10 Incompatible/important changes
11 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
13 - Adding disks with 'gnt-instance modify' now waits for the disks to sync per
14 default. Specify --no-wait-for-sync to override this behavior.
18 - KVM hypervisors can now access RBD storage directly without having to
19 go through a block device.
25 *(Released Tue, 1 Oct 2013)*
27 Incompatible/important changes
28 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
30 - hroller now also plans for capacity to move non-redundant instances off
31 any node to be rebooted; the old behavior of completely ignoring any
32 non-redundant instances can be restored by adding the --ignore-non-redundant
34 - The cluster option '--no-lvm-storage' was removed in favor of the new option
35 '--enabled-disk-templates'.
36 - On instance creation, disk templates no longer need to be specified
37 with '-t'. The default disk template will be taken from the list of
38 enabled disk templates.
39 - The monitoring daemon is now running as root, in order to be able to collect
40 information only available to root (such as the state of Xen instances).
41 - The ConfD client is now IPv6 compatible.
42 - File and shared file storage is no longer dis/enabled at configure time,
43 but using the option '--enabled-disk-templates' at cluster initialization and
45 - The default directories for file and shared file storage are not anymore
46 specified at configure time, but taken from the cluster's configuration.
47 They can be set at cluster initialization and modification with
48 '--file-storage-dir' and '--shared-file-storage-dir'.
49 - Cluster verification now includes stricter checks regarding the
50 default file and shared file storage directories. It now checks that
51 the directories are explicitely allowed in the 'file-storage-paths' file and
52 that the directories exist on all nodes.
53 - The list of allowed disk templates in the instance policy and the list
54 of cluster-wide enabled disk templates is now checked for consistency
55 on cluster or group modification. On cluster initialization, the ipolicy
56 disk templates are ensured to be a subset of the cluster-wide enabled
62 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
63 the correct command syntax. It is possible to use different DRBD versions
64 on different nodes as long as they are compatible to each other. This
65 enables rolling upgrades of DRBD with no downtime. As permanent operation
66 of different DRBD versions within a node group is discouraged,
67 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
68 - New "inst-status-xen" data collector for the monitoring daemon, providing
69 information about the state of the xen instances on the nodes.
70 - New "lv" data collector for the monitoring daemon, collecting data about the
71 logical volumes on the nodes, and pairing them with the name of the instances
73 - New "diskstats" data collector, collecting the data from /proc/diskstats and
74 presenting them over the monitoring daemon interface.
75 - The ConfD client is now IPv6 compatible.
79 The following new dependencies have been added.
83 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
84 for the unit tests (and only used for testing).
88 - ``hslogger`` (http://software.complete.org/hslogger) is now always
89 required, even if confd is not enabled.
95 - update of the documentation, in particular installation instructions
96 - merging of LD_* constants into DT_* constants
97 - python style changes to be compatible with newer versions of pylint
103 *(Released Thu, 29 Aug 2013)*
105 This was the first beta release of the 2.9 series. All important changes
106 are listed in the latest 2.9 entry.
112 *(Released Mon, 30 Sep 2013)*
114 Incompatible/important changes
115 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
117 - Instance policy can contain multiple instance specs, as described in
118 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
119 <design-partitioned>`. As a consequence, it's not possible to partially change
120 or override instance specs. Bounding specs (min and max) can be specified as a
121 whole using the new option ``--ipolicy-bounds-specs``, while standard
122 specs use the new option ``--ipolicy-std-specs``.
123 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
124 gnt-instance is a valid YAML object.
125 - hail now honors network restrictions when allocating nodes. This led to an
126 update of the IAllocator protocol. See the IAllocator documentation for
128 - confd now only answers static configuration request over the network. luxid
129 was extracted, listens on the local LUXI socket and responds to live queries.
130 This allows finer grained permissions if using separate users.
135 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
136 to always require authentication, ``--require-authentication``. It can
137 be specified in ``$sysconfdir/default/ganeti``.
138 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
139 be used to manage the disk templates to be used by instances in the cluster.
140 Initially, it will be set to a list that includes plain, drbd, if they were
141 enabled by specifying a volume group name, and file and sharedfile, if those
142 were enabled at configure time. Additionally, it will include all disk
143 templates that are currently used by instances. The order of disk templates
144 will be based on Ganeti's history of supporting them. In the future, the
145 first entry of the list will be used as a default disk template on instance
147 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
148 configuration back to the previous stable version.
149 - Disk templates in group ipolicy can be restored to the default value.
150 - Initial support for diskless instances and virtual clusters in QA.
151 - More QA and unit tests for instance policies.
152 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
153 describing why the opcode itself was executed.
154 - The monitoring daemon is now available. It allows users to query the cluster
155 for obtaining information about the status of the system. The daemon is only
156 responsible for providing the information over the network: the actual data
157 gathering is performed by data collectors (currently, only the DRBD status
158 collector is available).
159 - In order to help developers work on Ganeti, a new script
160 (``devel/build_chroot``) is provided, for building a chroot that contains all
161 the required development libraries and tools for compiling Ganeti on a Debian
163 - A new tool, ``harep``, for performing self-repair and recreation of instances
164 in Ganeti has been added.
165 - Split queries are enabled for tags, network, exports, cluster info, groups,
167 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
168 It prints the command line to set the current policies, to ease
170 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
171 devices for KVM virtio-net interfaces will get created with VNET_HDR
172 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
173 virtio-net interfaces, which prevents host kernel tainting and log
174 flooding, when dealing with broken or malicious virtio-net drivers.
175 It's set to true by default.
176 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
178 - Support 'viridian' parameter in Xen HVM
179 - Support DSA SSH keys in bootstrap
180 - To simplify the work of packaging frameworks that want to add the needed users
181 and groups in a split-user setup themselves, at build time three files in
182 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
183 the groups to be generated, the ``users`` file contains, one per line, the
184 users to be generated, optionally followed by their primary group, where
185 important. The ``groupmemberships`` file contains, one per line, additional
186 user-group membership relations that need to be established. The syntax of
187 these files will remain stable in all future versions.
192 The following new dependencies have been added:
195 - The ``curl`` library is not optional anymore for compiling the Haskell code.
196 - ``snap-server`` library (if monitoring is enabled).
199 - The minimum Python version needed to run Ganeti is now 2.6.
200 - ``yaml`` library (only for running the QA).
204 - Perform proper cleanup on termination of Haskell daemons
205 - Fix corner-case in handling of remaining retry time
211 *(Released Tue, 17 Sep 2013)*
213 - To simplify the work of packaging frameworks that want to add the needed users
214 and groups in a split-user setup themselves, at build time three files in
215 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
216 the groups to be generated, the ``users`` file contains, one per line, the
217 users to be generated, optionally followed by their primary group, where
218 important. The ``groupmemberships`` file contains, one per line, additional
219 user-group membership relations that need to be established. The syntax of
220 these files will remain stable in all future versions.
221 - Add a default to file-driver when unspecified over RAPI (Issue 571)
222 - Mark the DSA host pubkey as optional, and remove it during config downgrade
224 - Some documentation fixes
230 *(Released Tue, 27 Aug 2013)*
232 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
234 - Support 'viridian' parameter in Xen HVM (Issue 233)
235 - Include VCS version in ``gnt-cluster version``
236 - Support DSA SSH keys in bootstrap (Issue 338)
237 - Fix batch creation of instances
238 - Use FQDN to check master node status (Issue 551)
239 - Make the DRBD collector more failure-resilient
245 *(Released Fri, 2 Aug 2013)*
247 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
249 - Fix upgrading/downgrading from 2.7
250 - Increase maximum RAPI message size
251 - Documentation updates
252 - Split ``confd`` between ``luxid`` and ``confd``
253 - Merge 2.7 series up to the 2.7.1 release
254 - Allow the ``modify_etc_hosts`` option to be changed
255 - Add better debugging for ``luxid`` queries
256 - Expose bulk parameter for GetJobs in RAPI client
257 - Expose missing ``network`` fields in RAPI
258 - Add some ``cluster verify`` tests
259 - Some unittest fixes
260 - Fix a malfunction in ``hspace``'s tiered allocation
261 - Fix query compatibility between haskell and python implementations
262 - Add the ``vnet_hdr`` HV parameter for KVM
263 - Add ``--cleanup`` to instance failover
264 - Change the connected groups format in ``gnt-network info`` output; it
265 was previously displayed as a raw list by mistake. (Merged from 2.7)
271 *(Released Mon, 24 Jun 2013)*
273 This was the first beta release of the 2.8 series. All important changes
274 are listed in the latest 2.8 entry.
280 *(Released Thu, 26 Sep 2013)*
282 - Change the connected groups format in ``gnt-network info`` output; it
283 was previously displayed as a raw list by mistake
284 - Check disk template in right dict when copying
285 - Support multi-instance allocs without iallocator
286 - Fix some errors in the documentation
287 - Fix formatting of tuple in an error message
293 *(Released Thu, 25 Jul 2013)*
295 - Add logrotate functionality in daemon-util
296 - Add logrotate example file
297 - Add missing fields to network queries over rapi
298 - Fix network object timestamps
299 - Add support for querying network timestamps
300 - Fix a typo in the example crontab
301 - Fix a documentation typo
307 *(Released Thu, 04 Jul 2013)*
309 Incompatible/important changes
310 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
312 - Instance policies for disk size were documented to be on a per-disk
313 basis, but hail applied them to the sum of all disks. This has been
315 - ``hbal`` will now exit with status 0 if, during job execution over
316 LUXI, early exit has been requested and all jobs are successful;
317 before, exit status 1 was used, which cannot be differentiated from
319 - Compatibility with newer versions of rbd has been fixed
320 - ``gnt-instance batch-create`` has been changed to use the bulk create
321 opcode from Ganeti. This lead to incompatible changes in the format of
322 the JSON file. It's now not a custom dict anymore but a dict
323 compatible with the ``OpInstanceCreate`` opcode.
324 - Parent directories for file storage need to be listed in
325 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
326 write the file automatically based on old configuration values, but it
327 can not distribute it across all nodes and the file contents should be
328 verified. Use ``gnt-cluster copyfile
329 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
330 upgraded. The reason for requiring this list of paths now is that
331 before it would have been possible to inject new paths via RPC,
332 allowing files to be created in arbitrary locations. The RPC protocol
333 is protected using SSL/X.509 certificates, but as a design principle
334 Ganeti does not permit arbitrary paths to be passed.
335 - The parsing of the variants file for OSes (see
336 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
337 lines and comment lines (starting with ``#``) are ignored for better
339 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
340 longer available. ``gnt-node add`` now invokes a new tool on the
341 destination node, named ``prepare-node-join``, to configure the SSH
342 daemon. Paramiko is no longer necessary to configure nodes' SSH
343 daemons via ``gnt-node add``.
344 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
345 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
346 the flag is not reset after a master failover.
347 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
348 is recommended. 2.6 will be mandatory from the 2.8 series.
354 - New network management functionality to support automatic allocation
355 of IP addresses and managing of network parameters. See
356 :manpage:`gnt-network(8)` for more details.
357 - New external storage backend, to allow managing arbitrary storage
358 systems external to the cluster. See
359 :manpage:`ganeti-extstorage-interface(7)`.
360 - New ``exclusive-storage`` node parameter added, restricted to
361 nodegroup level. When it's set to true, physical disks are assigned in
362 an exclusive fashion to instances, as documented in :doc:`Partitioned
363 Ganeti <design-partitioned>`. Currently, only instances using the
364 ``plain`` disk template are supported.
365 - The KVM hypervisor has been updated with many new hypervisor
366 parameters, including a generic one for passing arbitrary command line
367 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
368 compatible up to qemu 1.4.
369 - A new tool, called ``mon-collector``, is the stand-alone executor of
370 the data collectors for a monitoring system. As of this version, it
371 just includes the DRBD data collector, that can be executed by calling
372 ``mon-collector`` using the ``drbd`` parameter. See
373 :manpage:`mon-collector(7)`.
374 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
375 for RAPI users. It allows granting permissions to query for
376 information to a specific user without giving
377 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
378 - A new tool named ``node-cleanup`` has been added. It cleans remains of
379 a cluster from a machine by stopping all daemons, removing
380 certificates and ssconf files. Unless the ``--no-backup`` option is
381 given, copies of the certificates are made.
382 - Instance creations now support the use of opportunistic locking,
383 potentially speeding up the (parallel) creation of multiple instances.
384 This feature is currently only available via the :doc:`RAPI
385 <rapi>` interface and when an instance allocator is used. If the
386 ``opportunistic_locking`` parameter is set the opcode will try to
387 acquire as many locks as possible, but will not wait for any locks
388 held by other opcodes. If not enough resources can be found to
389 allocate the instance, the temporary error code
390 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
391 retried thereafter, with or without opportunistic locking.
392 - New experimental linux-ha resource scripts.
393 - Restricted-commands support: ganeti can now be asked (via command line
394 or rapi) to perform commands on a node. These are passed via ganeti
395 RPC rather than ssh. This functionality is restricted to commands
396 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
397 reasons. The file is not copied automatically.
403 - Diskless instances are now externally mirrored (Issue 237). This for
404 now has only been tested in conjunction with explicit target nodes for
406 - Queries not needing locks or RPC access to the node can now be
407 performed by the confd daemon, making them independent from jobs, and
408 thus faster to execute. This is selectable at configure time.
409 - The functionality for allocating multiple instances at once has been
410 overhauled and is now also available through :doc:`RAPI <rapi>`.
412 There are no significant changes from version 2.7.0~rc3.
418 *(Released Tue, 25 Jun 2013)*
420 - Fix permissions on the confd query socket (Issue 477)
421 - Fix permissions on the job archive dir (Issue 498)
422 - Fix handling of an internal exception in replace-disks (Issue 472)
423 - Fix gnt-node info handling of shortened names (Issue 497)
424 - Fix gnt-instance grow-disk when wiping is enabled
425 - Documentation improvements, and support for newer pandoc
426 - Fix hspace honoring ipolicy for disks (Issue 484)
427 - Improve handling of the ``kvm_extra`` HV parameter
433 *(Released Fri, 24 May 2013)*
435 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
437 - Disks added through ``gnt-instance modify`` or created through
438 ``gnt-instance recreate-disks`` are wiped, if the
439 ``prealloc_wipe_disks`` flag is set.
440 - If wiping newly created disks fails, the disks are removed. Also,
441 partial failures in creating disks through ``gnt-instance modify``
442 triggers a cleanup of the partially-created disks.
443 - Removing the master IP address doesn't fail if the address has been
445 - Fix ownership of the OS log dir
446 - Workaround missing SO_PEERCRED constant (Issue 191)
452 *(Released Fri, 3 May 2013)*
454 This was the first release candidate of the 2.7 series. Since beta3:
456 - Fix kvm compatibility with qemu 1.4 (Issue 389)
457 - Documentation updates (admin guide, upgrade notes, install
458 instructions) (Issue 372)
459 - Fix gnt-group list nodes and instances count (Issue 436)
460 - Fix compilation without non-mandatory libraries (Issue 441)
461 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
462 - Make confd logging more verbose at INFO level (Issue 435)
463 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
464 - Fix failure path for instance storage type conversion (Issue 229)
465 - Update htools text backend documentation
466 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
467 - Disable inter-cluster instance move for file-based instances, because
468 it is dependant on instance export, which is not supported for
469 file-based instances. (Issue 414)
470 - Fix gnt-job crashes on non-ascii characters (Issue 427)
471 - Fix volume group checks on non-vm-capable nodes (Issue 432)
477 *(Released Mon, 22 Apr 2013)*
479 This was the third beta release of the 2.7 series. Since beta2:
481 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
482 - Fix data loss on wrong usage of ``gnt-instance move``
483 - Properly export errors in confd-based job queries
484 - Add ``users-setup`` tool
485 - Fix iallocator protocol to report 0 as a disk size for diskless
486 instances. This avoids hail breaking when a diskless instance is
488 - Fix job queue directory permission problem that made confd job queries
489 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
490 for access to archived jobs (Issue 406).
491 - Limit the sizes of networks supported by ``gnt-network`` to something
492 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
493 - Fix bugs in instance disk template conversion
494 - Fix GHC 7 compatibility
495 - Fix ``burnin`` install path (Issue 426).
496 - Allow very small disk grows (Issue 347).
497 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
498 that noded doesn't import masterd code (Issue 419).
499 - Make sure the default metavg at cluster init is the same as the vg, if
500 unspecified (Issue 358).
501 - Fix cleanup of partially created disks (part of Issue 416)
507 *(Released Tue, 2 Apr 2013)*
509 This was the second beta release of the 2.7 series. Since beta1:
511 - Networks no longer have a "type" slot, since this information was
512 unused in Ganeti: instead of it tags should be used.
513 - The rapi client now has a ``target_node`` option to MigrateInstance.
514 - Fix early exit return code for hbal (Issue 386).
515 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
516 - Fix ``rbd showmapped`` output parsing (Issue 312).
517 - Networks are now referenced indexed by UUID, rather than name. This
518 will require running cfgupgrade, from 2.7.0beta1, if networks are in
520 - The OS environment now includes network information.
521 - Deleting of a network is now disallowed if any instance nic is using
522 it, to prevent dangling references.
523 - External storage is now documented in man pages.
524 - The exclusive_storage flag can now only be set at nodegroup level.
525 - Hbal can now submit an explicit priority with its jobs.
526 - Many network related locking fixes.
527 - Bump up the required pylint version to 0.25.1.
528 - Fix the ``no_remember`` option in RAPI client.
529 - Many ipolicy related tests, qa, and fixes.
530 - Many documentation improvements and fixes.
531 - Fix building with ``--disable-file-storage``.
532 - Fix ``-q`` option in htools, which was broken if passed more than
534 - Some haskell/python interaction improvements and fixes.
535 - Fix iallocator in case of missing LVM storage.
536 - Fix confd config load in case of ``--no-lvm-storage``.
537 - The confd/query functionality is now mentioned in the security
544 *(Released Wed, 6 Feb 2013)*
546 This was the first beta release of the 2.7 series. All important changes
547 are listed in the latest 2.7 entry.
553 *(Released Fri, 21 Dec 2012)*
555 Important behaviour change: hbal won't rebalance anymore instances which
556 have the ``auto_balance`` attribute set to false. This was the intention
557 all along, but until now it only skipped those from the N+1 memory
558 reservation (DRBD-specific).
560 A significant number of bug fixes in this release:
562 - Fixed disk adoption interaction with ipolicy checks.
563 - Fixed networking issues when instances are started, stopped or
564 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
565 - Fixed the warning in cluster verify for shared storage instances not
567 - Fixed removal of storage directory on shared file storage (issue 262).
568 - Fixed validation of LVM volume group name in OpClusterSetParams
569 (``gnt-cluster modify``) (issue 285).
570 - Fixed runtime memory increases (``gnt-instance modify -m``).
571 - Fixed live migration under Xen's ``xl`` mode.
572 - Fixed ``gnt-instance console`` with ``xl``.
573 - Fixed building with newer Haskell compiler/libraries.
574 - Fixed PID file writing in Haskell daemons (confd); this prevents
575 restart issues if confd was launched manually (outside of
576 ``daemon-util``) while another copy of it was running
577 - Fixed a type error when doing live migrations with KVM (issue 297) and
578 the error messages for failing migrations have been improved.
579 - Fixed opcode validation for the out-of-band commands (``gnt-node
581 - Fixed a type error when unsetting OS hypervisor parameters (issue
582 311); now it's possible to unset all OS-specific hypervisor
584 - Fixed the ``dry-run`` mode for many operations: verification of
585 results was over-zealous but didn't take into account the ``dry-run``
586 operation, resulting in "wrong" failures.
587 - Fixed bash completion in ``gnt-job list`` when the job queue has
588 hundreds of entries; especially with older ``bash`` versions, this
589 results in significant CPU usage.
591 And lastly, a few other improvements have been made:
593 - Added option to force master-failover without voting (issue 282).
594 - Clarified error message on lock conflict (issue 287).
595 - Logging of newly submitted jobs has been improved (issue 290).
596 - Hostname checks have been made uniform between instance rename and
598 - The ``--submit`` option is now supported by ``gnt-debug delay``.
599 - Shutting down the master daemon by sending SIGTERM now stops it from
600 processing jobs waiting for locks; instead, those jobs will be started
601 once again after the master daemon is started the next time (issue
603 - Support for Xen's ``xl`` program has been improved (besides the fixes
605 - Reduced logging noise in the Haskell confd daemon (only show one log
606 entry for each config reload, instead of two).
607 - Several man page updates and typo fixes.
613 *(Released Fri, 12 Oct 2012)*
615 A small bugfix release. Among the bugs fixed:
617 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
618 made the command unusable.
619 - Commands that issue many jobs don't fail anymore just because some jobs
620 take so long that other jobs are archived.
621 - Failures during ``gnt-instance reinstall`` are reflected by the exit
623 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
625 - When ``always_failover`` is set, ``--allow-failover`` is not required
626 in migrate commands anymore.
627 - ``bash_completion`` works even if extglob is disabled.
628 - Fixed bug with locks that made failover for RDB-based instances fail.
629 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
630 a random node instead of one based on the allocator metric.
631 - Support for newer versions of pylint and pep8.
632 - Hail doesn't fail anymore when trying to add an instance of type
633 ``file``, ``sharedfile`` or ``rbd``.
634 - Added new Makefile target to rebuild the whole distribution, so that
635 all files are included.
641 *(Released Fri, 27 Jul 2012)*
644 .. attention:: The ``LUXI`` protocol has been made more consistent
645 regarding its handling of command arguments. This, however, leads to
646 incompatibility issues with previous versions. Please ensure that you
647 restart Ganeti daemons soon after the upgrade, otherwise most
648 ``LUXI`` calls (job submission, setting/resetting the drain flag,
649 pausing/resuming the watcher, cancelling and archiving jobs, querying
650 the cluster configuration) will fail.
659 The current ``admin_up`` field, which used to denote whether an instance
660 should be running or not, has been removed. Instead, ``admin_state`` is
661 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
663 The rational behind this is that an instance being “down” can have
666 - it could be down during a reboot
667 - it could be temporarily be down for a reinstall
668 - or it could be down because it is deprecated and kept just for its
671 The previous Boolean state was making it difficult to do capacity
672 calculations: should Ganeti reserve memory for a down instance? Now, the
673 tri-state field makes it clear:
675 - in ``up`` and ``down`` state, all resources are reserved for the
676 instance, and it can be at any time brought up if it is down
677 - in ``offline`` state, only disk space is reserved for it, but not
680 The field can have an extra use: since the transition between ``up`` and
681 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
682 transition between ``offline`` and ``down`` is done via ``gnt-instance
683 modify``, it is possible to given different rights to users. For
684 example, owners of an instance could be allowed to start/stop it, but
685 not transition it out of the offline state.
687 Instance policies and specs
688 +++++++++++++++++++++++++++
690 In previous Ganeti versions, an instance creation request was not
691 limited on the minimum size and on the maximum size just by the cluster
692 resources. As such, any policy could be implemented only in third-party
693 clients (RAPI clients, or shell wrappers over ``gnt-*``
694 tools). Furthermore, calculating cluster capacity via ``hspace`` again
695 required external input with regards to instance sizes.
697 In order to improve these workflows and to allow for example better
698 per-node group differentiation, we introduced instance specs, which
701 - minimum instance disk size, disk count, memory size, cpu count
702 - maximum values for the above metrics
703 - and “standard” values (used in ``hspace`` to calculate the standard
706 The minimum/maximum values can be also customised at node-group level,
707 for example allowing more powerful hardware to support bigger instance
710 Beside the instance specs, there are a few other settings belonging to
711 the instance policy framework. It is possible now to customise, per
712 cluster and node-group:
714 - the list of allowed disk templates
715 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
716 - the maximum ratio of instance to spindles (see below for more
717 information) for local storage
719 All these together should allow all tools that talk to Ganeti to know
720 what are the ranges of allowed values for instances and the
721 over-subscription that is allowed.
723 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
724 instance configuration, and the physical CPU configuration from the
725 node. For the spindle ratios however, we didn't track before these
726 values, so new parameters have been added:
728 - a new node parameter ``spindle_count``, defaults to 1, customisable at
729 node group or node level
730 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
732 Note that spindles in this context doesn't need to mean actual
733 mechanical hard-drives; it's just a relative number for both the node
734 I/O capacity and instance I/O consumption.
736 Instance migration behaviour
737 ++++++++++++++++++++++++++++
739 While live-migration is in general desirable over failover, it is
740 possible that for some workloads it is actually worse, due to the
741 variable time of the “suspend” phase during live migration.
743 To allow the tools to work consistently over such instances (without
744 having to hard-code instance names), a new backend parameter
745 ``always_failover`` has been added to control the migration/failover
746 behaviour. When set to True, all migration requests for an instance will
747 instead fall-back to failover.
749 Instance memory ballooning
750 ++++++++++++++++++++++++++
752 Initial support for memory ballooning has been added. The memory for an
753 instance is no longer fixed (backend parameter ``memory``), but instead
754 can vary between minimum and maximum values (backend parameters
755 ``minmem`` and ``maxmem``). Currently we only change an instance's
758 - live migrating or failing over and instance and the target node
759 doesn't have enough memory
760 - user requests changing the memory via ``gnt-instance modify
766 In order to control the use of specific CPUs by instance, support for
767 controlling CPU pinning has been added for the Xen, HVM and LXC
768 hypervisors. This is controlled by a new hypervisor parameter
769 ``cpu_mask``; details about possible values for this are in the
770 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
771 VCPU-to-CPU mapping) form will work well only when all nodes in your
772 cluster have the same amount of CPUs.
777 Another area in which Ganeti was not customisable were the parameters
778 used for storage configuration, e.g. how many stripes to use for LVM,
779 DRBD resync configuration, etc.
781 To improve this area, we've added disks parameters, which are
782 customisable at cluster and node group level, and which allow to
783 specify various parameters for disks (DRBD has the most parameters
784 currently), for example:
786 - DRBD resync algorithm and parameters (e.g. speed)
787 - the default VG for meta-data volumes for DRBD
788 - number of stripes for LVM (plain disk template)
791 These parameters can be modified via ``gnt-cluster modify -D …`` and
792 ``gnt-group modify -D …``, and are used at either instance creation (in
793 case of LVM stripes, for example) or at disk “activation” time
796 Rados block device support
797 ++++++++++++++++++++++++++
799 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
800 denoted by the ``rbd`` disk template type. This is considered
801 experimental, feedback is welcome. For details on configuring it, see
802 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
807 The existing master IP functionality works well only in simple setups (a
808 single network shared by all nodes); however, if nodes belong to
809 different networks, then the ``/32`` setup and lack of routing
810 information is not enough.
812 To allow the master IP to function well in more complex cases, the
813 system was reworked as follows:
815 - a master IP netmask setting has been added
816 - the master IP activation/turn-down code was moved from the node daemon
818 - whether to run the Ganeti-supplied master IP script or a user-supplied
819 on is a ``gnt-cluster init`` setting
821 Details about the location of the standard and custom setup scripts are
822 in the man page :manpage:`gnt-cluster(8)`; for information about the
823 setup script protocol, look at the Ganeti-supplied script.
828 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
831 It is now possible to use TLS-protected connections, and when renewing
832 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
833 it is now possible to specify spice or spice CA certificates. Also, it
834 is possible to configure a password for SPICE sessions via the
835 hypervisor parameter ``spice_password_file``.
837 There are also new parameters to control the compression and streaming
838 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
839 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
840 for the spice parameters.
842 Lastly, it is now possible to see the SPICE connection information via
843 ``gnt-instance console``.
848 A new tool (``tools/ovfconverter``) has been added that supports
849 conversion between Ganeti and the `Open Virtualization Format
850 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
853 This relies on the ``qemu-img`` tool to convert the disk formats, so the
854 actual compatibility with other virtualization solutions depends on it.
859 The configuration query daemon (``ganeti-confd``) is now optional, and
860 has been rewritten in Haskell; whether to use the daemon at all, use the
861 Python (default) or the Haskell version is selectable at configure time
862 via the ``--enable-confd`` parameter, which can take one of the
863 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
864 daemon will result in a smaller footprint; for larger systems, we
865 welcome feedback on the Haskell version which might become the default
868 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
869 daemon running. The Python version doesn't implement the new call.
872 User interface changes
873 ~~~~~~~~~~~~~~~~~~~~~~
875 We have replaced the ``--disks`` option of ``gnt-instance
876 replace-disks`` with a more flexible ``--disk`` option, which allows
877 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
878 disk size and mode can be changed upon recreation (via ``gnt-instance
879 recreate-disks``, which accepts the same ``--disk`` option).
881 As many people are used to a ``show`` command, we have added that as an
882 alias to ``info`` on all ``gnt-*`` commands.
884 The ``gnt-instance grow-disk`` command has a new mode in which it can
885 accept the target size of the disk, instead of the delta; this can be
886 more safe since two runs in absolute mode will be idempotent, and
887 sometimes it's also easier to specify the desired size directly.
889 Also the handling of instances with regard to offline secondaries has
890 been improved. Instance operations should not fail because one of it's
891 secondary nodes is offline, even though it's safe to proceed.
893 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
894 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
895 minors to instance name.
900 RAPI coverage has improved, with (for example) new resources for
901 recreate-disks, node power-cycle, etc.
906 There is partial support for ``xl`` in the Xen hypervisor; feedback is
909 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
910 whether to still support Python 2.4 or move to Python 2.6 as minimum
913 Support for Fedora has been slightly improved; the provided example
914 init.d script should work better on it and the INSTALL file should
915 document the needed dependencies.
920 The deprecated ``QueryLocks`` LUXI request has been removed. Use
921 ``Query(what=QR_LOCK, ...)`` instead.
923 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
924 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
925 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
926 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
927 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
929 RAPI client: ``CertificateError`` now derives from
930 ``GanetiApiError``. This should make it more easy to handle Ganeti
933 Deprecation warnings due to PyCrypto/paramiko import in
934 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
935 make sure to run an up-to-date paramiko version, if you use this tool.
937 The QA scripts now depend on Python 2.5 or above (the main code base
938 still works with Python 2.4).
940 The configuration file (``config.data``) is now written without
941 indentation for performance reasons; if you want to edit it, it can be
942 re-formatted via ``tools/fmtjson``.
944 A number of bugs has been fixed in the cluster merge tool.
946 ``x509`` certification verification (used in import-export) has been
947 changed to allow the same clock skew as permitted by the cluster
948 verification. This will remove some rare but hard to diagnose errors in
955 *(Released Thu, 19 Jul 2012)*
957 Very few changes from rc4 to the final release, only bugfixes:
959 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
960 instance, fix CDROM booting for KVM instances)
961 - fixed node group modification of node parameters
962 - fixed issue in LUClusterVerifyGroup with multi-group clusters
963 - fixed generation of bash completion to ensure a stable ordering
970 *(Released Fri, 13 Jul 2012)*
972 Third release candidate for 2.6. The following changes were done from
975 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
976 - Fixed an inconsistency in the LUXI protocol with the provided
977 arguments (NOT backwards compatible)
978 - Fixed a bug with node groups ipolicy where ``min`` was greater than
979 the cluster ``std`` value
980 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
981 easier instance debugging on nodes (requires ``hconfd`` to work)
987 *(Released Tue, 03 Jul 2012)*
989 Second release candidate for 2.6. The following changes were done from
992 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
994 - Fixed a RAPI regression on missing beparams/memory
995 - Fixed redistribution of files on offline nodes
996 - Added possibility to run activate-disks even though secondaries are
997 offline. With this change it relaxes also the strictness on some other
998 commands which use activate disks internally:
999 * ``gnt-instance start|reboot|rename|backup|export``
1000 - Made it possible to remove safely an instance if its secondaries are
1002 - Made it possible to reinstall even though secondaries are offline
1008 *(Released Mon, 25 Jun 2012)*
1010 First release candidate for 2.6. The following changes were done from
1013 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1015 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1016 behaviour. It supports now the modify operation
1017 - ``hcheck`` implemented to analyze cluster health and possibility of
1018 improving health by rebalance
1019 - ``hbal`` has been improved in dealing with split instances
1025 *(Released Mon, 11 Jun 2012)*
1027 Second beta release of 2.6. The following changes were done from beta2
1030 - Fixed ``daemon-util`` with non-root user models
1031 - Fixed creation of plain instances with ``--no-wait-for-sync``
1032 - Fix wrong iv_names when running ``cfgupgrade``
1033 - Export more information in RAPI group queries
1034 - Fixed bug when changing instance network interfaces
1035 - Extended burnin to do NIC changes
1036 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1037 - Changed default for DRBD barriers
1038 - Fixed DRBD error reporting for syncer rate
1039 - Verify the options on disk parameters
1041 And of course various fixes to documentation and improved unittests and
1048 *(Released Wed, 23 May 2012)*
1050 First beta release of 2.6. The following changes were done from beta1 to
1053 - integrated patch for distributions without ``start-stop-daemon``
1054 - adapted example init.d script to work on Fedora
1055 - fixed log handling in Haskell daemons
1056 - adapted checks in the watcher for pycurl linked against libnss
1057 - add partial support for ``xl`` instead of ``xm`` for Xen
1058 - fixed a type issue in cluster verification
1059 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1062 Plus integrated fixes from the 2.5 branch:
1064 - fixed ``kvm-ifup`` to use ``/bin/bash``
1065 - fixed parallel build failures
1066 - KVM live migration when using a custom keymap
1072 *(Released Tue, 24 Jul 2012)*
1074 A small bugfix release, with no new features:
1076 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1077 different default shell (e.g. Debian, Ubuntu)
1078 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1079 243 and Debian bug #650664)
1080 - fixed compatibility with KVM versions that don't support multiple boot
1081 devices (fixes Issue 230 and Debian bug #624256)
1083 Additionally, a few fixes were done to the build system (fixed parallel
1084 build failures) and to the unittests (fixed race condition in test for
1085 FileID functions, and the default enable/disable mode for QA test is now
1092 *(Released Fri, 11 May 2012)*
1094 A small bugfix release.
1096 The main issues solved are on the topic of compatibility with newer LVM
1099 - fixed parsing of ``lv_attr`` field
1100 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1101 the ``--force`` flag is needed
1103 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1104 to accept kernel 3.x too (was hardcoded to 2.6.*).
1106 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1107 of hook results and that also made display of other errors suboptimal
1108 was fixed; the code behaves now like 2.4 and earlier.
1110 Another change in 2.5, the cleanup of the OS scripts environment, is too
1111 aggressive: it removed even the ``PATH`` variable, which requires the OS
1112 scripts to *always* need to export it. Since this is a bit too strict,
1113 we now export a minimal PATH, the same that we export for hooks.
1115 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1116 integrated into this release.
1118 Finally, a few other miscellaneous changes were done (no new features,
1119 just small improvements):
1121 - Fix ``gnt-group --help`` display
1122 - Fix hardcoded Xen kernel path
1123 - Fix grow-disk handling of invalid units
1124 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1125 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1131 *(Released Thu, 12 Apr 2012)*
1133 Incompatible/important changes and bugfixes
1134 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1136 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1137 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1138 to match the underlying LUXI interface.
1139 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1140 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1141 not maintain backwards-compatibility as the underlying operation
1142 changed in an incompatible way. The RAPI client can talk to old
1143 servers, but it needs to be told so as the return value changed.
1144 - When creating file-based instances via RAPI, the ``file_driver``
1145 parameter no longer defaults to ``loop`` and must be specified.
1146 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1148 - Support for the undocumented and deprecated RAPI instance creation
1149 request format version 0 has been dropped. Use version 1, supported
1150 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1151 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1152 documentation <install>`.
1153 - The "cluster-verify" hooks are now executed per group by the
1154 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1155 if you just run ``gnt-cluster verify``, which generates one opcode per
1157 - The environment as passed to the OS scripts is cleared, and thus no
1158 environment variables defined in the node daemon's environment will be
1159 inherited by the scripts.
1160 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1162 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1163 support operations involving multiple node groups.
1164 - Offline nodes are ignored when failing over an instance.
1165 - Support for KVM version 1.0, which changed the version reporting format
1167 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1169 - ``Makefile`` is now compatible with Automake 1.11.2
1170 - Includes all bugfixes made in the 2.4 series
1175 - The ganeti-htools project has been merged into the ganeti-core source
1176 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1177 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1178 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1179 GPT-style partition tables (requires `parted
1180 <http://www.gnu.org/s/parted/>`_).
1181 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1182 - Allowed adding tags on instance creation.
1183 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1184 - Allow instances to be started in a paused state, enabling the user to
1185 see the complete console output on boot using the console.
1186 - Added new hypervisor flag to control default reboot behaviour
1187 (``reboot_behavior``).
1188 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1189 - Improved out-of-band management support:
1191 - Added ``gnt-node health`` command reporting the health status of
1193 - Added ``gnt-node power`` command to manage power status of nodes.
1194 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1196 - Instance migration can fall back to failover if instance is not
1198 - Filters can be used when listing nodes, instances, groups and locks;
1199 see :manpage:`ganeti(7)` manpage.
1200 - Added post-execution status as variables to :doc:`hooks <hooks>`
1202 - Instance tags are exported/imported together with the instance.
1203 - When given an explicit job ID, ``gnt-job info`` will work for archived
1205 - Jobs can define dependencies on other jobs (not yet supported via
1206 RAPI or command line, but used by internal commands and usable via
1209 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1212 - Instance failover is now available as a RAPI resource
1213 (``/2/instances/[instance_name]/failover``).
1214 - ``gnt-instance info`` defaults to static information if primary node
1216 - Opcodes have a new ``comment`` attribute.
1217 - Added basic SPICE support to KVM hypervisor.
1218 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1220 Node group improvements
1221 ~~~~~~~~~~~~~~~~~~~~~~~
1223 - ``gnt-cluster verify`` has been modified to check groups separately,
1224 thereby improving performance.
1225 - Node group support has been added to ``gnt-cluster verify-disks``,
1226 which now operates per node group.
1227 - Watcher has been changed to work better with node groups.
1229 - One process and state file per node group.
1230 - Slow watcher in one group doesn't block other group's watcher.
1232 - Added new command, ``gnt-group evacuate``, to move all instances in a
1233 node group to other groups.
1234 - Added ``gnt-instance change-group`` to move an instance to another
1236 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1237 per-group operations.
1238 - Node groups can be tagged.
1239 - Some operations switch from an exclusive to a shared lock as soon as
1241 - Instance's primary and secondary nodes' groups are now available as
1242 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1243 and ``snodes.group.uuid``).
1248 - Numerous updates to documentation and manpages.
1250 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1252 - Some opcode/job results are now also documented, see :doc:`RAPI
1255 - A lockset's internal lock is now also visible in lock monitor.
1256 - Log messages from job queue workers now contain information about the
1257 opcode they're processing.
1258 - ``gnt-instance console`` no longer requires the instance lock.
1259 - A short delay when waiting for job changes reduces the number of LUXI
1260 requests significantly.
1261 - DRBD metadata volumes are overwritten with zeros during disk creation.
1262 - Out-of-band commands no longer acquire the cluster lock in exclusive
1264 - ``devel/upload`` now uses correct permissions for directories.
1270 *(Released Fri, 23 Mar 2012)*
1272 This was the sixth release candidate of the 2.5 series.
1278 *(Released Mon, 9 Jan 2012)*
1280 This was the fifth release candidate of the 2.5 series.
1286 *(Released Thu, 27 Oct 2011)*
1288 This was the fourth release candidate of the 2.5 series.
1294 *(Released Wed, 26 Oct 2011)*
1296 This was the third release candidate of the 2.5 series.
1302 *(Released Tue, 18 Oct 2011)*
1304 This was the second release candidate of the 2.5 series.
1310 *(Released Tue, 4 Oct 2011)*
1312 This was the first release candidate of the 2.5 series.
1318 *(Released Wed, 31 Aug 2011)*
1320 This was the third beta release of the 2.5 series.
1326 *(Released Mon, 22 Aug 2011)*
1328 This was the second beta release of the 2.5 series.
1334 *(Released Fri, 12 Aug 2011)*
1336 This was the first beta release of the 2.5 series.
1342 *(Released Thu, 27 Oct 2011)*
1344 - Fixed bug when parsing command line parameter values ending in
1346 - Fixed assertion error after unclean master shutdown
1347 - Disable HTTP client pool for RPC, significantly reducing memory usage
1349 - Fixed queue archive creation with wrong permissions
1355 *(Released Tue, 23 Aug 2011)*
1359 - Fixed documentation for importing with ``--src-dir`` option
1360 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1361 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1367 *(Released Fri, 5 Aug 2011)*
1369 Many bug-fixes and a few small features:
1371 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1372 issues when you tried to add an instance with two MAC addresses in one
1374 - KVM: fixed per-instance stored UID value
1375 - KVM: configure bridged NICs at migration start
1376 - KVM: Fix a bug where instance will not start with never KVM versions
1378 - Added OS search path to ``gnt-cluster info``
1379 - Fixed an issue with ``file_storage_dir`` where you were forced to
1380 provide an absolute path, but the documentation states it is a
1381 relative path, the documentation was right
1382 - Added a new parameter to instance stop/start called ``--no-remember``
1383 that will make the state change to not be remembered
1384 - Implemented ``no_remember`` at RAPI level
1385 - Improved the documentation
1386 - Node evacuation: don't call IAllocator if node is already empty
1387 - Fixed bug in DRBD8 replace disks on current nodes
1388 - Fixed bug in recreate-disks for DRBD instances
1389 - Moved assertion checking locks in ``gnt-instance replace-disks``
1390 causing it to abort with not owning the right locks for some situation
1391 - Job queue: Fixed potential race condition when cancelling queued jobs
1392 - Fixed off-by-one bug in job serial generation
1393 - ``gnt-node volumes``: Fix instance names
1394 - Fixed aliases in bash completion
1395 - Fixed a bug in reopening log files after being sent a SIGHUP
1396 - Added a flag to burnin to allow specifying VCPU count
1397 - Bugfixes to non-root Ganeti configuration
1403 *(Released Thu, 12 May 2011)*
1405 Many bug-fixes and a few new small features:
1407 - Fixed a bug related to log opening failures
1408 - Fixed a bug in instance listing with orphan instances
1409 - Fixed a bug which prevented resetting the cluster-level node parameter
1410 ``oob_program`` to the default
1411 - Many fixes related to the ``cluster-merge`` tool
1412 - Fixed a race condition in the lock monitor, which caused failures
1413 during (at least) creation of many instances in parallel
1414 - Improved output for gnt-job info
1415 - Removed the quiet flag on some ssh calls which prevented debugging
1417 - Improved the N+1 failure messages in cluster verify by actually
1418 showing the memory values (needed and available)
1419 - Increased lock attempt timeouts so that when executing long operations
1420 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1421 too early and thus prevent the use of the 'fair' mechanism
1422 - Changed instance query data (``gnt-instance info``) to not acquire
1423 locks unless needed, thus allowing its use on locked instance if only
1424 static information is asked for
1425 - Improved behaviour with filesystems that do not support rename on an
1427 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1428 kept locks on all nodes during the wipe, which is unneeded
1429 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1430 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1431 10GiB) which caused the wipe to fail right at the end in some cases
1432 - Fixed master IP activation when doing master failover with no-voting
1433 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1434 the master node itself
1435 - Fixed potential data-loss in under disk full conditions, where Ganeti
1436 wouldn't check correctly the return code and would consider
1437 partially-written files 'correct'
1438 - Fixed bug related to multiple VGs and DRBD disk replacing
1439 - Added new disk parameter ``metavg`` that allows placement of the meta
1440 device for DRBD in a different volume group
1441 - Fixed error handling in the node daemon when the system libc doesn't
1442 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1443 - Fixed lock release during replace-disks, which kept cluster-wide locks
1444 when doing disk replaces with an iallocator script
1445 - Added check for missing bridges in cluster verify
1446 - Handle EPIPE errors while writing to the terminal better, so that
1447 piping the output to e.g. ``less`` doesn't cause a backtrace
1448 - Fixed rare case where a ^C during Luxi calls could have been
1449 interpreted as server errors, instead of simply terminating
1450 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1452 - Added a few more parameters to the KVM hypervisor, allowing a second
1453 CDROM, custom disk type for CDROMs and a floppy image
1454 - Removed redundant message in instance rename when the name is given
1456 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1457 disks on new nodes, allowing recreation when the original instance
1458 nodes are completely gone
1459 - Added option when converting disk templates to DRBD to skip waiting
1460 for the resync, in order to make the instance available sooner
1461 - Added two new variables to the OS scripts environment (containing the
1463 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1464 to allow use of ``pvgrub`` as bootloader
1465 - Changed the instance memory modifications to only check out-of-memory
1466 conditions on memory increases, and turned the secondary node warnings
1467 into errors (they can still be overridden via ``--force``)
1468 - Fixed the handling of a corner case when the Python installation gets
1469 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1470 to execute a command that doesn't exist
1471 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1472 node of the instance returned failures during instance shutdown; this
1473 adds the option ``--ignore-consistency`` to gnt-instance move
1475 And as usual, various improvements to the error messages, documentation
1482 *(Released Wed, 09 Mar 2011)*
1484 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1485 the RAPI users file if run twice (even with ``--dry-run``).
1487 The release fixes that bug (nothing else changed).
1493 *(Released Mon, 07 Mar 2011)*
1495 Final 2.4.0 release. Just a few small fixes:
1497 - Fixed RAPI node evacuate
1498 - Fixed the kvm-ifup script
1499 - Fixed internal error handling for special job cases
1500 - Updated man page to specify the escaping feature for options
1506 *(Released Mon, 28 Feb 2011)*
1508 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1509 that this feature wiped the disks of the wrong instance, leading to loss
1514 - Fixed title of query field containing instance name
1515 - Expanded the glossary in the documentation
1516 - Fixed one unittest (internal issue)
1522 *(Released Mon, 21 Feb 2011)*
1524 A number of bug fixes plus just a couple functionality changes.
1526 On the user-visible side, the ``gnt-* list`` command output has changed
1527 with respect to "special" field states. The current rc1 style of display
1528 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1529 the default output mode special fields states are displayed as follows:
1531 - Offline resource: ``*``
1532 - Unavailable/not applicable: ``-``
1533 - Data missing (RPC failure): ``?``
1534 - Unknown field: ``??``
1536 Another user-visible change is the addition of ``--force-join`` to
1541 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1542 - Fixed regression in RAPI/instance reinstall where all parameters were
1543 required (instead of optional)
1544 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1545 - Fixed iallocator usage (offline nodes were not considered offline)
1546 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1547 - Fixed hypervisor and OS parameter validation with respect to
1548 non-vm_capable nodes
1549 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1551 - Fixed ``tools/listrunner`` with respect to agent-based usage
1557 *(Released Fri, 4 Feb 2011)*
1559 Many changes and fixes since the beta1 release. While there were some
1560 internal changes, the code has been mostly stabilised for the RC
1563 Note: the dumb allocator was removed in this release, as it was not kept
1564 up-to-date with the IAllocator protocol changes. It is recommended to
1565 use the ``hail`` command from the ganeti-htools package.
1567 Note: the 2.4 and up versions of Ganeti are not compatible with the
1568 0.2.x branch of ganeti-htools. You need to upgrade to
1569 ganeti-htools-0.3.0 (or later).
1571 Regressions fixed from 2.3
1572 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1574 - Fixed the ``gnt-cluster verify-disks`` command
1575 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1577 - Fixed disk adoption breakage
1578 - Fixed wrong headers in instance listing for field aliases
1583 - Fixed corner case in KVM handling of NICs
1584 - Fixed many cases of wrong handling of non-vm_capable nodes
1585 - Fixed a bug where a missing instance symlink was not possible to
1586 recreate with any ``gnt-*`` command (now ``gnt-instance
1587 activate-disks`` does it)
1588 - Fixed the volume group name as reported by ``gnt-cluster
1590 - Increased timeouts for the import-export code, hopefully leading to
1591 fewer aborts due network or instance timeouts
1592 - Fixed bug in ``gnt-node list-storage``
1593 - Fixed bug where not all daemons were started on cluster
1594 initialisation, but only at the first watcher run
1595 - Fixed many bugs in the OOB implementation
1596 - Fixed watcher behaviour in presence of instances with offline
1598 - Fixed instance list output for instances running on the wrong node
1599 - a few fixes to the cluster-merge tool, but it still cannot merge
1600 multi-node groups (currently it is not recommended to use this tool)
1606 - Improved network configuration for the KVM hypervisor
1607 - Added e1000 as a supported NIC for Xen-HVM
1608 - Improved the lvmstrap tool to also be able to use partitions, as
1609 opposed to full disks
1610 - Improved speed of disk wiping (the cluster parameter
1611 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1612 of instance creations
1613 - Added documentation for the OS parameters
1614 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1615 hypervisor is not responding
1616 - Added display of blacklisted and hidden OS information in
1617 ``gnt-cluster info``
1618 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1619 NIC and node parameters, which might create problems with currently
1620 invalid (but undetected) configuration files, but prevents validation
1621 failures when unrelated parameters are modified
1622 - Changed cluster initialisation to wait for the master daemon to become
1624 - Expanded the RAPI interface:
1626 - Added config redistribution resource
1627 - Added activation/deactivation of instance disks
1628 - Added export of console information
1630 - Implemented log file reopening on SIGHUP, which allows using
1631 logrotate(8) for the Ganeti log files
1632 - Added a basic OOB helper script as an example
1638 *(Released Fri, 14 Jan 2011)*
1643 - Fixed timezone issues when formatting timestamps
1644 - Added support for node groups, available via ``gnt-group`` and other
1646 - Added out-of-band framework and management, see :doc:`design
1647 document <design-oob>`
1648 - Removed support for roman numbers from ``gnt-node list`` and
1649 ``gnt-instance list``.
1650 - Allowed modification of master network interface via ``gnt-cluster
1651 modify --master-netdev``
1652 - Accept offline secondaries while shutting down instance disks
1653 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1654 - Added support for multiple LVM volume groups
1655 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1657 - Added commands to list available fields:
1659 - ``gnt-node list-fields``
1660 - ``gnt-group list-fields``
1661 - ``gnt-instance list-fields``
1663 - Updated documentation and man pages
1668 - Moved ``rapi_users`` file into separate directory, now named
1669 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1671 - Added new tool for running commands on many machines,
1672 ``tools/ganeti-listrunner``
1673 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1674 improving the ``gnt-instance console`` output
1675 - Allowed customisation of disk index separator at ``configure`` time
1676 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1677 - Added support for non-partitioned md disks in ``lvmstrap``
1678 - Added script to gracefully power off KVM instances
1679 - Split ``utils`` module into smaller parts
1680 - Changed query operations to return more detailed information, e.g.
1681 whether an information is unavailable due to an offline node. To use
1682 this new functionality, the LUXI call ``Query`` must be used. Field
1683 information is now stored by the master daemon and can be retrieved
1684 using ``QueryFields``. Instances, nodes and groups can also be queried
1685 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1686 exposed via RAPI). The following commands make use of this
1687 infrastructure change:
1689 - ``gnt-group list``
1690 - ``gnt-group list-fields``
1692 - ``gnt-node list-fields``
1693 - ``gnt-instance list``
1694 - ``gnt-instance list-fields``
1695 - ``gnt-debug locks``
1700 - New RAPI resources (see :doc:`rapi`):
1704 - ``/2/groups/[group_name]``
1705 - ``/2/groups/[group_name]/assign-nodes``
1706 - ``/2/groups/[group_name]/modify``
1707 - ``/2/groups/[group_name]/rename``
1708 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1712 - Implemented ``no_install`` for instance creation
1713 - Implemented OS parameters for instance reinstallation, allowing
1714 use of special settings on reinstallation (e.g. for preserving data)
1719 - Added IPv6 support in import/export
1720 - Pause DRBD synchronization while wiping disks on instance creation
1721 - Updated unittests and QA scripts
1722 - Improved network parameters passed to KVM
1723 - Converted man pages from docbook to reStructuredText
1729 *(Released Mon, 20 Dec 2010)*
1731 Released version 2.3.1~rc1 without any changes.
1737 *(Released Wed, 1 Dec 2010)*
1739 - impexpd: Disable OpenSSL compression in socat if possible (backport
1740 from master, commit e90739d625b, see :doc:`installation guide
1741 <install-quick>` for details)
1742 - Changed unittest coverage report to exclude test scripts
1743 - Added script to check version format
1749 *(Released Wed, 1 Dec 2010)*
1751 Released version 2.3.0~rc1 without any changes.
1757 *(Released Fri, 19 Nov 2010)*
1759 A number of bugfixes and documentation updates:
1761 - Update ganeti-os-interface documentation
1762 - Fixed a bug related to duplicate MACs or similar items which should be
1764 - Fix breakage in OS state modify
1765 - Reinstall instance: disallow offline secondaries (fixes bug related to
1766 OS changing but reinstall failing)
1767 - plus all the other fixes between 2.2.1 and 2.2.2
1773 *(Released Tue, 2 Nov 2010)*
1775 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1776 - Fixed master failover race with watcher
1777 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1779 - Accept previously stopped instance for export with instance removal
1780 - Simplify and extend the environment variables for instance OS scripts
1781 - Added new node flags, ``master_capable`` and ``vm_capable``
1782 - Added optional instance disk wiping prior during allocation. This is a
1783 cluster-wide option and can be set/modified using
1784 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1785 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1786 :doc:`install-quick`
1787 - Added a new watcher option (``--ignore-pause``)
1788 - Added option to ignore offline node on instance start/stop
1789 (``--ignore-offline``)
1790 - Allow overriding OS parameters with ``gnt-instance reinstall``
1791 - Added ability to change node's secondary IP address using ``gnt-node
1793 - Implemented privilege separation for all daemons except
1794 ``ganeti-noded``, see ``configure`` options
1795 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1797 - Implemented job priorities (see ``ganeti(7)`` manpage)
1798 - Ignore failures while shutting down instances during failover from
1800 - Exit daemon's bootstrap process only once daemon is ready
1801 - Export more information via ``LUInstanceQuery``/remote API
1802 - Improved documentation, QA and unittests
1803 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1804 restart if the file was created or changed
1805 - Added LUXI protocol version sent with each request and response,
1806 allowing detection of server/client mismatches
1807 - Moved the Python scripts among gnt-* and ganeti-* into modules
1808 - Moved all code related to setting up SSH to an external script,
1810 - Infrastructure changes for node group support in future versions
1816 *(Released Fri, 19 Nov 2010)*
1818 A few small bugs fixed, and some improvements to the build system:
1820 - Fix documentation regarding conversion to drbd
1821 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1823 - Fix error handling in node modify with multiple changes
1824 - Allow remote imports without checked names
1830 *(Released Tue, 19 Oct 2010)*
1832 - Disable SSL session ID cache in RPC client
1838 *(Released Thu, 14 Oct 2010)*
1840 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1841 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1843 - Fix problems with interaction between Curl and Python's HTTP server,
1844 resulting in increased speed in many RPC calls
1845 - Improve our release script to prevent breakage with older aclocal and
1852 *(Released Thu, 7 Oct 2010)*
1854 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1855 value retrieved from master
1856 - Added support for blacklisted or hidden OS definitions
1857 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1858 - Added support for -mem-path in KVM hypervisor abstraction layer
1859 - Allow overriding instance parameters in tool for inter-cluster
1860 instance moves (``tools/move-instance``)
1861 - Improved opcode summaries (e.g. in ``gnt-job list``)
1862 - Improve consistency of OS listing by sorting it
1863 - Documentation updates
1869 *(Released Fri, 8 Oct 2010)*
1871 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1877 *(Released Mon, 4 Oct 2010)*
1879 - Fixed regression in ``gnt-instance rename``
1885 *(Released Wed, 22 Sep 2010)*
1887 - Fixed OS_VARIANT variable for OS scripts
1888 - Fixed cluster tag operations via RAPI
1889 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1890 - Disabled RAPI CA checks in watcher
1896 *(Released Mon, 23 Aug 2010)*
1898 - Support DRBD versions of the format "a.b.c.d"
1900 - Re-introduce support for usage from multiple threads in RAPI client
1901 - Instance renames and modify via RAPI
1902 - Work around race condition between processing and archival in job
1904 - Mark opcodes following failed one as failed, too
1905 - Job field ``lock_status`` was removed due to difficulties making it
1906 work with the changed job queue in Ganeti 2.2; a better way to monitor
1907 locks is expected for a later 2.2.x release
1908 - Fixed dry-run behaviour with many commands
1909 - Support ``ssh-agent`` again when adding nodes
1910 - Many additional bugfixes
1916 *(Released Fri, 30 Jul 2010)*
1918 Important change: the internal RPC mechanism between Ganeti nodes has
1919 changed from using a home-grown http library (based on the Python base
1920 libraries) to use the PycURL library. This requires that PycURL is
1921 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1922 against GnuTLS by default. cURL's support for GnuTLS had known issues
1923 before cURL 7.21.0 and we recommend using the latest cURL release or
1924 linking against OpenSSL. Most other distributions already link PycURL
1925 and cURL against OpenSSL. The command::
1927 python -c 'import pycurl; print pycurl.version'
1929 can be used to determine the libraries PycURL and cURL are linked
1932 Other significant changes:
1934 - Rewrote much of the internals of the job queue, in order to achieve
1935 better parallelism; this decouples job query operations from the job
1936 processing, and it should allow much nicer behaviour of the master
1937 daemon under load, and it also has uncovered some long-standing bugs
1938 related to the job serialisation (now fixed)
1939 - Added a default iallocator setting to the cluster parameters,
1940 eliminating the need to always pass nodes or an iallocator for
1941 operations that require selection of new node(s)
1942 - Added experimental support for the LXC virtualization method
1943 - Added support for OS parameters, which allows the installation of
1944 instances to pass parameter to OS scripts in order to customise the
1946 - Added a hypervisor parameter controlling the migration type (live or
1947 non-live), since hypervisors have various levels of reliability; this
1948 has renamed the 'live' parameter to 'mode'
1949 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1950 logical volumes, meaning that cluster verify will ignore them and not
1951 flag their presence as errors
1952 - The watcher will now reset the error count for failed instances after
1953 8 hours, thus allowing self-healing if the problem that caused the
1954 instances to be down/fail to start has cleared in the meantime
1955 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1956 check for, and warn, if the drbd module parameter ``usermode_helper``
1957 is not consistent with the cluster-wide setting; this is needed to
1958 make diagnose easier of failed drbd creations
1959 - Started adding base IPv6 support, but this is not yet
1960 enabled/available for use
1961 - Rename operations (cluster, instance) will now return the new name,
1962 which is especially useful if a short name was passed in
1963 - Added support for instance migration in RAPI
1964 - Added a tool to pre-configure nodes for the SSH setup, before joining
1965 them to the cluster; this will allow in the future a simplified model
1966 for node joining (but not yet fully enabled in 2.2); this needs the
1967 paramiko python library
1968 - Fixed handling of name-resolving errors
1969 - Fixed consistency of job results on the error path
1970 - Fixed master-failover race condition when executed multiple times in
1972 - Fixed many bugs related to the job queue (mostly introduced during the
1973 2.2 development cycle, so not all are impacting 2.1)
1974 - Fixed instance migration with missing disk symlinks
1975 - Fixed handling of unknown jobs in ``gnt-job archive``
1976 - And many other small fixes/improvements
1980 - Enhanced both the unittest and the QA coverage
1981 - Switched the opcode validation to a generic model, and extended the
1982 validation to all opcode parameters
1983 - Changed more parts of the code that write shell scripts to use the
1985 - Switched the master daemon to use the asyncore library for the Luxi
1992 *(Released Thu, 17 Jun 2010)*
1994 - Added tool (``move-instance``) and infrastructure to move instances
1995 between separate clusters (see :doc:`separate documentation
1996 <move-instance>` and :doc:`design document <design-2.2>`)
1997 - Added per-request RPC timeout
1998 - RAPI now requires a Content-Type header for requests with a body (e.g.
1999 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2000 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2001 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2003 - Implemented initial support for running Ganeti daemons as separate
2004 users, see configure-time flags ``--with-user-prefix`` and
2005 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2007 - Instances can be removed after export (``gnt-backup export
2008 --remove-instance``)
2009 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2010 key (instead of 1024 bit)
2011 - Added new cluster configuration file for cluster domain secret
2012 - Import/export now use SSL instead of SSH
2013 - Added support for showing estimated time when exporting an instance,
2014 see the ``ganeti-os-interface(7)`` manpage and look for
2021 *(Released Tue, 16 Nov 2010)*
2023 Some more bugfixes. Unless critical bugs occur, this will be the last
2026 - Fix case of MAC special-values
2027 - Fix mac checker regex
2028 - backend: Fix typo causing "out of range" error
2029 - Add missing --units in gnt-instance list man page
2035 *(Released Tue, 24 Aug 2010)*
2038 - Don't ignore secondary node silently on non-mirrored disk templates
2040 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2041 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2042 - Properly document the usb_mouse parameter
2043 - Fix path in ganeti-rapi(8) (issue 116)
2044 - Adjust error message when the ganeti user's .ssh directory is
2046 - Add same-node-check when changing the disk template to drbd
2052 *(Released Fri, 16 Jul 2010)*
2055 - Add an option to only select some reboot types during qa/burnin.
2056 (on some hypervisors consequent reboots are not supported)
2057 - Fix infrequent race condition in master failover. Sometimes the old
2058 master ip address would be still detected as up for a short time
2059 after it was removed, causing failover to fail.
2060 - Decrease mlockall warnings when the ctypes module is missing. On
2061 Python 2.4 we support running even if no ctypes module is installed,
2062 but we were too verbose about this issue.
2063 - Fix building on old distributions, on which man doesn't have a
2065 - Fix RAPI not to ignore the MAC address on instance creation
2066 - Implement the old instance creation format in the RAPI client.
2072 *(Released Thu, 01 Jul 2010)*
2074 A small bugfix release:
2075 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2076 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2077 a non-existing option
2078 - Fix instance create: the --force-variant option was ignored
2079 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2080 - Fix modify node storage with non-FQDN arguments
2081 - Fix RAPI client to authenticate under Python 2.6 when used
2082 for more than 5 requests needing authentication
2083 - Fix gnt-instance modify -t (storage) giving a wrong error message
2084 when converting a non-shutdown drbd instance to plain
2090 *(Released Fri, 18 Jun 2010)*
2092 A small bugfix release:
2094 - Fix live migration of KVM instances started with older Ganeti
2095 versions which had fewer hypervisor parameters
2096 - Fix gnt-instance grow-disk on down instances
2097 - Fix an error-reporting bug during instance migration
2098 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2099 silently ignoring broken ones
2100 - Fix an RPC error reporting bug affecting, for example, RAPI client
2102 - Fix bug triggered by different API version os-es on different nodes
2103 - Fix a bug in instance startup with custom hvparams: OS level
2104 parameters would fail to be applied.
2105 - Fix the RAPI client under Python 2.6 (but more work is needed to
2106 make it work completely well with OpenSSL)
2107 - Fix handling of errors when resolving names from DNS
2113 *(Released Thu, 3 Jun 2010)*
2115 A medium sized development cycle. Some new features, and some
2116 fixes/small improvements/cleanups.
2118 Significant features
2119 ~~~~~~~~~~~~~~~~~~~~
2121 The node deamon now tries to mlock itself into memory, unless the
2122 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2123 its logs, and falls back to console logging. This allows emergency
2124 features such as ``gnt-node powercycle`` to work even in the event of a
2125 broken node disk (tested offlining the disk hosting the node's
2126 filesystem and dropping its memory caches; don't try this at home)
2128 KVM: add vhost-net acceleration support. It can be tested with a new
2129 enough version of the kernel and of qemu-kvm.
2131 KVM: Add instance chrooting feature. If you use privilege dropping for
2132 your VMs you can also now force them to chroot to an empty directory,
2133 before starting the emulated guest.
2135 KVM: Add maximum migration bandwith and maximum downtime tweaking
2136 support (requires a new-enough version of qemu-kvm).
2138 Cluster verify will now warn if the master node doesn't have the master
2139 ip configured on it.
2141 Add a new (incompatible) instance creation request format to RAPI which
2142 supports all parameters (previously only a subset was supported, and it
2143 wasn't possible to extend the old format to accomodate all the new
2144 features. The old format is still supported, and a client can check for
2145 this feature, before using it, by checking for its presence in the
2146 ``features`` RAPI resource.
2148 Now with ancient latin support. Try it passing the ``--roman`` option to
2149 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2150 (requires the python-roman module to be installed, in order to work).
2155 As usual many internal code refactorings, documentation updates, and
2158 - Lots of improvements and cleanups to the experimental Remote API
2159 (RAPI) client library.
2160 - A new unit test suite for the core daemon libraries.
2161 - A fix to creating missing directories makes sure the umask is not
2162 applied anymore. This enforces the same directory permissions
2164 - Better handling terminating daemons with ctrl+c (used when running
2165 them in debugging mode).
2166 - Fix a race condition in live migrating a KVM instance, when stat()
2167 on the old proc status file returned EINVAL, which is an unexpected
2169 - Fixed manpage checking with newer man and utf-8 charachters. But now
2170 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2176 *(Released Fri, 7 May 2010)*
2178 Fix a bug which prevented untagged KVM instances from starting.
2184 *(Released Fri, 7 May 2010)*
2186 Another release with a long development cycle, during which many
2187 different features were added.
2189 Significant features
2190 ~~~~~~~~~~~~~~~~~~~~
2192 The KVM hypervisor now can run the individual instances as non-root, to
2193 reduce the impact of a VM being hijacked due to bugs in the
2194 hypervisor. It is possible to run all instances as a single (non-root)
2195 user, to manually specify a user for each instance, or to dynamically
2196 allocate a user out of a cluster-wide pool to each instance, with the
2197 guarantee that no two instances will run under the same user ID on any
2200 An experimental RAPI client library, that can be used standalone
2201 (without the other Ganeti libraries), is provided in the source tree as
2202 ``lib/rapi/client.py``. Note this client might change its interface in
2203 the future, as we iterate on its capabilities.
2205 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2206 replace the cluster's certificates and crypto keys. This might help in
2207 case they have been compromised, or have simply expired.
2209 A new disk option for instance creation has been added that allows one
2210 to "adopt" currently existing logical volumes, with data
2211 preservation. This should allow easier migration to Ganeti from
2212 unmanaged (or managed via other software) instances.
2214 Another disk improvement is the possibility to convert between redundant
2215 (DRBD) and plain (LVM) disk configuration for an instance. This should
2216 allow better scalability (starting with one node and growing the
2217 cluster, or shrinking a two-node cluster to one node).
2219 A new feature that could help with automated node failovers has been
2220 implemented: if a node sees itself as offline (by querying the master
2221 candidates), it will try to shutdown (hard) all instances and any active
2222 DRBD devices. This reduces the risk of duplicate instances if an
2223 external script automatically failovers the instances on such nodes. To
2224 enable this, the cluster parameter ``maintain_node_health`` should be
2225 enabled; in the future this option (per the name) will enable other
2226 automatic maintenance features.
2228 Instance export/import now will reuse the original instance
2229 specifications for all parameters; that means exporting an instance,
2230 deleting it and the importing it back should give an almost identical
2231 instance. Note that the default import behaviour has changed from
2232 before, where it created only one NIC; now it recreates the original
2235 Cluster verify has added a few new checks: SSL certificates validity,
2236 /etc/hosts consistency across the cluster, etc.
2241 As usual, many internal changes were done, documentation fixes,
2244 - Fixed cluster initialization with disabled cluster storage (regression
2245 introduced in 2.1.1)
2246 - File-based storage supports growing the disks
2247 - Fixed behaviour of node role changes
2248 - Fixed cluster verify for some corner cases, plus a general rewrite of
2249 cluster verify to allow future extension with more checks
2250 - Fixed log spamming by watcher and node daemon (regression introduced
2252 - Fixed possible validation issues when changing the list of enabled
2254 - Fixed cleanup of /etc/hosts during node removal
2255 - Fixed RAPI response for invalid methods
2256 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2257 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2258 from ide disks, etc.)
2259 - Allow OS changes without re-installation (to record a changed OS
2260 outside of Ganeti, or to allow OS renames)
2261 - Allow instance creation without OS installation (useful for example if
2262 the OS will be installed manually, or restored from a backup not in
2264 - Implemented option to make cluster ``copyfile`` use the replication
2266 - Added list of enabled hypervisors to ssconf (possibly useful for
2268 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2270 - A partial form of node re-IP is possible via node readd, which now
2271 allows changed node primary IP
2272 - Command line utilities now show an informational message if the job is
2274 - The logs of the master daemon now show the PID/UID/GID of the
2281 *(Released Fri, 12 Mar 2010)*
2283 During the 2.1.0 long release candidate cycle, a lot of improvements and
2284 changes have accumulated with were released later as 2.1.1.
2289 The node evacuate command (``gnt-node evacuate``) was significantly
2290 rewritten, and as such the IAllocator protocol was changed - a new
2291 request type has been added. This unfortunate change during a stable
2292 series is designed to improve performance of node evacuations; on
2293 clusters with more than about five nodes and which are well-balanced,
2294 evacuation should proceed in parallel for all instances of the node
2295 being evacuated. As such, any existing IAllocator scripts need to be
2296 updated, otherwise the above command will fail due to the unknown
2297 request. The provided "dumb" allocator has not been updated; but the
2298 ganeti-htools package supports the new protocol since version 0.2.4.
2300 Another important change is increased validation of node and instance
2301 names. This might create problems in special cases, if invalid host
2302 names are being used.
2304 Also, a new layer of hypervisor parameters has been added, that sits at
2305 OS level between the cluster defaults and the instance ones. This allows
2306 customisation of virtualization parameters depending on the installed
2307 OS. For example instances with OS 'X' may have a different KVM kernel
2308 (or any other parameter) than the cluster defaults. This is intended to
2309 help managing a multiple OSes on the same cluster, without manual
2310 modification of each instance's parameters.
2312 A tool for merging clusters, ``cluster-merge``, has been added in the
2313 tools sub-directory.
2318 - Improved the int/float conversions that should make the code more
2319 robust in face of errors from the node daemons
2320 - Fixed the remove node code in case of internal configuration errors
2321 - Fixed the node daemon behaviour in face of inconsistent queue
2322 directory (e.g. read-only file-system where we can't open the files
2324 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2325 now it either aborts cleanly or, if given the new "auto_promote"
2326 parameter, will automatically promote other nodes as needed
2327 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2328 completely prevent Ganeti from working
2329 - Fixed bug for instance export when not all disks were successfully
2331 - Fixed behaviour of node add when the new node is slow in starting up
2333 - Fixed handling of signals in the LUXI client, which should improve
2334 behaviour of command-line scripts
2335 - Added checks for invalid node/instance names in the configuration (now
2336 flagged during cluster verify)
2337 - Fixed watcher behaviour for disk activation errors
2338 - Fixed two potentially endless loops in http library, which led to the
2339 RAPI daemon hanging and consuming 100% CPU in some cases
2340 - Fixed bug in RAPI daemon related to hashed passwords
2341 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2343 - Enhanced compatibility with non-Debian OSes, but not using absolute
2344 path in some commands and allowing customisation of the ssh
2345 configuration directory
2346 - Fixed possible future issue with new Python versions by abiding to the
2347 proper use of ``__slots__`` attribute on classes
2348 - Added checks that should prevent directory traversal attacks
2349 - Many documentation fixes based on feedback from users
2354 - Added an "early_release" more for instance replace disks and node
2355 evacuate, where we release locks earlier and thus allow higher
2356 parallelism within the cluster
2357 - Added watcher hooks, intended to allow the watcher to restart other
2358 daemons (e.g. from the ganeti-nbma project), but they can be used of
2359 course for any other purpose
2360 - Added a compile-time disable for DRBD barriers, to increase
2361 performance if the administrator trusts the power supply or the
2362 storage system to not lose writes
2363 - Added the option of using syslog for logging instead of, or in
2364 addition to, Ganeti's own log files
2365 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2366 can indeed boot from a paravirtual NIC
2367 - Added a generic debug level for many operations; while this is not
2368 used widely yet, it allows one to pass the debug value all the way to
2370 - Enhanced the hooks environment for instance moves (failovers,
2371 migrations) where the primary/secondary nodes changed during the
2372 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2373 - Enhanced data validations for many user-supplied values; one important
2374 item is the restrictions imposed on instance and node names, which
2375 might reject some (invalid) host names
2376 - Add a configure-time option to disable file-based storage, if it's not
2377 needed; this allows greater security separation between the master
2378 node and the other nodes from the point of view of the inter-node RPC
2380 - Added user notification in interactive tools if job is waiting in the
2381 job queue or trying to acquire locks
2382 - Added log messages when a job is waiting for locks
2383 - Added filtering by node tags in instance operations which admit
2384 multiple instances (start, stop, reboot, reinstall)
2385 - Added a new tool for cluster mergers, ``cluster-merge``
2386 - Parameters from command line which are of the form ``a=b,c=d`` can now
2387 use backslash escapes to pass in values which contain commas,
2388 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2390 - For KVM, the instance name is the first parameter passed to KVM, so
2391 that it's more visible in the process list
2397 *(Released Tue, 2 Mar 2010)*
2399 Ganeti 2.1 brings many improvements with it. Major changes:
2401 - Added infrastructure to ease automated disk repairs
2402 - Added new daemon to export configuration data in a cheaper way than
2403 using the remote API
2404 - Instance NICs can now be routed instead of being associated with a
2406 - Improved job locking logic to reduce impact of jobs acquiring multiple
2407 locks waiting for other long-running jobs
2409 In-depth implementation details can be found in the Ganeti 2.1 design
2415 - Added chroot hypervisor
2416 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2418 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2419 ``bootloader_path`` and ``bootloader_args``)
2420 - Added the ``use_localtime`` option for the xen-hvm and kvm
2421 hypervisors, and the default value for this has changed to false (in
2422 2.0 xen-hvm always enabled it)
2423 - Added luxi call to submit multiple jobs in one go
2424 - Added cluster initialization option to not modify ``/etc/hosts``
2426 - Added network interface parameters
2427 - Added dry run mode to some LUs
2428 - Added RAPI resources:
2430 - ``/2/instances/[instance_name]/info``
2431 - ``/2/instances/[instance_name]/replace-disks``
2432 - ``/2/nodes/[node_name]/evacuate``
2433 - ``/2/nodes/[node_name]/migrate``
2434 - ``/2/nodes/[node_name]/role``
2435 - ``/2/nodes/[node_name]/storage``
2436 - ``/2/nodes/[node_name]/storage/modify``
2437 - ``/2/nodes/[node_name]/storage/repair``
2439 - Added OpCodes to evacuate or migrate all instances on a node
2440 - Added new command to list storage elements on nodes (``gnt-node
2441 list-storage``) and modify them (``gnt-node modify-storage``)
2442 - Added new ssconf files with master candidate IP address
2443 (``ssconf_master_candidates_ips``), node primary IP address
2444 (``ssconf_node_primary_ips``) and node secondary IP address
2445 (``ssconf_node_secondary_ips``)
2446 - Added ``ganeti-confd`` and a client library to query the Ganeti
2447 configuration via UDP
2448 - Added ability to run hooks after cluster initialization and before
2450 - Added automatic mode for disk replace (``gnt-instance replace-disks
2452 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2453 after catastrophic data-loss
2454 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2456 - Added ``gnt-instance move`` command to move instances
2457 - Added ``gnt-cluster watcher`` command to control watcher
2458 - Added ``gnt-node powercycle`` command to powercycle nodes
2459 - Added new job status field ``lock_status``
2460 - Added parseable error codes to cluster verification (``gnt-cluster
2461 verify --error-codes``) and made output less verbose (use
2462 ``--verbose`` to restore previous behaviour)
2463 - Added UUIDs to the main config entities (cluster, nodes, instances)
2464 - Added support for OS variants
2465 - Added support for hashed passwords in the Ganeti remote API users file
2467 - Added option to specify maximum timeout on instance shutdown
2468 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2469 - Added new helper script to start and stop Ganeti daemons
2470 (``daemon-util``), with the intent to reduce the work necessary to
2471 adjust Ganeti for non-Debian distributions and to start/stop daemons
2473 - Added more unittests
2474 - Fixed critical bug in ganeti-masterd startup
2475 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2476 now customisable at the cluster level for both the KVM and Xen
2477 hypervisors using the new ``migration_port`` parameter
2478 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2479 reinstalling an instance
2480 - Allowed ``@`` in tag names
2481 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2482 - Many documentation updates
2483 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2484 - ``gnt-instance reinstall`` can now reinstall multiple instances
2485 - Updated many command line parameters
2486 - Introduced new OS API version 15
2487 - No longer support a default hypervisor
2488 - Treat virtual LVs as inexistent
2489 - Improved job locking logic to reduce lock contention
2490 - Match instance and node names case insensitively
2491 - Reimplemented bash completion script to be more complete
2498 *(Released Thu, 4 Feb 2010)*
2500 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2501 - Fix a string formatting bug
2502 - Improve safety of the code in some error paths
2503 - Improve data validation in the master of values returned from nodes
2509 *(Released Thu, 17 Dec 2009)*
2511 - Fix security issue due to missing validation of iallocator names; this
2512 allows local and remote execution of arbitrary executables
2513 - Fix failure of gnt-node list during instance removal
2514 - Ship the RAPI documentation in the archive
2520 *(Released Wed, 30 Sep 2009)*
2522 - Fixed many wrong messages
2523 - Fixed a few bugs related to the locking library
2524 - Fixed MAC checking at instance creation time
2525 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2526 - Fixed a few issues related to signal handling in both daemons and
2528 - Fixed the example startup script provided
2529 - Fixed insserv dependencies in the example startup script (patch from
2531 - Fixed handling of drained nodes in the iallocator framework
2532 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2534 - Fixed error related to invalid job IDs in job polling
2535 - Fixed job/opcode persistence on unclean master shutdown
2536 - Fixed handling of partial job processing after unclean master
2538 - Fixed error reporting from LUs, previously all errors were converted
2539 into execution errors
2540 - Fixed error reporting from burnin
2541 - Decreased significantly the memory usage of the job queue
2542 - Optimised slightly multi-job submission
2543 - Optimised slightly opcode loading
2544 - Backported the multi-job submit framework from the development
2545 branch; multi-instance start and stop should be faster
2546 - Added script to clean archived jobs after 21 days; this will reduce
2547 the size of the queue directory
2548 - Added some extra checks in disk size tracking
2549 - Added an example ethers hook script
2550 - Added a cluster parameter that prevents Ganeti from modifying of
2552 - Added more node information to RAPI responses
2553 - Added a ``gnt-job watch`` command that allows following the ouput of a
2555 - Added a bind-address option to ganeti-rapi
2556 - Added more checks to the configuration verify
2557 - Enhanced the burnin script such that some operations can be retried
2559 - Converted instance reinstall to multi-instance model
2565 *(Released Fri, 7 Aug 2009)*
2567 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2568 to allow using the pre-2.0.2 behaviour in activation, if any existing
2569 instances have mismatched disk sizes in the configuration
2570 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2571 any configuration mismatches for disk sizes
2572 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2573 failover to work on two-node clusters
2574 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2576 - Fixed detection of OS script errors in ``gnt-backup export``
2577 - Fixed exit code of ``gnt-backup export``
2583 *(Released Fri, 17 Jul 2009)*
2585 - Added experimental support for stripped logical volumes; this should
2586 enhance performance but comes with a higher complexity in the block
2587 device handling; stripping is only enabled when passing
2588 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2589 affected even in the non-stripped mode
2590 - Improved resiliency against transient failures at the end of DRBD
2591 resyncs, and in general of DRBD resync checks
2592 - Fixed a couple of issues with exports and snapshot errors
2593 - Fixed a couple of issues in instance listing
2594 - Added display of the disk size in ``gnt-instance info``
2595 - Fixed checking for valid OSes in instance creation
2596 - Fixed handling of the "vcpus" parameter in instance listing and in
2597 general of invalid parameters
2598 - Fixed http server library, and thus RAPI, to handle invalid
2599 username/password combinations correctly; this means that now they
2600 report unauthorized for queries too, not only for modifications,
2601 allowing earlier detect of configuration problems
2602 - Added a new "role" node list field, equivalent to the master/master
2603 candidate/drained/offline flags combinations
2604 - Fixed cluster modify and changes of candidate pool size
2605 - Fixed cluster verify error messages for wrong files on regular nodes
2606 - Fixed a couple of issues with node demotion from master candidate role
2607 - Fixed node readd issues
2608 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2609 - Added a new ``--no-voting`` option for masterfailover to fix failover
2610 on two-nodes clusters when the former master node is unreachable
2611 - Added instance reinstall over RAPI
2617 *(Released Tue, 16 Jun 2009)*
2619 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2620 allow re-adding the start in single-user option (regression from 1.2)
2621 - the watcher writes the instance status to a file, to allow monitoring
2622 to report the instance status (from the master) based on cached
2623 results of the watcher's queries; while this can get stale if the
2624 watcher is being locked due to other work on the cluster, this is
2625 still an improvement
2626 - the watcher now also restarts the node daemon and the rapi daemon if
2628 - fixed the watcher to handle full and drained queue cases
2629 - hooks export more instance data in the environment, which helps if
2630 hook scripts need to take action based on the instance's properties
2631 (no longer need to query back into ganeti)
2632 - instance failovers when the instance is stopped do not check for free
2633 RAM, so that failing over a stopped instance is possible in low memory
2635 - rapi uses queries for tags instead of jobs (for less job traffic), and
2636 for cluster tags it won't talk to masterd at all but read them from
2638 - a couple of error handling fixes in RAPI
2639 - drbd handling: improved the error handling of inconsistent disks after
2640 resync to reduce the frequency of "there are some degraded disks for
2641 this instance" messages
2642 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2643 error handling path called a wrong function name)
2649 *(Released Wed, 27 May 2009)*
2651 - no changes from rc5
2657 *(Released Wed, 20 May 2009)*
2659 - fix a couple of bugs (validation, argument checks)
2660 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2661 - some small improvements to RAPI and IAllocator
2662 - make watcher automatically start the master daemon if down
2668 *(Released Mon, 27 Apr 2009)*
2670 - change the OS list to not require locks; this helps with big clusters
2671 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2672 volume group is broken
2673 - ``gnt-instance info``, without any arguments, doesn't run for all
2674 instances anymore; either pass ``--all`` or pass the desired
2675 instances; this helps against mistakes on big clusters where listing
2676 the information for all instances takes a long time
2677 - miscellaneous doc and man pages fixes
2683 *(Released Wed, 8 Apr 2009)*
2685 - Change the internal locking model of some ``gnt-node`` commands, in
2686 order to reduce contention (and blocking of master daemon) when
2687 batching many creation/reinstall jobs
2688 - Fixes to Xen soft reboot
2689 - No longer build documentation at build time, instead distribute it in
2690 the archive, in order to reduce the need for the whole docbook/rst
2697 *(Released Fri, 27 Mar 2009)*
2699 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2700 - Fix watcher startup sequence, improves the behaviour of busy clusters
2701 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2702 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2703 bind address and other places
2704 - Some documentation fixes and updates
2710 *(Released Mon, 2 Mar 2009)*
2712 - More documentation updates, now all docs should be more-or-less
2714 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2716 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2718 - Fix an issue related to $libdir/run/ganeti and cluster creation
2724 *(Released Thu, 19 Feb 2009)*
2726 - Xen PVM and KVM have switched the default value for the instance root
2727 disk to the first partition on the first drive, instead of the whole
2728 drive; this means that the OS installation scripts must be changed
2730 - Man pages have been updated
2731 - RAPI has been switched by default to HTTPS, and the exported functions
2732 should all work correctly
2733 - RAPI v1 has been removed
2734 - Many improvements to the KVM hypervisor
2735 - Block device errors are now better reported
2736 - Many other bugfixes and small improvements
2742 *(Released Mon, 26 Jan 2009)*
2744 - Version 2 is a general rewrite of the code and therefore the
2745 differences are too many to list, see the design document for 2.0 in
2746 the ``doc/`` subdirectory for more details
2747 - In this beta version there is not yet a migration path from 1.2 (there
2748 will be one in the final 2.0 release)
2749 - A few significant changes are:
2751 - all commands are executed by a daemon (``ganeti-masterd``) and the
2752 various ``gnt-*`` commands are just front-ends to it
2753 - all the commands are entered into, and executed from a job queue,
2754 see the ``gnt-job(8)`` manpage
2755 - the RAPI daemon supports read-write operations, secured by basic
2756 HTTP authentication on top of HTTPS
2757 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2758 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2759 to migrate to DRBD 8 first while still running Ganeti 1.2)
2760 - DRBD devices are using statically allocated minor numbers, which
2761 will be assigned to existing instances during the migration process
2762 - there is support for both Xen PVM and Xen HVM instances running on
2764 - KVM virtualization is supported too
2765 - file-based storage has been implemented, which means that it is
2766 possible to run the cluster without LVM and DRBD storage, for
2767 example using a shared filesystem exported from shared storage (and
2768 still have live migration)
2774 *(Released Tue, 13 Jan 2009)*
2776 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2777 - Reuse the old instance mac address by default on instance import, if
2778 the instance name is the same.
2779 - Handle situations in which the node info rpc returns incomplete
2781 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2782 - Improved version of batcher:
2784 - state file support
2785 - instance mac address support
2786 - support for HVM clusters/instances
2788 - Add an option to show the number of cpu sockets and nodes in
2790 - Support OSes that handle more than one version of the OS api (but do
2791 not change the current API in any other way)
2792 - Fix ``gnt-node migrate``
2793 - ``gnt-debug`` man page
2794 - Fixes various more typos and small issues
2795 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2801 *(Released Wed, 24 Sep 2008)*
2803 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2804 type of disk exported to fully virtualized instances.
2805 - provide access to the serial console of HVM instances
2806 - instance auto_balance flag, set by default. If turned off it will
2807 avoid warnings on cluster verify if there is not enough memory to fail
2808 over an instance. in the future it will prevent automatically failing
2809 it over when we will support that.
2810 - batcher tool for instance creation, see ``tools/README.batcher``
2811 - ``gnt-instance reinstall --select-os`` to interactively select a new
2812 operating system when reinstalling an instance.
2813 - when changing the memory amount on instance modify a check has been
2814 added that the instance will be able to start. also warnings are
2815 emitted if the instance will not be able to fail over, if auto_balance
2817 - documentation fixes
2818 - sync fields between ``gnt-instance list/modify/add/import``
2819 - fix a race condition in drbd when the sync speed was set after giving
2820 the device a remote peer.
2826 *(Released Tue, 22 Jul 2008)*
2828 - note: the allowed size and number of tags per object were reduced
2829 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2830 - fixed twisted 8.x compatibility
2831 - fixed ``gnt-instance replace-disks`` with iallocator
2832 - add TCP keepalives on twisted connections to detect restarted nodes
2833 - disk increase support, see ``gnt-instance grow-disk``
2834 - implement bulk node/instance query for RAPI
2835 - add tags in node/instance listing (optional)
2836 - experimental migration (and live migration) support, read the man page
2837 for ``gnt-instance migrate``
2838 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2839 has some small improvements in handling its state file
2845 *(Released Fri, 13 Jun 2008)*
2847 - Experimental readonly, REST-based remote API implementation;
2848 automatically started on master node, TCP port 5080, if enabled by
2849 ``--enable-rapi`` parameter to configure script.
2850 - Instance allocator support. Add and import instance accept a
2851 ``--iallocator`` parameter, and call that instance allocator to decide
2852 which node to use for the instance. The iallocator document describes
2853 what's expected from an allocator script.
2854 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2855 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2856 node is lost there is still enough memory to fail over the instances
2858 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2859 ``gnt-cluster verify``, to check for site-specific compliance. All the
2860 hooks will run, and their output, if any, will be displayed. Any
2861 failing hook will make the verification return an error value.
2862 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2863 primary and secondary interfaces
2864 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2865 that is still declared as part of the cluster and has failed.
2866 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2867 specifying output fields, that just adds the chosen fields to the
2869 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2870 export from the filesystem.
2871 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2872 have been added. Using them you can enable/disable acpi and pae
2873 support, and specify a path for a cd image to be exported to the
2874 instance. These parameters as the name suggest only work on HVM
2876 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2877 PAE support will be set to the previously hardcoded values, but the
2878 (previously hardcoded) path to the CDROM ISO image will be unset and
2879 if required, needs to be set manually with ``gnt-instance modify``
2881 - The address to which an instance's VNC console is bound is now
2882 selectable per-instance, rather than being cluster wide. Of course
2883 this only applies to instances controlled via VNC, so currently just
2884 applies to HVM clusters.
2890 *(Released Mon, 18 Feb 2008)*
2892 - more tweaks to the disk activation code (especially helpful for DRBD)
2893 - change the default ``gnt-instance list`` output format, now there is
2894 one combined status field (see the manpage for the exact values this
2896 - some more fixes for the mac export to hooks change
2897 - make Ganeti not break with DRBD 8.2.x (which changed the version
2898 format in ``/proc/drbd``) (issue 24)
2899 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2900 template, allowing migration from DRBD0.7+MD to DRBD8
2906 *(Released Wed, 30 Jan 2008)*
2908 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2910 - add command aliases infrastructure and a few aliases
2911 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2912 man pages and the ``--help`` option of ``gnt-node
2913 list``/``gnt-instance list``
2914 - fix ``gnt-backup list`` with down nodes (issue 21)
2915 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2916 - fix the dist archive and add a check for including svn/git files in
2918 - some developer-related changes: improve the burnin and the QA suite,
2919 add an upload script for testing during development
2925 *(Released Wed, 16 Jan 2008)*
2927 - experimental HVM support, read the install document, section
2928 "Initializing the cluster"
2929 - allow for the PVM hypervisor per-instance kernel and initrd paths
2930 - add a new command ``gnt-cluster verify-disks`` which uses a new
2931 algorithm to improve the reconnection of the DRBD pairs if the device
2932 on the secondary node has gone away
2933 - make logical volume code auto-activate LVs at disk activation time
2934 - slightly improve the speed of activating disks
2935 - allow specification of the MAC address at instance creation time, and
2936 changing it later via ``gnt-instance modify``
2937 - fix handling of external commands that generate lots of output on
2939 - update documentation with regard to minimum version of DRBD8 supported
2945 *(Released Tue, 4 Dec 2007)*
2947 - Log the ``xm create`` output to the node daemon log on failure (to
2948 help diagnosing the error)
2949 - In debug mode, log all external commands output if failed to the logs
2950 - Change parsing of lvm commands to ignore stderr
2956 *(Released Wed, 28 Nov 2007)*
2958 - Another round of updates to the DRBD 8 code to deal with more failures
2959 in the replace secondary node operation
2960 - Some more logging of failures in disk operations (lvm, drbd)
2961 - A few documentation updates
2968 *(Released Tue, 13 Nov 2007)*
2970 - Change configuration file format from Python's Pickle to JSON.
2971 Upgrading is possible using the cfgupgrade utility.
2972 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2973 faster replace disks and is more stable (DRBD 8 has many improvements
2974 compared to DRBD 0.7)
2975 - Added command line tags support (see man pages for ``gnt-instance``,
2976 ``gnt-node``, ``gnt-cluster``)
2977 - Added instance rename support
2978 - Added multi-instance startup/shutdown
2979 - Added cluster rename support
2980 - Added ``gnt-node evacuate`` to simplify some node operations
2981 - Added instance reboot operation that can speedup reboot as compared to
2983 - Soften the requirement that hostnames are in FQDN format
2984 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2986 - Removed dependency on debian's patched fping that uses the
2987 non-standard ``-S`` option
2988 - Now the OS definitions are searched for in multiple, configurable
2989 paths (easier for distros to package)
2990 - Some changes to the hooks infrastructure (especially the new
2991 post-configuration update hook)
2992 - Other small bugfixes
2994 .. vim: set textwidth=72 syntax=rst :