issue 472)
447 - Fix gnt-node info handling of shortened names (Issue 497)
448 - Fix gnt-instance grow-disk when wiping is enabled
449 - Documentation improvements, and support for newer pandoc
450 - Fix hspace honoring ipolicy for disks (Issue 484)
451 - Improve handling of the ``kvm_extra`` HV parameter
457 *(Released Fri, 24 May 2013)*
459 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
461 - Disks added through ``gnt-instance modify`` or created through
462 ``gnt-instance recreate-disks`` are wiped, if the
463 ``prealloc_wipe_disks`` flag is set.
464 - If wiping newly created disks fails, the disks are removed. Also,
465 partial failures in creating disks through ``gnt-instance modify``
466 triggers a cleanup of the partially-created disks.
467 - Removing the master IP address doesn't fail if the address has been
469 - Fix ownership of the OS log dir
470 - Workaround missing SO_PEERCRED constant (Issue 191)
476 *(Released Fri, 3 May 2013)*
478 This was the first release candidate of the 2.7 series. Since beta3:
480 - Fix kvm compatibility with qemu 1.4 (Issue 389)
481 - Documentation updates (admin guide, upgrade notes, install
482 instructions) (Issue 372)
483 - Fix gnt-group list nodes and instances count (Issue 436)
484 - Fix compilation without non-mandatory libraries (Issue 441)
485 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
486 - Make confd logging more verbose at INFO level (Issue 435)
487 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
488 - Fix failure path for instance storage type conversion (Issue 229)
489 - Update htools text backend documentation
490 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
491 - Disable inter-cluster instance move for file-based instances, because
492 it is dependant on instance export, which is not supported for
493 file-based instances. (Issue 414)
494 - Fix gnt-job crashes on non-ascii characters (Issue 427)
495 - Fix volume group checks on non-vm-capable nodes (Issue 432)
501 *(Released Mon, 22 Apr 2013)*
503 This was the third beta release of the 2.7 series. Since beta2:
505 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
506 - Fix data loss on wrong usage of ``gnt-instance move``
507 - Properly export errors in confd-based job queries
508 - Add ``users-setup`` tool
509 - Fix iallocator protocol to report 0 as a disk size for diskless
510 instances. This avoids hail breaking when a diskless instance is
512 - Fix job queue directory permission problem that made confd job queries
513 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
514 for access to archived jobs (Issue 406).
515 - Limit the sizes of networks supported by ``gnt-network`` to something
516 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
517 - Fix bugs in instance disk template conversion
518 - Fix GHC 7 compatibility
519 - Fix ``burnin`` install path (Issue 426).
520 - Allow very small disk grows (Issue 347).
521 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
522 that noded doesn't import masterd code (Issue 419).
523 - Make sure the default metavg at cluster init is the same as the vg, if
524 unspecified (Issue 358).
525 - Fix cleanup of partially created disks (part of Issue 416)
531 *(Released Tue, 2 Apr 2013)*
533 This was the second beta release of the 2.7 series. Since beta1:
535 - Networks no longer have a "type" slot, since this information was
536 unused in Ganeti: instead of it tags should be used.
537 - The rapi client now has a ``target_node`` option to MigrateInstance.
538 - Fix early exit return code for hbal (Issue 386).
539 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
540 - Fix ``rbd showmapped`` output parsing (Issue 312).
541 - Networks are now referenced indexed by UUID, rather than name. This
542 will require running cfgupgrade, from 2.7.0beta1, if networks are in
544 - The OS environment now includes network information.
545 - Deleting of a network is now disallowed if any instance nic is using
546 it, to prevent dangling references.
547 - External storage is now documented in man pages.
548 - The exclusive_storage flag can now only be set at nodegroup level.
549 - Hbal can now submit an explicit priority with its jobs.
550 - Many network related locking fixes.
551 - Bump up the required pylint version to 0.25.1.
552 - Fix the ``no_remember`` option in RAPI client.
553 - Many ipolicy related tests, qa, and fixes.
554 - Many documentation improvements and fixes.
555 - Fix building with ``--disable-file-storage``.
556 - Fix ``-q`` option in htools, which was broken if passed more than
558 - Some haskell/python interaction improvements and fixes.
559 - Fix iallocator in case of missing LVM storage.
560 - Fix confd config load in case of ``--no-lvm-storage``.
561 - The confd/query functionality is now mentioned in the security
568 *(Released Wed, 6 Feb 2013)*
570 This was the first beta release of the 2.7 series. All important changes
571 are listed in the latest 2.7 entry.
577 *(Released Fri, 21 Dec 2012)*
579 Important behaviour change: hbal won't rebalance anymore instances which
580 have the ``auto_balance`` attribute set to false. This was the intention
581 all along, but until now it only skipped those from the N+1 memory
582 reservation (DRBD-specific).
584 A significant number of bug fixes in this release:
586 - Fixed disk adoption interaction with ipolicy checks.
587 - Fixed networking issues when instances are started, stopped or
588 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
589 - Fixed the warning in cluster verify for shared storage instances not
591 - Fixed removal of storage directory on shared file storage (issue 262).
592 - Fixed validation of LVM volume group name in OpClusterSetParams
593 (``gnt-cluster modify``) (issue 285).
594 - Fixed runtime memory increases (``gnt-instance modify -m``).
595 - Fixed live migration under Xen's ``xl`` mode.
596 - Fixed ``gnt-instance console`` with ``xl``.
597 - Fixed building with newer Haskell compiler/libraries.
598 - Fixed PID file writing in Haskell daemons (confd); this prevents
599 restart issues if confd was launched manually (outside of
600 ``daemon-util``) while another copy of it was running
601 - Fixed a type error when doing live migrations with KVM (issue 297) and
602 the error messages for failing migrations have been improved.
603 - Fixed opcode validation for the out-of-band commands (``gnt-node
605 - Fixed a type error when unsetting OS hypervisor parameters (issue
606 311); now it's possible to unset all OS-specific hypervisor
608 - Fixed the ``dry-run`` mode for many operations: verification of
609 results was over-zealous but didn't take into account the ``dry-run``
610 operation, resulting in "wrong" failures.
611 - Fixed bash completion in ``gnt-job list`` when the job queue has
612 hundreds of entries; especially with older ``bash`` versions, this
613 results in significant CPU usage.
615 And lastly, a few other improvements have been made:
617 - Added option to force master-failover without voting (issue 282).
618 - Clarified error message on lock conflict (issue 287).
619 - Logging of newly submitted jobs has been improved (issue 290).
620 - Hostname checks have been made uniform between instance rename and
622 - The ``--submit`` option is now supported by ``gnt-debug delay``.
623 - Shutting down the master daemon by sending SIGTERM now stops it from
624 processing jobs waiting for locks; instead, those jobs will be started
625 once again after the master daemon is started the next time (issue
627 - Support for Xen's ``xl`` program has been improved (besides the fixes
629 - Reduced logging noise in the Haskell confd daemon (only show one log
630 entry for each config reload, instead of two).
631 - Several man page updates and typo fixes.
637 *(Released Fri, 12 Oct 2012)*
639 A small bugfix release. Among the bugs fixed:
641 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
642 made the command unusable.
643 - Commands that issue many jobs don't fail anymore just because some jobs
644 take so long that other jobs are archived.
645 - Failures during ``gnt-instance reinstall`` are reflected by the exit
647 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
649 - When ``always_failover`` is set, ``--allow-failover`` is not required
650 in migrate commands anymore.
651 - ``bash_completion`` works even if extglob is disabled.
652 - Fixed bug with locks that made failover for RDB-based instances fail.
653 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
654 a random node instead of one based on the allocator metric.
655 - Support for newer versions of pylint and pep8.
656 - Hail doesn't fail anymore when trying to add an instance of type
657 ``file``, ``sharedfile`` or ``rbd``.
658 - Added new Makefile target to rebuild the whole distribution, so that
659 all files are included.
665 *(Released Fri, 27 Jul 2012)*
668 .. attention:: The ``LUXI`` protocol has been made more consistent
669 regarding its handling of command arguments. This, however, leads to
670 incompatibility issues with previous versions. Please ensure that you
671 restart Ganeti daemons soon after the upgrade, otherwise most
672 ``LUXI`` calls (job submission, setting/resetting the drain flag,
673 pausing/resuming the watcher, cancelling and archiving jobs, querying
674 the cluster configuration) will fail.
683 The current ``admin_up`` field, which used to denote whether an instance
684 should be running or not, has been removed. Instead, ``admin_state`` is
685 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
687 The rational behind this is that an instance being “down” can have
690 - it could be down during a reboot
691 - it could be temporarily be down for a reinstall
692 - or it could be down because it is deprecated and kept just for its
695 The previous Boolean state was making it difficult to do capacity
696 calculations: should Ganeti reserve memory for a down instance? Now, the
697 tri-state field makes it clear:
699 - in ``up`` and ``down`` state, all resources are reserved for the
700 instance, and it can be at any time brought up if it is down
701 - in ``offline`` state, only disk space is reserved for it, but not
704 The field can have an extra use: since the transition between ``up`` and
705 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
706 transition between ``offline`` and ``down`` is done via ``gnt-instance
707 modify``, it is possible to given different rights to users. For
708 example, owners of an instance could be allowed to start/stop it, but
709 not transition it out of the offline state.
711 Instance policies and specs
712 +++++++++++++++++++++++++++
714 In previous Ganeti versions, an instance creation request was not
715 limited on the minimum size and on the maximum size just by the cluster
716 resources. As such, any policy could be implemented only in third-party
717 clients (RAPI clients, or shell wrappers over ``gnt-*``
718 tools). Furthermore, calculating cluster capacity via ``hspace`` again
719 required external input with regards to instance sizes.
721 In order to improve these workflows and to allow for example better
722 per-node group differentiation, we introduced instance specs, which
725 - minimum instance disk size, disk count, memory size, cpu count
726 - maximum values for the above metrics
727 - and “standard” values (used in ``hspace`` to calculate the standard
730 The minimum/maximum values can be also customised at node-group level,
731 for example allowing more powerful hardware to support bigger instance
734 Beside the instance specs, there are a few other settings belonging to
735 the instance policy framework. It is possible now to customise, per
736 cluster and node-group:
738 - the list of allowed disk templates
739 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
740 - the maximum ratio of instance to spindles (see below for more
741 information) for local storage
743 All these together should allow all tools that talk to Ganeti to know
744 what are the ranges of allowed values for instances and the
745 over-subscription that is allowed.
747 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
748 instance configuration, and the physical CPU configuration from the
749 node. For the spindle ratios however, we didn't track before these
750 values, so new parameters have been added:
752 - a new node parameter ``spindle_count``, defaults to 1, customisable at
753 node group or node level
754 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
756 Note that spindles in this context doesn't need to mean actual
757 mechanical hard-drives; it's just a relative number for both the node
758 I/O capacity and instance I/O consumption.
760 Instance migration behaviour
761 ++++++++++++++++++++++++++++
763 While live-migration is in general desirable over failover, it is
764 possible that for some workloads it is actually worse, due to the
765 variable time of the “suspend” phase during live migration.
767 To allow the tools to work consistently over such instances (without
768 having to hard-code instance names), a new backend parameter
769 ``always_failover`` has been added to control the migration/failover
770 behaviour. When set to True, all migration requests for an instance will
771 instead fall-back to failover.
773 Instance memory ballooning
774 ++++++++++++++++++++++++++
776 Initial support for memory ballooning has been added. The memory for an
777 instance is no longer fixed (backend parameter ``memory``), but instead
778 can vary between minimum and maximum values (backend parameters
779 ``minmem`` and ``maxmem``). Currently we only change an instance's
782 - live migrating or failing over and instance and the target node
783 doesn't have enough memory
784 - user requests changing the memory via ``gnt-instance modify
790 In order to control the use of specific CPUs by instance, support for
791 controlling CPU pinning has been added for the Xen, HVM and LXC
792 hypervisors. This is controlled by a new hypervisor parameter
793 ``cpu_mask``; details about possible values for this are in the
794 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
795 VCPU-to-CPU mapping) form will work well only when all nodes in your
796 cluster have the same amount of CPUs.
801 Another area in which Ganeti was not customisable were the parameters
802 used for storage configuration, e.g. how many stripes to use for LVM,
803 DRBD resync configuration, etc.
805 To improve this area, we've added disks parameters, which are
806 customisable at cluster and node group level, and which allow to
807 specify various parameters for disks (DRBD has the most parameters
808 currently), for example:
810 - DRBD resync algorithm and parameters (e.g. speed)
811 - the default VG for meta-data volumes for DRBD
812 - number of stripes for LVM (plain disk template)
815 These parameters can be modified via ``gnt-cluster modify -D …`` and
816 ``gnt-group modify -D …``, and are used at either instance creation (in
817 case of LVM stripes, for example) or at disk “activation” time
820 Rados block device support
821 ++++++++++++++++++++++++++
823 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
824 denoted by the ``rbd`` disk template type. This is considered
825 experimental, feedback is welcome. For details on configuring it, see
826 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
831 The existing master IP functionality works well only in simple setups (a
832 single network shared by all nodes); however, if nodes belong to
833 different networks, then the ``/32`` setup and lack of routing
834 information is not enough.
836 To allow the master IP to function well in more complex cases, the
837 system was reworked as follows:
839 - a master IP netmask setting has been added
840 - the master IP activation/turn-down code was moved from the node daemon
842 - whether to run the Ganeti-supplied master IP script or a user-supplied
843 on is a ``gnt-cluster init`` setting
845 Details about the location of the standard and custom setup scripts are
846 in the man page :manpage:`gnt-cluster(8)`; for information about the
847 setup script protocol, look at the Ganeti-supplied script.
852 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
855 It is now possible to use TLS-protected connections, and when renewing
856 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
857 it is now possible to specify spice or spice CA certificates. Also, it
858 is possible to configure a password for SPICE sessions via the
859 hypervisor parameter ``spice_password_file``.
861 There are also new parameters to control the compression and streaming
862 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
863 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
864 for the spice parameters.
866 Lastly, it is now possible to see the SPICE connection information via
867 ``gnt-instance console``.
872 A new tool (``tools/ovfconverter``) has been added that supports
873 conversion between Ganeti and the `Open Virtualization Format
874 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
877 This relies on the ``qemu-img`` tool to convert the disk formats, so the
878 actual compatibility with other virtualization solutions depends on it.
883 The configuration query daemon (``ganeti-confd``) is now optional, and
884 has been rewritten in Haskell; whether to use the daemon at all, use the
885 Python (default) or the Haskell version is selectable at configure time
886 via the ``--enable-confd`` parameter, which can take one of the
887 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
888 daemon will result in a smaller footprint; for larger systems, we
889 welcome feedback on the Haskell version which might become the default
892 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
893 daemon running. The Python version doesn't implement the new call.
896 User interface changes
897 ~~~~~~~~~~~~~~~~~~~~~~
899 We have replaced the ``--disks`` option of ``gnt-instance
900 replace-disks`` with a more flexible ``--disk`` option, which allows
901 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
902 disk size and mode can be changed upon recreation (via ``gnt-instance
903 recreate-disks``, which accepts the same ``--disk`` option).
905 As many people are used to a ``show`` command, we have added that as an
906 alias to ``info`` on all ``gnt-*`` commands.
908 The ``gnt-instance grow-disk`` command has a new mode in which it can
909 accept the target size of the disk, instead of the delta; this can be
910 more safe since two runs in absolute mode will be idempotent, and
911 sometimes it's also easier to specify the desired size directly.
913 Also the handling of instances with regard to offline secondaries has
914 been improved. Instance operations should not fail because one of it's
915 secondary nodes is offline, even though it's safe to proceed.
917 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
918 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
919 minors to instance name.
924 RAPI coverage has improved, with (for example) new resources for
925 recreate-disks, node power-cycle, etc.
930 There is partial support for ``xl`` in the Xen hypervisor; feedback is
933 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
934 whether to still support Python 2.4 or move to Python 2.6 as minimum
937 Support for Fedora has been slightly improved; the provided example
938 init.d script should work better on it and the INSTALL file should
939 document the needed dependencies.
944 The deprecated ``QueryLocks`` LUXI request has been removed. Use
945 ``Query(what=QR_LOCK, ...)`` instead.
947 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
948 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
949 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
950 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
951 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
953 RAPI client: ``CertificateError`` now derives from
954 ``GanetiApiError``. This should make it more easy to handle Ganeti
957 Deprecation warnings due to PyCrypto/paramiko import in
958 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
959 make sure to run an up-to-date paramiko version, if you use this tool.
961 The QA scripts now depend on Python 2.5 or above (the main code base
962 still works with Python 2.4).
964 The configuration file (``config.data``) is now written without
965 indentation for performance reasons; if you want to edit it, it can be
966 re-formatted via ``tools/fmtjson``.
968 A number of bugs has been fixed in the cluster merge tool.
970 ``x509`` certification verification (used in import-export) has been
971 changed to allow the same clock skew as permitted by the cluster
972 verification. This will remove some rare but hard to diagnose errors in
979 *(Released Thu, 19 Jul 2012)*
981 Very few changes from rc4 to the final release, only bugfixes:
983 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
984 instance, fix CDROM booting for KVM instances)
985 - fixed node group modification of node parameters
986 - fixed issue in LUClusterVerifyGroup with multi-group clusters
987 - fixed generation of bash completion to ensure a stable ordering
994 *(Released Fri, 13 Jul 2012)*
996 Third release candidate for 2.6. The following changes were done from
999 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1000 - Fixed an inconsistency in the LUXI protocol with the provided
1001 arguments (NOT backwards compatible)
1002 - Fixed a bug with node groups ipolicy where ``min`` was greater than
1003 the cluster ``std`` value
1004 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1005 easier instance debugging on nodes (requires ``hconfd`` to work)
1011 *(Released Tue, 03 Jul 2012)*
1013 Second release candidate for 2.6. The following changes were done from
1016 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1018 - Fixed a RAPI regression on missing beparams/memory
1019 - Fixed redistribution of files on offline nodes
1020 - Added possibility to run activate-disks even though secondaries are
1021 offline. With this change it relaxes also the strictness on some other
1022 commands which use activate disks internally:
1023 * ``gnt-instance start|reboot|rename|backup|export``
1024 - Made it possible to remove safely an instance if its secondaries are
1026 - Made it possible to reinstall even though secondaries are offline
1032 *(Released Mon, 25 Jun 2012)*
1034 First release candidate for 2.6. The following changes were done from
1037 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1039 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1040 behaviour. It supports now the modify operation
1041 - ``hcheck`` implemented to analyze cluster health and possibility of
1042 improving health by rebalance
1043 - ``hbal`` has been improved in dealing with split instances
1049 *(Released Mon, 11 Jun 2012)*
1051 Second beta release of 2.6. The following changes were done from beta2
1054 - Fixed ``daemon-util`` with non-root user models
1055 - Fixed creation of plain instances with ``--no-wait-for-sync``
1056 - Fix wrong iv_names when running ``cfgupgrade``
1057 - Export more information in RAPI group queries
1058 - Fixed bug when changing instance network interfaces
1059 - Extended burnin to do NIC changes
1060 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1061 - Changed default for DRBD barriers
1062 - Fixed DRBD error reporting for syncer rate
1063 - Verify the options on disk parameters
1065 And of course various fixes to documentation and improved unittests and
1072 *(Released Wed, 23 May 2012)*
1074 First beta release of 2.6. The following changes were done from beta1 to
1077 - integrated patch for distributions without ``start-stop-daemon``
1078 - adapted example init.d script to work on Fedora
1079 - fixed log handling in Haskell daemons
1080 - adapted checks in the watcher for pycurl linked against libnss
1081 - add partial support for ``xl`` instead of ``xm`` for Xen
1082 - fixed a type issue in cluster verification
1083 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1086 Plus integrated fixes from the 2.5 branch:
1088 - fixed ``kvm-ifup`` to use ``/bin/bash``
1089 - fixed parallel build failures
1090 - KVM live migration when using a custom keymap
1096 *(Released Tue, 24 Jul 2012)*
1098 A small bugfix release, with no new features:
1100 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1101 different default shell (e.g. Debian, Ubuntu)
1102 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1103 243 and Debian bug #650664)
1104 - fixed compatibility with KVM versions that don't support multiple boot
1105 devices (fixes Issue 230 and Debian bug #624256)
1107 Additionally, a few fixes were done to the build system (fixed parallel
1108 build failures) and to the unittests (fixed race condition in test for
1109 FileID functions, and the default enable/disable mode for QA test is now
1116 *(Released Fri, 11 May 2012)*
1118 A small bugfix release.
1120 The main issues solved are on the topic of compatibility with newer LVM
1123 - fixed parsing of ``lv_attr`` field
1124 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1125 the ``--force`` flag is needed
1127 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1128 to accept kernel 3.x too (was hardcoded to 2.6.*).
1130 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1131 of hook results and that also made display of other errors suboptimal
1132 was fixed; the code behaves now like 2.4 and earlier.
1134 Another change in 2.5, the cleanup of the OS scripts environment, is too
1135 aggressive: it removed even the ``PATH`` variable, which requires the OS
1136 scripts to *always* need to export it. Since this is a bit too strict,
1137 we now export a minimal PATH, the same that we export for hooks.
1139 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1140 integrated into this release.
1142 Finally, a few other miscellaneous changes were done (no new features,
1143 just small improvements):
1145 - Fix ``gnt-group --help`` display
1146 - Fix hardcoded Xen kernel path
1147 - Fix grow-disk handling of invalid units
1148 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1149 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1155 *(Released Thu, 12 Apr 2012)*
1157 Incompatible/important changes and bugfixes
1158 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1160 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1161 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1162 to match the underlying LUXI interface.
1163 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1164 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1165 not maintain backwards-compatibility as the underlying operation
1166 changed in an incompatible way. The RAPI client can talk to old
1167 servers, but it needs to be told so as the return value changed.
1168 - When creating file-based instances via RAPI, the ``file_driver``
1169 parameter no longer defaults to ``loop`` and must be specified.
1170 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1172 - Support for the undocumented and deprecated RAPI instance creation
1173 request format version 0 has been dropped. Use version 1, supported
1174 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1175 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1176 documentation <install>`.
1177 - The "cluster-verify" hooks are now executed per group by the
1178 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1179 if you just run ``gnt-cluster verify``, which generates one opcode per
1181 - The environment as passed to the OS scripts is cleared, and thus no
1182 environment variables defined in the node daemon's environment will be
1183 inherited by the scripts.
1184 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1186 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1187 support operations involving multiple node groups.
1188 - Offline nodes are ignored when failing over an instance.
1189 - Support for KVM version 1.0, which changed the version reporting format
1191 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1193 - ``Makefile`` is now compatible with Automake 1.11.2
1194 - Includes all bugfixes made in the 2.4 series
1199 - The ganeti-htools project has been merged into the ganeti-core source
1200 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1201 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1202 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1203 GPT-style partition tables (requires `parted
1204 <http://www.gnu.org/s/parted/>`_).
1205 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1206 - Allowed adding tags on instance creation.
1207 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1208 - Allow instances to be started in a paused state, enabling the user to
1209 see the complete console output on boot using the console.
1210 - Added new hypervisor flag to control default reboot behaviour
1211 (``reboot_behavior``).
1212 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1213 - Improved out-of-band management support:
1215 - Added ``gnt-node health`` command reporting the health status of
1217 - Added ``gnt-node power`` command to manage power status of nodes.
1218 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1220 - Instance migration can fall back to failover if instance is not
1222 - Filters can be used when listing nodes, instances, groups and locks;
1223 see :manpage:`ganeti(7)` manpage.
1224 - Added post-execution status as variables to :doc:`hooks <hooks>`
1226 - Instance tags are exported/imported together with the instance.
1227 - When given an explicit job ID, ``gnt-job info`` will work for archived
1229 - Jobs can define dependencies on other jobs (not yet supported via
1230 RAPI or command line, but used by internal commands and usable via
1233 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1236 - Instance failover is now available as a RAPI resource
1237 (``/2/instances/[instance_name]/failover``).
1238 - ``gnt-instance info`` defaults to static information if primary node
1240 - Opcodes have a new ``comment`` attribute.
1241 - Added basic SPICE support to KVM hypervisor.
1242 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1244 Node group improvements
1245 ~~~~~~~~~~~~~~~~~~~~~~~
1247 - ``gnt-cluster verify`` has been modified to check groups separately,
1248 thereby improving performance.
1249 - Node group support has been added to ``gnt-cluster verify-disks``,
1250 which now operates per node group.
1251 - Watcher has been changed to work better with node groups.
1253 - One process and state file per node group.
1254 - Slow watcher in one group doesn't block other group's watcher.
1256 - Added new command, ``gnt-group evacuate``, to move all instances in a
1257 node group to other groups.
1258 - Added ``gnt-instance change-group`` to move an instance to another
1260 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1261 per-group operations.
1262 - Node groups can be tagged.
1263 - Some operations switch from an exclusive to a shared lock as soon as
1265 - Instance's primary and secondary nodes' groups are now available as
1266 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1267 and ``snodes.group.uuid``).
1272 - Numerous updates to documentation and manpages.
1274 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1276 - Some opcode/job results are now also documented, see :doc:`RAPI
1279 - A lockset's internal lock is now also visible in lock monitor.
1280 - Log messages from job queue workers now contain information about the
1281 opcode they're processing.
1282 - ``gnt-instance console`` no longer requires the instance lock.
1283 - A short delay when waiting for job changes reduces the number of LUXI
1284 requests significantly.
1285 - DRBD metadata volumes are overwritten with zeros during disk creation.
1286 - Out-of-band commands no longer acquire the cluster lock in exclusive
1288 - ``devel/upload`` now uses correct permissions for directories.
1294 *(Released Fri, 23 Mar 2012)*
1296 This was the sixth release candidate of the 2.5 series.
1302 *(Released Mon, 9 Jan 2012)*
1304 This was the fifth release candidate of the 2.5 series.
1310 *(Released Thu, 27 Oct 2011)*
1312 This was the fourth release candidate of the 2.5 series.
1318 *(Released Wed, 26 Oct 2011)*
1320 This was the third release candidate of the 2.5 series.
1326 *(Released Tue, 18 Oct 2011)*
1328 This was the second release candidate of the 2.5 series.
1334 *(Released Tue, 4 Oct 2011)*
1336 This was the first release candidate of the 2.5 series.
1342 *(Released Wed, 31 Aug 2011)*
1344 This was the third beta release of the 2.5 series.
1350 *(Released Mon, 22 Aug 2011)*
1352 This was the second beta release of the 2.5 series.
1358 *(Released Fri, 12 Aug 2011)*
1360 This was the first beta release of the 2.5 series.
1366 *(Released Thu, 27 Oct 2011)*
1368 - Fixed bug when parsing command line parameter values ending in
1370 - Fixed assertion error after unclean master shutdown
1371 - Disable HTTP client pool for RPC, significantly reducing memory usage
1373 - Fixed queue archive creation with wrong permissions
1379 *(Released Tue, 23 Aug 2011)*
1383 - Fixed documentation for importing with ``--src-dir`` option
1384 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1385 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1391 *(Released Fri, 5 Aug 2011)*
1393 Many bug-fixes and a few small features:
1395 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1396 issues when you tried to add an instance with two MAC addresses in one
1398 - KVM: fixed per-instance stored UID value
1399 - KVM: configure bridged NICs at migration start
1400 - KVM: Fix a bug where instance will not start with never KVM versions
1402 - Added OS search path to ``gnt-cluster info``
1403 - Fixed an issue with ``file_storage_dir`` where you were forced to
1404 provide an absolute path, but the documentation states it is a
1405 relative path, the documentation was right
1406 - Added a new parameter to instance stop/start called ``--no-remember``
1407 that will make the state change to not be remembered
1408 - Implemented ``no_remember`` at RAPI level
1409 - Improved the documentation
1410 - Node evacuation: don't call IAllocator if node is already empty
1411 - Fixed bug in DRBD8 replace disks on current nodes
1412 - Fixed bug in recreate-disks for DRBD instances
1413 - Moved assertion checking locks in ``gnt-instance replace-disks``
1414 causing it to abort with not owning the right locks for some situation
1415 - Job queue: Fixed potential race condition when cancelling queued jobs
1416 - Fixed off-by-one bug in job serial generation
1417 - ``gnt-node volumes``: Fix instance names
1418 - Fixed aliases in bash completion
1419 - Fixed a bug in reopening log files after being sent a SIGHUP
1420 - Added a flag to burnin to allow specifying VCPU count
1421 - Bugfixes to non-root Ganeti configuration
1427 *(Released Thu, 12 May 2011)*
1429 Many bug-fixes and a few new small features:
1431 - Fixed a bug related to log opening failures
1432 - Fixed a bug in instance listing with orphan instances
1433 - Fixed a bug which prevented resetting the cluster-level node parameter
1434 ``oob_program`` to the default
1435 - Many fixes related to the ``cluster-merge`` tool
1436 - Fixed a race condition in the lock monitor, which caused failures
1437 during (at least) creation of many instances in parallel
1438 - Improved output for gnt-job info
1439 - Removed the quiet flag on some ssh calls which prevented debugging
1441 - Improved the N+1 failure messages in cluster verify by actually
1442 showing the memory values (needed and available)
1443 - Increased lock attempt timeouts so that when executing long operations
1444 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1445 too early and thus prevent the use of the 'fair' mechanism
1446 - Changed instance query data (``gnt-instance info``) to not acquire
1447 locks unless needed, thus allowing its use on locked instance if only
1448 static information is asked for
1449 - Improved behaviour with filesystems that do not support rename on an
1451 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1452 kept locks on all nodes during the wipe, which is unneeded
1453 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1454 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1455 10GiB) which caused the wipe to fail right at the end in some cases
1456 - Fixed master IP activation when doing master failover with no-voting
1457 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1458 the master node itself
1459 - Fixed potential data-loss in under disk full conditions, where Ganeti
1460 wouldn't check correctly the return code and would consider
1461 partially-written files 'correct'
1462 - Fixed bug related to multiple VGs and DRBD disk replacing
1463 - Added new disk parameter ``metavg`` that allows placement of the meta
1464 device for DRBD in a different volume group
1465 - Fixed error handling in the node daemon when the system libc doesn't
1466 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1467 - Fixed lock release during replace-disks, which kept cluster-wide locks
1468 when doing disk replaces with an iallocator script
1469 - Added check for missing bridges in cluster verify
1470 - Handle EPIPE errors while writing to the terminal better, so that
1471 piping the output to e.g. ``less`` doesn't cause a backtrace
1472 - Fixed rare case where a ^C during Luxi calls could have been
1473 interpreted as server errors, instead of simply terminating
1474 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1476 - Added a few more parameters to the KVM hypervisor, allowing a second
1477 CDROM, custom disk type for CDROMs and a floppy image
1478 - Removed redundant message in instance rename when the name is given
1480 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1481 disks on new nodes, allowing recreation when the original instance
1482 nodes are completely gone
1483 - Added option when converting disk templates to DRBD to skip waiting
1484 for the resync, in order to make the instance available sooner
1485 - Added two new variables to the OS scripts environment (containing the
1487 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1488 to allow use of ``pvgrub`` as bootloader
1489 - Changed the instance memory modifications to only check out-of-memory
1490 conditions on memory increases, and turned the secondary node warnings
1491 into errors (they can still be overridden via ``--force``)
1492 - Fixed the handling of a corner case when the Python installation gets
1493 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1494 to execute a command that doesn't exist
1495 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1496 node of the instance returned failures during instance shutdown; this
1497 adds the option ``--ignore-consistency`` to gnt-instance move
1499 And as usual, various improvements to the error messages, documentation
1506 *(Released Wed, 09 Mar 2011)*
1508 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1509 the RAPI users file if run twice (even with ``--dry-run``).
1511 The release fixes that bug (nothing else changed).
1517 *(Released Mon, 07 Mar 2011)*
1519 Final 2.4.0 release. Just a few small fixes:
1521 - Fixed RAPI node evacuate
1522 - Fixed the kvm-ifup script
1523 - Fixed internal error handling for special job cases
1524 - Updated man page to specify the escaping feature for options
1530 *(Released Mon, 28 Feb 2011)*
1532 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1533 that this feature wiped the disks of the wrong instance, leading to loss
1538 - Fixed title of query field containing instance name
1539 - Expanded the glossary in the documentation
1540 - Fixed one unittest (internal issue)
1546 *(Released Mon, 21 Feb 2011)*
1548 A number of bug fixes plus just a couple functionality changes.
1550 On the user-visible side, the ``gnt-* list`` command output has changed
1551 with respect to "special" field states. The current rc1 style of display
1552 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1553 the default output mode special fields states are displayed as follows:
1555 - Offline resource: ``*``
1556 - Unavailable/not applicable: ``-``
1557 - Data missing (RPC failure): ``?``
1558 - Unknown field: ``??``
1560 Another user-visible change is the addition of ``--force-join`` to
1565 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1566 - Fixed regression in RAPI/instance reinstall where all parameters were
1567 required (instead of optional)
1568 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1569 - Fixed iallocator usage (offline nodes were not considered offline)
1570 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1571 - Fixed hypervisor and OS parameter validation with respect to
1572 non-vm_capable nodes
1573 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1575 - Fixed ``tools/listrunner`` with respect to agent-based usage
1581 *(Released Fri, 4 Feb 2011)*
1583 Many changes and fixes since the beta1 release. While there were some
1584 internal changes, the code has been mostly stabilised for the RC
1587 Note: the dumb allocator was removed in this release, as it was not kept
1588 up-to-date with the IAllocator protocol changes. It is recommended to
1589 use the ``hail`` command from the ganeti-htools package.
1591 Note: the 2.4 and up versions of Ganeti are not compatible with the
1592 0.2.x branch of ganeti-htools. You need to upgrade to
1593 ganeti-htools-0.3.0 (or later).
1595 Regressions fixed from 2.3
1596 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1598 - Fixed the ``gnt-cluster verify-disks`` command
1599 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1601 - Fixed disk adoption breakage
1602 - Fixed wrong headers in instance listing for field aliases
1607 - Fixed corner case in KVM handling of NICs
1608 - Fixed many cases of wrong handling of non-vm_capable nodes
1609 - Fixed a bug where a missing instance symlink was not possible to
1610 recreate with any ``gnt-*`` command (now ``gnt-instance
1611 activate-disks`` does it)
1612 - Fixed the volume group name as reported by ``gnt-cluster
1614 - Increased timeouts for the import-export code, hopefully leading to
1615 fewer aborts due network or instance timeouts
1616 - Fixed bug in ``gnt-node list-storage``
1617 - Fixed bug where not all daemons were started on cluster
1618 initialisation, but only at the first watcher run
1619 - Fixed many bugs in the OOB implementation
1620 - Fixed watcher behaviour in presence of instances with offline
1622 - Fixed instance list output for instances running on the wrong node
1623 - a few fixes to the cluster-merge tool, but it still cannot merge
1624 multi-node groups (currently it is not recommended to use this tool)
1630 - Improved network configuration for the KVM hypervisor
1631 - Added e1000 as a supported NIC for Xen-HVM
1632 - Improved the lvmstrap tool to also be able to use partitions, as
1633 opposed to full disks
1634 - Improved speed of disk wiping (the cluster parameter
1635 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1636 of instance creations
1637 - Added documentation for the OS parameters
1638 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1639 hypervisor is not responding
1640 - Added display of blacklisted and hidden OS information in
1641 ``gnt-cluster info``
1642 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1643 NIC and node parameters, which might create problems with currently
1644 invalid (but undetected) configuration files, but prevents validation
1645 failures when unrelated parameters are modified
1646 - Changed cluster initialisation to wait for the master daemon to become
1648 - Expanded the RAPI interface:
1650 - Added config redistribution resource
1651 - Added activation/deactivation of instance disks
1652 - Added export of console information
1654 - Implemented log file reopening on SIGHUP, which allows using
1655 logrotate(8) for the Ganeti log files
1656 - Added a basic OOB helper script as an example
1662 *(Released Fri, 14 Jan 2011)*
1667 - Fixed timezone issues when formatting timestamps
1668 - Added support for node groups, available via ``gnt-group`` and other
1670 - Added out-of-band framework and management, see :doc:`design
1671 document <design-oob>`
1672 - Removed support for roman numbers from ``gnt-node list`` and
1673 ``gnt-instance list``.
1674 - Allowed modification of master network interface via ``gnt-cluster
1675 modify --master-netdev``
1676 - Accept offline secondaries while shutting down instance disks
1677 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1678 - Added support for multiple LVM volume groups
1679 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1681 - Added commands to list available fields:
1683 - ``gnt-node list-fields``
1684 - ``gnt-group list-fields``
1685 - ``gnt-instance list-fields``
1687 - Updated documentation and man pages
1692 - Moved ``rapi_users`` file into separate directory, now named
1693 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1695 - Added new tool for running commands on many machines,
1696 ``tools/ganeti-listrunner``
1697 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1698 improving the ``gnt-instance console`` output
1699 - Allowed customisation of disk index separator at ``configure`` time
1700 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1701 - Added support for non-partitioned md disks in ``lvmstrap``
1702 - Added script to gracefully power off KVM instances
1703 - Split ``utils`` module into smaller parts
1704 - Changed query operations to return more detailed information, e.g.
1705 whether an information is unavailable due to an offline node. To use
1706 this new functionality, the LUXI call ``Query`` must be used. Field
1707 information is now stored by the master daemon and can be retrieved
1708 using ``QueryFields``. Instances, nodes and groups can also be queried
1709 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1710 exposed via RAPI). The following commands make use of this
1711 infrastructure change:
1713 - ``gnt-group list``
1714 - ``gnt-group list-fields``
1716 - ``gnt-node list-fields``
1717 - ``gnt-instance list``
1718 - ``gnt-instance list-fields``
1719 - ``gnt-debug locks``
1724 - New RAPI resources (see :doc:`rapi`):
1728 - ``/2/groups/[group_name]``
1729 - ``/2/groups/[group_name]/assign-nodes``
1730 - ``/2/groups/[group_name]/modify``
1731 - ``/2/groups/[group_name]/rename``
1732 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1736 - Implemented ``no_install`` for instance creation
1737 - Implemented OS parameters for instance reinstallation, allowing
1738 use of special settings on reinstallation (e.g. for preserving data)
1743 - Added IPv6 support in import/export
1744 - Pause DRBD synchronization while wiping disks on instance creation
1745 - Updated unittests and QA scripts
1746 - Improved network parameters passed to KVM
1747 - Converted man pages from docbook to reStructuredText
1753 *(Released Mon, 20 Dec 2010)*
1755 Released version 2.3.1~rc1 without any changes.
1761 *(Released Wed, 1 Dec 2010)*
1763 - impexpd: Disable OpenSSL compression in socat if possible (backport
1764 from master, commit e90739d625b, see :doc:`installation guide
1765 <install-quick>` for details)
1766 - Changed unittest coverage report to exclude test scripts
1767 - Added script to check version format
1773 *(Released Wed, 1 Dec 2010)*
1775 Released version 2.3.0~rc1 without any changes.
1781 *(Released Fri, 19 Nov 2010)*
1783 A number of bugfixes and documentation updates:
1785 - Update ganeti-os-interface documentation
1786 - Fixed a bug related to duplicate MACs or similar items which should be
1788 - Fix breakage in OS state modify
1789 - Reinstall instance: disallow offline secondaries (fixes bug related to
1790 OS changing but reinstall failing)
1791 - plus all the other fixes between 2.2.1 and 2.2.2
1797 *(Released Tue, 2 Nov 2010)*
1799 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1800 - Fixed master failover race with watcher
1801 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1803 - Accept previously stopped instance for export with instance removal
1804 - Simplify and extend the environment variables for instance OS scripts
1805 - Added new node flags, ``master_capable`` and ``vm_capable``
1806 - Added optional instance disk wiping prior during allocation. This is a
1807 cluster-wide option and can be set/modified using
1808 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1809 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1810 :doc:`install-quick`
1811 - Added a new watcher option (``--ignore-pause``)
1812 - Added option to ignore offline node on instance start/stop
1813 (``--ignore-offline``)
1814 - Allow overriding OS parameters with ``gnt-instance reinstall``
1815 - Added ability to change node's secondary IP address using ``gnt-node
1817 - Implemented privilege separation for all daemons except
1818 ``ganeti-noded``, see ``configure`` options
1819 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1821 - Implemented job priorities (see ``ganeti(7)`` manpage)
1822 - Ignore failures while shutting down instances during failover from
1824 - Exit daemon's bootstrap process only once daemon is ready
1825 - Export more information via ``LUInstanceQuery``/remote API
1826 - Improved documentation, QA and unittests
1827 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1828 restart if the file was created or changed
1829 - Added LUXI protocol version sent with each request and response,
1830 allowing detection of server/client mismatches
1831 - Moved the Python scripts among gnt-* and ganeti-* into modules
1832 - Moved all code related to setting up SSH to an external script,
1834 - Infrastructure changes for node group support in future versions
1840 *(Released Fri, 19 Nov 2010)*
1842 A few small bugs fixed, and some improvements to the build system:
1844 - Fix documentation regarding conversion to drbd
1845 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1847 - Fix error handling in node modify with multiple changes
1848 - Allow remote imports without checked names
1854 *(Released Tue, 19 Oct 2010)*
1856 - Disable SSL session ID cache in RPC client
1862 *(Released Thu, 14 Oct 2010)*
1864 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1865 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1867 - Fix problems with interaction between Curl and Python's HTTP server,
1868 resulting in increased speed in many RPC calls
1869 - Improve our release script to prevent breakage with older aclocal and
1876 *(Released Thu, 7 Oct 2010)*
1878 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1879 value retrieved from master
1880 - Added support for blacklisted or hidden OS definitions
1881 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1882 - Added support for -mem-path in KVM hypervisor abstraction layer
1883 - Allow overriding instance parameters in tool for inter-cluster
1884 instance moves (``tools/move-instance``)
1885 - Improved opcode summaries (e.g. in ``gnt-job list``)
1886 - Improve consistency of OS listing by sorting it
1887 - Documentation updates
1893 *(Released Fri, 8 Oct 2010)*
1895 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1901 *(Released Mon, 4 Oct 2010)*
1903 - Fixed regression in ``gnt-instance rename``
1909 *(Released Wed, 22 Sep 2010)*
1911 - Fixed OS_VARIANT variable for OS scripts
1912 - Fixed cluster tag operations via RAPI
1913 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1914 - Disabled RAPI CA checks in watcher
1920 *(Released Mon, 23 Aug 2010)*
1922 - Support DRBD versions of the format "a.b.c.d"
1924 - Re-introduce support for usage from multiple threads in RAPI client
1925 - Instance renames and modify via RAPI
1926 - Work around race condition between processing and archival in job
1928 - Mark opcodes following failed one as failed, too
1929 - Job field ``lock_status`` was removed due to difficulties making it
1930 work with the changed job queue in Ganeti 2.2; a better way to monitor
1931 locks is expected for a later 2.2.x release
1932 - Fixed dry-run behaviour with many commands
1933 - Support ``ssh-agent`` again when adding nodes
1934 - Many additional bugfixes
1940 *(Released Fri, 30 Jul 2010)*
1942 Important change: the internal RPC mechanism between Ganeti nodes has
1943 changed from using a home-grown http library (based on the Python base
1944 libraries) to use the PycURL library. This requires that PycURL is
1945 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1946 against GnuTLS by default. cURL's support for GnuTLS had known issues
1947 before cURL 7.21.0 and we recommend using the latest cURL release or
1948 linking against OpenSSL. Most other distributions already link PycURL
1949 and cURL against OpenSSL. The command::
1951 python -c 'import pycurl; print pycurl.version'
1953 can be used to determine the libraries PycURL and cURL are linked
1956 Other significant changes:
1958 - Rewrote much of the internals of the job queue, in order to achieve
1959 better parallelism; this decouples job query operations from the job
1960 processing, and it should allow much nicer behaviour of the master
1961 daemon under load, and it also has uncovered some long-standing bugs
1962 related to the job serialisation (now fixed)
1963 - Added a default iallocator setting to the cluster parameters,
1964 eliminating the need to always pass nodes or an iallocator for
1965 operations that require selection of new node(s)
1966 - Added experimental support for the LXC virtualization method
1967 - Added support for OS parameters, which allows the installation of
1968 instances to pass parameter to OS scripts in order to customise the
1970 - Added a hypervisor parameter controlling the migration type (live or
1971 non-live), since hypervisors have various levels of reliability; this
1972 has renamed the 'live' parameter to 'mode'
1973 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1974 logical volumes, meaning that cluster verify will ignore them and not
1975 flag their presence as errors
1976 - The watcher will now reset the error count for failed instances after
1977 8 hours, thus allowing self-healing if the problem that caused the
1978 instances to be down/fail to start has cleared in the meantime
1979 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1980 check for, and warn, if the drbd module parameter ``usermode_helper``
1981 is not consistent with the cluster-wide setting; this is needed to
1982 make diagnose easier of failed drbd creations
1983 - Started adding base IPv6 support, but this is not yet
1984 enabled/available for use
1985 - Rename operations (cluster, instance) will now return the new name,
1986 which is especially useful if a short name was passed in
1987 - Added support for instance migration in RAPI
1988 - Added a tool to pre-configure nodes for the SSH setup, before joining
1989 them to the cluster; this will allow in the future a simplified model
1990 for node joining (but not yet fully enabled in 2.2); this needs the
1991 paramiko python library
1992 - Fixed handling of name-resolving errors
1993 - Fixed consistency of job results on the error path
1994 - Fixed master-failover race condition when executed multiple times in
1996 - Fixed many bugs related to the job queue (mostly introduced during the
1997 2.2 development cycle, so not all are impacting 2.1)
1998 - Fixed instance migration with missing disk symlinks
1999 - Fixed handling of unknown jobs in ``gnt-job archive``
2000 - And many other small fixes/improvements
2004 - Enhanced both the unittest and the QA coverage
2005 - Switched the opcode validation to a generic model, and extended the
2006 validation to all opcode parameters
2007 - Changed more parts of the code that write shell scripts to use the
2009 - Switched the master daemon to use the asyncore library for the Luxi
2016 *(Released Thu, 17 Jun 2010)*
2018 - Added tool (``move-instance``) and infrastructure to move instances
2019 between separate clusters (see :doc:`separate documentation
2020 <move-instance>` and :doc:`design document <design-2.2>`)
2021 - Added per-request RPC timeout
2022 - RAPI now requires a Content-Type header for requests with a body (e.g.
2023 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2024 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2025 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2027 - Implemented initial support for running Ganeti daemons as separate
2028 users, see configure-time flags ``--with-user-prefix`` and
2029 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2031 - Instances can be removed after export (``gnt-backup export
2032 --remove-instance``)
2033 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2034 key (instead of 1024 bit)
2035 - Added new cluster configuration file for cluster domain secret
2036 - Import/export now use SSL instead of SSH
2037 - Added support for showing estimated time when exporting an instance,
2038 see the ``ganeti-os-interface(7)`` manpage and look for
2045 *(Released Tue, 16 Nov 2010)*
2047 Some more bugfixes. Unless critical bugs occur, this will be the last
2050 - Fix case of MAC special-values
2051 - Fix mac checker regex
2052 - backend: Fix typo causing "out of range" error
2053 - Add missing --units in gnt-instance list man page
2059 *(Released Tue, 24 Aug 2010)*
2062 - Don't ignore secondary node silently on non-mirrored disk templates
2064 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2065 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2066 - Properly document the usb_mouse parameter
2067 - Fix path in ganeti-rapi(8) (issue 116)
2068 - Adjust error message when the ganeti user's .ssh directory is
2070 - Add same-node-check when changing the disk template to drbd
2076 *(Released Fri, 16 Jul 2010)*
2079 - Add an option to only select some reboot types during qa/burnin.
2080 (on some hypervisors consequent reboots are not supported)
2081 - Fix infrequent race condition in master failover. Sometimes the old
2082 master ip address would be still detected as up for a short time
2083 after it was removed, causing failover to fail.
2084 - Decrease mlockall warnings when the ctypes module is missing. On
2085 Python 2.4 we support running even if no ctypes module is installed,
2086 but we were too verbose about this issue.
2087 - Fix building on old distributions, on which man doesn't have a
2089 - Fix RAPI not to ignore the MAC address on instance creation
2090 - Implement the old instance creation format in the RAPI client.
2096 *(Released Thu, 01 Jul 2010)*
2098 A small bugfix release:
2099 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2100 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2101 a non-existing option
2102 - Fix instance create: the --force-variant option was ignored
2103 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2104 - Fix modify node storage with non-FQDN arguments
2105 - Fix RAPI client to authenticate under Python 2.6 when used
2106 for more than 5 requests needing authentication
2107 - Fix gnt-instance modify -t (storage) giving a wrong error message
2108 when converting a non-shutdown drbd instance to plain
2114 *(Released Fri, 18 Jun 2010)*
2116 A small bugfix release:
2118 - Fix live migration of KVM instances started with older Ganeti
2119 versions which had fewer hypervisor parameters
2120 - Fix gnt-instance grow-disk on down instances
2121 - Fix an error-reporting bug during instance migration
2122 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2123 silently ignoring broken ones
2124 - Fix an RPC error reporting bug affecting, for example, RAPI client
2126 - Fix bug triggered by different API version os-es on different nodes
2127 - Fix a bug in instance startup with custom hvparams: OS level
2128 parameters would fail to be applied.
2129 - Fix the RAPI client under Python 2.6 (but more work is needed to
2130 make it work completely well with OpenSSL)
2131 - Fix handling of errors when resolving names from DNS
2137 *(Released Thu, 3 Jun 2010)*
2139 A medium sized development cycle. Some new features, and some
2140 fixes/small improvements/cleanups.
2142 Significant features
2143 ~~~~~~~~~~~~~~~~~~~~
2145 The node deamon now tries to mlock itself into memory, unless the
2146 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2147 its logs, and falls back to console logging. This allows emergency
2148 features such as ``gnt-node powercycle`` to work even in the event of a
2149 broken node disk (tested offlining the disk hosting the node's
2150 filesystem and dropping its memory caches; don't try this at home)
2152 KVM: add vhost-net acceleration support. It can be tested with a new
2153 enough version of the kernel and of qemu-kvm.
2155 KVM: Add instance chrooting feature. If you use privilege dropping for
2156 your VMs you can also now force them to chroot to an empty directory,
2157 before starting the emulated guest.
2159 KVM: Add maximum migration bandwith and maximum downtime tweaking
2160 support (requires a new-enough version of qemu-kvm).
2162 Cluster verify will now warn if the master node doesn't have the master
2163 ip configured on it.
2165 Add a new (incompatible) instance creation request format to RAPI which
2166 supports all parameters (previously only a subset was supported, and it
2167 wasn't possible to extend the old format to accomodate all the new
2168 features. The old format is still supported, and a client can check for
2169 this feature, before using it, by checking for its presence in the
2170 ``features`` RAPI resource.
2172 Now with ancient latin support. Try it passing the ``--roman`` option to
2173 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2174 (requires the python-roman module to be installed, in order to work).
2179 As usual many internal code refactorings, documentation updates, and
2182 - Lots of improvements and cleanups to the experimental Remote API
2183 (RAPI) client library.
2184 - A new unit test suite for the core daemon libraries.
2185 - A fix to creating missing directories makes sure the umask is not
2186 applied anymore. This enforces the same directory permissions
2188 - Better handling terminating daemons with ctrl+c (used when running
2189 them in debugging mode).
2190 - Fix a race condition in live migrating a KVM instance, when stat()
2191 on the old proc status file returned EINVAL, which is an unexpected
2193 - Fixed manpage checking with newer man and utf-8 charachters. But now
2194 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2200 *(Released Fri, 7 May 2010)*
2202 Fix a bug which prevented untagged KVM instances from starting.
2208 *(Released Fri, 7 May 2010)*
2210 Another release with a long development cycle, during which many
2211 different features were added.
2213 Significant features
2214 ~~~~~~~~~~~~~~~~~~~~
2216 The KVM hypervisor now can run the individual instances as non-root, to
2217 reduce the impact of a VM being hijacked due to bugs in the
2218 hypervisor. It is possible to run all instances as a single (non-root)
2219 user, to manually specify a user for each instance, or to dynamically
2220 allocate a user out of a cluster-wide pool to each instance, with the
2221 guarantee that no two instances will run under the same user ID on any
2224 An experimental RAPI client library, that can be used standalone
2225 (without the other Ganeti libraries), is provided in the source tree as
2226 ``lib/rapi/client.py``. Note this client might change its interface in
2227 the future, as we iterate on its capabilities.
2229 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2230 replace the cluster's certificates and crypto keys. This might help in
2231 case they have been compromised, or have simply expired.
2233 A new disk option for instance creation has been added that allows one
2234 to "adopt" currently existing logical volumes, with data
2235 preservation. This should allow easier migration to Ganeti from
2236 unmanaged (or managed via other software) instances.
2238 Another disk improvement is the possibility to convert between redundant
2239 (DRBD) and plain (LVM) disk configuration for an instance. This should
2240 allow better scalability (starting with one node and growing the
2241 cluster, or shrinking a two-node cluster to one node).
2243 A new feature that could help with automated node failovers has been
2244 implemented: if a node sees itself as offline (by querying the master
2245 candidates), it will try to shutdown (hard) all instances and any active
2246 DRBD devices. This reduces the risk of duplicate instances if an
2247 external script automatically failovers the instances on such nodes. To
2248 enable this, the cluster parameter ``maintain_node_health`` should be
2249 enabled; in the future this option (per the name) will enable other
2250 automatic maintenance features.
2252 Instance export/import now will reuse the original instance
2253 specifications for all parameters; that means exporting an instance,
2254 deleting it and the importing it back should give an almost identical
2255 instance. Note that the default import behaviour has changed from
2256 before, where it created only one NIC; now it recreates the original
2259 Cluster verify has added a few new checks: SSL certificates validity,
2260 /etc/hosts consistency across the cluster, etc.
2265 As usual, many internal changes were done, documentation fixes,
2268 - Fixed cluster initialization with disabled cluster storage (regression
2269 introduced in 2.1.1)
2270 - File-based storage supports growing the disks
2271 - Fixed behaviour of node role changes
2272 - Fixed cluster verify for some corner cases, plus a general rewrite of
2273 cluster verify to allow future extension with more checks
2274 - Fixed log spamming by watcher and node daemon (regression introduced
2276 - Fixed possible validation issues when changing the list of enabled
2278 - Fixed cleanup of /etc/hosts during node removal
2279 - Fixed RAPI response for invalid methods
2280 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2281 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2282 from ide disks, etc.)
2283 - Allow OS changes without re-installation (to record a changed OS
2284 outside of Ganeti, or to allow OS renames)
2285 - Allow instance creation without OS installation (useful for example if
2286 the OS will be installed manually, or restored from a backup not in
2288 - Implemented option to make cluster ``copyfile`` use the replication
2290 - Added list of enabled hypervisors to ssconf (possibly useful for
2292 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2294 - A partial form of node re-IP is possible via node readd, which now
2295 allows changed node primary IP
2296 - Command line utilities now show an informational message if the job is
2298 - The logs of the master daemon now show the PID/UID/GID of the
2305 *(Released Fri, 12 Mar 2010)*
2307 During the 2.1.0 long release candidate cycle, a lot of improvements and
2308 changes have accumulated with were released later as 2.1.1.
2313 The node evacuate command (``gnt-node evacuate``) was significantly
2314 rewritten, and as such the IAllocator protocol was changed - a new
2315 request type has been added. This unfortunate change during a stable
2316 series is designed to improve performance of node evacuations; on
2317 clusters with more than about five nodes and which are well-balanced,
2318 evacuation should proceed in parallel for all instances of the node
2319 being evacuated. As such, any existing IAllocator scripts need to be
2320 updated, otherwise the above command will fail due to the unknown
2321 request. The provided "dumb" allocator has not been updated; but the
2322 ganeti-htools package supports the new protocol since version 0.2.4.
2324 Another important change is increased validation of node and instance
2325 names. This might create problems in special cases, if invalid host
2326 names are being used.
2328 Also, a new layer of hypervisor parameters has been added, that sits at
2329 OS level between the cluster defaults and the instance ones. This allows
2330 customisation of virtualization parameters depending on the installed
2331 OS. For example instances with OS 'X' may have a different KVM kernel
2332 (or any other parameter) than the cluster defaults. This is intended to
2333 help managing a multiple OSes on the same cluster, without manual
2334 modification of each instance's parameters.
2336 A tool for merging clusters, ``cluster-merge``, has been added in the
2337 tools sub-directory.
2342 - Improved the int/float conversions that should make the code more
2343 robust in face of errors from the node daemons
2344 - Fixed the remove node code in case of internal configuration errors
2345 - Fixed the node daemon behaviour in face of inconsistent queue
2346 directory (e.g. read-only file-system where we can't open the files
2348 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2349 now it either aborts cleanly or, if given the new "auto_promote"
2350 parameter, will automatically promote other nodes as needed
2351 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2352 completely prevent Ganeti from working
2353 - Fixed bug for instance export when not all disks were successfully
2355 - Fixed behaviour of node add when the new node is slow in starting up
2357 - Fixed handling of signals in the LUXI client, which should improve
2358 behaviour of command-line scripts
2359 - Added checks for invalid node/instance names in the configuration (now
2360 flagged during cluster verify)
2361 - Fixed watcher behaviour for disk activation errors
2362 - Fixed two potentially endless loops in http library, which led to the
2363 RAPI daemon hanging and consuming 100% CPU in some cases
2364 - Fixed bug in RAPI daemon related to hashed passwords
2365 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2367 - Enhanced compatibility with non-Debian OSes, but not using absolute
2368 path in some commands and allowing customisation of the ssh
2369 configuration directory
2370 - Fixed possible future issue with new Python versions by abiding to the
2371 proper use of ``__slots__`` attribute on classes
2372 - Added checks that should prevent directory traversal attacks
2373 - Many documentation fixes based on feedback from users
2378 - Added an "early_release" more for instance replace disks and node
2379 evacuate, where we release locks earlier and thus allow higher
2380 parallelism within the cluster
2381 - Added watcher hooks, intended to allow the watcher to restart other
2382 daemons (e.g. from the ganeti-nbma project), but they can be used of
2383 course for any other purpose
2384 - Added a compile-time disable for DRBD barriers, to increase
2385 performance if the administrator trusts the power supply or the
2386 storage system to not lose writes
2387 - Added the option of using syslog for logging instead of, or in
2388 addition to, Ganeti's own log files
2389 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2390 can indeed boot from a paravirtual NIC
2391 - Added a generic debug level for many operations; while this is not
2392 used widely yet, it allows one to pass the debug value all the way to
2394 - Enhanced the hooks environment for instance moves (failovers,
2395 migrations) where the primary/secondary nodes changed during the
2396 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2397 - Enhanced data validations for many user-supplied values; one important
2398 item is the restrictions imposed on instance and node names, which
2399 might reject some (invalid) host names
2400 - Add a configure-time option to disable file-based storage, if it's not
2401 needed; this allows greater security separation between the master
2402 node and the other nodes from the point of view of the inter-node RPC
2404 - Added user notification in interactive tools if job is waiting in the
2405 job queue or trying to acquire locks
2406 - Added log messages when a job is waiting for locks
2407 - Added filtering by node tags in instance operations which admit
2408 multiple instances (start, stop, reboot, reinstall)
2409 - Added a new tool for cluster mergers, ``cluster-merge``
2410 - Parameters from command line which are of the form ``a=b,c=d`` can now
2411 use backslash escapes to pass in values which contain commas,
2412 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2414 - For KVM, the instance name is the first parameter passed to KVM, so
2415 that it's more visible in the process list
2421 *(Released Tue, 2 Mar 2010)*
2423 Ganeti 2.1 brings many improvements with it. Major changes:
2425 - Added infrastructure to ease automated disk repairs
2426 - Added new daemon to export configuration data in a cheaper way than
2427 using the remote API
2428 - Instance NICs can now be routed instead of being associated with a
2430 - Improved job locking logic to reduce impact of jobs acquiring multiple
2431 locks waiting for other long-running jobs
2433 In-depth implementation details can be found in the Ganeti 2.1 design
2439 - Added chroot hypervisor
2440 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2442 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2443 ``bootloader_path`` and ``bootloader_args``)
2444 - Added the ``use_localtime`` option for the xen-hvm and kvm
2445 hypervisors, and the default value for this has changed to false (in
2446 2.0 xen-hvm always enabled it)
2447 - Added luxi call to submit multiple jobs in one go
2448 - Added cluster initialization option to not modify ``/etc/hosts``
2450 - Added network interface parameters
2451 - Added dry run mode to some LUs
2452 - Added RAPI resources:
2454 - ``/2/instances/[instance_name]/info``
2455 - ``/2/instances/[instance_name]/replace-disks``
2456 - ``/2/nodes/[node_name]/evacuate``
2457 - ``/2/nodes/[node_name]/migrate``
2458 - ``/2/nodes/[node_name]/role``
2459 - ``/2/nodes/[node_name]/storage``
2460 - ``/2/nodes/[node_name]/storage/modify``
2461 - ``/2/nodes/[node_name]/storage/repair``
2463 - Added OpCodes to evacuate or migrate all instances on a node
2464 - Added new command to list storage elements on nodes (``gnt-node
2465 list-storage``) and modify them (``gnt-node modify-storage``)
2466 - Added new ssconf files with master candidate IP address
2467 (``ssconf_master_candidates_ips``), node primary IP address
2468 (``ssconf_node_primary_ips``) and node secondary IP address
2469 (``ssconf_node_secondary_ips``)
2470 - Added ``ganeti-confd`` and a client library to query the Ganeti
2471 configuration via UDP
2472 - Added ability to run hooks after cluster initialization and before
2474 - Added automatic mode for disk replace (``gnt-instance replace-disks
2476 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2477 after catastrophic data-loss
2478 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2480 - Added ``gnt-instance move`` command to move instances
2481 - Added ``gnt-cluster watcher`` command to control watcher
2482 - Added ``gnt-node powercycle`` command to powercycle nodes
2483 - Added new job status field ``lock_status``
2484 - Added parseable error codes to cluster verification (``gnt-cluster
2485 verify --error-codes``) and made output less verbose (use
2486 ``--verbose`` to restore previous behaviour)
2487 - Added UUIDs to the main config entities (cluster, nodes, instances)
2488 - Added support for OS variants
2489 - Added support for hashed passwords in the Ganeti remote API users file
2491 - Added option to specify maximum timeout on instance shutdown
2492 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2493 - Added new helper script to start and stop Ganeti daemons
2494 (``daemon-util``), with the intent to reduce the work necessary to
2495 adjust Ganeti for non-Debian distributions and to start/stop daemons
2497 - Added more unittests
2498 - Fixed critical bug in ganeti-masterd startup
2499 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2500 now customisable at the cluster level for both the KVM and Xen
2501 hypervisors using the new ``migration_port`` parameter
2502 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2503 reinstalling an instance
2504 - Allowed ``@`` in tag names
2505 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2506 - Many documentation updates
2507 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2508 - ``gnt-instance reinstall`` can now reinstall multiple instances
2509 - Updated many command line parameters
2510 - Introduced new OS API version 15
2511 - No longer support a default hypervisor
2512 - Treat virtual LVs as inexistent
2513 - Improved job locking logic to reduce lock contention
2514 - Match instance and node names case insensitively
2515 - Reimplemented bash completion script to be more complete
2522 *(Released Thu, 4 Feb 2010)*
2524 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2525 - Fix a string formatting bug
2526 - Improve safety of the code in some error paths
2527 - Improve data validation in the master of values returned from nodes
2533 *(Released Thu, 17 Dec 2009)*
2535 - Fix security issue due to missing validation of iallocator names; this
2536 allows local and remote execution of arbitrary executables
2537 - Fix failure of gnt-node list during instance removal
2538 - Ship the RAPI documentation in the archive
2544 *(Released Wed, 30 Sep 2009)*
2546 - Fixed many wrong messages
2547 - Fixed a few bugs related to the locking library
2548 - Fixed MAC checking at instance creation time
2549 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2550 - Fixed a few issues related to signal handling in both daemons and
2552 - Fixed the example startup script provided
2553 - Fixed insserv dependencies in the example startup script (patch from
2555 - Fixed handling of drained nodes in the iallocator framework
2556 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2558 - Fixed error related to invalid job IDs in job polling
2559 - Fixed job/opcode persistence on unclean master shutdown
2560 - Fixed handling of partial job processing after unclean master
2562 - Fixed error reporting from LUs, previously all errors were converted
2563 into execution errors
2564 - Fixed error reporting from burnin
2565 - Decreased significantly the memory usage of the job queue
2566 - Optimised slightly multi-job submission
2567 - Optimised slightly opcode loading
2568 - Backported the multi-job submit framework from the development
2569 branch; multi-instance start and stop should be faster
2570 - Added script to clean archived jobs after 21 days; this will reduce
2571 the size of the queue directory
2572 - Added some extra checks in disk size tracking
2573 - Added an example ethers hook script
2574 - Added a cluster parameter that prevents Ganeti from modifying of
2576 - Added more node information to RAPI responses
2577 - Added a ``gnt-job watch`` command that allows following the ouput of a
2579 - Added a bind-address option to ganeti-rapi
2580 - Added more checks to the configuration verify
2581 - Enhanced the burnin script such that some operations can be retried
2583 - Converted instance reinstall to multi-instance model
2589 *(Released Fri, 7 Aug 2009)*
2591 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2592 to allow using the pre-2.0.2 behaviour in activation, if any existing
2593 instances have mismatched disk sizes in the configuration
2594 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2595 any configuration mismatches for disk sizes
2596 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2597 failover to work on two-node clusters
2598 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2600 - Fixed detection of OS script errors in ``gnt-backup export``
2601 - Fixed exit code of ``gnt-backup export``
2607 *(Released Fri, 17 Jul 2009)*
2609 - Added experimental support for stripped logical volumes; this should
2610 enhance performance but comes with a higher complexity in the block
2611 device handling; stripping is only enabled when passing
2612 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2613 affected even in the non-stripped mode
2614 - Improved resiliency against transient failures at the end of DRBD
2615 resyncs, and in general of DRBD resync checks
2616 - Fixed a couple of issues with exports and snapshot errors
2617 - Fixed a couple of issues in instance listing
2618 - Added display of the disk size in ``gnt-instance info``
2619 - Fixed checking for valid OSes in instance creation
2620 - Fixed handling of the "vcpus" parameter in instance listing and in
2621 general of invalid parameters
2622 - Fixed http server library, and thus RAPI, to handle invalid
2623 username/password combinations correctly; this means that now they
2624 report unauthorized for queries too, not only for modifications,
2625 allowing earlier detect of configuration problems
2626 - Added a new "role" node list field, equivalent to the master/master
2627 candidate/drained/offline flags combinations
2628 - Fixed cluster modify and changes of candidate pool size
2629 - Fixed cluster verify error messages for wrong files on regular nodes
2630 - Fixed a couple of issues with node demotion from master candidate role
2631 - Fixed node readd issues
2632 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2633 - Added a new ``--no-voting`` option for masterfailover to fix failover
2634 on two-nodes clusters when the former master node is unreachable
2635 - Added instance reinstall over RAPI
2641 *(Released Tue, 16 Jun 2009)*
2643 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2644 allow re-adding the start in single-user option (regression from 1.2)
2645 - the watcher writes the instance status to a file, to allow monitoring
2646 to report the instance status (from the master) based on cached
2647 results of the watcher's queries; while this can get stale if the
2648 watcher is being locked due to other work on the cluster, this is
2649 still an improvement
2650 - the watcher now also restarts the node daemon and the rapi daemon if
2652 - fixed the watcher to handle full and drained queue cases
2653 - hooks export more instance data in the environment, which helps if
2654 hook scripts need to take action based on the instance's properties
2655 (no longer need to query back into ganeti)
2656 - instance failovers when the instance is stopped do not check for free
2657 RAM, so that failing over a stopped instance is possible in low memory
2659 - rapi uses queries for tags instead of jobs (for less job traffic), and
2660 for cluster tags it won't talk to masterd at all but read them from
2662 - a couple of error handling fixes in RAPI
2663 - drbd handling: improved the error handling of inconsistent disks after
2664 resync to reduce the frequency of "there are some degraded disks for
2665 this instance" messages
2666 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2667 error handling path called a wrong function name)
2673 *(Released Wed, 27 May 2009)*
2675 - no changes from rc5
2681 *(Released Wed, 20 May 2009)*
2683 - fix a couple of bugs (validation, argument checks)
2684 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2685 - some small improvements to RAPI and IAllocator
2686 - make watcher automatically start the master daemon if down
2692 *(Released Mon, 27 Apr 2009)*
2694 - change the OS list to not require locks; this helps with big clusters
2695 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2696 volume group is broken
2697 - ``gnt-instance info``, without any arguments, doesn't run for all
2698 instances anymore; either pass ``--all`` or pass the desired
2699 instances; this helps against mistakes on big clusters where listing
2700 the information for all instances takes a long time
2701 - miscellaneous doc and man pages fixes
2707 *(Released Wed, 8 Apr 2009)*
2709 - Change the internal locking model of some ``gnt-node`` commands, in
2710 order to reduce contention (and blocking of master daemon) when
2711 batching many creation/reinstall jobs
2712 - Fixes to Xen soft reboot
2713 - No longer build documentation at build time, instead distribute it in
2714 the archive, in order to reduce the need for the whole docbook/rst
2721 *(Released Fri, 27 Mar 2009)*
2723 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2724 - Fix watcher startup sequence, improves the behaviour of busy clusters
2725 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2726 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2727 bind address and other places
2728 - Some documentation fixes and updates
2734 *(Released Mon, 2 Mar 2009)*
2736 - More documentation updates, now all docs should be more-or-less
2738 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2740 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2742 - Fix an issue related to $libdir/run/ganeti and cluster creation
2748 *(Released Thu, 19 Feb 2009)*
2750 - Xen PVM and KVM have switched the default value for the instance root
2751 disk to the first partition on the first drive, instead of the whole
2752 drive; this means that the OS installation scripts must be changed
2754 - Man pages have been updated
2755 - RAPI has been switched by default to HTTPS, and the exported functions
2756 should all work correctly
2757 - RAPI v1 has been removed
2758 - Many improvements to the KVM hypervisor
2759 - Block device errors are now better reported
2760 - Many other bugfixes and small improvements
2766 *(Released Mon, 26 Jan 2009)*
2768 - Version 2 is a general rewrite of the code and therefore the
2769 differences are too many to list, see the design document for 2.0 in
2770 the ``doc/`` subdirectory for more details
2771 - In this beta version there is not yet a migration path from 1.2 (there
2772 will be one in the final 2.0 release)
2773 - A few significant changes are:
2775 - all commands are executed by a daemon (``ganeti-masterd``) and the
2776 various ``gnt-*`` commands are just front-ends to it
2777 - all the commands are entered into, and executed from a job queue,
2778 see the ``gnt-job(8)`` manpage
2779 - the RAPI daemon supports read-write operations, secured by basic
2780 HTTP authentication on top of HTTPS
2781 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2782 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2783 to migrate to DRBD 8 first while still running Ganeti 1.2)
2784 - DRBD devices are using statically allocated minor numbers, which
2785 will be assigned to existing instances during the migration process
2786 - there is support for both Xen PVM and Xen HVM instances running on
2788 - KVM virtualization is supported too
2789 - file-based storage has been implemented, which means that it is
2790 possible to run the cluster without LVM and DRBD storage, for
2791 example using a shared filesystem exported from shared storage (and
2792 still have live migration)
2798 *(Released Tue, 13 Jan 2009)*
2800 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2801 - Reuse the old instance mac address by default on instance import, if
2802 the instance name is the same.
2803 - Handle situations in which the node info rpc returns incomplete
2805 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2806 - Improved version of batcher:
2808 - state file support
2809 - instance mac address support
2810 - support for HVM clusters/instances
2812 - Add an option to show the number of cpu sockets and nodes in
2814 - Support OSes that handle more than one version of the OS api (but do
2815 not change the current API in any other way)
2816 - Fix ``gnt-node migrate``
2817 - ``gnt-debug`` man page
2818 - Fixes various more typos and small issues
2819 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2825 *(Released Wed, 24 Sep 2008)*
2827 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2828 type of disk exported to fully virtualized instances.
2829 - provide access to the serial console of HVM instances
2830 - instance auto_balance flag, set by default. If turned off it will
2831 avoid warnings on cluster verify if there is not enough memory to fail
2832 over an instance. in the future it will prevent automatically failing
2833 it over when we will support that.
2834 - batcher tool for instance creation, see ``tools/README.batcher``
2835 - ``gnt-instance reinstall --select-os`` to interactively select a new
2836 operating system when reinstalling an instance.
2837 - when changing the memory amount on instance modify a check has been
2838 added that the instance will be able to start. also warnings are
2839 emitted if the instance will not be able to fail over, if auto_balance
2841 - documentation fixes
2842 - sync fields between ``gnt-instance list/modify/add/import``
2843 - fix a race condition in drbd when the sync speed was set after giving
2844 the device a remote peer.
2850 *(Released Tue, 22 Jul 2008)*
2852 - note: the allowed size and number of tags per object were reduced
2853 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2854 - fixed twisted 8.x compatibility
2855 - fixed ``gnt-instance replace-disks`` with iallocator
2856 - add TCP keepalives on twisted connections to detect restarted nodes
2857 - disk increase support, see ``gnt-instance grow-disk``
2858 - implement bulk node/instance query for RAPI
2859 - add tags in node/instance listing (optional)
2860 - experimental migration (and live migration) support, read the man page
2861 for ``gnt-instance migrate``
2862 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2863 has some small improvements in handling its state file
2869 *(Released Fri, 13 Jun 2008)*
2871 - Experimental readonly, REST-based remote API implementation;
2872 automatically started on master node, TCP port 5080, if enabled by
2873 ``--enable-rapi`` parameter to configure script.
2874 - Instance allocator support. Add and import instance accept a
2875 ``--iallocator`` parameter, and call that instance allocator to decide
2876 which node to use for the instance. The iallocator document describes
2877 what's expected from an allocator script.
2878 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2879 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2880 node is lost there is still enough memory to fail over the instances
2882 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2883 ``gnt-cluster verify``, to check for site-specific compliance. All the
2884 hooks will run, and their output, if any, will be displayed. Any
2885 failing hook will make the verification return an error value.
2886 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2887 primary and secondary interfaces
2888 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2889 that is still declared as part of the cluster and has failed.
2890 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2891 specifying output fields, that just adds the chosen fields to the
2893 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2894 export from the filesystem.
2895 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2896 have been added. Using them you can enable/disable acpi and pae
2897 support, and specify a path for a cd image to be exported to the
2898 instance. These parameters as the name suggest only work on HVM
2900 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2901 PAE support will be set to the previously hardcoded values, but the
2902 (previously hardcoded) path to the CDROM ISO image will be unset and
2903 if required, needs to be set manually with ``gnt-instance modify``
2905 - The address to which an instance's VNC console is bound is now
2906 selectable per-instance, rather than being cluster wide. Of course
2907 this only applies to instances controlled via VNC, so currently just
2908 applies to HVM clusters.
2914 *(Released Mon, 18 Feb 2008)*
2916 - more tweaks to the disk activation code (especially helpful for DRBD)
2917 - change the default ``gnt-instance list`` output format, now there is
2918 one combined status field (see the manpage for the exact values this
2920 - some more fixes for the mac export to hooks change
2921 - make Ganeti not break with DRBD 8.2.x (which changed the version
2922 format in ``/proc/drbd``) (issue 24)
2923 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2924 template, allowing migration from DRBD0.7+MD to DRBD8
2930 *(Released Wed, 30 Jan 2008)*
2932 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2934 - add command aliases infrastructure and a few aliases
2935 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2936 man pages and the ``--help`` option of ``gnt-node
2937 list``/``gnt-instance list``
2938 - fix ``gnt-backup list`` with down nodes (issue 21)
2939 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2940 - fix the dist archive and add a check for including svn/git files in
2942 - some developer-related changes: improve the burnin and the QA suite,
2943 add an upload script for testing during development
2949 *(Released Wed, 16 Jan 2008)*
2951 - experimental HVM support, read the install document, section
2952 "Initializing the cluster"
2953 - allow for the PVM hypervisor per-instance kernel and initrd paths
2954 - add a new command ``gnt-cluster verify-disks`` which uses a new
2955 algorithm to improve the reconnection of the DRBD pairs if the device
2956 on the secondary node has gone away
2957 - make logical volume code auto-activate LVs at disk activation time
2958 - slightly improve the speed of activating disks
2959 - allow specification of the MAC address at instance creation time, and
2960 changing it later via ``gnt-instance modify``
2961 - fix handling of external commands that generate lots of output on
2963 - update documentation with regard to minimum version of DRBD8 supported
2969 *(Released Tue, 4 Dec 2007)*
2971 - Log the ``xm create`` output to the node daemon log on failure (to
2972 help diagnosing the error)
2973 - In debug mode, log all external commands output if failed to the logs
2974 - Change parsing of lvm commands to ignore stderr
2980 *(Released Wed, 28 Nov 2007)*
2982 - Another round of updates to the DRBD 8 code to deal with more failures
2983 in the replace secondary node operation
2984 - Some more logging of failures in disk operations (lvm, drbd)
2985 - A few documentation updates
2992 *(Released Tue, 13 Nov 2007)*
2994 - Change configuration file format from Python's Pickle to JSON.
2995 Upgrading is possible using the cfgupgrade utility.
2996 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2997 faster replace disks and is more stable (DRBD 8 has many improvements
2998 compared to DRBD 0.7)
2999 - Added command line tags support (see man pages for ``gnt-instance``,
3000 ``gnt-node``, ``gnt-cluster``)
3001 - Added instance rename support
3002 - Added multi-instance startup/shutdown
3003 - Added cluster rename support
3004 - Added ``gnt-node evacuate`` to simplify some node operations
3005 - Added instance reboot operation that can speedup reboot as compared to
3007 - Soften the requirement that hostnames are in FQDN format
3008 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3010 - Removed dependency on debian's patched fping that uses the
3011 non-standard ``-S`` option
3012 - Now the OS definitions are searched for in multiple, configurable
3013 paths (easier for distros to package)
3014 - Some changes to the hooks infrastructure (especially the new
3015 post-configuration update hook)
3016 - Other small bugfixes
3018 .. vim: set textwidth=72 syntax=rst :