8 *(Released Tue, 15 Oct 2013)*
10 Incompatible/important changes
11 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
13 - hroller now also plans for capacity to move non-redundant instances off
14 any node to be rebooted; the old behavior of completely ignoring any
15 non-redundant instances can be restored by adding the --ignore-non-redundant
17 - The cluster option '--no-lvm-storage' was removed in favor of the new option
18 '--enabled-disk-templates'.
19 - On instance creation, disk templates no longer need to be specified
20 with '-t'. The default disk template will be taken from the list of
21 enabled disk templates.
22 - The monitoring daemon is now running as root, in order to be able to collect
23 information only available to root (such as the state of Xen instances).
24 - The ConfD client is now IPv6 compatible.
25 - File and shared file storage is no longer dis/enabled at configure time,
26 but using the option '--enabled-disk-templates' at cluster initialization and
28 - The default directories for file and shared file storage are not anymore
29 specified at configure time, but taken from the cluster's configuration.
30 They can be set at cluster initialization and modification with
31 '--file-storage-dir' and '--shared-file-storage-dir'.
32 - Cluster verification now includes stricter checks regarding the
33 default file and shared file storage directories. It now checks that
34 the directories are explicitely allowed in the 'file-storage-paths' file and
35 that the directories exist on all nodes.
36 - The list of allowed disk templates in the instance policy and the list
37 of cluster-wide enabled disk templates is now checked for consistency
38 on cluster or group modification. On cluster initialization, the ipolicy
39 disk templates are ensured to be a subset of the cluster-wide enabled
45 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
46 the correct command syntax. It is possible to use different DRBD versions
47 on different nodes as long as they are compatible to each other. This
48 enables rolling upgrades of DRBD with no downtime. As permanent operation
49 of different DRBD versions within a node group is discouraged,
50 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
51 - New "inst-status-xen" data collector for the monitoring daemon, providing
52 information about the state of the xen instances on the nodes.
53 - New "lv" data collector for the monitoring daemon, collecting data about the
54 logical volumes on the nodes, and pairing them with the name of the instances
56 - New "diskstats" data collector, collecting the data from /proc/diskstats and
57 presenting them over the monitoring daemon interface.
58 - The ConfD client is now IPv6 compatible.
62 The following new dependencies have been added.
66 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
67 for the unit tests (and only used for testing).
71 - ``hslogger`` (http://software.complete.org/hslogger) is now always
72 required, even if confd is not enabled.
77 - in implicit configuration upgrade, match ipolicy with enabled disk templates
78 - improved harep documentation (inherited from stable-2.8)
84 *(Released Wed, 9 Oct 2013)*
86 The second release candidate in the 2.9 series. Since 2.9.0 rc1:
88 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
89 at least one DRBD instance.
90 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
97 *(Released Tue, 1 Oct 2013)*
99 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
102 - update of the documentation, in particular installation instructions
103 - merging of LD_* constants into DT_* constants
104 - python style changes to be compatible with newer versions of pylint
110 *(Released Thu, 29 Aug 2013)*
112 This was the first beta release of the 2.9 series. All important changes
113 are listed in the latest 2.9 entry.
119 *(Released Mon, 30 Sep 2013)*
121 Incompatible/important changes
122 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
124 - Instance policy can contain multiple instance specs, as described in
125 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
126 <design-partitioned>`. As a consequence, it's not possible to partially change
127 or override instance specs. Bounding specs (min and max) can be specified as a
128 whole using the new option ``--ipolicy-bounds-specs``, while standard
129 specs use the new option ``--ipolicy-std-specs``.
130 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
131 gnt-instance is a valid YAML object.
132 - hail now honors network restrictions when allocating nodes. This led to an
133 update of the IAllocator protocol. See the IAllocator documentation for
135 - confd now only answers static configuration request over the network. luxid
136 was extracted, listens on the local LUXI socket and responds to live queries.
137 This allows finer grained permissions if using separate users.
142 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
143 to always require authentication, ``--require-authentication``. It can
144 be specified in ``$sysconfdir/default/ganeti``.
145 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
146 be used to manage the disk templates to be used by instances in the cluster.
147 Initially, it will be set to a list that includes plain, drbd, if they were
148 enabled by specifying a volume group name, and file and sharedfile, if those
149 were enabled at configure time. Additionally, it will include all disk
150 templates that are currently used by instances. The order of disk templates
151 will be based on Ganeti's history of supporting them. In the future, the
152 first entry of the list will be used as a default disk template on instance
154 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
155 configuration back to the previous stable version.
156 - Disk templates in group ipolicy can be restored to the default value.
157 - Initial support for diskless instances and virtual clusters in QA.
158 - More QA and unit tests for instance policies.
159 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
160 describing why the opcode itself was executed.
161 - The monitoring daemon is now available. It allows users to query the cluster
162 for obtaining information about the status of the system. The daemon is only
163 responsible for providing the information over the network: the actual data
164 gathering is performed by data collectors (currently, only the DRBD status
165 collector is available).
166 - In order to help developers work on Ganeti, a new script
167 (``devel/build_chroot``) is provided, for building a chroot that contains all
168 the required development libraries and tools for compiling Ganeti on a Debian
170 - A new tool, ``harep``, for performing self-repair and recreation of instances
171 in Ganeti has been added.
172 - Split queries are enabled for tags, network, exports, cluster info, groups,
174 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
175 It prints the command line to set the current policies, to ease
177 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
178 devices for KVM virtio-net interfaces will get created with VNET_HDR
179 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
180 virtio-net interfaces, which prevents host kernel tainting and log
181 flooding, when dealing with broken or malicious virtio-net drivers.
182 It's set to true by default.
183 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
185 - Support 'viridian' parameter in Xen HVM
186 - Support DSA SSH keys in bootstrap
187 - To simplify the work of packaging frameworks that want to add the needed users
188 and groups in a split-user setup themselves, at build time three files in
189 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
190 the groups to be generated, the ``users`` file contains, one per line, the
191 users to be generated, optionally followed by their primary group, where
192 important. The ``groupmemberships`` file contains, one per line, additional
193 user-group membership relations that need to be established. The syntax of
194 these files will remain stable in all future versions.
199 The following new dependencies have been added:
202 - The ``curl`` library is not optional anymore for compiling the Haskell code.
203 - ``snap-server`` library (if monitoring is enabled).
206 - The minimum Python version needed to run Ganeti is now 2.6.
207 - ``yaml`` library (only for running the QA).
211 - Perform proper cleanup on termination of Haskell daemons
212 - Fix corner-case in handling of remaining retry time
218 *(Released Tue, 17 Sep 2013)*
220 - To simplify the work of packaging frameworks that want to add the needed users
221 and groups in a split-user setup themselves, at build time three files in
222 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
223 the groups to be generated, the ``users`` file contains, one per line, the
224 users to be generated, optionally followed by their primary group, where
225 important. The ``groupmemberships`` file contains, one per line, additional
226 user-group membership relations that need to be established. The syntax of
227 these files will remain stable in all future versions.
228 - Add a default to file-driver when unspecified over RAPI (Issue 571)
229 - Mark the DSA host pubkey as optional, and remove it during config downgrade
231 - Some documentation fixes
237 *(Released Tue, 27 Aug 2013)*
239 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
241 - Support 'viridian' parameter in Xen HVM (Issue 233)
242 - Include VCS version in ``gnt-cluster version``
243 - Support DSA SSH keys in bootstrap (Issue 338)
244 - Fix batch creation of instances
245 - Use FQDN to check master node status (Issue 551)
246 - Make the DRBD collector more failure-resilient
252 *(Released Fri, 2 Aug 2013)*
254 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
256 - Fix upgrading/downgrading from 2.7
257 - Increase maximum RAPI message size
258 - Documentation updates
259 - Split ``confd`` between ``luxid`` and ``confd``
260 - Merge 2.7 series up to the 2.7.1 release
261 - Allow the ``modify_etc_hosts`` option to be changed
262 - Add better debugging for ``luxid`` queries
263 - Expose bulk parameter for GetJobs in RAPI client
264 - Expose missing ``network`` fields in RAPI
265 - Add some ``cluster verify`` tests
266 - Some unittest fixes
267 - Fix a malfunction in ``hspace``'s tiered allocation
268 - Fix query compatibility between haskell and python implementations
269 - Add the ``vnet_hdr`` HV parameter for KVM
270 - Add ``--cleanup`` to instance failover
271 - Change the connected groups format in ``gnt-network info`` output; it
272 was previously displayed as a raw list by mistake. (Merged from 2.7)
278 *(Released Mon, 24 Jun 2013)*
280 This was the first beta release of the 2.8 series. All important changes
281 are listed in the latest 2.8 entry.
287 *(Released Thu, 26 Sep 2013)*
289 - Change the connected groups format in ``gnt-network info`` output; it
290 was previously displayed as a raw list by mistake
291 - Check disk template in right dict when copying
292 - Support multi-instance allocs without iallocator
293 - Fix some errors in the documentation
294 - Fix formatting of tuple in an error message
300 *(Released Thu, 25 Jul 2013)*
302 - Add logrotate functionality in daemon-util
303 - Add logrotate example file
304 - Add missing fields to network queries over rapi
305 - Fix network object timestamps
306 - Add support for querying network timestamps
307 - Fix a typo in the example crontab
308 - Fix a documentation typo
314 *(Released Thu, 04 Jul 2013)*
316 Incompatible/important changes
317 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
319 - Instance policies for disk size were documented to be on a per-disk
320 basis, but hail applied them to the sum of all disks. This has been
322 - ``hbal`` will now exit with status 0 if, during job execution over
323 LUXI, early exit has been requested and all jobs are successful;
324 before, exit status 1 was used, which cannot be differentiated from
326 - Compatibility with newer versions of rbd has been fixed
327 - ``gnt-instance batch-create`` has been changed to use the bulk create
328 opcode from Ganeti. This lead to incompatible changes in the format of
329 the JSON file. It's now not a custom dict anymore but a dict
330 compatible with the ``OpInstanceCreate`` opcode.
331 - Parent directories for file storage need to be listed in
332 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
333 write the file automatically based on old configuration values, but it
334 can not distribute it across all nodes and the file contents should be
335 verified. Use ``gnt-cluster copyfile
336 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
337 upgraded. The reason for requiring this list of paths now is that
338 before it would have been possible to inject new paths via RPC,
339 allowing files to be created in arbitrary locations. The RPC protocol
340 is protected using SSL/X.509 certificates, but as a design principle
341 Ganeti does not permit arbitrary paths to be passed.
342 - The parsing of the variants file for OSes (see
343 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
344 lines and comment lines (starting with ``#``) are ignored for better
346 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
347 longer available. ``gnt-node add`` now invokes a new tool on the
348 destination node, named ``prepare-node-join``, to configure the SSH
349 daemon. Paramiko is no longer necessary to configure nodes' SSH
350 daemons via ``gnt-node add``.
351 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
352 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
353 the flag is not reset after a master failover.
354 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
355 is recommended. 2.6 will be mandatory from the 2.8 series.
361 - New network management functionality to support automatic allocation
362 of IP addresses and managing of network parameters. See
363 :manpage:`gnt-network(8)` for more details.
364 - New external storage backend, to allow managing arbitrary storage
365 systems external to the cluster. See
366 :manpage:`ganeti-extstorage-interface(7)`.
367 - New ``exclusive-storage`` node parameter added, restricted to
368 nodegroup level. When it's set to true, physical disks are assigned in
369 an exclusive fashion to instances, as documented in :doc:`Partitioned
370 Ganeti <design-partitioned>`. Currently, only instances using the
371 ``plain`` disk template are supported.
372 - The KVM hypervisor has been updated with many new hypervisor
373 parameters, including a generic one for passing arbitrary command line
374 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
375 compatible up to qemu 1.4.
376 - A new tool, called ``mon-collector``, is the stand-alone executor of
377 the data collectors for a monitoring system. As of this version, it
378 just includes the DRBD data collector, that can be executed by calling
379 ``mon-collector`` using the ``drbd`` parameter. See
380 :manpage:`mon-collector(7)`.
381 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
382 for RAPI users. It allows granting permissions to query for
383 information to a specific user without giving
384 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
385 - A new tool named ``node-cleanup`` has been added. It cleans remains of
386 a cluster from a machine by stopping all daemons, removing
387 certificates and ssconf files. Unless the ``--no-backup`` option is
388 given, copies of the certificates are made.
389 - Instance creations now support the use of opportunistic locking,
390 potentially speeding up the (parallel) creation of multiple instances.
391 This feature is currently only available via the :doc:`RAPI
392 <rapi>` interface and when an instance allocator is used. If the
393 ``opportunistic_locking`` parameter is set the opcode will try to
394 acquire as many locks as possible, but will not wait for any locks
395 held by other opcodes. If not enough resources can be found to
396 allocate the instance, the temporary error code
397 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
398 retried thereafter, with or without opportunistic locking.
399 - New experimental linux-ha resource scripts.
400 - Restricted-commands support: ganeti can now be asked (via command line
401 or rapi) to perform commands on a node. These are passed via ganeti
402 RPC rather than ssh. This functionality is restricted to commands
403 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
404 reasons. The file is not copied automatically.
410 - Diskless instances are now externally mirrored (Issue 237). This for
411 now has only been tested in conjunction with explicit target nodes for
413 - Queries not needing locks or RPC access to the node can now be
414 performed by the confd daemon, making them independent from jobs, and
415 thus faster to execute. This is selectable at configure time.
416 - The functionality for allocating multiple instances at once has been
417 overhauled and is now also available through :doc:`RAPI <rapi>`.
419 There are no significant changes from version 2.7.0~rc3.
425 *(Released Tue, 25 Jun 2013)*
427 - Fix permissions on the confd query socket (Issue 477)
428 - Fix permissions on the job archive dir (Issue 498)
429 - Fix handling of an internal exception in replace-disks (Issue 472)
430 - Fix gnt-node info handling of shortened names (Issue 497)
431 - Fix gnt-instance grow-disk when wiping is enabled
432 - Documentation improvements, and support for newer pandoc
433 - Fix hspace honoring ipolicy for disks (Issue 484)
434 - Improve handling of the ``kvm_extra`` HV parameter
440 *(Released Fri, 24 May 2013)*
442 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
444 - Disks added through ``gnt-instance modify`` or created through
445 ``gnt-instance recreate-disks`` are wiped, if the
446 ``prealloc_wipe_disks`` flag is set.
447 - If wiping newly created disks fails, the disks are removed. Also,
448 partial failures in creating disks through ``gnt-instance modify``
449 triggers a cleanup of the partially-created disks.
450 - Removing the master IP address doesn't fail if the address has been
452 - Fix ownership of the OS log dir
453 - Workaround missing SO_PEERCRED constant (Issue 191)
459 *(Released Fri, 3 May 2013)*
461 This was the first release candidate of the 2.7 series. Since beta3:
463 - Fix kvm compatibility with qemu 1.4 (Issue 389)
464 - Documentation updates (admin guide, upgrade notes, install
465 instructions) (Issue 372)
466 - Fix gnt-group list nodes and instances count (Issue 436)
467 - Fix compilation without non-mandatory libraries (Issue 441)
468 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
469 - Make confd logging more verbose at INFO level (Issue 435)
470 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
471 - Fix failure path for instance storage type conversion (Issue 229)
472 - Update htools text backend documentation
473 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
474 - Disable inter-cluster instance move for file-based instances, because
475 it is dependant on instance export, which is not supported for
476 file-based instances. (Issue 414)
477 - Fix gnt-job crashes on non-ascii characters (Issue 427)
478 - Fix volume group checks on non-vm-capable nodes (Issue 432)
484 *(Released Mon, 22 Apr 2013)*
486 This was the third beta release of the 2.7 series. Since beta2:
488 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
489 - Fix data loss on wrong usage of ``gnt-instance move``
490 - Properly export errors in confd-based job queries
491 - Add ``users-setup`` tool
492 - Fix iallocator protocol to report 0 as a disk size for diskless
493 instances. This avoids hail breaking when a diskless instance is
495 - Fix job queue directory permission problem that made confd job queries
496 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
497 for access to archived jobs (Issue 406).
498 - Limit the sizes of networks supported by ``gnt-network`` to something
499 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
500 - Fix bugs in instance disk template conversion
501 - Fix GHC 7 compatibility
502 - Fix ``burnin`` install path (Issue 426).
503 - Allow very small disk grows (Issue 347).
504 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
505 that noded doesn't import masterd code (Issue 419).
506 - Make sure the default metavg at cluster init is the same as the vg, if
507 unspecified (Issue 358).
508 - Fix cleanup of partially created disks (part of Issue 416)
514 *(Released Tue, 2 Apr 2013)*
516 This was the second beta release of the 2.7 series. Since beta1:
518 - Networks no longer have a "type" slot, since this information was
519 unused in Ganeti: instead of it tags should be used.
520 - The rapi client now has a ``target_node`` option to MigrateInstance.
521 - Fix early exit return code for hbal (Issue 386).
522 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
523 - Fix ``rbd showmapped`` output parsing (Issue 312).
524 - Networks are now referenced indexed by UUID, rather than name. This
525 will require running cfgupgrade, from 2.7.0beta1, if networks are in
527 - The OS environment now includes network information.
528 - Deleting of a network is now disallowed if any instance nic is using
529 it, to prevent dangling references.
530 - External storage is now documented in man pages.
531 - The exclusive_storage flag can now only be set at nodegroup level.
532 - Hbal can now submit an explicit priority with its jobs.
533 - Many network related locking fixes.
534 - Bump up the required pylint version to 0.25.1.
535 - Fix the ``no_remember`` option in RAPI client.
536 - Many ipolicy related tests, qa, and fixes.
537 - Many documentation improvements and fixes.
538 - Fix building with ``--disable-file-storage``.
539 - Fix ``-q`` option in htools, which was broken if passed more than
541 - Some haskell/python interaction improvements and fixes.
542 - Fix iallocator in case of missing LVM storage.
543 - Fix confd config load in case of ``--no-lvm-storage``.
544 - The confd/query functionality is now mentioned in the security
551 *(Released Wed, 6 Feb 2013)*
553 This was the first beta release of the 2.7 series. All important changes
554 are listed in the latest 2.7 entry.
560 *(Released Fri, 21 Dec 2012)*
562 Important behaviour change: hbal won't rebalance anymore instances which
563 have the ``auto_balance`` attribute set to false. This was the intention
564 all along, but until now it only skipped those from the N+1 memory
565 reservation (DRBD-specific).
567 A significant number of bug fixes in this release:
569 - Fixed disk adoption interaction with ipolicy checks.
570 - Fixed networking issues when instances are started, stopped or
571 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
572 - Fixed the warning in cluster verify for shared storage instances not
574 - Fixed removal of storage directory on shared file storage (issue 262).
575 - Fixed validation of LVM volume group name in OpClusterSetParams
576 (``gnt-cluster modify``) (issue 285).
577 - Fixed runtime memory increases (``gnt-instance modify -m``).
578 - Fixed live migration under Xen's ``xl`` mode.
579 - Fixed ``gnt-instance console`` with ``xl``.
580 - Fixed building with newer Haskell compiler/libraries.
581 - Fixed PID file writing in Haskell daemons (confd); this prevents
582 restart issues if confd was launched manually (outside of
583 ``daemon-util``) while another copy of it was running
584 - Fixed a type error when doing live migrations with KVM (issue 297) and
585 the error messages for failing migrations have been improved.
586 - Fixed opcode validation for the out-of-band commands (``gnt-node
588 - Fixed a type error when unsetting OS hypervisor parameters (issue
589 311); now it's possible to unset all OS-specific hypervisor
591 - Fixed the ``dry-run`` mode for many operations: verification of
592 results was over-zealous but didn't take into account the ``dry-run``
593 operation, resulting in "wrong" failures.
594 - Fixed bash completion in ``gnt-job list`` when the job queue has
595 hundreds of entries; especially with older ``bash`` versions, this
596 results in significant CPU usage.
598 And lastly, a few other improvements have been made:
600 - Added option to force master-failover without voting (issue 282).
601 - Clarified error message on lock conflict (issue 287).
602 - Logging of newly submitted jobs has been improved (issue 290).
603 - Hostname checks have been made uniform between instance rename and
605 - The ``--submit`` option is now supported by ``gnt-debug delay``.
606 - Shutting down the master daemon by sending SIGTERM now stops it from
607 processing jobs waiting for locks; instead, those jobs will be started
608 once again after the master daemon is started the next time (issue
610 - Support for Xen's ``xl`` program has been improved (besides the fixes
612 - Reduced logging noise in the Haskell confd daemon (only show one log
613 entry for each config reload, instead of two).
614 - Several man page updates and typo fixes.
620 *(Released Fri, 12 Oct 2012)*
622 A small bugfix release. Among the bugs fixed:
624 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
625 made the command unusable.
626 - Commands that issue many jobs don't fail anymore just because some jobs
627 take so long that other jobs are archived.
628 - Failures during ``gnt-instance reinstall`` are reflected by the exit
630 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
632 - When ``always_failover`` is set, ``--allow-failover`` is not required
633 in migrate commands anymore.
634 - ``bash_completion`` works even if extglob is disabled.
635 - Fixed bug with locks that made failover for RDB-based instances fail.
636 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
637 a random node instead of one based on the allocator metric.
638 - Support for newer versions of pylint and pep8.
639 - Hail doesn't fail anymore when trying to add an instance of type
640 ``file``, ``sharedfile`` or ``rbd``.
641 - Added new Makefile target to rebuild the whole distribution, so that
642 all files are included.
648 *(Released Fri, 27 Jul 2012)*
651 .. attention:: The ``LUXI`` protocol has been made more consistent
652 regarding its handling of command arguments. This, however, leads to
653 incompatibility issues with previous versions. Please ensure that you
654 restart Ganeti daemons soon after the upgrade, otherwise most
655 ``LUXI`` calls (job submission, setting/resetting the drain flag,
656 pausing/resuming the watcher, cancelling and archiving jobs, querying
657 the cluster configuration) will fail.
666 The current ``admin_up`` field, which used to denote whether an instance
667 should be running or not, has been removed. Instead, ``admin_state`` is
668 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
670 The rational behind this is that an instance being “down” can have
673 - it could be down during a reboot
674 - it could be temporarily be down for a reinstall
675 - or it could be down because it is deprecated and kept just for its
678 The previous Boolean state was making it difficult to do capacity
679 calculations: should Ganeti reserve memory for a down instance? Now, the
680 tri-state field makes it clear:
682 - in ``up`` and ``down`` state, all resources are reserved for the
683 instance, and it can be at any time brought up if it is down
684 - in ``offline`` state, only disk space is reserved for it, but not
687 The field can have an extra use: since the transition between ``up`` and
688 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
689 transition between ``offline`` and ``down`` is done via ``gnt-instance
690 modify``, it is possible to given different rights to users. For
691 example, owners of an instance could be allowed to start/stop it, but
692 not transition it out of the offline state.
694 Instance policies and specs
695 +++++++++++++++++++++++++++
697 In previous Ganeti versions, an instance creation request was not
698 limited on the minimum size and on the maximum size just by the cluster
699 resources. As such, any policy could be implemented only in third-party
700 clients (RAPI clients, or shell wrappers over ``gnt-*``
701 tools). Furthermore, calculating cluster capacity via ``hspace`` again
702 required external input with regards to instance sizes.
704 In order to improve these workflows and to allow for example better
705 per-node group differentiation, we introduced instance specs, which
708 - minimum instance disk size, disk count, memory size, cpu count
709 - maximum values for the above metrics
710 - and “standard” values (used in ``hspace`` to calculate the standard
713 The minimum/maximum values can be also customised at node-group level,
714 for example allowing more powerful hardware to support bigger instance
717 Beside the instance specs, there are a few other settings belonging to
718 the instance policy framework. It is possible now to customise, per
719 cluster and node-group:
721 - the list of allowed disk templates
722 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
723 - the maximum ratio of instance to spindles (see below for more
724 information) for local storage
726 All these together should allow all tools that talk to Ganeti to know
727 what are the ranges of allowed values for instances and the
728 over-subscription that is allowed.
730 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
731 instance configuration, and the physical CPU configuration from the
732 node. For the spindle ratios however, we didn't track before these
733 values, so new parameters have been added:
735 - a new node parameter ``spindle_count``, defaults to 1, customisable at
736 node group or node level
737 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
739 Note that spindles in this context doesn't need to mean actual
740 mechanical hard-drives; it's just a relative number for both the node
741 I/O capacity and instance I/O consumption.
743 Instance migration behaviour
744 ++++++++++++++++++++++++++++
746 While live-migration is in general desirable over failover, it is
747 possible that for some workloads it is actually worse, due to the
748 variable time of the “suspend” phase during live migration.
750 To allow the tools to work consistently over such instances (without
751 having to hard-code instance names), a new backend parameter
752 ``always_failover`` has been added to control the migration/failover
753 behaviour. When set to True, all migration requests for an instance will
754 instead fall-back to failover.
756 Instance memory ballooning
757 ++++++++++++++++++++++++++
759 Initial support for memory ballooning has been added. The memory for an
760 instance is no longer fixed (backend parameter ``memory``), but instead
761 can vary between minimum and maximum values (backend parameters
762 ``minmem`` and ``maxmem``). Currently we only change an instance's
765 - live migrating or failing over and instance and the target node
766 doesn't have enough memory
767 - user requests changing the memory via ``gnt-instance modify
773 In order to control the use of specific CPUs by instance, support for
774 controlling CPU pinning has been added for the Xen, HVM and LXC
775 hypervisors. This is controlled by a new hypervisor parameter
776 ``cpu_mask``; details about possible values for this are in the
777 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
778 VCPU-to-CPU mapping) form will work well only when all nodes in your
779 cluster have the same amount of CPUs.
784 Another area in which Ganeti was not customisable were the parameters
785 used for storage configuration, e.g. how many stripes to use for LVM,
786 DRBD resync configuration, etc.
788 To improve this area, we've added disks parameters, which are
789 customisable at cluster and node group level, and which allow to
790 specify various parameters for disks (DRBD has the most parameters
791 currently), for example:
793 - DRBD resync algorithm and parameters (e.g. speed)
794 - the default VG for meta-data volumes for DRBD
795 - number of stripes for LVM (plain disk template)
798 These parameters can be modified via ``gnt-cluster modify -D …`` and
799 ``gnt-group modify -D …``, and are used at either instance creation (in
800 case of LVM stripes, for example) or at disk “activation” time
803 Rados block device support
804 ++++++++++++++++++++++++++
806 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
807 denoted by the ``rbd`` disk template type. This is considered
808 experimental, feedback is welcome. For details on configuring it, see
809 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
814 The existing master IP functionality works well only in simple setups (a
815 single network shared by all nodes); however, if nodes belong to
816 different networks, then the ``/32`` setup and lack of routing
817 information is not enough.
819 To allow the master IP to function well in more complex cases, the
820 system was reworked as follows:
822 - a master IP netmask setting has been added
823 - the master IP activation/turn-down code was moved from the node daemon
825 - whether to run the Ganeti-supplied master IP script or a user-supplied
826 on is a ``gnt-cluster init`` setting
828 Details about the location of the standard and custom setup scripts are
829 in the man page :manpage:`gnt-cluster(8)`; for information about the
830 setup script protocol, look at the Ganeti-supplied script.
835 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
838 It is now possible to use TLS-protected connections, and when renewing
839 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
840 it is now possible to specify spice or spice CA certificates. Also, it
841 is possible to configure a password for SPICE sessions via the
842 hypervisor parameter ``spice_password_file``.
844 There are also new parameters to control the compression and streaming
845 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
846 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
847 for the spice parameters.
849 Lastly, it is now possible to see the SPICE connection information via
850 ``gnt-instance console``.
855 A new tool (``tools/ovfconverter``) has been added that supports
856 conversion between Ganeti and the `Open Virtualization Format
857 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
860 This relies on the ``qemu-img`` tool to convert the disk formats, so the
861 actual compatibility with other virtualization solutions depends on it.
866 The configuration query daemon (``ganeti-confd``) is now optional, and
867 has been rewritten in Haskell; whether to use the daemon at all, use the
868 Python (default) or the Haskell version is selectable at configure time
869 via the ``--enable-confd`` parameter, which can take one of the
870 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
871 daemon will result in a smaller footprint; for larger systems, we
872 welcome feedback on the Haskell version which might become the default
875 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
876 daemon running. The Python version doesn't implement the new call.
879 User interface changes
880 ~~~~~~~~~~~~~~~~~~~~~~
882 We have replaced the ``--disks`` option of ``gnt-instance
883 replace-disks`` with a more flexible ``--disk`` option, which allows
884 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
885 disk size and mode can be changed upon recreation (via ``gnt-instance
886 recreate-disks``, which accepts the same ``--disk`` option).
888 As many people are used to a ``show`` command, we have added that as an
889 alias to ``info`` on all ``gnt-*`` commands.
891 The ``gnt-instance grow-disk`` command has a new mode in which it can
892 accept the target size of the disk, instead of the delta; this can be
893 more safe since two runs in absolute mode will be idempotent, and
894 sometimes it's also easier to specify the desired size directly.
896 Also the handling of instances with regard to offline secondaries has
897 been improved. Instance operations should not fail because one of it's
898 secondary nodes is offline, even though it's safe to proceed.
900 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
901 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
902 minors to instance name.
907 RAPI coverage has improved, with (for example) new resources for
908 recreate-disks, node power-cycle, etc.
913 There is partial support for ``xl`` in the Xen hypervisor; feedback is
916 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
917 whether to still support Python 2.4 or move to Python 2.6 as minimum
920 Support for Fedora has been slightly improved; the provided example
921 init.d script should work better on it and the INSTALL file should
922 document the needed dependencies.
927 The deprecated ``QueryLocks`` LUXI request has been removed. Use
928 ``Query(what=QR_LOCK, ...)`` instead.
930 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
931 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
932 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
933 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
934 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
936 RAPI client: ``CertificateError`` now derives from
937 ``GanetiApiError``. This should make it more easy to handle Ganeti
940 Deprecation warnings due to PyCrypto/paramiko import in
941 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
942 make sure to run an up-to-date paramiko version, if you use this tool.
944 The QA scripts now depend on Python 2.5 or above (the main code base
945 still works with Python 2.4).
947 The configuration file (``config.data``) is now written without
948 indentation for performance reasons; if you want to edit it, it can be
949 re-formatted via ``tools/fmtjson``.
951 A number of bugs has been fixed in the cluster merge tool.
953 ``x509`` certification verification (used in import-export) has been
954 changed to allow the same clock skew as permitted by the cluster
955 verification. This will remove some rare but hard to diagnose errors in
962 *(Released Thu, 19 Jul 2012)*
964 Very few changes from rc4 to the final release, only bugfixes:
966 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
967 instance, fix CDROM booting for KVM instances)
968 - fixed node group modification of node parameters
969 - fixed issue in LUClusterVerifyGroup with multi-group clusters
970 - fixed generation of bash completion to ensure a stable ordering
977 *(Released Fri, 13 Jul 2012)*
979 Third release candidate for 2.6. The following changes were done from
982 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
983 - Fixed an inconsistency in the LUXI protocol with the provided
984 arguments (NOT backwards compatible)
985 - Fixed a bug with node groups ipolicy where ``min`` was greater than
986 the cluster ``std`` value
987 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
988 easier instance debugging on nodes (requires ``hconfd`` to work)
994 *(Released Tue, 03 Jul 2012)*
996 Second release candidate for 2.6. The following changes were done from
999 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1001 - Fixed a RAPI regression on missing beparams/memory
1002 - Fixed redistribution of files on offline nodes
1003 - Added possibility to run activate-disks even though secondaries are
1004 offline. With this change it relaxes also the strictness on some other
1005 commands which use activate disks internally:
1006 * ``gnt-instance start|reboot|rename|backup|export``
1007 - Made it possible to remove safely an instance if its secondaries are
1009 - Made it possible to reinstall even though secondaries are offline
1015 *(Released Mon, 25 Jun 2012)*
1017 First release candidate for 2.6. The following changes were done from
1020 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1022 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1023 behaviour. It supports now the modify operation
1024 - ``hcheck`` implemented to analyze cluster health and possibility of
1025 improving health by rebalance
1026 - ``hbal`` has been improved in dealing with split instances
1032 *(Released Mon, 11 Jun 2012)*
1034 Second beta release of 2.6. The following changes were done from beta2
1037 - Fixed ``daemon-util`` with non-root user models
1038 - Fixed creation of plain instances with ``--no-wait-for-sync``
1039 - Fix wrong iv_names when running ``cfgupgrade``
1040 - Export more information in RAPI group queries
1041 - Fixed bug when changing instance network interfaces
1042 - Extended burnin to do NIC changes
1043 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1044 - Changed default for DRBD barriers
1045 - Fixed DRBD error reporting for syncer rate
1046 - Verify the options on disk parameters
1048 And of course various fixes to documentation and improved unittests and
1055 *(Released Wed, 23 May 2012)*
1057 First beta release of 2.6. The following changes were done from beta1 to
1060 - integrated patch for distributions without ``start-stop-daemon``
1061 - adapted example init.d script to work on Fedora
1062 - fixed log handling in Haskell daemons
1063 - adapted checks in the watcher for pycurl linked against libnss
1064 - add partial support for ``xl`` instead of ``xm`` for Xen
1065 - fixed a type issue in cluster verification
1066 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1069 Plus integrated fixes from the 2.5 branch:
1071 - fixed ``kvm-ifup`` to use ``/bin/bash``
1072 - fixed parallel build failures
1073 - KVM live migration when using a custom keymap
1079 *(Released Tue, 24 Jul 2012)*
1081 A small bugfix release, with no new features:
1083 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1084 different default shell (e.g. Debian, Ubuntu)
1085 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1086 243 and Debian bug #650664)
1087 - fixed compatibility with KVM versions that don't support multiple boot
1088 devices (fixes Issue 230 and Debian bug #624256)
1090 Additionally, a few fixes were done to the build system (fixed parallel
1091 build failures) and to the unittests (fixed race condition in test for
1092 FileID functions, and the default enable/disable mode for QA test is now
1099 *(Released Fri, 11 May 2012)*
1101 A small bugfix release.
1103 The main issues solved are on the topic of compatibility with newer LVM
1106 - fixed parsing of ``lv_attr`` field
1107 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1108 the ``--force`` flag is needed
1110 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1111 to accept kernel 3.x too (was hardcoded to 2.6.*).
1113 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1114 of hook results and that also made display of other errors suboptimal
1115 was fixed; the code behaves now like 2.4 and earlier.
1117 Another change in 2.5, the cleanup of the OS scripts environment, is too
1118 aggressive: it removed even the ``PATH`` variable, which requires the OS
1119 scripts to *always* need to export it. Since this is a bit too strict,
1120 we now export a minimal PATH, the same that we export for hooks.
1122 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1123 integrated into this release.
1125 Finally, a few other miscellaneous changes were done (no new features,
1126 just small improvements):
1128 - Fix ``gnt-group --help`` display
1129 - Fix hardcoded Xen kernel path
1130 - Fix grow-disk handling of invalid units
1131 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1132 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1138 *(Released Thu, 12 Apr 2012)*
1140 Incompatible/important changes and bugfixes
1141 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1143 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1144 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1145 to match the underlying LUXI interface.
1146 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1147 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1148 not maintain backwards-compatibility as the underlying operation
1149 changed in an incompatible way. The RAPI client can talk to old
1150 servers, but it needs to be told so as the return value changed.
1151 - When creating file-based instances via RAPI, the ``file_driver``
1152 parameter no longer defaults to ``loop`` and must be specified.
1153 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1155 - Support for the undocumented and deprecated RAPI instance creation
1156 request format version 0 has been dropped. Use version 1, supported
1157 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1158 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1159 documentation <install>`.
1160 - The "cluster-verify" hooks are now executed per group by the
1161 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1162 if you just run ``gnt-cluster verify``, which generates one opcode per
1164 - The environment as passed to the OS scripts is cleared, and thus no
1165 environment variables defined in the node daemon's environment will be
1166 inherited by the scripts.
1167 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1169 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1170 support operations involving multiple node groups.
1171 - Offline nodes are ignored when failing over an instance.
1172 - Support for KVM version 1.0, which changed the version reporting format
1174 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1176 - ``Makefile`` is now compatible with Automake 1.11.2
1177 - Includes all bugfixes made in the 2.4 series
1182 - The ganeti-htools project has been merged into the ganeti-core source
1183 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1184 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1185 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1186 GPT-style partition tables (requires `parted
1187 <http://www.gnu.org/s/parted/>`_).
1188 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1189 - Allowed adding tags on instance creation.
1190 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1191 - Allow instances to be started in a paused state, enabling the user to
1192 see the complete console output on boot using the console.
1193 - Added new hypervisor flag to control default reboot behaviour
1194 (``reboot_behavior``).
1195 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1196 - Improved out-of-band management support:
1198 - Added ``gnt-node health`` command reporting the health status of
1200 - Added ``gnt-node power`` command to manage power status of nodes.
1201 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1203 - Instance migration can fall back to failover if instance is not
1205 - Filters can be used when listing nodes, instances, groups and locks;
1206 see :manpage:`ganeti(7)` manpage.
1207 - Added post-execution status as variables to :doc:`hooks <hooks>`
1209 - Instance tags are exported/imported together with the instance.
1210 - When given an explicit job ID, ``gnt-job info`` will work for archived
1212 - Jobs can define dependencies on other jobs (not yet supported via
1213 RAPI or command line, but used by internal commands and usable via
1216 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1219 - Instance failover is now available as a RAPI resource
1220 (``/2/instances/[instance_name]/failover``).
1221 - ``gnt-instance info`` defaults to static information if primary node
1223 - Opcodes have a new ``comment`` attribute.
1224 - Added basic SPICE support to KVM hypervisor.
1225 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1227 Node group improvements
1228 ~~~~~~~~~~~~~~~~~~~~~~~
1230 - ``gnt-cluster verify`` has been modified to check groups separately,
1231 thereby improving performance.
1232 - Node group support has been added to ``gnt-cluster verify-disks``,
1233 which now operates per node group.
1234 - Watcher has been changed to work better with node groups.
1236 - One process and state file per node group.
1237 - Slow watcher in one group doesn't block other group's watcher.
1239 - Added new command, ``gnt-group evacuate``, to move all instances in a
1240 node group to other groups.
1241 - Added ``gnt-instance change-group`` to move an instance to another
1243 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1244 per-group operations.
1245 - Node groups can be tagged.
1246 - Some operations switch from an exclusive to a shared lock as soon as
1248 - Instance's primary and secondary nodes' groups are now available as
1249 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1250 and ``snodes.group.uuid``).
1255 - Numerous updates to documentation and manpages.
1257 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1259 - Some opcode/job results are now also documented, see :doc:`RAPI
1262 - A lockset's internal lock is now also visible in lock monitor.
1263 - Log messages from job queue workers now contain information about the
1264 opcode they're processing.
1265 - ``gnt-instance console`` no longer requires the instance lock.
1266 - A short delay when waiting for job changes reduces the number of LUXI
1267 requests significantly.
1268 - DRBD metadata volumes are overwritten with zeros during disk creation.
1269 - Out-of-band commands no longer acquire the cluster lock in exclusive
1271 - ``devel/upload`` now uses correct permissions for directories.
1277 *(Released Fri, 23 Mar 2012)*
1279 This was the sixth release candidate of the 2.5 series.
1285 *(Released Mon, 9 Jan 2012)*
1287 This was the fifth release candidate of the 2.5 series.
1293 *(Released Thu, 27 Oct 2011)*
1295 This was the fourth release candidate of the 2.5 series.
1301 *(Released Wed, 26 Oct 2011)*
1303 This was the third release candidate of the 2.5 series.
1309 *(Released Tue, 18 Oct 2011)*
1311 This was the second release candidate of the 2.5 series.
1317 *(Released Tue, 4 Oct 2011)*
1319 This was the first release candidate of the 2.5 series.
1325 *(Released Wed, 31 Aug 2011)*
1327 This was the third beta release of the 2.5 series.
1333 *(Released Mon, 22 Aug 2011)*
1335 This was the second beta release of the 2.5 series.
1341 *(Released Fri, 12 Aug 2011)*
1343 This was the first beta release of the 2.5 series.
1349 *(Released Thu, 27 Oct 2011)*
1351 - Fixed bug when parsing command line parameter values ending in
1353 - Fixed assertion error after unclean master shutdown
1354 - Disable HTTP client pool for RPC, significantly reducing memory usage
1356 - Fixed queue archive creation with wrong permissions
1362 *(Released Tue, 23 Aug 2011)*
1366 - Fixed documentation for importing with ``--src-dir`` option
1367 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1368 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1374 *(Released Fri, 5 Aug 2011)*
1376 Many bug-fixes and a few small features:
1378 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1379 issues when you tried to add an instance with two MAC addresses in one
1381 - KVM: fixed per-instance stored UID value
1382 - KVM: configure bridged NICs at migration start
1383 - KVM: Fix a bug where instance will not start with never KVM versions
1385 - Added OS search path to ``gnt-cluster info``
1386 - Fixed an issue with ``file_storage_dir`` where you were forced to
1387 provide an absolute path, but the documentation states it is a
1388 relative path, the documentation was right
1389 - Added a new parameter to instance stop/start called ``--no-remember``
1390 that will make the state change to not be remembered
1391 - Implemented ``no_remember`` at RAPI level
1392 - Improved the documentation
1393 - Node evacuation: don't call IAllocator if node is already empty
1394 - Fixed bug in DRBD8 replace disks on current nodes
1395 - Fixed bug in recreate-disks for DRBD instances
1396 - Moved assertion checking locks in ``gnt-instance replace-disks``
1397 causing it to abort with not owning the right locks for some situation
1398 - Job queue: Fixed potential race condition when cancelling queued jobs
1399 - Fixed off-by-one bug in job serial generation
1400 - ``gnt-node volumes``: Fix instance names
1401 - Fixed aliases in bash completion
1402 - Fixed a bug in reopening log files after being sent a SIGHUP
1403 - Added a flag to burnin to allow specifying VCPU count
1404 - Bugfixes to non-root Ganeti configuration
1410 *(Released Thu, 12 May 2011)*
1412 Many bug-fixes and a few new small features:
1414 - Fixed a bug related to log opening failures
1415 - Fixed a bug in instance listing with orphan instances
1416 - Fixed a bug which prevented resetting the cluster-level node parameter
1417 ``oob_program`` to the default
1418 - Many fixes related to the ``cluster-merge`` tool
1419 - Fixed a race condition in the lock monitor, which caused failures
1420 during (at least) creation of many instances in parallel
1421 - Improved output for gnt-job info
1422 - Removed the quiet flag on some ssh calls which prevented debugging
1424 - Improved the N+1 failure messages in cluster verify by actually
1425 showing the memory values (needed and available)
1426 - Increased lock attempt timeouts so that when executing long operations
1427 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1428 too early and thus prevent the use of the 'fair' mechanism
1429 - Changed instance query data (``gnt-instance info``) to not acquire
1430 locks unless needed, thus allowing its use on locked instance if only
1431 static information is asked for
1432 - Improved behaviour with filesystems that do not support rename on an
1434 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1435 kept locks on all nodes during the wipe, which is unneeded
1436 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1437 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1438 10GiB) which caused the wipe to fail right at the end in some cases
1439 - Fixed master IP activation when doing master failover with no-voting
1440 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1441 the master node itself
1442 - Fixed potential data-loss in under disk full conditions, where Ganeti
1443 wouldn't check correctly the return code and would consider
1444 partially-written files 'correct'
1445 - Fixed bug related to multiple VGs and DRBD disk replacing
1446 - Added new disk parameter ``metavg`` that allows placement of the meta
1447 device for DRBD in a different volume group
1448 - Fixed error handling in the node daemon when the system libc doesn't
1449 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1450 - Fixed lock release during replace-disks, which kept cluster-wide locks
1451 when doing disk replaces with an iallocator script
1452 - Added check for missing bridges in cluster verify
1453 - Handle EPIPE errors while writing to the terminal better, so that
1454 piping the output to e.g. ``less`` doesn't cause a backtrace
1455 - Fixed rare case where a ^C during Luxi calls could have been
1456 interpreted as server errors, instead of simply terminating
1457 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1459 - Added a few more parameters to the KVM hypervisor, allowing a second
1460 CDROM, custom disk type for CDROMs and a floppy image
1461 - Removed redundant message in instance rename when the name is given
1463 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1464 disks on new nodes, allowing recreation when the original instance
1465 nodes are completely gone
1466 - Added option when converting disk templates to DRBD to skip waiting
1467 for the resync, in order to make the instance available sooner
1468 - Added two new variables to the OS scripts environment (containing the
1470 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1471 to allow use of ``pvgrub`` as bootloader
1472 - Changed the instance memory modifications to only check out-of-memory
1473 conditions on memory increases, and turned the secondary node warnings
1474 into errors (they can still be overridden via ``--force``)
1475 - Fixed the handling of a corner case when the Python installation gets
1476 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1477 to execute a command that doesn't exist
1478 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1479 node of the instance returned failures during instance shutdown; this
1480 adds the option ``--ignore-consistency`` to gnt-instance move
1482 And as usual, various improvements to the error messages, documentation
1489 *(Released Wed, 09 Mar 2011)*
1491 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1492 the RAPI users file if run twice (even with ``--dry-run``).
1494 The release fixes that bug (nothing else changed).
1500 *(Released Mon, 07 Mar 2011)*
1502 Final 2.4.0 release. Just a few small fixes:
1504 - Fixed RAPI node evacuate
1505 - Fixed the kvm-ifup script
1506 - Fixed internal error handling for special job cases
1507 - Updated man page to specify the escaping feature for options
1513 *(Released Mon, 28 Feb 2011)*
1515 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1516 that this feature wiped the disks of the wrong instance, leading to loss
1521 - Fixed title of query field containing instance name
1522 - Expanded the glossary in the documentation
1523 - Fixed one unittest (internal issue)
1529 *(Released Mon, 21 Feb 2011)*
1531 A number of bug fixes plus just a couple functionality changes.
1533 On the user-visible side, the ``gnt-* list`` command output has changed
1534 with respect to "special" field states. The current rc1 style of display
1535 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1536 the default output mode special fields states are displayed as follows:
1538 - Offline resource: ``*``
1539 - Unavailable/not applicable: ``-``
1540 - Data missing (RPC failure): ``?``
1541 - Unknown field: ``??``
1543 Another user-visible change is the addition of ``--force-join`` to
1548 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1549 - Fixed regression in RAPI/instance reinstall where all parameters were
1550 required (instead of optional)
1551 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1552 - Fixed iallocator usage (offline nodes were not considered offline)
1553 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1554 - Fixed hypervisor and OS parameter validation with respect to
1555 non-vm_capable nodes
1556 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1558 - Fixed ``tools/listrunner`` with respect to agent-based usage
1564 *(Released Fri, 4 Feb 2011)*
1566 Many changes and fixes since the beta1 release. While there were some
1567 internal changes, the code has been mostly stabilised for the RC
1570 Note: the dumb allocator was removed in this release, as it was not kept
1571 up-to-date with the IAllocator protocol changes. It is recommended to
1572 use the ``hail`` command from the ganeti-htools package.
1574 Note: the 2.4 and up versions of Ganeti are not compatible with the
1575 0.2.x branch of ganeti-htools. You need to upgrade to
1576 ganeti-htools-0.3.0 (or later).
1578 Regressions fixed from 2.3
1579 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1581 - Fixed the ``gnt-cluster verify-disks`` command
1582 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1584 - Fixed disk adoption breakage
1585 - Fixed wrong headers in instance listing for field aliases
1590 - Fixed corner case in KVM handling of NICs
1591 - Fixed many cases of wrong handling of non-vm_capable nodes
1592 - Fixed a bug where a missing instance symlink was not possible to
1593 recreate with any ``gnt-*`` command (now ``gnt-instance
1594 activate-disks`` does it)
1595 - Fixed the volume group name as reported by ``gnt-cluster
1597 - Increased timeouts for the import-export code, hopefully leading to
1598 fewer aborts due network or instance timeouts
1599 - Fixed bug in ``gnt-node list-storage``
1600 - Fixed bug where not all daemons were started on cluster
1601 initialisation, but only at the first watcher run
1602 - Fixed many bugs in the OOB implementation
1603 - Fixed watcher behaviour in presence of instances with offline
1605 - Fixed instance list output for instances running on the wrong node
1606 - a few fixes to the cluster-merge tool, but it still cannot merge
1607 multi-node groups (currently it is not recommended to use this tool)
1613 - Improved network configuration for the KVM hypervisor
1614 - Added e1000 as a supported NIC for Xen-HVM
1615 - Improved the lvmstrap tool to also be able to use partitions, as
1616 opposed to full disks
1617 - Improved speed of disk wiping (the cluster parameter
1618 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1619 of instance creations
1620 - Added documentation for the OS parameters
1621 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1622 hypervisor is not responding
1623 - Added display of blacklisted and hidden OS information in
1624 ``gnt-cluster info``
1625 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1626 NIC and node parameters, which might create problems with currently
1627 invalid (but undetected) configuration files, but prevents validation
1628 failures when unrelated parameters are modified
1629 - Changed cluster initialisation to wait for the master daemon to become
1631 - Expanded the RAPI interface:
1633 - Added config redistribution resource
1634 - Added activation/deactivation of instance disks
1635 - Added export of console information
1637 - Implemented log file reopening on SIGHUP, which allows using
1638 logrotate(8) for the Ganeti log files
1639 - Added a basic OOB helper script as an example
1645 *(Released Fri, 14 Jan 2011)*
1650 - Fixed timezone issues when formatting timestamps
1651 - Added support for node groups, available via ``gnt-group`` and other
1653 - Added out-of-band framework and management, see :doc:`design
1654 document <design-oob>`
1655 - Removed support for roman numbers from ``gnt-node list`` and
1656 ``gnt-instance list``.
1657 - Allowed modification of master network interface via ``gnt-cluster
1658 modify --master-netdev``
1659 - Accept offline secondaries while shutting down instance disks
1660 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1661 - Added support for multiple LVM volume groups
1662 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1664 - Added commands to list available fields:
1666 - ``gnt-node list-fields``
1667 - ``gnt-group list-fields``
1668 - ``gnt-instance list-fields``
1670 - Updated documentation and man pages
1675 - Moved ``rapi_users`` file into separate directory, now named
1676 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1678 - Added new tool for running commands on many machines,
1679 ``tools/ganeti-listrunner``
1680 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1681 improving the ``gnt-instance console`` output
1682 - Allowed customisation of disk index separator at ``configure`` time
1683 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1684 - Added support for non-partitioned md disks in ``lvmstrap``
1685 - Added script to gracefully power off KVM instances
1686 - Split ``utils`` module into smaller parts
1687 - Changed query operations to return more detailed information, e.g.
1688 whether an information is unavailable due to an offline node. To use
1689 this new functionality, the LUXI call ``Query`` must be used. Field
1690 information is now stored by the master daemon and can be retrieved
1691 using ``QueryFields``. Instances, nodes and groups can also be queried
1692 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1693 exposed via RAPI). The following commands make use of this
1694 infrastructure change:
1696 - ``gnt-group list``
1697 - ``gnt-group list-fields``
1699 - ``gnt-node list-fields``
1700 - ``gnt-instance list``
1701 - ``gnt-instance list-fields``
1702 - ``gnt-debug locks``
1707 - New RAPI resources (see :doc:`rapi`):
1711 - ``/2/groups/[group_name]``
1712 - ``/2/groups/[group_name]/assign-nodes``
1713 - ``/2/groups/[group_name]/modify``
1714 - ``/2/groups/[group_name]/rename``
1715 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1719 - Implemented ``no_install`` for instance creation
1720 - Implemented OS parameters for instance reinstallation, allowing
1721 use of special settings on reinstallation (e.g. for preserving data)
1726 - Added IPv6 support in import/export
1727 - Pause DRBD synchronization while wiping disks on instance creation
1728 - Updated unittests and QA scripts
1729 - Improved network parameters passed to KVM
1730 - Converted man pages from docbook to reStructuredText
1736 *(Released Mon, 20 Dec 2010)*
1738 Released version 2.3.1~rc1 without any changes.
1744 *(Released Wed, 1 Dec 2010)*
1746 - impexpd: Disable OpenSSL compression in socat if possible (backport
1747 from master, commit e90739d625b, see :doc:`installation guide
1748 <install-quick>` for details)
1749 - Changed unittest coverage report to exclude test scripts
1750 - Added script to check version format
1756 *(Released Wed, 1 Dec 2010)*
1758 Released version 2.3.0~rc1 without any changes.
1764 *(Released Fri, 19 Nov 2010)*
1766 A number of bugfixes and documentation updates:
1768 - Update ganeti-os-interface documentation
1769 - Fixed a bug related to duplicate MACs or similar items which should be
1771 - Fix breakage in OS state modify
1772 - Reinstall instance: disallow offline secondaries (fixes bug related to
1773 OS changing but reinstall failing)
1774 - plus all the other fixes between 2.2.1 and 2.2.2
1780 *(Released Tue, 2 Nov 2010)*
1782 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1783 - Fixed master failover race with watcher
1784 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1786 - Accept previously stopped instance for export with instance removal
1787 - Simplify and extend the environment variables for instance OS scripts
1788 - Added new node flags, ``master_capable`` and ``vm_capable``
1789 - Added optional instance disk wiping prior during allocation. This is a
1790 cluster-wide option and can be set/modified using
1791 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1792 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1793 :doc:`install-quick`
1794 - Added a new watcher option (``--ignore-pause``)
1795 - Added option to ignore offline node on instance start/stop
1796 (``--ignore-offline``)
1797 - Allow overriding OS parameters with ``gnt-instance reinstall``
1798 - Added ability to change node's secondary IP address using ``gnt-node
1800 - Implemented privilege separation for all daemons except
1801 ``ganeti-noded``, see ``configure`` options
1802 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1804 - Implemented job priorities (see ``ganeti(7)`` manpage)
1805 - Ignore failures while shutting down instances during failover from
1807 - Exit daemon's bootstrap process only once daemon is ready
1808 - Export more information via ``LUInstanceQuery``/remote API
1809 - Improved documentation, QA and unittests
1810 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1811 restart if the file was created or changed
1812 - Added LUXI protocol version sent with each request and response,
1813 allowing detection of server/client mismatches
1814 - Moved the Python scripts among gnt-* and ganeti-* into modules
1815 - Moved all code related to setting up SSH to an external script,
1817 - Infrastructure changes for node group support in future versions
1823 *(Released Fri, 19 Nov 2010)*
1825 A few small bugs fixed, and some improvements to the build system:
1827 - Fix documentation regarding conversion to drbd
1828 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1830 - Fix error handling in node modify with multiple changes
1831 - Allow remote imports without checked names
1837 *(Released Tue, 19 Oct 2010)*
1839 - Disable SSL session ID cache in RPC client
1845 *(Released Thu, 14 Oct 2010)*
1847 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1848 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1850 - Fix problems with interaction between Curl and Python's HTTP server,
1851 resulting in increased speed in many RPC calls
1852 - Improve our release script to prevent breakage with older aclocal and
1859 *(Released Thu, 7 Oct 2010)*
1861 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1862 value retrieved from master
1863 - Added support for blacklisted or hidden OS definitions
1864 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1865 - Added support for -mem-path in KVM hypervisor abstraction layer
1866 - Allow overriding instance parameters in tool for inter-cluster
1867 instance moves (``tools/move-instance``)
1868 - Improved opcode summaries (e.g. in ``gnt-job list``)
1869 - Improve consistency of OS listing by sorting it
1870 - Documentation updates
1876 *(Released Fri, 8 Oct 2010)*
1878 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1884 *(Released Mon, 4 Oct 2010)*
1886 - Fixed regression in ``gnt-instance rename``
1892 *(Released Wed, 22 Sep 2010)*
1894 - Fixed OS_VARIANT variable for OS scripts
1895 - Fixed cluster tag operations via RAPI
1896 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1897 - Disabled RAPI CA checks in watcher
1903 *(Released Mon, 23 Aug 2010)*
1905 - Support DRBD versions of the format "a.b.c.d"
1907 - Re-introduce support for usage from multiple threads in RAPI client
1908 - Instance renames and modify via RAPI
1909 - Work around race condition between processing and archival in job
1911 - Mark opcodes following failed one as failed, too
1912 - Job field ``lock_status`` was removed due to difficulties making it
1913 work with the changed job queue in Ganeti 2.2; a better way to monitor
1914 locks is expected for a later 2.2.x release
1915 - Fixed dry-run behaviour with many commands
1916 - Support ``ssh-agent`` again when adding nodes
1917 - Many additional bugfixes
1923 *(Released Fri, 30 Jul 2010)*
1925 Important change: the internal RPC mechanism between Ganeti nodes has
1926 changed from using a home-grown http library (based on the Python base
1927 libraries) to use the PycURL library. This requires that PycURL is
1928 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1929 against GnuTLS by default. cURL's support for GnuTLS had known issues
1930 before cURL 7.21.0 and we recommend using the latest cURL release or
1931 linking against OpenSSL. Most other distributions already link PycURL
1932 and cURL against OpenSSL. The command::
1934 python -c 'import pycurl; print pycurl.version'
1936 can be used to determine the libraries PycURL and cURL are linked
1939 Other significant changes:
1941 - Rewrote much of the internals of the job queue, in order to achieve
1942 better parallelism; this decouples job query operations from the job
1943 processing, and it should allow much nicer behaviour of the master
1944 daemon under load, and it also has uncovered some long-standing bugs
1945 related to the job serialisation (now fixed)
1946 - Added a default iallocator setting to the cluster parameters,
1947 eliminating the need to always pass nodes or an iallocator for
1948 operations that require selection of new node(s)
1949 - Added experimental support for the LXC virtualization method
1950 - Added support for OS parameters, which allows the installation of
1951 instances to pass parameter to OS scripts in order to customise the
1953 - Added a hypervisor parameter controlling the migration type (live or
1954 non-live), since hypervisors have various levels of reliability; this
1955 has renamed the 'live' parameter to 'mode'
1956 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1957 logical volumes, meaning that cluster verify will ignore them and not
1958 flag their presence as errors
1959 - The watcher will now reset the error count for failed instances after
1960 8 hours, thus allowing self-healing if the problem that caused the
1961 instances to be down/fail to start has cleared in the meantime
1962 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1963 check for, and warn, if the drbd module parameter ``usermode_helper``
1964 is not consistent with the cluster-wide setting; this is needed to
1965 make diagnose easier of failed drbd creations
1966 - Started adding base IPv6 support, but this is not yet
1967 enabled/available for use
1968 - Rename operations (cluster, instance) will now return the new name,
1969 which is especially useful if a short name was passed in
1970 - Added support for instance migration in RAPI
1971 - Added a tool to pre-configure nodes for the SSH setup, before joining
1972 them to the cluster; this will allow in the future a simplified model
1973 for node joining (but not yet fully enabled in 2.2); this needs the
1974 paramiko python library
1975 - Fixed handling of name-resolving errors
1976 - Fixed consistency of job results on the error path
1977 - Fixed master-failover race condition when executed multiple times in
1979 - Fixed many bugs related to the job queue (mostly introduced during the
1980 2.2 development cycle, so not all are impacting 2.1)
1981 - Fixed instance migration with missing disk symlinks
1982 - Fixed handling of unknown jobs in ``gnt-job archive``
1983 - And many other small fixes/improvements
1987 - Enhanced both the unittest and the QA coverage
1988 - Switched the opcode validation to a generic model, and extended the
1989 validation to all opcode parameters
1990 - Changed more parts of the code that write shell scripts to use the
1992 - Switched the master daemon to use the asyncore library for the Luxi
1999 *(Released Thu, 17 Jun 2010)*
2001 - Added tool (``move-instance``) and infrastructure to move instances
2002 between separate clusters (see :doc:`separate documentation
2003 <move-instance>` and :doc:`design document <design-2.2>`)
2004 - Added per-request RPC timeout
2005 - RAPI now requires a Content-Type header for requests with a body (e.g.
2006 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2007 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2008 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2010 - Implemented initial support for running Ganeti daemons as separate
2011 users, see configure-time flags ``--with-user-prefix`` and
2012 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2014 - Instances can be removed after export (``gnt-backup export
2015 --remove-instance``)
2016 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2017 key (instead of 1024 bit)
2018 - Added new cluster configuration file for cluster domain secret
2019 - Import/export now use SSL instead of SSH
2020 - Added support for showing estimated time when exporting an instance,
2021 see the ``ganeti-os-interface(7)`` manpage and look for
2028 *(Released Tue, 16 Nov 2010)*
2030 Some more bugfixes. Unless critical bugs occur, this will be the last
2033 - Fix case of MAC special-values
2034 - Fix mac checker regex
2035 - backend: Fix typo causing "out of range" error
2036 - Add missing --units in gnt-instance list man page
2042 *(Released Tue, 24 Aug 2010)*
2045 - Don't ignore secondary node silently on non-mirrored disk templates
2047 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2048 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2049 - Properly document the usb_mouse parameter
2050 - Fix path in ganeti-rapi(8) (issue 116)
2051 - Adjust error message when the ganeti user's .ssh directory is
2053 - Add same-node-check when changing the disk template to drbd
2059 *(Released Fri, 16 Jul 2010)*
2062 - Add an option to only select some reboot types during qa/burnin.
2063 (on some hypervisors consequent reboots are not supported)
2064 - Fix infrequent race condition in master failover. Sometimes the old
2065 master ip address would be still detected as up for a short time
2066 after it was removed, causing failover to fail.
2067 - Decrease mlockall warnings when the ctypes module is missing. On
2068 Python 2.4 we support running even if no ctypes module is installed,
2069 but we were too verbose about this issue.
2070 - Fix building on old distributions, on which man doesn't have a
2072 - Fix RAPI not to ignore the MAC address on instance creation
2073 - Implement the old instance creation format in the RAPI client.
2079 *(Released Thu, 01 Jul 2010)*
2081 A small bugfix release:
2082 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2083 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2084 a non-existing option
2085 - Fix instance create: the --force-variant option was ignored
2086 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2087 - Fix modify node storage with non-FQDN arguments
2088 - Fix RAPI client to authenticate under Python 2.6 when used
2089 for more than 5 requests needing authentication
2090 - Fix gnt-instance modify -t (storage) giving a wrong error message
2091 when converting a non-shutdown drbd instance to plain
2097 *(Released Fri, 18 Jun 2010)*
2099 A small bugfix release:
2101 - Fix live migration of KVM instances started with older Ganeti
2102 versions which had fewer hypervisor parameters
2103 - Fix gnt-instance grow-disk on down instances
2104 - Fix an error-reporting bug during instance migration
2105 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2106 silently ignoring broken ones
2107 - Fix an RPC error reporting bug affecting, for example, RAPI client
2109 - Fix bug triggered by different API version os-es on different nodes
2110 - Fix a bug in instance startup with custom hvparams: OS level
2111 parameters would fail to be applied.
2112 - Fix the RAPI client under Python 2.6 (but more work is needed to
2113 make it work completely well with OpenSSL)
2114 - Fix handling of errors when resolving names from DNS
2120 *(Released Thu, 3 Jun 2010)*
2122 A medium sized development cycle. Some new features, and some
2123 fixes/small improvements/cleanups.
2125 Significant features
2126 ~~~~~~~~~~~~~~~~~~~~
2128 The node deamon now tries to mlock itself into memory, unless the
2129 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2130 its logs, and falls back to console logging. This allows emergency
2131 features such as ``gnt-node powercycle`` to work even in the event of a
2132 broken node disk (tested offlining the disk hosting the node's
2133 filesystem and dropping its memory caches; don't try this at home)
2135 KVM: add vhost-net acceleration support. It can be tested with a new
2136 enough version of the kernel and of qemu-kvm.
2138 KVM: Add instance chrooting feature. If you use privilege dropping for
2139 your VMs you can also now force them to chroot to an empty directory,
2140 before starting the emulated guest.
2142 KVM: Add maximum migration bandwith and maximum downtime tweaking
2143 support (requires a new-enough version of qemu-kvm).
2145 Cluster verify will now warn if the master node doesn't have the master
2146 ip configured on it.
2148 Add a new (incompatible) instance creation request format to RAPI which
2149 supports all parameters (previously only a subset was supported, and it
2150 wasn't possible to extend the old format to accomodate all the new
2151 features. The old format is still supported, and a client can check for
2152 this feature, before using it, by checking for its presence in the
2153 ``features`` RAPI resource.
2155 Now with ancient latin support. Try it passing the ``--roman`` option to
2156 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2157 (requires the python-roman module to be installed, in order to work).
2162 As usual many internal code refactorings, documentation updates, and
2165 - Lots of improvements and cleanups to the experimental Remote API
2166 (RAPI) client library.
2167 - A new unit test suite for the core daemon libraries.
2168 - A fix to creating missing directories makes sure the umask is not
2169 applied anymore. This enforces the same directory permissions
2171 - Better handling terminating daemons with ctrl+c (used when running
2172 them in debugging mode).
2173 - Fix a race condition in live migrating a KVM instance, when stat()
2174 on the old proc status file returned EINVAL, which is an unexpected
2176 - Fixed manpage checking with newer man and utf-8 charachters. But now
2177 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2183 *(Released Fri, 7 May 2010)*
2185 Fix a bug which prevented untagged KVM instances from starting.
2191 *(Released Fri, 7 May 2010)*
2193 Another release with a long development cycle, during which many
2194 different features were added.
2196 Significant features
2197 ~~~~~~~~~~~~~~~~~~~~
2199 The KVM hypervisor now can run the individual instances as non-root, to
2200 reduce the impact of a VM being hijacked due to bugs in the
2201 hypervisor. It is possible to run all instances as a single (non-root)
2202 user, to manually specify a user for each instance, or to dynamically
2203 allocate a user out of a cluster-wide pool to each instance, with the
2204 guarantee that no two instances will run under the same user ID on any
2207 An experimental RAPI client library, that can be used standalone
2208 (without the other Ganeti libraries), is provided in the source tree as
2209 ``lib/rapi/client.py``. Note this client might change its interface in
2210 the future, as we iterate on its capabilities.
2212 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2213 replace the cluster's certificates and crypto keys. This might help in
2214 case they have been compromised, or have simply expired.
2216 A new disk option for instance creation has been added that allows one
2217 to "adopt" currently existing logical volumes, with data
2218 preservation. This should allow easier migration to Ganeti from
2219 unmanaged (or managed via other software) instances.
2221 Another disk improvement is the possibility to convert between redundant
2222 (DRBD) and plain (LVM) disk configuration for an instance. This should
2223 allow better scalability (starting with one node and growing the
2224 cluster, or shrinking a two-node cluster to one node).
2226 A new feature that could help with automated node failovers has been
2227 implemented: if a node sees itself as offline (by querying the master
2228 candidates), it will try to shutdown (hard) all instances and any active
2229 DRBD devices. This reduces the risk of duplicate instances if an
2230 external script automatically failovers the instances on such nodes. To
2231 enable this, the cluster parameter ``maintain_node_health`` should be
2232 enabled; in the future this option (per the name) will enable other
2233 automatic maintenance features.
2235 Instance export/import now will reuse the original instance
2236 specifications for all parameters; that means exporting an instance,
2237 deleting it and the importing it back should give an almost identical
2238 instance. Note that the default import behaviour has changed from
2239 before, where it created only one NIC; now it recreates the original
2242 Cluster verify has added a few new checks: SSL certificates validity,
2243 /etc/hosts consistency across the cluster, etc.
2248 As usual, many internal changes were done, documentation fixes,
2251 - Fixed cluster initialization with disabled cluster storage (regression
2252 introduced in 2.1.1)
2253 - File-based storage supports growing the disks
2254 - Fixed behaviour of node role changes
2255 - Fixed cluster verify for some corner cases, plus a general rewrite of
2256 cluster verify to allow future extension with more checks
2257 - Fixed log spamming by watcher and node daemon (regression introduced
2259 - Fixed possible validation issues when changing the list of enabled
2261 - Fixed cleanup of /etc/hosts during node removal
2262 - Fixed RAPI response for invalid methods
2263 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2264 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2265 from ide disks, etc.)
2266 - Allow OS changes without re-installation (to record a changed OS
2267 outside of Ganeti, or to allow OS renames)
2268 - Allow instance creation without OS installation (useful for example if
2269 the OS will be installed manually, or restored from a backup not in
2271 - Implemented option to make cluster ``copyfile`` use the replication
2273 - Added list of enabled hypervisors to ssconf (possibly useful for
2275 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2277 - A partial form of node re-IP is possible via node readd, which now
2278 allows changed node primary IP
2279 - Command line utilities now show an informational message if the job is
2281 - The logs of the master daemon now show the PID/UID/GID of the
2288 *(Released Fri, 12 Mar 2010)*
2290 During the 2.1.0 long release candidate cycle, a lot of improvements and
2291 changes have accumulated with were released later as 2.1.1.
2296 The node evacuate command (``gnt-node evacuate``) was significantly
2297 rewritten, and as such the IAllocator protocol was changed - a new
2298 request type has been added. This unfortunate change during a stable
2299 series is designed to improve performance of node evacuations; on
2300 clusters with more than about five nodes and which are well-balanced,
2301 evacuation should proceed in parallel for all instances of the node
2302 being evacuated. As such, any existing IAllocator scripts need to be
2303 updated, otherwise the above command will fail due to the unknown
2304 request. The provided "dumb" allocator has not been updated; but the
2305 ganeti-htools package supports the new protocol since version 0.2.4.
2307 Another important change is increased validation of node and instance
2308 names. This might create problems in special cases, if invalid host
2309 names are being used.
2311 Also, a new layer of hypervisor parameters has been added, that sits at
2312 OS level between the cluster defaults and the instance ones. This allows
2313 customisation of virtualization parameters depending on the installed
2314 OS. For example instances with OS 'X' may have a different KVM kernel
2315 (or any other parameter) than the cluster defaults. This is intended to
2316 help managing a multiple OSes on the same cluster, without manual
2317 modification of each instance's parameters.
2319 A tool for merging clusters, ``cluster-merge``, has been added in the
2320 tools sub-directory.
2325 - Improved the int/float conversions that should make the code more
2326 robust in face of errors from the node daemons
2327 - Fixed the remove node code in case of internal configuration errors
2328 - Fixed the node daemon behaviour in face of inconsistent queue
2329 directory (e.g. read-only file-system where we can't open the files
2331 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2332 now it either aborts cleanly or, if given the new "auto_promote"
2333 parameter, will automatically promote other nodes as needed
2334 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2335 completely prevent Ganeti from working
2336 - Fixed bug for instance export when not all disks were successfully
2338 - Fixed behaviour of node add when the new node is slow in starting up
2340 - Fixed handling of signals in the LUXI client, which should improve
2341 behaviour of command-line scripts
2342 - Added checks for invalid node/instance names in the configuration (now
2343 flagged during cluster verify)
2344 - Fixed watcher behaviour for disk activation errors
2345 - Fixed two potentially endless loops in http library, which led to the
2346 RAPI daemon hanging and consuming 100% CPU in some cases
2347 - Fixed bug in RAPI daemon related to hashed passwords
2348 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2350 - Enhanced compatibility with non-Debian OSes, but not using absolute
2351 path in some commands and allowing customisation of the ssh
2352 configuration directory
2353 - Fixed possible future issue with new Python versions by abiding to the
2354 proper use of ``__slots__`` attribute on classes
2355 - Added checks that should prevent directory traversal attacks
2356 - Many documentation fixes based on feedback from users
2361 - Added an "early_release" more for instance replace disks and node
2362 evacuate, where we release locks earlier and thus allow higher
2363 parallelism within the cluster
2364 - Added watcher hooks, intended to allow the watcher to restart other
2365 daemons (e.g. from the ganeti-nbma project), but they can be used of
2366 course for any other purpose
2367 - Added a compile-time disable for DRBD barriers, to increase
2368 performance if the administrator trusts the power supply or the
2369 storage system to not lose writes
2370 - Added the option of using syslog for logging instead of, or in
2371 addition to, Ganeti's own log files
2372 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2373 can indeed boot from a paravirtual NIC
2374 - Added a generic debug level for many operations; while this is not
2375 used widely yet, it allows one to pass the debug value all the way to
2377 - Enhanced the hooks environment for instance moves (failovers,
2378 migrations) where the primary/secondary nodes changed during the
2379 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2380 - Enhanced data validations for many user-supplied values; one important
2381 item is the restrictions imposed on instance and node names, which
2382 might reject some (invalid) host names
2383 - Add a configure-time option to disable file-based storage, if it's not
2384 needed; this allows greater security separation between the master
2385 node and the other nodes from the point of view of the inter-node RPC
2387 - Added user notification in interactive tools if job is waiting in the
2388 job queue or trying to acquire locks
2389 - Added log messages when a job is waiting for locks
2390 - Added filtering by node tags in instance operations which admit
2391 multiple instances (start, stop, reboot, reinstall)
2392 - Added a new tool for cluster mergers, ``cluster-merge``
2393 - Parameters from command line which are of the form ``a=b,c=d`` can now
2394 use backslash escapes to pass in values which contain commas,
2395 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2397 - For KVM, the instance name is the first parameter passed to KVM, so
2398 that it's more visible in the process list
2404 *(Released Tue, 2 Mar 2010)*
2406 Ganeti 2.1 brings many improvements with it. Major changes:
2408 - Added infrastructure to ease automated disk repairs
2409 - Added new daemon to export configuration data in a cheaper way than
2410 using the remote API
2411 - Instance NICs can now be routed instead of being associated with a
2413 - Improved job locking logic to reduce impact of jobs acquiring multiple
2414 locks waiting for other long-running jobs
2416 In-depth implementation details can be found in the Ganeti 2.1 design
2422 - Added chroot hypervisor
2423 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2425 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2426 ``bootloader_path`` and ``bootloader_args``)
2427 - Added the ``use_localtime`` option for the xen-hvm and kvm
2428 hypervisors, and the default value for this has changed to false (in
2429 2.0 xen-hvm always enabled it)
2430 - Added luxi call to submit multiple jobs in one go
2431 - Added cluster initialization option to not modify ``/etc/hosts``
2433 - Added network interface parameters
2434 - Added dry run mode to some LUs
2435 - Added RAPI resources:
2437 - ``/2/instances/[instance_name]/info``
2438 - ``/2/instances/[instance_name]/replace-disks``
2439 - ``/2/nodes/[node_name]/evacuate``
2440 - ``/2/nodes/[node_name]/migrate``
2441 - ``/2/nodes/[node_name]/role``
2442 - ``/2/nodes/[node_name]/storage``
2443 - ``/2/nodes/[node_name]/storage/modify``
2444 - ``/2/nodes/[node_name]/storage/repair``
2446 - Added OpCodes to evacuate or migrate all instances on a node
2447 - Added new command to list storage elements on nodes (``gnt-node
2448 list-storage``) and modify them (``gnt-node modify-storage``)
2449 - Added new ssconf files with master candidate IP address
2450 (``ssconf_master_candidates_ips``), node primary IP address
2451 (``ssconf_node_primary_ips``) and node secondary IP address
2452 (``ssconf_node_secondary_ips``)
2453 - Added ``ganeti-confd`` and a client library to query the Ganeti
2454 configuration via UDP
2455 - Added ability to run hooks after cluster initialization and before
2457 - Added automatic mode for disk replace (``gnt-instance replace-disks
2459 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2460 after catastrophic data-loss
2461 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2463 - Added ``gnt-instance move`` command to move instances
2464 - Added ``gnt-cluster watcher`` command to control watcher
2465 - Added ``gnt-node powercycle`` command to powercycle nodes
2466 - Added new job status field ``lock_status``
2467 - Added parseable error codes to cluster verification (``gnt-cluster
2468 verify --error-codes``) and made output less verbose (use
2469 ``--verbose`` to restore previous behaviour)
2470 - Added UUIDs to the main config entities (cluster, nodes, instances)
2471 - Added support for OS variants
2472 - Added support for hashed passwords in the Ganeti remote API users file
2474 - Added option to specify maximum timeout on instance shutdown
2475 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2476 - Added new helper script to start and stop Ganeti daemons
2477 (``daemon-util``), with the intent to reduce the work necessary to
2478 adjust Ganeti for non-Debian distributions and to start/stop daemons
2480 - Added more unittests
2481 - Fixed critical bug in ganeti-masterd startup
2482 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2483 now customisable at the cluster level for both the KVM and Xen
2484 hypervisors using the new ``migration_port`` parameter
2485 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2486 reinstalling an instance
2487 - Allowed ``@`` in tag names
2488 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2489 - Many documentation updates
2490 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2491 - ``gnt-instance reinstall`` can now reinstall multiple instances
2492 - Updated many command line parameters
2493 - Introduced new OS API version 15
2494 - No longer support a default hypervisor
2495 - Treat virtual LVs as inexistent
2496 - Improved job locking logic to reduce lock contention
2497 - Match instance and node names case insensitively
2498 - Reimplemented bash completion script to be more complete
2505 *(Released Thu, 4 Feb 2010)*
2507 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2508 - Fix a string formatting bug
2509 - Improve safety of the code in some error paths
2510 - Improve data validation in the master of values returned from nodes
2516 *(Released Thu, 17 Dec 2009)*
2518 - Fix security issue due to missing validation of iallocator names; this
2519 allows local and remote execution of arbitrary executables
2520 - Fix failure of gnt-node list during instance removal
2521 - Ship the RAPI documentation in the archive
2527 *(Released Wed, 30 Sep 2009)*
2529 - Fixed many wrong messages
2530 - Fixed a few bugs related to the locking library
2531 - Fixed MAC checking at instance creation time
2532 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2533 - Fixed a few issues related to signal handling in both daemons and
2535 - Fixed the example startup script provided
2536 - Fixed insserv dependencies in the example startup script (patch from
2538 - Fixed handling of drained nodes in the iallocator framework
2539 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2541 - Fixed error related to invalid job IDs in job polling
2542 - Fixed job/opcode persistence on unclean master shutdown
2543 - Fixed handling of partial job processing after unclean master
2545 - Fixed error reporting from LUs, previously all errors were converted
2546 into execution errors
2547 - Fixed error reporting from burnin
2548 - Decreased significantly the memory usage of the job queue
2549 - Optimised slightly multi-job submission
2550 - Optimised slightly opcode loading
2551 - Backported the multi-job submit framework from the development
2552 branch; multi-instance start and stop should be faster
2553 - Added script to clean archived jobs after 21 days; this will reduce
2554 the size of the queue directory
2555 - Added some extra checks in disk size tracking
2556 - Added an example ethers hook script
2557 - Added a cluster parameter that prevents Ganeti from modifying of
2559 - Added more node information to RAPI responses
2560 - Added a ``gnt-job watch`` command that allows following the ouput of a
2562 - Added a bind-address option to ganeti-rapi
2563 - Added more checks to the configuration verify
2564 - Enhanced the burnin script such that some operations can be retried
2566 - Converted instance reinstall to multi-instance model
2572 *(Released Fri, 7 Aug 2009)*
2574 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2575 to allow using the pre-2.0.2 behaviour in activation, if any existing
2576 instances have mismatched disk sizes in the configuration
2577 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2578 any configuration mismatches for disk sizes
2579 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2580 failover to work on two-node clusters
2581 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2583 - Fixed detection of OS script errors in ``gnt-backup export``
2584 - Fixed exit code of ``gnt-backup export``
2590 *(Released Fri, 17 Jul 2009)*
2592 - Added experimental support for stripped logical volumes; this should
2593 enhance performance but comes with a higher complexity in the block
2594 device handling; stripping is only enabled when passing
2595 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2596 affected even in the non-stripped mode
2597 - Improved resiliency against transient failures at the end of DRBD
2598 resyncs, and in general of DRBD resync checks
2599 - Fixed a couple of issues with exports and snapshot errors
2600 - Fixed a couple of issues in instance listing
2601 - Added display of the disk size in ``gnt-instance info``
2602 - Fixed checking for valid OSes in instance creation
2603 - Fixed handling of the "vcpus" parameter in instance listing and in
2604 general of invalid parameters
2605 - Fixed http server library, and thus RAPI, to handle invalid
2606 username/password combinations correctly; this means that now they
2607 report unauthorized for queries too, not only for modifications,
2608 allowing earlier detect of configuration problems
2609 - Added a new "role" node list field, equivalent to the master/master
2610 candidate/drained/offline flags combinations
2611 - Fixed cluster modify and changes of candidate pool size
2612 - Fixed cluster verify error messages for wrong files on regular nodes
2613 - Fixed a couple of issues with node demotion from master candidate role
2614 - Fixed node readd issues
2615 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2616 - Added a new ``--no-voting`` option for masterfailover to fix failover
2617 on two-nodes clusters when the former master node is unreachable
2618 - Added instance reinstall over RAPI
2624 *(Released Tue, 16 Jun 2009)*
2626 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2627 allow re-adding the start in single-user option (regression from 1.2)
2628 - the watcher writes the instance status to a file, to allow monitoring
2629 to report the instance status (from the master) based on cached
2630 results of the watcher's queries; while this can get stale if the
2631 watcher is being locked due to other work on the cluster, this is
2632 still an improvement
2633 - the watcher now also restarts the node daemon and the rapi daemon if
2635 - fixed the watcher to handle full and drained queue cases
2636 - hooks export more instance data in the environment, which helps if
2637 hook scripts need to take action based on the instance's properties
2638 (no longer need to query back into ganeti)
2639 - instance failovers when the instance is stopped do not check for free
2640 RAM, so that failing over a stopped instance is possible in low memory
2642 - rapi uses queries for tags instead of jobs (for less job traffic), and
2643 for cluster tags it won't talk to masterd at all but read them from
2645 - a couple of error handling fixes in RAPI
2646 - drbd handling: improved the error handling of inconsistent disks after
2647 resync to reduce the frequency of "there are some degraded disks for
2648 this instance" messages
2649 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2650 error handling path called a wrong function name)
2656 *(Released Wed, 27 May 2009)*
2658 - no changes from rc5
2664 *(Released Wed, 20 May 2009)*
2666 - fix a couple of bugs (validation, argument checks)
2667 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2668 - some small improvements to RAPI and IAllocator
2669 - make watcher automatically start the master daemon if down
2675 *(Released Mon, 27 Apr 2009)*
2677 - change the OS list to not require locks; this helps with big clusters
2678 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2679 volume group is broken
2680 - ``gnt-instance info``, without any arguments, doesn't run for all
2681 instances anymore; either pass ``--all`` or pass the desired
2682 instances; this helps against mistakes on big clusters where listing
2683 the information for all instances takes a long time
2684 - miscellaneous doc and man pages fixes
2690 *(Released Wed, 8 Apr 2009)*
2692 - Change the internal locking model of some ``gnt-node`` commands, in
2693 order to reduce contention (and blocking of master daemon) when
2694 batching many creation/reinstall jobs
2695 - Fixes to Xen soft reboot
2696 - No longer build documentation at build time, instead distribute it in
2697 the archive, in order to reduce the need for the whole docbook/rst
2704 *(Released Fri, 27 Mar 2009)*
2706 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2707 - Fix watcher startup sequence, improves the behaviour of busy clusters
2708 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2709 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2710 bind address and other places
2711 - Some documentation fixes and updates
2717 *(Released Mon, 2 Mar 2009)*
2719 - More documentation updates, now all docs should be more-or-less
2721 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2723 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2725 - Fix an issue related to $libdir/run/ganeti and cluster creation
2731 *(Released Thu, 19 Feb 2009)*
2733 - Xen PVM and KVM have switched the default value for the instance root
2734 disk to the first partition on the first drive, instead of the whole
2735 drive; this means that the OS installation scripts must be changed
2737 - Man pages have been updated
2738 - RAPI has been switched by default to HTTPS, and the exported functions
2739 should all work correctly
2740 - RAPI v1 has been removed
2741 - Many improvements to the KVM hypervisor
2742 - Block device errors are now better reported
2743 - Many other bugfixes and small improvements
2749 *(Released Mon, 26 Jan 2009)*
2751 - Version 2 is a general rewrite of the code and therefore the
2752 differences are too many to list, see the design document for 2.0 in
2753 the ``doc/`` subdirectory for more details
2754 - In this beta version there is not yet a migration path from 1.2 (there
2755 will be one in the final 2.0 release)
2756 - A few significant changes are:
2758 - all commands are executed by a daemon (``ganeti-masterd``) and the
2759 various ``gnt-*`` commands are just front-ends to it
2760 - all the commands are entered into, and executed from a job queue,
2761 see the ``gnt-job(8)`` manpage
2762 - the RAPI daemon supports read-write operations, secured by basic
2763 HTTP authentication on top of HTTPS
2764 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2765 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2766 to migrate to DRBD 8 first while still running Ganeti 1.2)
2767 - DRBD devices are using statically allocated minor numbers, which
2768 will be assigned to existing instances during the migration process
2769 - there is support for both Xen PVM and Xen HVM instances running on
2771 - KVM virtualization is supported too
2772 - file-based storage has been implemented, which means that it is
2773 possible to run the cluster without LVM and DRBD storage, for
2774 example using a shared filesystem exported from shared storage (and
2775 still have live migration)
2781 *(Released Tue, 13 Jan 2009)*
2783 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2784 - Reuse the old instance mac address by default on instance import, if
2785 the instance name is the same.
2786 - Handle situations in which the node info rpc returns incomplete
2788 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2789 - Improved version of batcher:
2791 - state file support
2792 - instance mac address support
2793 - support for HVM clusters/instances
2795 - Add an option to show the number of cpu sockets and nodes in
2797 - Support OSes that handle more than one version of the OS api (but do
2798 not change the current API in any other way)
2799 - Fix ``gnt-node migrate``
2800 - ``gnt-debug`` man page
2801 - Fixes various more typos and small issues
2802 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2808 *(Released Wed, 24 Sep 2008)*
2810 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2811 type of disk exported to fully virtualized instances.
2812 - provide access to the serial console of HVM instances
2813 - instance auto_balance flag, set by default. If turned off it will
2814 avoid warnings on cluster verify if there is not enough memory to fail
2815 over an instance. in the future it will prevent automatically failing
2816 it over when we will support that.
2817 - batcher tool for instance creation, see ``tools/README.batcher``
2818 - ``gnt-instance reinstall --select-os`` to interactively select a new
2819 operating system when reinstalling an instance.
2820 - when changing the memory amount on instance modify a check has been
2821 added that the instance will be able to start. also warnings are
2822 emitted if the instance will not be able to fail over, if auto_balance
2824 - documentation fixes
2825 - sync fields between ``gnt-instance list/modify/add/import``
2826 - fix a race condition in drbd when the sync speed was set after giving
2827 the device a remote peer.
2833 *(Released Tue, 22 Jul 2008)*
2835 - note: the allowed size and number of tags per object were reduced
2836 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2837 - fixed twisted 8.x compatibility
2838 - fixed ``gnt-instance replace-disks`` with iallocator
2839 - add TCP keepalives on twisted connections to detect restarted nodes
2840 - disk increase support, see ``gnt-instance grow-disk``
2841 - implement bulk node/instance query for RAPI
2842 - add tags in node/instance listing (optional)
2843 - experimental migration (and live migration) support, read the man page
2844 for ``gnt-instance migrate``
2845 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2846 has some small improvements in handling its state file
2852 *(Released Fri, 13 Jun 2008)*
2854 - Experimental readonly, REST-based remote API implementation;
2855 automatically started on master node, TCP port 5080, if enabled by
2856 ``--enable-rapi`` parameter to configure script.
2857 - Instance allocator support. Add and import instance accept a
2858 ``--iallocator`` parameter, and call that instance allocator to decide
2859 which node to use for the instance. The iallocator document describes
2860 what's expected from an allocator script.
2861 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2862 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2863 node is lost there is still enough memory to fail over the instances
2865 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2866 ``gnt-cluster verify``, to check for site-specific compliance. All the
2867 hooks will run, and their output, if any, will be displayed. Any
2868 failing hook will make the verification return an error value.
2869 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2870 primary and secondary interfaces
2871 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2872 that is still declared as part of the cluster and has failed.
2873 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2874 specifying output fields, that just adds the chosen fields to the
2876 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2877 export from the filesystem.
2878 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2879 have been added. Using them you can enable/disable acpi and pae
2880 support, and specify a path for a cd image to be exported to the
2881 instance. These parameters as the name suggest only work on HVM
2883 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2884 PAE support will be set to the previously hardcoded values, but the
2885 (previously hardcoded) path to the CDROM ISO image will be unset and
2886 if required, needs to be set manually with ``gnt-instance modify``
2888 - The address to which an instance's VNC console is bound is now
2889 selectable per-instance, rather than being cluster wide. Of course
2890 this only applies to instances controlled via VNC, so currently just
2891 applies to HVM clusters.
2897 *(Released Mon, 18 Feb 2008)*
2899 - more tweaks to the disk activation code (especially helpful for DRBD)
2900 - change the default ``gnt-instance list`` output format, now there is
2901 one combined status field (see the manpage for the exact values this
2903 - some more fixes for the mac export to hooks change
2904 - make Ganeti not break with DRBD 8.2.x (which changed the version
2905 format in ``/proc/drbd``) (issue 24)
2906 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2907 template, allowing migration from DRBD0.7+MD to DRBD8
2913 *(Released Wed, 30 Jan 2008)*
2915 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2917 - add command aliases infrastructure and a few aliases
2918 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2919 man pages and the ``--help`` option of ``gnt-node
2920 list``/``gnt-instance list``
2921 - fix ``gnt-backup list`` with down nodes (issue 21)
2922 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2923 - fix the dist archive and add a check for including svn/git files in
2925 - some developer-related changes: improve the burnin and the QA suite,
2926 add an upload script for testing during development
2932 *(Released Wed, 16 Jan 2008)*
2934 - experimental HVM support, read the install document, section
2935 "Initializing the cluster"
2936 - allow for the PVM hypervisor per-instance kernel and initrd paths
2937 - add a new command ``gnt-cluster verify-disks`` which uses a new
2938 algorithm to improve the reconnection of the DRBD pairs if the device
2939 on the secondary node has gone away
2940 - make logical volume code auto-activate LVs at disk activation time
2941 - slightly improve the speed of activating disks
2942 - allow specification of the MAC address at instance creation time, and
2943 changing it later via ``gnt-instance modify``
2944 - fix handling of external commands that generate lots of output on
2946 - update documentation with regard to minimum version of DRBD8 supported
2952 *(Released Tue, 4 Dec 2007)*
2954 - Log the ``xm create`` output to the node daemon log on failure (to
2955 help diagnosing the error)
2956 - In debug mode, log all external commands output if failed to the logs
2957 - Change parsing of lvm commands to ignore stderr
2963 *(Released Wed, 28 Nov 2007)*
2965 - Another round of updates to the DRBD 8 code to deal with more failures
2966 in the replace secondary node operation
2967 - Some more logging of failures in disk operations (lvm, drbd)
2968 - A few documentation updates
2975 *(Released Tue, 13 Nov 2007)*
2977 - Change configuration file format from Python's Pickle to JSON.
2978 Upgrading is possible using the cfgupgrade utility.
2979 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2980 faster replace disks and is more stable (DRBD 8 has many improvements
2981 compared to DRBD 0.7)
2982 - Added command line tags support (see man pages for ``gnt-instance``,
2983 ``gnt-node``, ``gnt-cluster``)
2984 - Added instance rename support
2985 - Added multi-instance startup/shutdown
2986 - Added cluster rename support
2987 - Added ``gnt-node evacuate`` to simplify some node operations
2988 - Added instance reboot operation that can speedup reboot as compared to
2990 - Soften the requirement that hostnames are in FQDN format
2991 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2993 - Removed dependency on debian's patched fping that uses the
2994 non-standard ``-S`` option
2995 - Now the OS definitions are searched for in multiple, configurable
2996 paths (easier for distros to package)
2997 - Some changes to the hooks infrastructure (especially the new
2998 post-configuration update hook)
2999 - Other small bugfixes
3001 .. vim: set textwidth=72 syntax=rst :