8 *(Released Wed, 9 Oct 2013)*
10 Incompatible/important changes
11 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
13 - hroller now also plans for capacity to move non-redundant instances off
14 any node to be rebooted; the old behavior of completely ignoring any
15 non-redundant instances can be restored by adding the --ignore-non-redundant
17 - The cluster option '--no-lvm-storage' was removed in favor of the new option
18 '--enabled-disk-templates'.
19 - On instance creation, disk templates no longer need to be specified
20 with '-t'. The default disk template will be taken from the list of
21 enabled disk templates.
22 - The monitoring daemon is now running as root, in order to be able to collect
23 information only available to root (such as the state of Xen instances).
24 - The ConfD client is now IPv6 compatible.
25 - File and shared file storage is no longer dis/enabled at configure time,
26 but using the option '--enabled-disk-templates' at cluster initialization and
28 - The default directories for file and shared file storage are not anymore
29 specified at configure time, but taken from the cluster's configuration.
30 They can be set at cluster initialization and modification with
31 '--file-storage-dir' and '--shared-file-storage-dir'.
32 - Cluster verification now includes stricter checks regarding the
33 default file and shared file storage directories. It now checks that
34 the directories are explicitely allowed in the 'file-storage-paths' file and
35 that the directories exist on all nodes.
36 - The list of allowed disk templates in the instance policy and the list
37 of cluster-wide enabled disk templates is now checked for consistency
38 on cluster or group modification. On cluster initialization, the ipolicy
39 disk templates are ensured to be a subset of the cluster-wide enabled
45 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
46 the correct command syntax. It is possible to use different DRBD versions
47 on different nodes as long as they are compatible to each other. This
48 enables rolling upgrades of DRBD with no downtime. As permanent operation
49 of different DRBD versions within a node group is discouraged,
50 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
51 - New "inst-status-xen" data collector for the monitoring daemon, providing
52 information about the state of the xen instances on the nodes.
53 - New "lv" data collector for the monitoring daemon, collecting data about the
54 logical volumes on the nodes, and pairing them with the name of the instances
56 - New "diskstats" data collector, collecting the data from /proc/diskstats and
57 presenting them over the monitoring daemon interface.
58 - The ConfD client is now IPv6 compatible.
62 The following new dependencies have been added.
66 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
67 for the unit tests (and only used for testing).
71 - ``hslogger`` (http://software.complete.org/hslogger) is now always
72 required, even if confd is not enabled.
77 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
78 at least one DRBD instance.
79 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
86 *(Released Tue, 1 Oct 2013)*
88 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
91 - update of the documentation, in particular installation instructions
92 - merging of LD_* constants into DT_* constants
93 - python style changes to be compatible with newer versions of pylint
99 *(Released Thu, 29 Aug 2013)*
101 This was the first beta release of the 2.9 series. All important changes
102 are listed in the latest 2.9 entry.
108 *(Released Mon, 30 Sep 2013)*
110 Incompatible/important changes
111 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
113 - Instance policy can contain multiple instance specs, as described in
114 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
115 <design-partitioned>`. As a consequence, it's not possible to partially change
116 or override instance specs. Bounding specs (min and max) can be specified as a
117 whole using the new option ``--ipolicy-bounds-specs``, while standard
118 specs use the new option ``--ipolicy-std-specs``.
119 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
120 gnt-instance is a valid YAML object.
121 - hail now honors network restrictions when allocating nodes. This led to an
122 update of the IAllocator protocol. See the IAllocator documentation for
124 - confd now only answers static configuration request over the network. luxid
125 was extracted, listens on the local LUXI socket and responds to live queries.
126 This allows finer grained permissions if using separate users.
131 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
132 to always require authentication, ``--require-authentication``. It can
133 be specified in ``$sysconfdir/default/ganeti``.
134 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
135 be used to manage the disk templates to be used by instances in the cluster.
136 Initially, it will be set to a list that includes plain, drbd, if they were
137 enabled by specifying a volume group name, and file and sharedfile, if those
138 were enabled at configure time. Additionally, it will include all disk
139 templates that are currently used by instances. The order of disk templates
140 will be based on Ganeti's history of supporting them. In the future, the
141 first entry of the list will be used as a default disk template on instance
143 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
144 configuration back to the previous stable version.
145 - Disk templates in group ipolicy can be restored to the default value.
146 - Initial support for diskless instances and virtual clusters in QA.
147 - More QA and unit tests for instance policies.
148 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
149 describing why the opcode itself was executed.
150 - The monitoring daemon is now available. It allows users to query the cluster
151 for obtaining information about the status of the system. The daemon is only
152 responsible for providing the information over the network: the actual data
153 gathering is performed by data collectors (currently, only the DRBD status
154 collector is available).
155 - In order to help developers work on Ganeti, a new script
156 (``devel/build_chroot``) is provided, for building a chroot that contains all
157 the required development libraries and tools for compiling Ganeti on a Debian
159 - A new tool, ``harep``, for performing self-repair and recreation of instances
160 in Ganeti has been added.
161 - Split queries are enabled for tags, network, exports, cluster info, groups,
163 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
164 It prints the command line to set the current policies, to ease
166 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
167 devices for KVM virtio-net interfaces will get created with VNET_HDR
168 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
169 virtio-net interfaces, which prevents host kernel tainting and log
170 flooding, when dealing with broken or malicious virtio-net drivers.
171 It's set to true by default.
172 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
174 - Support 'viridian' parameter in Xen HVM
175 - Support DSA SSH keys in bootstrap
176 - To simplify the work of packaging frameworks that want to add the needed users
177 and groups in a split-user setup themselves, at build time three files in
178 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
179 the groups to be generated, the ``users`` file contains, one per line, the
180 users to be generated, optionally followed by their primary group, where
181 important. The ``groupmemberships`` file contains, one per line, additional
182 user-group membership relations that need to be established. The syntax of
183 these files will remain stable in all future versions.
188 The following new dependencies have been added:
191 - The ``curl`` library is not optional anymore for compiling the Haskell code.
192 - ``snap-server`` library (if monitoring is enabled).
195 - The minimum Python version needed to run Ganeti is now 2.6.
196 - ``yaml`` library (only for running the QA).
200 - Perform proper cleanup on termination of Haskell daemons
201 - Fix corner-case in handling of remaining retry time
207 *(Released Tue, 17 Sep 2013)*
209 - To simplify the work of packaging frameworks that want to add the needed users
210 and groups in a split-user setup themselves, at build time three files in
211 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
212 the groups to be generated, the ``users`` file contains, one per line, the
213 users to be generated, optionally followed by their primary group, where
214 important. The ``groupmemberships`` file contains, one per line, additional
215 user-group membership relations that need to be established. The syntax of
216 these files will remain stable in all future versions.
217 - Add a default to file-driver when unspecified over RAPI (Issue 571)
218 - Mark the DSA host pubkey as optional, and remove it during config downgrade
220 - Some documentation fixes
226 *(Released Tue, 27 Aug 2013)*
228 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
230 - Support 'viridian' parameter in Xen HVM (Issue 233)
231 - Include VCS version in ``gnt-cluster version``
232 - Support DSA SSH keys in bootstrap (Issue 338)
233 - Fix batch creation of instances
234 - Use FQDN to check master node status (Issue 551)
235 - Make the DRBD collector more failure-resilient
241 *(Released Fri, 2 Aug 2013)*
243 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
245 - Fix upgrading/downgrading from 2.7
246 - Increase maximum RAPI message size
247 - Documentation updates
248 - Split ``confd`` between ``luxid`` and ``confd``
249 - Merge 2.7 series up to the 2.7.1 release
250 - Allow the ``modify_etc_hosts`` option to be changed
251 - Add better debugging for ``luxid`` queries
252 - Expose bulk parameter for GetJobs in RAPI client
253 - Expose missing ``network`` fields in RAPI
254 - Add some ``cluster verify`` tests
255 - Some unittest fixes
256 - Fix a malfunction in ``hspace``'s tiered allocation
257 - Fix query compatibility between haskell and python implementations
258 - Add the ``vnet_hdr`` HV parameter for KVM
259 - Add ``--cleanup`` to instance failover
260 - Change the connected groups format in ``gnt-network info`` output; it
261 was previously displayed as a raw list by mistake. (Merged from 2.7)
267 *(Released Mon, 24 Jun 2013)*
269 This was the first beta release of the 2.8 series. All important changes
270 are listed in the latest 2.8 entry.
276 *(Released Thu, 26 Sep 2013)*
278 - Change the connected groups format in ``gnt-network info`` output; it
279 was previously displayed as a raw list by mistake
280 - Check disk template in right dict when copying
281 - Support multi-instance allocs without iallocator
282 - Fix some errors in the documentation
283 - Fix formatting of tuple in an error message
289 *(Released Thu, 25 Jul 2013)*
291 - Add logrotate functionality in daemon-util
292 - Add logrotate example file
293 - Add missing fields to network queries over rapi
294 - Fix network object timestamps
295 - Add support for querying network timestamps
296 - Fix a typo in the example crontab
297 - Fix a documentation typo
303 *(Released Thu, 04 Jul 2013)*
305 Incompatible/important changes
306 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
308 - Instance policies for disk size were documented to be on a per-disk
309 basis, but hail applied them to the sum of all disks. This has been
311 - ``hbal`` will now exit with status 0 if, during job execution over
312 LUXI, early exit has been requested and all jobs are successful;
313 before, exit status 1 was used, which cannot be differentiated from
315 - Compatibility with newer versions of rbd has been fixed
316 - ``gnt-instance batch-create`` has been changed to use the bulk create
317 opcode from Ganeti. This lead to incompatible changes in the format of
318 the JSON file. It's now not a custom dict anymore but a dict
319 compatible with the ``OpInstanceCreate`` opcode.
320 - Parent directories for file storage need to be listed in
321 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
322 write the file automatically based on old configuration values, but it
323 can not distribute it across all nodes and the file contents should be
324 verified. Use ``gnt-cluster copyfile
325 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
326 upgraded. The reason for requiring this list of paths now is that
327 before it would have been possible to inject new paths via RPC,
328 allowing files to be created in arbitrary locations. The RPC protocol
329 is protected using SSL/X.509 certificates, but as a design principle
330 Ganeti does not permit arbitrary paths to be passed.
331 - The parsing of the variants file for OSes (see
332 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
333 lines and comment lines (starting with ``#``) are ignored for better
335 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
336 longer available. ``gnt-node add`` now invokes a new tool on the
337 destination node, named ``prepare-node-join``, to configure the SSH
338 daemon. Paramiko is no longer necessary to configure nodes' SSH
339 daemons via ``gnt-node add``.
340 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
341 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
342 the flag is not reset after a master failover.
343 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
344 is recommended. 2.6 will be mandatory from the 2.8 series.
350 - New network management functionality to support automatic allocation
351 of IP addresses and managing of network parameters. See
352 :manpage:`gnt-network(8)` for more details.
353 - New external storage backend, to allow managing arbitrary storage
354 systems external to the cluster. See
355 :manpage:`ganeti-extstorage-interface(7)`.
356 - New ``exclusive-storage`` node parameter added, restricted to
357 nodegroup level. When it's set to true, physical disks are assigned in
358 an exclusive fashion to instances, as documented in :doc:`Partitioned
359 Ganeti <design-partitioned>`. Currently, only instances using the
360 ``plain`` disk template are supported.
361 - The KVM hypervisor has been updated with many new hypervisor
362 parameters, including a generic one for passing arbitrary command line
363 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
364 compatible up to qemu 1.4.
365 - A new tool, called ``mon-collector``, is the stand-alone executor of
366 the data collectors for a monitoring system. As of this version, it
367 just includes the DRBD data collector, that can be executed by calling
368 ``mon-collector`` using the ``drbd`` parameter. See
369 :manpage:`mon-collector(7)`.
370 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
371 for RAPI users. It allows granting permissions to query for
372 information to a specific user without giving
373 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
374 - A new tool named ``node-cleanup`` has been added. It cleans remains of
375 a cluster from a machine by stopping all daemons, removing
376 certificates and ssconf files. Unless the ``--no-backup`` option is
377 given, copies of the certificates are made.
378 - Instance creations now support the use of opportunistic locking,
379 potentially speeding up the (parallel) creation of multiple instances.
380 This feature is currently only available via the :doc:`RAPI
381 <rapi>` interface and when an instance allocator is used. If the
382 ``opportunistic_locking`` parameter is set the opcode will try to
383 acquire as many locks as possible, but will not wait for any locks
384 held by other opcodes. If not enough resources can be found to
385 allocate the instance, the temporary error code
386 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
387 retried thereafter, with or without opportunistic locking.
388 - New experimental linux-ha resource scripts.
389 - Restricted-commands support: ganeti can now be asked (via command line
390 or rapi) to perform commands on a node. These are passed via ganeti
391 RPC rather than ssh. This functionality is restricted to commands
392 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
393 reasons. The file is not copied automatically.
399 - Diskless instances are now externally mirrored (Issue 237). This for
400 now has only been tested in conjunction with explicit target nodes for
402 - Queries not needing locks or RPC access to the node can now be
403 performed by the confd daemon, making them independent from jobs, and
404 thus faster to execute. This is selectable at configure time.
405 - The functionality for allocating multiple instances at once has been
406 overhauled and is now also available through :doc:`RAPI <rapi>`.
408 There are no significant changes from version 2.7.0~rc3.
414 *(Released Tue, 25 Jun 2013)*
416 - Fix permissions on the confd query socket (Issue 477)
417 - Fix permissions on the job archive dir (Issue 498)
418 - Fix handling of an internal exception in replace-disks (Issue 472)
419 - Fix gnt-node info handling of shortened names (Issue 497)
420 - Fix gnt-instance grow-disk when wiping is enabled
421 - Documentation improvements, and support for newer pandoc
422 - Fix hspace honoring ipolicy for disks (Issue 484)
423 - Improve handling of the ``kvm_extra`` HV parameter
429 *(Released Fri, 24 May 2013)*
431 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
433 - Disks added through ``gnt-instance modify`` or created through
434 ``gnt-instance recreate-disks`` are wiped, if the
435 ``prealloc_wipe_disks`` flag is set.
436 - If wiping newly created disks fails, the disks are removed. Also,
437 partial failures in creating disks through ``gnt-instance modify``
438 triggers a cleanup of the partially-created disks.
439 - Removing the master IP address doesn't fail if the address has been
441 - Fix ownership of the OS log dir
442 - Workaround missing SO_PEERCRED constant (Issue 191)
448 *(Released Fri, 3 May 2013)*
450 This was the first release candidate of the 2.7 series. Since beta3:
452 - Fix kvm compatibility with qemu 1.4 (Issue 389)
453 - Documentation updates (admin guide, upgrade notes, install
454 instructions) (Issue 372)
455 - Fix gnt-group list nodes and instances count (Issue 436)
456 - Fix compilation without non-mandatory libraries (Issue 441)
457 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
458 - Make confd logging more verbose at INFO level (Issue 435)
459 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
460 - Fix failure path for instance storage type conversion (Issue 229)
461 - Update htools text backend documentation
462 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
463 - Disable inter-cluster instance move for file-based instances, because
464 it is dependant on instance export, which is not supported for
465 file-based instances. (Issue 414)
466 - Fix gnt-job crashes on non-ascii characters (Issue 427)
467 - Fix volume group checks on non-vm-capable nodes (Issue 432)
473 *(Released Mon, 22 Apr 2013)*
475 This was the third beta release of the 2.7 series. Since beta2:
477 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
478 - Fix data loss on wrong usage of ``gnt-instance move``
479 - Properly export errors in confd-based job queries
480 - Add ``users-setup`` tool
481 - Fix iallocator protocol to report 0 as a disk size for diskless
482 instances. This avoids hail breaking when a diskless instance is
484 - Fix job queue directory permission problem that made confd job queries
485 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
486 for access to archived jobs (Issue 406).
487 - Limit the sizes of networks supported by ``gnt-network`` to something
488 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
489 - Fix bugs in instance disk template conversion
490 - Fix GHC 7 compatibility
491 - Fix ``burnin`` install path (Issue 426).
492 - Allow very small disk grows (Issue 347).
493 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
494 that noded doesn't import masterd code (Issue 419).
495 - Make sure the default metavg at cluster init is the same as the vg, if
496 unspecified (Issue 358).
497 - Fix cleanup of partially created disks (part of Issue 416)
503 *(Released Tue, 2 Apr 2013)*
505 This was the second beta release of the 2.7 series. Since beta1:
507 - Networks no longer have a "type" slot, since this information was
508 unused in Ganeti: instead of it tags should be used.
509 - The rapi client now has a ``target_node`` option to MigrateInstance.
510 - Fix early exit return code for hbal (Issue 386).
511 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
512 - Fix ``rbd showmapped`` output parsing (Issue 312).
513 - Networks are now referenced indexed by UUID, rather than name. This
514 will require running cfgupgrade, from 2.7.0beta1, if networks are in
516 - The OS environment now includes network information.
517 - Deleting of a network is now disallowed if any instance nic is using
518 it, to prevent dangling references.
519 - External storage is now documented in man pages.
520 - The exclusive_storage flag can now only be set at nodegroup level.
521 - Hbal can now submit an explicit priority with its jobs.
522 - Many network related locking fixes.
523 - Bump up the required pylint version to 0.25.1.
524 - Fix the ``no_remember`` option in RAPI client.
525 - Many ipolicy related tests, qa, and fixes.
526 - Many documentation improvements and fixes.
527 - Fix building with ``--disable-file-storage``.
528 - Fix ``-q`` option in htools, which was broken if passed more than
530 - Some haskell/python interaction improvements and fixes.
531 - Fix iallocator in case of missing LVM storage.
532 - Fix confd config load in case of ``--no-lvm-storage``.
533 - The confd/query functionality is now mentioned in the security
540 *(Released Wed, 6 Feb 2013)*
542 This was the first beta release of the 2.7 series. All important changes
543 are listed in the latest 2.7 entry.
549 *(Released Fri, 21 Dec 2012)*
551 Important behaviour change: hbal won't rebalance anymore instances which
552 have the ``auto_balance`` attribute set to false. This was the intention
553 all along, but until now it only skipped those from the N+1 memory
554 reservation (DRBD-specific).
556 A significant number of bug fixes in this release:
558 - Fixed disk adoption interaction with ipolicy checks.
559 - Fixed networking issues when instances are started, stopped or
560 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
561 - Fixed the warning in cluster verify for shared storage instances not
563 - Fixed removal of storage directory on shared file storage (issue 262).
564 - Fixed validation of LVM volume group name in OpClusterSetParams
565 (``gnt-cluster modify``) (issue 285).
566 - Fixed runtime memory increases (``gnt-instance modify -m``).
567 - Fixed live migration under Xen's ``xl`` mode.
568 - Fixed ``gnt-instance console`` with ``xl``.
569 - Fixed building with newer Haskell compiler/libraries.
570 - Fixed PID file writing in Haskell daemons (confd); this prevents
571 restart issues if confd was launched manually (outside of
572 ``daemon-util``) while another copy of it was running
573 - Fixed a type error when doing live migrations with KVM (issue 297) and
574 the error messages for failing migrations have been improved.
575 - Fixed opcode validation for the out-of-band commands (``gnt-node
577 - Fixed a type error when unsetting OS hypervisor parameters (issue
578 311); now it's possible to unset all OS-specific hypervisor
580 - Fixed the ``dry-run`` mode for many operations: verification of
581 results was over-zealous but didn't take into account the ``dry-run``
582 operation, resulting in "wrong" failures.
583 - Fixed bash completion in ``gnt-job list`` when the job queue has
584 hundreds of entries; especially with older ``bash`` versions, this
585 results in significant CPU usage.
587 And lastly, a few other improvements have been made:
589 - Added option to force master-failover without voting (issue 282).
590 - Clarified error message on lock conflict (issue 287).
591 - Logging of newly submitted jobs has been improved (issue 290).
592 - Hostname checks have been made uniform between instance rename and
594 - The ``--submit`` option is now supported by ``gnt-debug delay``.
595 - Shutting down the master daemon by sending SIGTERM now stops it from
596 processing jobs waiting for locks; instead, those jobs will be started
597 once again after the master daemon is started the next time (issue
599 - Support for Xen's ``xl`` program has been improved (besides the fixes
601 - Reduced logging noise in the Haskell confd daemon (only show one log
602 entry for each config reload, instead of two).
603 - Several man page updates and typo fixes.
609 *(Released Fri, 12 Oct 2012)*
611 A small bugfix release. Among the bugs fixed:
613 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
614 made the command unusable.
615 - Commands that issue many jobs don't fail anymore just because some jobs
616 take so long that other jobs are archived.
617 - Failures during ``gnt-instance reinstall`` are reflected by the exit
619 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
621 - When ``always_failover`` is set, ``--allow-failover`` is not required
622 in migrate commands anymore.
623 - ``bash_completion`` works even if extglob is disabled.
624 - Fixed bug with locks that made failover for RDB-based instances fail.
625 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
626 a random node instead of one based on the allocator metric.
627 - Support for newer versions of pylint and pep8.
628 - Hail doesn't fail anymore when trying to add an instance of type
629 ``file``, ``sharedfile`` or ``rbd``.
630 - Added new Makefile target to rebuild the whole distribution, so that
631 all files are included.
637 *(Released Fri, 27 Jul 2012)*
640 .. attention:: The ``LUXI`` protocol has been made more consistent
641 regarding its handling of command arguments. This, however, leads to
642 incompatibility issues with previous versions. Please ensure that you
643 restart Ganeti daemons soon after the upgrade, otherwise most
644 ``LUXI`` calls (job submission, setting/resetting the drain flag,
645 pausing/resuming the watcher, cancelling and archiving jobs, querying
646 the cluster configuration) will fail.
655 The current ``admin_up`` field, which used to denote whether an instance
656 should be running or not, has been removed. Instead, ``admin_state`` is
657 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
659 The rational behind this is that an instance being “down” can have
662 - it could be down during a reboot
663 - it could be temporarily be down for a reinstall
664 - or it could be down because it is deprecated and kept just for its
667 The previous Boolean state was making it difficult to do capacity
668 calculations: should Ganeti reserve memory for a down instance? Now, the
669 tri-state field makes it clear:
671 - in ``up`` and ``down`` state, all resources are reserved for the
672 instance, and it can be at any time brought up if it is down
673 - in ``offline`` state, only disk space is reserved for it, but not
676 The field can have an extra use: since the transition between ``up`` and
677 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
678 transition between ``offline`` and ``down`` is done via ``gnt-instance
679 modify``, it is possible to given different rights to users. For
680 example, owners of an instance could be allowed to start/stop it, but
681 not transition it out of the offline state.
683 Instance policies and specs
684 +++++++++++++++++++++++++++
686 In previous Ganeti versions, an instance creation request was not
687 limited on the minimum size and on the maximum size just by the cluster
688 resources. As such, any policy could be implemented only in third-party
689 clients (RAPI clients, or shell wrappers over ``gnt-*``
690 tools). Furthermore, calculating cluster capacity via ``hspace`` again
691 required external input with regards to instance sizes.
693 In order to improve these workflows and to allow for example better
694 per-node group differentiation, we introduced instance specs, which
697 - minimum instance disk size, disk count, memory size, cpu count
698 - maximum values for the above metrics
699 - and “standard” values (used in ``hspace`` to calculate the standard
702 The minimum/maximum values can be also customised at node-group level,
703 for example allowing more powerful hardware to support bigger instance
706 Beside the instance specs, there are a few other settings belonging to
707 the instance policy framework. It is possible now to customise, per
708 cluster and node-group:
710 - the list of allowed disk templates
711 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
712 - the maximum ratio of instance to spindles (see below for more
713 information) for local storage
715 All these together should allow all tools that talk to Ganeti to know
716 what are the ranges of allowed values for instances and the
717 over-subscription that is allowed.
719 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
720 instance configuration, and the physical CPU configuration from the
721 node. For the spindle ratios however, we didn't track before these
722 values, so new parameters have been added:
724 - a new node parameter ``spindle_count``, defaults to 1, customisable at
725 node group or node level
726 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
728 Note that spindles in this context doesn't need to mean actual
729 mechanical hard-drives; it's just a relative number for both the node
730 I/O capacity and instance I/O consumption.
732 Instance migration behaviour
733 ++++++++++++++++++++++++++++
735 While live-migration is in general desirable over failover, it is
736 possible that for some workloads it is actually worse, due to the
737 variable time of the “suspend” phase during live migration.
739 To allow the tools to work consistently over such instances (without
740 having to hard-code instance names), a new backend parameter
741 ``always_failover`` has been added to control the migration/failover
742 behaviour. When set to True, all migration requests for an instance will
743 instead fall-back to failover.
745 Instance memory ballooning
746 ++++++++++++++++++++++++++
748 Initial support for memory ballooning has been added. The memory for an
749 instance is no longer fixed (backend parameter ``memory``), but instead
750 can vary between minimum and maximum values (backend parameters
751 ``minmem`` and ``maxmem``). Currently we only change an instance's
754 - live migrating or failing over and instance and the target node
755 doesn't have enough memory
756 - user requests changing the memory via ``gnt-instance modify
762 In order to control the use of specific CPUs by instance, support for
763 controlling CPU pinning has been added for the Xen, HVM and LXC
764 hypervisors. This is controlled by a new hypervisor parameter
765 ``cpu_mask``; details about possible values for this are in the
766 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
767 VCPU-to-CPU mapping) form will work well only when all nodes in your
768 cluster have the same amount of CPUs.
773 Another area in which Ganeti was not customisable were the parameters
774 used for storage configuration, e.g. how many stripes to use for LVM,
775 DRBD resync configuration, etc.
777 To improve this area, we've added disks parameters, which are
778 customisable at cluster and node group level, and which allow to
779 specify various parameters for disks (DRBD has the most parameters
780 currently), for example:
782 - DRBD resync algorithm and parameters (e.g. speed)
783 - the default VG for meta-data volumes for DRBD
784 - number of stripes for LVM (plain disk template)
787 These parameters can be modified via ``gnt-cluster modify -D …`` and
788 ``gnt-group modify -D …``, and are used at either instance creation (in
789 case of LVM stripes, for example) or at disk “activation” time
792 Rados block device support
793 ++++++++++++++++++++++++++
795 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
796 denoted by the ``rbd`` disk template type. This is considered
797 experimental, feedback is welcome. For details on configuring it, see
798 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
803 The existing master IP functionality works well only in simple setups (a
804 single network shared by all nodes); however, if nodes belong to
805 different networks, then the ``/32`` setup and lack of routing
806 information is not enough.
808 To allow the master IP to function well in more complex cases, the
809 system was reworked as follows:
811 - a master IP netmask setting has been added
812 - the master IP activation/turn-down code was moved from the node daemon
814 - whether to run the Ganeti-supplied master IP script or a user-supplied
815 on is a ``gnt-cluster init`` setting
817 Details about the location of the standard and custom setup scripts are
818 in the man page :manpage:`gnt-cluster(8)`; for information about the
819 setup script protocol, look at the Ganeti-supplied script.
824 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
827 It is now possible to use TLS-protected connections, and when renewing
828 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
829 it is now possible to specify spice or spice CA certificates. Also, it
830 is possible to configure a password for SPICE sessions via the
831 hypervisor parameter ``spice_password_file``.
833 There are also new parameters to control the compression and streaming
834 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
835 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
836 for the spice parameters.
838 Lastly, it is now possible to see the SPICE connection information via
839 ``gnt-instance console``.
844 A new tool (``tools/ovfconverter``) has been added that supports
845 conversion between Ganeti and the `Open Virtualization Format
846 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
849 This relies on the ``qemu-img`` tool to convert the disk formats, so the
850 actual compatibility with other virtualization solutions depends on it.
855 The configuration query daemon (``ganeti-confd``) is now optional, and
856 has been rewritten in Haskell; whether to use the daemon at all, use the
857 Python (default) or the Haskell version is selectable at configure time
858 via the ``--enable-confd`` parameter, which can take one of the
859 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
860 daemon will result in a smaller footprint; for larger systems, we
861 welcome feedback on the Haskell version which might become the default
864 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
865 daemon running. The Python version doesn't implement the new call.
868 User interface changes
869 ~~~~~~~~~~~~~~~~~~~~~~
871 We have replaced the ``--disks`` option of ``gnt-instance
872 replace-disks`` with a more flexible ``--disk`` option, which allows
873 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
874 disk size and mode can be changed upon recreation (via ``gnt-instance
875 recreate-disks``, which accepts the same ``--disk`` option).
877 As many people are used to a ``show`` command, we have added that as an
878 alias to ``info`` on all ``gnt-*`` commands.
880 The ``gnt-instance grow-disk`` command has a new mode in which it can
881 accept the target size of the disk, instead of the delta; this can be
882 more safe since two runs in absolute mode will be idempotent, and
883 sometimes it's also easier to specify the desired size directly.
885 Also the handling of instances with regard to offline secondaries has
886 been improved. Instance operations should not fail because one of it's
887 secondary nodes is offline, even though it's safe to proceed.
889 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
890 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
891 minors to instance name.
896 RAPI coverage has improved, with (for example) new resources for
897 recreate-disks, node power-cycle, etc.
902 There is partial support for ``xl`` in the Xen hypervisor; feedback is
905 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
906 whether to still support Python 2.4 or move to Python 2.6 as minimum
909 Support for Fedora has been slightly improved; the provided example
910 init.d script should work better on it and the INSTALL file should
911 document the needed dependencies.
916 The deprecated ``QueryLocks`` LUXI request has been removed. Use
917 ``Query(what=QR_LOCK, ...)`` instead.
919 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
920 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
921 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
922 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
923 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
925 RAPI client: ``CertificateError`` now derives from
926 ``GanetiApiError``. This should make it more easy to handle Ganeti
929 Deprecation warnings due to PyCrypto/paramiko import in
930 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
931 make sure to run an up-to-date paramiko version, if you use this tool.
933 The QA scripts now depend on Python 2.5 or above (the main code base
934 still works with Python 2.4).
936 The configuration file (``config.data``) is now written without
937 indentation for performance reasons; if you want to edit it, it can be
938 re-formatted via ``tools/fmtjson``.
940 A number of bugs has been fixed in the cluster merge tool.
942 ``x509`` certification verification (used in import-export) has been
943 changed to allow the same clock skew as permitted by the cluster
944 verification. This will remove some rare but hard to diagnose errors in
951 *(Released Thu, 19 Jul 2012)*
953 Very few changes from rc4 to the final release, only bugfixes:
955 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
956 instance, fix CDROM booting for KVM instances)
957 - fixed node group modification of node parameters
958 - fixed issue in LUClusterVerifyGroup with multi-group clusters
959 - fixed generation of bash completion to ensure a stable ordering
966 *(Released Fri, 13 Jul 2012)*
968 Third release candidate for 2.6. The following changes were done from
971 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
972 - Fixed an inconsistency in the LUXI protocol with the provided
973 arguments (NOT backwards compatible)
974 - Fixed a bug with node groups ipolicy where ``min`` was greater than
975 the cluster ``std`` value
976 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
977 easier instance debugging on nodes (requires ``hconfd`` to work)
983 *(Released Tue, 03 Jul 2012)*
985 Second release candidate for 2.6. The following changes were done from
988 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
990 - Fixed a RAPI regression on missing beparams/memory
991 - Fixed redistribution of files on offline nodes
992 - Added possibility to run activate-disks even though secondaries are
993 offline. With this change it relaxes also the strictness on some other
994 commands which use activate disks internally:
995 * ``gnt-instance start|reboot|rename|backup|export``
996 - Made it possible to remove safely an instance if its secondaries are
998 - Made it possible to reinstall even though secondaries are offline
1004 *(Released Mon, 25 Jun 2012)*
1006 First release candidate for 2.6. The following changes were done from
1009 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1011 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1012 behaviour. It supports now the modify operation
1013 - ``hcheck`` implemented to analyze cluster health and possibility of
1014 improving health by rebalance
1015 - ``hbal`` has been improved in dealing with split instances
1021 *(Released Mon, 11 Jun 2012)*
1023 Second beta release of 2.6. The following changes were done from beta2
1026 - Fixed ``daemon-util`` with non-root user models
1027 - Fixed creation of plain instances with ``--no-wait-for-sync``
1028 - Fix wrong iv_names when running ``cfgupgrade``
1029 - Export more information in RAPI group queries
1030 - Fixed bug when changing instance network interfaces
1031 - Extended burnin to do NIC changes
1032 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1033 - Changed default for DRBD barriers
1034 - Fixed DRBD error reporting for syncer rate
1035 - Verify the options on disk parameters
1037 And of course various fixes to documentation and improved unittests and
1044 *(Released Wed, 23 May 2012)*
1046 First beta release of 2.6. The following changes were done from beta1 to
1049 - integrated patch for distributions without ``start-stop-daemon``
1050 - adapted example init.d script to work on Fedora
1051 - fixed log handling in Haskell daemons
1052 - adapted checks in the watcher for pycurl linked against libnss
1053 - add partial support for ``xl`` instead of ``xm`` for Xen
1054 - fixed a type issue in cluster verification
1055 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1058 Plus integrated fixes from the 2.5 branch:
1060 - fixed ``kvm-ifup`` to use ``/bin/bash``
1061 - fixed parallel build failures
1062 - KVM live migration when using a custom keymap
1068 *(Released Tue, 24 Jul 2012)*
1070 A small bugfix release, with no new features:
1072 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1073 different default shell (e.g. Debian, Ubuntu)
1074 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1075 243 and Debian bug #650664)
1076 - fixed compatibility with KVM versions that don't support multiple boot
1077 devices (fixes Issue 230 and Debian bug #624256)
1079 Additionally, a few fixes were done to the build system (fixed parallel
1080 build failures) and to the unittests (fixed race condition in test for
1081 FileID functions, and the default enable/disable mode for QA test is now
1088 *(Released Fri, 11 May 2012)*
1090 A small bugfix release.
1092 The main issues solved are on the topic of compatibility with newer LVM
1095 - fixed parsing of ``lv_attr`` field
1096 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1097 the ``--force`` flag is needed
1099 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1100 to accept kernel 3.x too (was hardcoded to 2.6.*).
1102 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1103 of hook results and that also made display of other errors suboptimal
1104 was fixed; the code behaves now like 2.4 and earlier.
1106 Another change in 2.5, the cleanup of the OS scripts environment, is too
1107 aggressive: it removed even the ``PATH`` variable, which requires the OS
1108 scripts to *always* need to export it. Since this is a bit too strict,
1109 we now export a minimal PATH, the same that we export for hooks.
1111 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1112 integrated into this release.
1114 Finally, a few other miscellaneous changes were done (no new features,
1115 just small improvements):
1117 - Fix ``gnt-group --help`` display
1118 - Fix hardcoded Xen kernel path
1119 - Fix grow-disk handling of invalid units
1120 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1121 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1127 *(Released Thu, 12 Apr 2012)*
1129 Incompatible/important changes and bugfixes
1130 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1132 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1133 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1134 to match the underlying LUXI interface.
1135 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1136 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1137 not maintain backwards-compatibility as the underlying operation
1138 changed in an incompatible way. The RAPI client can talk to old
1139 servers, but it needs to be told so as the return value changed.
1140 - When creating file-based instances via RAPI, the ``file_driver``
1141 parameter no longer defaults to ``loop`` and must be specified.
1142 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1144 - Support for the undocumented and deprecated RAPI instance creation
1145 request format version 0 has been dropped. Use version 1, supported
1146 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1147 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1148 documentation <install>`.
1149 - The "cluster-verify" hooks are now executed per group by the
1150 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1151 if you just run ``gnt-cluster verify``, which generates one opcode per
1153 - The environment as passed to the OS scripts is cleared, and thus no
1154 environment variables defined in the node daemon's environment will be
1155 inherited by the scripts.
1156 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1158 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1159 support operations involving multiple node groups.
1160 - Offline nodes are ignored when failing over an instance.
1161 - Support for KVM version 1.0, which changed the version reporting format
1163 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1165 - ``Makefile`` is now compatible with Automake 1.11.2
1166 - Includes all bugfixes made in the 2.4 series
1171 - The ganeti-htools project has been merged into the ganeti-core source
1172 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1173 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1174 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1175 GPT-style partition tables (requires `parted
1176 <http://www.gnu.org/s/parted/>`_).
1177 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1178 - Allowed adding tags on instance creation.
1179 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1180 - Allow instances to be started in a paused state, enabling the user to
1181 see the complete console output on boot using the console.
1182 - Added new hypervisor flag to control default reboot behaviour
1183 (``reboot_behavior``).
1184 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1185 - Improved out-of-band management support:
1187 - Added ``gnt-node health`` command reporting the health status of
1189 - Added ``gnt-node power`` command to manage power status of nodes.
1190 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1192 - Instance migration can fall back to failover if instance is not
1194 - Filters can be used when listing nodes, instances, groups and locks;
1195 see :manpage:`ganeti(7)` manpage.
1196 - Added post-execution status as variables to :doc:`hooks <hooks>`
1198 - Instance tags are exported/imported together with the instance.
1199 - When given an explicit job ID, ``gnt-job info`` will work for archived
1201 - Jobs can define dependencies on other jobs (not yet supported via
1202 RAPI or command line, but used by internal commands and usable via
1205 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1208 - Instance failover is now available as a RAPI resource
1209 (``/2/instances/[instance_name]/failover``).
1210 - ``gnt-instance info`` defaults to static information if primary node
1212 - Opcodes have a new ``comment`` attribute.
1213 - Added basic SPICE support to KVM hypervisor.
1214 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1216 Node group improvements
1217 ~~~~~~~~~~~~~~~~~~~~~~~
1219 - ``gnt-cluster verify`` has been modified to check groups separately,
1220 thereby improving performance.
1221 - Node group support has been added to ``gnt-cluster verify-disks``,
1222 which now operates per node group.
1223 - Watcher has been changed to work better with node groups.
1225 - One process and state file per node group.
1226 - Slow watcher in one group doesn't block other group's watcher.
1228 - Added new command, ``gnt-group evacuate``, to move all instances in a
1229 node group to other groups.
1230 - Added ``gnt-instance change-group`` to move an instance to another
1232 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1233 per-group operations.
1234 - Node groups can be tagged.
1235 - Some operations switch from an exclusive to a shared lock as soon as
1237 - Instance's primary and secondary nodes' groups are now available as
1238 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1239 and ``snodes.group.uuid``).
1244 - Numerous updates to documentation and manpages.
1246 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1248 - Some opcode/job results are now also documented, see :doc:`RAPI
1251 - A lockset's internal lock is now also visible in lock monitor.
1252 - Log messages from job queue workers now contain information about the
1253 opcode they're processing.
1254 - ``gnt-instance console`` no longer requires the instance lock.
1255 - A short delay when waiting for job changes reduces the number of LUXI
1256 requests significantly.
1257 - DRBD metadata volumes are overwritten with zeros during disk creation.
1258 - Out-of-band commands no longer acquire the cluster lock in exclusive
1260 - ``devel/upload`` now uses correct permissions for directories.
1266 *(Released Fri, 23 Mar 2012)*
1268 This was the sixth release candidate of the 2.5 series.
1274 *(Released Mon, 9 Jan 2012)*
1276 This was the fifth release candidate of the 2.5 series.
1282 *(Released Thu, 27 Oct 2011)*
1284 This was the fourth release candidate of the 2.5 series.
1290 *(Released Wed, 26 Oct 2011)*
1292 This was the third release candidate of the 2.5 series.
1298 *(Released Tue, 18 Oct 2011)*
1300 This was the second release candidate of the 2.5 series.
1306 *(Released Tue, 4 Oct 2011)*
1308 This was the first release candidate of the 2.5 series.
1314 *(Released Wed, 31 Aug 2011)*
1316 This was the third beta release of the 2.5 series.
1322 *(Released Mon, 22 Aug 2011)*
1324 This was the second beta release of the 2.5 series.
1330 *(Released Fri, 12 Aug 2011)*
1332 This was the first beta release of the 2.5 series.
1338 *(Released Thu, 27 Oct 2011)*
1340 - Fixed bug when parsing command line parameter values ending in
1342 - Fixed assertion error after unclean master shutdown
1343 - Disable HTTP client pool for RPC, significantly reducing memory usage
1345 - Fixed queue archive creation with wrong permissions
1351 *(Released Tue, 23 Aug 2011)*
1355 - Fixed documentation for importing with ``--src-dir`` option
1356 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1357 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1363 *(Released Fri, 5 Aug 2011)*
1365 Many bug-fixes and a few small features:
1367 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1368 issues when you tried to add an instance with two MAC addresses in one
1370 - KVM: fixed per-instance stored UID value
1371 - KVM: configure bridged NICs at migration start
1372 - KVM: Fix a bug where instance will not start with never KVM versions
1374 - Added OS search path to ``gnt-cluster info``
1375 - Fixed an issue with ``file_storage_dir`` where you were forced to
1376 provide an absolute path, but the documentation states it is a
1377 relative path, the documentation was right
1378 - Added a new parameter to instance stop/start called ``--no-remember``
1379 that will make the state change to not be remembered
1380 - Implemented ``no_remember`` at RAPI level
1381 - Improved the documentation
1382 - Node evacuation: don't call IAllocator if node is already empty
1383 - Fixed bug in DRBD8 replace disks on current nodes
1384 - Fixed bug in recreate-disks for DRBD instances
1385 - Moved assertion checking locks in ``gnt-instance replace-disks``
1386 causing it to abort with not owning the right locks for some situation
1387 - Job queue: Fixed potential race condition when cancelling queued jobs
1388 - Fixed off-by-one bug in job serial generation
1389 - ``gnt-node volumes``: Fix instance names
1390 - Fixed aliases in bash completion
1391 - Fixed a bug in reopening log files after being sent a SIGHUP
1392 - Added a flag to burnin to allow specifying VCPU count
1393 - Bugfixes to non-root Ganeti configuration
1399 *(Released Thu, 12 May 2011)*
1401 Many bug-fixes and a few new small features:
1403 - Fixed a bug related to log opening failures
1404 - Fixed a bug in instance listing with orphan instances
1405 - Fixed a bug which prevented resetting the cluster-level node parameter
1406 ``oob_program`` to the default
1407 - Many fixes related to the ``cluster-merge`` tool
1408 - Fixed a race condition in the lock monitor, which caused failures
1409 during (at least) creation of many instances in parallel
1410 - Improved output for gnt-job info
1411 - Removed the quiet flag on some ssh calls which prevented debugging
1413 - Improved the N+1 failure messages in cluster verify by actually
1414 showing the memory values (needed and available)
1415 - Increased lock attempt timeouts so that when executing long operations
1416 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1417 too early and thus prevent the use of the 'fair' mechanism
1418 - Changed instance query data (``gnt-instance info``) to not acquire
1419 locks unless needed, thus allowing its use on locked instance if only
1420 static information is asked for
1421 - Improved behaviour with filesystems that do not support rename on an
1423 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1424 kept locks on all nodes during the wipe, which is unneeded
1425 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1426 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1427 10GiB) which caused the wipe to fail right at the end in some cases
1428 - Fixed master IP activation when doing master failover with no-voting
1429 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1430 the master node itself
1431 - Fixed potential data-loss in under disk full conditions, where Ganeti
1432 wouldn't check correctly the return code and would consider
1433 partially-written files 'correct'
1434 - Fixed bug related to multiple VGs and DRBD disk replacing
1435 - Added new disk parameter ``metavg`` that allows placement of the meta
1436 device for DRBD in a different volume group
1437 - Fixed error handling in the node daemon when the system libc doesn't
1438 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1439 - Fixed lock release during replace-disks, which kept cluster-wide locks
1440 when doing disk replaces with an iallocator script
1441 - Added check for missing bridges in cluster verify
1442 - Handle EPIPE errors while writing to the terminal better, so that
1443 piping the output to e.g. ``less`` doesn't cause a backtrace
1444 - Fixed rare case where a ^C during Luxi calls could have been
1445 interpreted as server errors, instead of simply terminating
1446 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1448 - Added a few more parameters to the KVM hypervisor, allowing a second
1449 CDROM, custom disk type for CDROMs and a floppy image
1450 - Removed redundant message in instance rename when the name is given
1452 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1453 disks on new nodes, allowing recreation when the original instance
1454 nodes are completely gone
1455 - Added option when converting disk templates to DRBD to skip waiting
1456 for the resync, in order to make the instance available sooner
1457 - Added two new variables to the OS scripts environment (containing the
1459 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1460 to allow use of ``pvgrub`` as bootloader
1461 - Changed the instance memory modifications to only check out-of-memory
1462 conditions on memory increases, and turned the secondary node warnings
1463 into errors (they can still be overridden via ``--force``)
1464 - Fixed the handling of a corner case when the Python installation gets
1465 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1466 to execute a command that doesn't exist
1467 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1468 node of the instance returned failures during instance shutdown; this
1469 adds the option ``--ignore-consistency`` to gnt-instance move
1471 And as usual, various improvements to the error messages, documentation
1478 *(Released Wed, 09 Mar 2011)*
1480 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1481 the RAPI users file if run twice (even with ``--dry-run``).
1483 The release fixes that bug (nothing else changed).
1489 *(Released Mon, 07 Mar 2011)*
1491 Final 2.4.0 release. Just a few small fixes:
1493 - Fixed RAPI node evacuate
1494 - Fixed the kvm-ifup script
1495 - Fixed internal error handling for special job cases
1496 - Updated man page to specify the escaping feature for options
1502 *(Released Mon, 28 Feb 2011)*
1504 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1505 that this feature wiped the disks of the wrong instance, leading to loss
1510 - Fixed title of query field containing instance name
1511 - Expanded the glossary in the documentation
1512 - Fixed one unittest (internal issue)
1518 *(Released Mon, 21 Feb 2011)*
1520 A number of bug fixes plus just a couple functionality changes.
1522 On the user-visible side, the ``gnt-* list`` command output has changed
1523 with respect to "special" field states. The current rc1 style of display
1524 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1525 the default output mode special fields states are displayed as follows:
1527 - Offline resource: ``*``
1528 - Unavailable/not applicable: ``-``
1529 - Data missing (RPC failure): ``?``
1530 - Unknown field: ``??``
1532 Another user-visible change is the addition of ``--force-join`` to
1537 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1538 - Fixed regression in RAPI/instance reinstall where all parameters were
1539 required (instead of optional)
1540 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1541 - Fixed iallocator usage (offline nodes were not considered offline)
1542 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1543 - Fixed hypervisor and OS parameter validation with respect to
1544 non-vm_capable nodes
1545 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1547 - Fixed ``tools/listrunner`` with respect to agent-based usage
1553 *(Released Fri, 4 Feb 2011)*
1555 Many changes and fixes since the beta1 release. While there were some
1556 internal changes, the code has been mostly stabilised for the RC
1559 Note: the dumb allocator was removed in this release, as it was not kept
1560 up-to-date with the IAllocator protocol changes. It is recommended to
1561 use the ``hail`` command from the ganeti-htools package.
1563 Note: the 2.4 and up versions of Ganeti are not compatible with the
1564 0.2.x branch of ganeti-htools. You need to upgrade to
1565 ganeti-htools-0.3.0 (or later).
1567 Regressions fixed from 2.3
1568 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1570 - Fixed the ``gnt-cluster verify-disks`` command
1571 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1573 - Fixed disk adoption breakage
1574 - Fixed wrong headers in instance listing for field aliases
1579 - Fixed corner case in KVM handling of NICs
1580 - Fixed many cases of wrong handling of non-vm_capable nodes
1581 - Fixed a bug where a missing instance symlink was not possible to
1582 recreate with any ``gnt-*`` command (now ``gnt-instance
1583 activate-disks`` does it)
1584 - Fixed the volume group name as reported by ``gnt-cluster
1586 - Increased timeouts for the import-export code, hopefully leading to
1587 fewer aborts due network or instance timeouts
1588 - Fixed bug in ``gnt-node list-storage``
1589 - Fixed bug where not all daemons were started on cluster
1590 initialisation, but only at the first watcher run
1591 - Fixed many bugs in the OOB implementation
1592 - Fixed watcher behaviour in presence of instances with offline
1594 - Fixed instance list output for instances running on the wrong node
1595 - a few fixes to the cluster-merge tool, but it still cannot merge
1596 multi-node groups (currently it is not recommended to use this tool)
1602 - Improved network configuration for the KVM hypervisor
1603 - Added e1000 as a supported NIC for Xen-HVM
1604 - Improved the lvmstrap tool to also be able to use partitions, as
1605 opposed to full disks
1606 - Improved speed of disk wiping (the cluster parameter
1607 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1608 of instance creations
1609 - Added documentation for the OS parameters
1610 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1611 hypervisor is not responding
1612 - Added display of blacklisted and hidden OS information in
1613 ``gnt-cluster info``
1614 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1615 NIC and node parameters, which might create problems with currently
1616 invalid (but undetected) configuration files, but prevents validation
1617 failures when unrelated parameters are modified
1618 - Changed cluster initialisation to wait for the master daemon to become
1620 - Expanded the RAPI interface:
1622 - Added config redistribution resource
1623 - Added activation/deactivation of instance disks
1624 - Added export of console information
1626 - Implemented log file reopening on SIGHUP, which allows using
1627 logrotate(8) for the Ganeti log files
1628 - Added a basic OOB helper script as an example
1634 *(Released Fri, 14 Jan 2011)*
1639 - Fixed timezone issues when formatting timestamps
1640 - Added support for node groups, available via ``gnt-group`` and other
1642 - Added out-of-band framework and management, see :doc:`design
1643 document <design-oob>`
1644 - Removed support for roman numbers from ``gnt-node list`` and
1645 ``gnt-instance list``.
1646 - Allowed modification of master network interface via ``gnt-cluster
1647 modify --master-netdev``
1648 - Accept offline secondaries while shutting down instance disks
1649 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1650 - Added support for multiple LVM volume groups
1651 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1653 - Added commands to list available fields:
1655 - ``gnt-node list-fields``
1656 - ``gnt-group list-fields``
1657 - ``gnt-instance list-fields``
1659 - Updated documentation and man pages
1664 - Moved ``rapi_users`` file into separate directory, now named
1665 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1667 - Added new tool for running commands on many machines,
1668 ``tools/ganeti-listrunner``
1669 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1670 improving the ``gnt-instance console`` output
1671 - Allowed customisation of disk index separator at ``configure`` time
1672 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1673 - Added support for non-partitioned md disks in ``lvmstrap``
1674 - Added script to gracefully power off KVM instances
1675 - Split ``utils`` module into smaller parts
1676 - Changed query operations to return more detailed information, e.g.
1677 whether an information is unavailable due to an offline node. To use
1678 this new functionality, the LUXI call ``Query`` must be used. Field
1679 information is now stored by the master daemon and can be retrieved
1680 using ``QueryFields``. Instances, nodes and groups can also be queried
1681 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1682 exposed via RAPI). The following commands make use of this
1683 infrastructure change:
1685 - ``gnt-group list``
1686 - ``gnt-group list-fields``
1688 - ``gnt-node list-fields``
1689 - ``gnt-instance list``
1690 - ``gnt-instance list-fields``
1691 - ``gnt-debug locks``
1696 - New RAPI resources (see :doc:`rapi`):
1700 - ``/2/groups/[group_name]``
1701 - ``/2/groups/[group_name]/assign-nodes``
1702 - ``/2/groups/[group_name]/modify``
1703 - ``/2/groups/[group_name]/rename``
1704 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1708 - Implemented ``no_install`` for instance creation
1709 - Implemented OS parameters for instance reinstallation, allowing
1710 use of special settings on reinstallation (e.g. for preserving data)
1715 - Added IPv6 support in import/export
1716 - Pause DRBD synchronization while wiping disks on instance creation
1717 - Updated unittests and QA scripts
1718 - Improved network parameters passed to KVM
1719 - Converted man pages from docbook to reStructuredText
1725 *(Released Mon, 20 Dec 2010)*
1727 Released version 2.3.1~rc1 without any changes.
1733 *(Released Wed, 1 Dec 2010)*
1735 - impexpd: Disable OpenSSL compression in socat if possible (backport
1736 from master, commit e90739d625b, see :doc:`installation guide
1737 <install-quick>` for details)
1738 - Changed unittest coverage report to exclude test scripts
1739 - Added script to check version format
1745 *(Released Wed, 1 Dec 2010)*
1747 Released version 2.3.0~rc1 without any changes.
1753 *(Released Fri, 19 Nov 2010)*
1755 A number of bugfixes and documentation updates:
1757 - Update ganeti-os-interface documentation
1758 - Fixed a bug related to duplicate MACs or similar items which should be
1760 - Fix breakage in OS state modify
1761 - Reinstall instance: disallow offline secondaries (fixes bug related to
1762 OS changing but reinstall failing)
1763 - plus all the other fixes between 2.2.1 and 2.2.2
1769 *(Released Tue, 2 Nov 2010)*
1771 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1772 - Fixed master failover race with watcher
1773 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1775 - Accept previously stopped instance for export with instance removal
1776 - Simplify and extend the environment variables for instance OS scripts
1777 - Added new node flags, ``master_capable`` and ``vm_capable``
1778 - Added optional instance disk wiping prior during allocation. This is a
1779 cluster-wide option and can be set/modified using
1780 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1781 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1782 :doc:`install-quick`
1783 - Added a new watcher option (``--ignore-pause``)
1784 - Added option to ignore offline node on instance start/stop
1785 (``--ignore-offline``)
1786 - Allow overriding OS parameters with ``gnt-instance reinstall``
1787 - Added ability to change node's secondary IP address using ``gnt-node
1789 - Implemented privilege separation for all daemons except
1790 ``ganeti-noded``, see ``configure`` options
1791 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1793 - Implemented job priorities (see ``ganeti(7)`` manpage)
1794 - Ignore failures while shutting down instances during failover from
1796 - Exit daemon's bootstrap process only once daemon is ready
1797 - Export more information via ``LUInstanceQuery``/remote API
1798 - Improved documentation, QA and unittests
1799 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1800 restart if the file was created or changed
1801 - Added LUXI protocol version sent with each request and response,
1802 allowing detection of server/client mismatches
1803 - Moved the Python scripts among gnt-* and ganeti-* into modules
1804 - Moved all code related to setting up SSH to an external script,
1806 - Infrastructure changes for node group support in future versions
1812 *(Released Fri, 19 Nov 2010)*
1814 A few small bugs fixed, and some improvements to the build system:
1816 - Fix documentation regarding conversion to drbd
1817 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1819 - Fix error handling in node modify with multiple changes
1820 - Allow remote imports without checked names
1826 *(Released Tue, 19 Oct 2010)*
1828 - Disable SSL session ID cache in RPC client
1834 *(Released Thu, 14 Oct 2010)*
1836 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1837 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1839 - Fix problems with interaction between Curl and Python's HTTP server,
1840 resulting in increased speed in many RPC calls
1841 - Improve our release script to prevent breakage with older aclocal and
1848 *(Released Thu, 7 Oct 2010)*
1850 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1851 value retrieved from master
1852 - Added support for blacklisted or hidden OS definitions
1853 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1854 - Added support for -mem-path in KVM hypervisor abstraction layer
1855 - Allow overriding instance parameters in tool for inter-cluster
1856 instance moves (``tools/move-instance``)
1857 - Improved opcode summaries (e.g. in ``gnt-job list``)
1858 - Improve consistency of OS listing by sorting it
1859 - Documentation updates
1865 *(Released Fri, 8 Oct 2010)*
1867 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1873 *(Released Mon, 4 Oct 2010)*
1875 - Fixed regression in ``gnt-instance rename``
1881 *(Released Wed, 22 Sep 2010)*
1883 - Fixed OS_VARIANT variable for OS scripts
1884 - Fixed cluster tag operations via RAPI
1885 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1886 - Disabled RAPI CA checks in watcher
1892 *(Released Mon, 23 Aug 2010)*
1894 - Support DRBD versions of the format "a.b.c.d"
1896 - Re-introduce support for usage from multiple threads in RAPI client
1897 - Instance renames and modify via RAPI
1898 - Work around race condition between processing and archival in job
1900 - Mark opcodes following failed one as failed, too
1901 - Job field ``lock_status`` was removed due to difficulties making it
1902 work with the changed job queue in Ganeti 2.2; a better way to monitor
1903 locks is expected for a later 2.2.x release
1904 - Fixed dry-run behaviour with many commands
1905 - Support ``ssh-agent`` again when adding nodes
1906 - Many additional bugfixes
1912 *(Released Fri, 30 Jul 2010)*
1914 Important change: the internal RPC mechanism between Ganeti nodes has
1915 changed from using a home-grown http library (based on the Python base
1916 libraries) to use the PycURL library. This requires that PycURL is
1917 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1918 against GnuTLS by default. cURL's support for GnuTLS had known issues
1919 before cURL 7.21.0 and we recommend using the latest cURL release or
1920 linking against OpenSSL. Most other distributions already link PycURL
1921 and cURL against OpenSSL. The command::
1923 python -c 'import pycurl; print pycurl.version'
1925 can be used to determine the libraries PycURL and cURL are linked
1928 Other significant changes:
1930 - Rewrote much of the internals of the job queue, in order to achieve
1931 better parallelism; this decouples job query operations from the job
1932 processing, and it should allow much nicer behaviour of the master
1933 daemon under load, and it also has uncovered some long-standing bugs
1934 related to the job serialisation (now fixed)
1935 - Added a default iallocator setting to the cluster parameters,
1936 eliminating the need to always pass nodes or an iallocator for
1937 operations that require selection of new node(s)
1938 - Added experimental support for the LXC virtualization method
1939 - Added support for OS parameters, which allows the installation of
1940 instances to pass parameter to OS scripts in order to customise the
1942 - Added a hypervisor parameter controlling the migration type (live or
1943 non-live), since hypervisors have various levels of reliability; this
1944 has renamed the 'live' parameter to 'mode'
1945 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1946 logical volumes, meaning that cluster verify will ignore them and not
1947 flag their presence as errors
1948 - The watcher will now reset the error count for failed instances after
1949 8 hours, thus allowing self-healing if the problem that caused the
1950 instances to be down/fail to start has cleared in the meantime
1951 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1952 check for, and warn, if the drbd module parameter ``usermode_helper``
1953 is not consistent with the cluster-wide setting; this is needed to
1954 make diagnose easier of failed drbd creations
1955 - Started adding base IPv6 support, but this is not yet
1956 enabled/available for use
1957 - Rename operations (cluster, instance) will now return the new name,
1958 which is especially useful if a short name was passed in
1959 - Added support for instance migration in RAPI
1960 - Added a tool to pre-configure nodes for the SSH setup, before joining
1961 them to the cluster; this will allow in the future a simplified model
1962 for node joining (but not yet fully enabled in 2.2); this needs the
1963 paramiko python library
1964 - Fixed handling of name-resolving errors
1965 - Fixed consistency of job results on the error path
1966 - Fixed master-failover race condition when executed multiple times in
1968 - Fixed many bugs related to the job queue (mostly introduced during the
1969 2.2 development cycle, so not all are impacting 2.1)
1970 - Fixed instance migration with missing disk symlinks
1971 - Fixed handling of unknown jobs in ``gnt-job archive``
1972 - And many other small fixes/improvements
1976 - Enhanced both the unittest and the QA coverage
1977 - Switched the opcode validation to a generic model, and extended the
1978 validation to all opcode parameters
1979 - Changed more parts of the code that write shell scripts to use the
1981 - Switched the master daemon to use the asyncore library for the Luxi
1988 *(Released Thu, 17 Jun 2010)*
1990 - Added tool (``move-instance``) and infrastructure to move instances
1991 between separate clusters (see :doc:`separate documentation
1992 <move-instance>` and :doc:`design document <design-2.2>`)
1993 - Added per-request RPC timeout
1994 - RAPI now requires a Content-Type header for requests with a body (e.g.
1995 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1996 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1997 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1999 - Implemented initial support for running Ganeti daemons as separate
2000 users, see configure-time flags ``--with-user-prefix`` and
2001 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2003 - Instances can be removed after export (``gnt-backup export
2004 --remove-instance``)
2005 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2006 key (instead of 1024 bit)
2007 - Added new cluster configuration file for cluster domain secret
2008 - Import/export now use SSL instead of SSH
2009 - Added support for showing estimated time when exporting an instance,
2010 see the ``ganeti-os-interface(7)`` manpage and look for
2017 *(Released Tue, 16 Nov 2010)*
2019 Some more bugfixes. Unless critical bugs occur, this will be the last
2022 - Fix case of MAC special-values
2023 - Fix mac checker regex
2024 - backend: Fix typo causing "out of range" error
2025 - Add missing --units in gnt-instance list man page
2031 *(Released Tue, 24 Aug 2010)*
2034 - Don't ignore secondary node silently on non-mirrored disk templates
2036 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2037 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2038 - Properly document the usb_mouse parameter
2039 - Fix path in ganeti-rapi(8) (issue 116)
2040 - Adjust error message when the ganeti user's .ssh directory is
2042 - Add same-node-check when changing the disk template to drbd
2048 *(Released Fri, 16 Jul 2010)*
2051 - Add an option to only select some reboot types during qa/burnin.
2052 (on some hypervisors consequent reboots are not supported)
2053 - Fix infrequent race condition in master failover. Sometimes the old
2054 master ip address would be still detected as up for a short time
2055 after it was removed, causing failover to fail.
2056 - Decrease mlockall warnings when the ctypes module is missing. On
2057 Python 2.4 we support running even if no ctypes module is installed,
2058 but we were too verbose about this issue.
2059 - Fix building on old distributions, on which man doesn't have a
2061 - Fix RAPI not to ignore the MAC address on instance creation
2062 - Implement the old instance creation format in the RAPI client.
2068 *(Released Thu, 01 Jul 2010)*
2070 A small bugfix release:
2071 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2072 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2073 a non-existing option
2074 - Fix instance create: the --force-variant option was ignored
2075 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2076 - Fix modify node storage with non-FQDN arguments
2077 - Fix RAPI client to authenticate under Python 2.6 when used
2078 for more than 5 requests needing authentication
2079 - Fix gnt-instance modify -t (storage) giving a wrong error message
2080 when converting a non-shutdown drbd instance to plain
2086 *(Released Fri, 18 Jun 2010)*
2088 A small bugfix release:
2090 - Fix live migration of KVM instances started with older Ganeti
2091 versions which had fewer hypervisor parameters
2092 - Fix gnt-instance grow-disk on down instances
2093 - Fix an error-reporting bug during instance migration
2094 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2095 silently ignoring broken ones
2096 - Fix an RPC error reporting bug affecting, for example, RAPI client
2098 - Fix bug triggered by different API version os-es on different nodes
2099 - Fix a bug in instance startup with custom hvparams: OS level
2100 parameters would fail to be applied.
2101 - Fix the RAPI client under Python 2.6 (but more work is needed to
2102 make it work completely well with OpenSSL)
2103 - Fix handling of errors when resolving names from DNS
2109 *(Released Thu, 3 Jun 2010)*
2111 A medium sized development cycle. Some new features, and some
2112 fixes/small improvements/cleanups.
2114 Significant features
2115 ~~~~~~~~~~~~~~~~~~~~
2117 The node deamon now tries to mlock itself into memory, unless the
2118 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2119 its logs, and falls back to console logging. This allows emergency
2120 features such as ``gnt-node powercycle`` to work even in the event of a
2121 broken node disk (tested offlining the disk hosting the node's
2122 filesystem and dropping its memory caches; don't try this at home)
2124 KVM: add vhost-net acceleration support. It can be tested with a new
2125 enough version of the kernel and of qemu-kvm.
2127 KVM: Add instance chrooting feature. If you use privilege dropping for
2128 your VMs you can also now force them to chroot to an empty directory,
2129 before starting the emulated guest.
2131 KVM: Add maximum migration bandwith and maximum downtime tweaking
2132 support (requires a new-enough version of qemu-kvm).
2134 Cluster verify will now warn if the master node doesn't have the master
2135 ip configured on it.
2137 Add a new (incompatible) instance creation request format to RAPI which
2138 supports all parameters (previously only a subset was supported, and it
2139 wasn't possible to extend the old format to accomodate all the new
2140 features. The old format is still supported, and a client can check for
2141 this feature, before using it, by checking for its presence in the
2142 ``features`` RAPI resource.
2144 Now with ancient latin support. Try it passing the ``--roman`` option to
2145 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2146 (requires the python-roman module to be installed, in order to work).
2151 As usual many internal code refactorings, documentation updates, and
2154 - Lots of improvements and cleanups to the experimental Remote API
2155 (RAPI) client library.
2156 - A new unit test suite for the core daemon libraries.
2157 - A fix to creating missing directories makes sure the umask is not
2158 applied anymore. This enforces the same directory permissions
2160 - Better handling terminating daemons with ctrl+c (used when running
2161 them in debugging mode).
2162 - Fix a race condition in live migrating a KVM instance, when stat()
2163 on the old proc status file returned EINVAL, which is an unexpected
2165 - Fixed manpage checking with newer man and utf-8 charachters. But now
2166 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2172 *(Released Fri, 7 May 2010)*
2174 Fix a bug which prevented untagged KVM instances from starting.
2180 *(Released Fri, 7 May 2010)*
2182 Another release with a long development cycle, during which many
2183 different features were added.
2185 Significant features
2186 ~~~~~~~~~~~~~~~~~~~~
2188 The KVM hypervisor now can run the individual instances as non-root, to
2189 reduce the impact of a VM being hijacked due to bugs in the
2190 hypervisor. It is possible to run all instances as a single (non-root)
2191 user, to manually specify a user for each instance, or to dynamically
2192 allocate a user out of a cluster-wide pool to each instance, with the
2193 guarantee that no two instances will run under the same user ID on any
2196 An experimental RAPI client library, that can be used standalone
2197 (without the other Ganeti libraries), is provided in the source tree as
2198 ``lib/rapi/client.py``. Note this client might change its interface in
2199 the future, as we iterate on its capabilities.
2201 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2202 replace the cluster's certificates and crypto keys. This might help in
2203 case they have been compromised, or have simply expired.
2205 A new disk option for instance creation has been added that allows one
2206 to "adopt" currently existing logical volumes, with data
2207 preservation. This should allow easier migration to Ganeti from
2208 unmanaged (or managed via other software) instances.
2210 Another disk improvement is the possibility to convert between redundant
2211 (DRBD) and plain (LVM) disk configuration for an instance. This should
2212 allow better scalability (starting with one node and growing the
2213 cluster, or shrinking a two-node cluster to one node).
2215 A new feature that could help with automated node failovers has been
2216 implemented: if a node sees itself as offline (by querying the master
2217 candidates), it will try to shutdown (hard) all instances and any active
2218 DRBD devices. This reduces the risk of duplicate instances if an
2219 external script automatically failovers the instances on such nodes. To
2220 enable this, the cluster parameter ``maintain_node_health`` should be
2221 enabled; in the future this option (per the name) will enable other
2222 automatic maintenance features.
2224 Instance export/import now will reuse the original instance
2225 specifications for all parameters; that means exporting an instance,
2226 deleting it and the importing it back should give an almost identical
2227 instance. Note that the default import behaviour has changed from
2228 before, where it created only one NIC; now it recreates the original
2231 Cluster verify has added a few new checks: SSL certificates validity,
2232 /etc/hosts consistency across the cluster, etc.
2237 As usual, many internal changes were done, documentation fixes,
2240 - Fixed cluster initialization with disabled cluster storage (regression
2241 introduced in 2.1.1)
2242 - File-based storage supports growing the disks
2243 - Fixed behaviour of node role changes
2244 - Fixed cluster verify for some corner cases, plus a general rewrite of
2245 cluster verify to allow future extension with more checks
2246 - Fixed log spamming by watcher and node daemon (regression introduced
2248 - Fixed possible validation issues when changing the list of enabled
2250 - Fixed cleanup of /etc/hosts during node removal
2251 - Fixed RAPI response for invalid methods
2252 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2253 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2254 from ide disks, etc.)
2255 - Allow OS changes without re-installation (to record a changed OS
2256 outside of Ganeti, or to allow OS renames)
2257 - Allow instance creation without OS installation (useful for example if
2258 the OS will be installed manually, or restored from a backup not in
2260 - Implemented option to make cluster ``copyfile`` use the replication
2262 - Added list of enabled hypervisors to ssconf (possibly useful for
2264 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2266 - A partial form of node re-IP is possible via node readd, which now
2267 allows changed node primary IP
2268 - Command line utilities now show an informational message if the job is
2270 - The logs of the master daemon now show the PID/UID/GID of the
2277 *(Released Fri, 12 Mar 2010)*
2279 During the 2.1.0 long release candidate cycle, a lot of improvements and
2280 changes have accumulated with were released later as 2.1.1.
2285 The node evacuate command (``gnt-node evacuate``) was significantly
2286 rewritten, and as such the IAllocator protocol was changed - a new
2287 request type has been added. This unfortunate change during a stable
2288 series is designed to improve performance of node evacuations; on
2289 clusters with more than about five nodes and which are well-balanced,
2290 evacuation should proceed in parallel for all instances of the node
2291 being evacuated. As such, any existing IAllocator scripts need to be
2292 updated, otherwise the above command will fail due to the unknown
2293 request. The provided "dumb" allocator has not been updated; but the
2294 ganeti-htools package supports the new protocol since version 0.2.4.
2296 Another important change is increased validation of node and instance
2297 names. This might create problems in special cases, if invalid host
2298 names are being used.
2300 Also, a new layer of hypervisor parameters has been added, that sits at
2301 OS level between the cluster defaults and the instance ones. This allows
2302 customisation of virtualization parameters depending on the installed
2303 OS. For example instances with OS 'X' may have a different KVM kernel
2304 (or any other parameter) than the cluster defaults. This is intended to
2305 help managing a multiple OSes on the same cluster, without manual
2306 modification of each instance's parameters.
2308 A tool for merging clusters, ``cluster-merge``, has been added in the
2309 tools sub-directory.
2314 - Improved the int/float conversions that should make the code more
2315 robust in face of errors from the node daemons
2316 - Fixed the remove node code in case of internal configuration errors
2317 - Fixed the node daemon behaviour in face of inconsistent queue
2318 directory (e.g. read-only file-system where we can't open the files
2320 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2321 now it either aborts cleanly or, if given the new "auto_promote"
2322 parameter, will automatically promote other nodes as needed
2323 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2324 completely prevent Ganeti from working
2325 - Fixed bug for instance export when not all disks were successfully
2327 - Fixed behaviour of node add when the new node is slow in starting up
2329 - Fixed handling of signals in the LUXI client, which should improve
2330 behaviour of command-line scripts
2331 - Added checks for invalid node/instance names in the configuration (now
2332 flagged during cluster verify)
2333 - Fixed watcher behaviour for disk activation errors
2334 - Fixed two potentially endless loops in http library, which led to the
2335 RAPI daemon hanging and consuming 100% CPU in some cases
2336 - Fixed bug in RAPI daemon related to hashed passwords
2337 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2339 - Enhanced compatibility with non-Debian OSes, but not using absolute
2340 path in some commands and allowing customisation of the ssh
2341 configuration directory
2342 - Fixed possible future issue with new Python versions by abiding to the
2343 proper use of ``__slots__`` attribute on classes
2344 - Added checks that should prevent directory traversal attacks
2345 - Many documentation fixes based on feedback from users
2350 - Added an "early_release" more for instance replace disks and node
2351 evacuate, where we release locks earlier and thus allow higher
2352 parallelism within the cluster
2353 - Added watcher hooks, intended to allow the watcher to restart other
2354 daemons (e.g. from the ganeti-nbma project), but they can be used of
2355 course for any other purpose
2356 - Added a compile-time disable for DRBD barriers, to increase
2357 performance if the administrator trusts the power supply or the
2358 storage system to not lose writes
2359 - Added the option of using syslog for logging instead of, or in
2360 addition to, Ganeti's own log files
2361 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2362 can indeed boot from a paravirtual NIC
2363 - Added a generic debug level for many operations; while this is not
2364 used widely yet, it allows one to pass the debug value all the way to
2366 - Enhanced the hooks environment for instance moves (failovers,
2367 migrations) where the primary/secondary nodes changed during the
2368 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2369 - Enhanced data validations for many user-supplied values; one important
2370 item is the restrictions imposed on instance and node names, which
2371 might reject some (invalid) host names
2372 - Add a configure-time option to disable file-based storage, if it's not
2373 needed; this allows greater security separation between the master
2374 node and the other nodes from the point of view of the inter-node RPC
2376 - Added user notification in interactive tools if job is waiting in the
2377 job queue or trying to acquire locks
2378 - Added log messages when a job is waiting for locks
2379 - Added filtering by node tags in instance operations which admit
2380 multiple instances (start, stop, reboot, reinstall)
2381 - Added a new tool for cluster mergers, ``cluster-merge``
2382 - Parameters from command line which are of the form ``a=b,c=d`` can now
2383 use backslash escapes to pass in values which contain commas,
2384 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2386 - For KVM, the instance name is the first parameter passed to KVM, so
2387 that it's more visible in the process list
2393 *(Released Tue, 2 Mar 2010)*
2395 Ganeti 2.1 brings many improvements with it. Major changes:
2397 - Added infrastructure to ease automated disk repairs
2398 - Added new daemon to export configuration data in a cheaper way than
2399 using the remote API
2400 - Instance NICs can now be routed instead of being associated with a
2402 - Improved job locking logic to reduce impact of jobs acquiring multiple
2403 locks waiting for other long-running jobs
2405 In-depth implementation details can be found in the Ganeti 2.1 design
2411 - Added chroot hypervisor
2412 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2414 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2415 ``bootloader_path`` and ``bootloader_args``)
2416 - Added the ``use_localtime`` option for the xen-hvm and kvm
2417 hypervisors, and the default value for this has changed to false (in
2418 2.0 xen-hvm always enabled it)
2419 - Added luxi call to submit multiple jobs in one go
2420 - Added cluster initialization option to not modify ``/etc/hosts``
2422 - Added network interface parameters
2423 - Added dry run mode to some LUs
2424 - Added RAPI resources:
2426 - ``/2/instances/[instance_name]/info``
2427 - ``/2/instances/[instance_name]/replace-disks``
2428 - ``/2/nodes/[node_name]/evacuate``
2429 - ``/2/nodes/[node_name]/migrate``
2430 - ``/2/nodes/[node_name]/role``
2431 - ``/2/nodes/[node_name]/storage``
2432 - ``/2/nodes/[node_name]/storage/modify``
2433 - ``/2/nodes/[node_name]/storage/repair``
2435 - Added OpCodes to evacuate or migrate all instances on a node
2436 - Added new command to list storage elements on nodes (``gnt-node
2437 list-storage``) and modify them (``gnt-node modify-storage``)
2438 - Added new ssconf files with master candidate IP address
2439 (``ssconf_master_candidates_ips``), node primary IP address
2440 (``ssconf_node_primary_ips``) and node secondary IP address
2441 (``ssconf_node_secondary_ips``)
2442 - Added ``ganeti-confd`` and a client library to query the Ganeti
2443 configuration via UDP
2444 - Added ability to run hooks after cluster initialization and before
2446 - Added automatic mode for disk replace (``gnt-instance replace-disks
2448 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2449 after catastrophic data-loss
2450 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2452 - Added ``gnt-instance move`` command to move instances
2453 - Added ``gnt-cluster watcher`` command to control watcher
2454 - Added ``gnt-node powercycle`` command to powercycle nodes
2455 - Added new job status field ``lock_status``
2456 - Added parseable error codes to cluster verification (``gnt-cluster
2457 verify --error-codes``) and made output less verbose (use
2458 ``--verbose`` to restore previous behaviour)
2459 - Added UUIDs to the main config entities (cluster, nodes, instances)
2460 - Added support for OS variants
2461 - Added support for hashed passwords in the Ganeti remote API users file
2463 - Added option to specify maximum timeout on instance shutdown
2464 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2465 - Added new helper script to start and stop Ganeti daemons
2466 (``daemon-util``), with the intent to reduce the work necessary to
2467 adjust Ganeti for non-Debian distributions and to start/stop daemons
2469 - Added more unittests
2470 - Fixed critical bug in ganeti-masterd startup
2471 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2472 now customisable at the cluster level for both the KVM and Xen
2473 hypervisors using the new ``migration_port`` parameter
2474 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2475 reinstalling an instance
2476 - Allowed ``@`` in tag names
2477 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2478 - Many documentation updates
2479 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2480 - ``gnt-instance reinstall`` can now reinstall multiple instances
2481 - Updated many command line parameters
2482 - Introduced new OS API version 15
2483 - No longer support a default hypervisor
2484 - Treat virtual LVs as inexistent
2485 - Improved job locking logic to reduce lock contention
2486 - Match instance and node names case insensitively
2487 - Reimplemented bash completion script to be more complete
2494 *(Released Thu, 4 Feb 2010)*
2496 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2497 - Fix a string formatting bug
2498 - Improve safety of the code in some error paths
2499 - Improve data validation in the master of values returned from nodes
2505 *(Released Thu, 17 Dec 2009)*
2507 - Fix security issue due to missing validation of iallocator names; this
2508 allows local and remote execution of arbitrary executables
2509 - Fix failure of gnt-node list during instance removal
2510 - Ship the RAPI documentation in the archive
2516 *(Released Wed, 30 Sep 2009)*
2518 - Fixed many wrong messages
2519 - Fixed a few bugs related to the locking library
2520 - Fixed MAC checking at instance creation time
2521 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2522 - Fixed a few issues related to signal handling in both daemons and
2524 - Fixed the example startup script provided
2525 - Fixed insserv dependencies in the example startup script (patch from
2527 - Fixed handling of drained nodes in the iallocator framework
2528 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2530 - Fixed error related to invalid job IDs in job polling
2531 - Fixed job/opcode persistence on unclean master shutdown
2532 - Fixed handling of partial job processing after unclean master
2534 - Fixed error reporting from LUs, previously all errors were converted
2535 into execution errors
2536 - Fixed error reporting from burnin
2537 - Decreased significantly the memory usage of the job queue
2538 - Optimised slightly multi-job submission
2539 - Optimised slightly opcode loading
2540 - Backported the multi-job submit framework from the development
2541 branch; multi-instance start and stop should be faster
2542 - Added script to clean archived jobs after 21 days; this will reduce
2543 the size of the queue directory
2544 - Added some extra checks in disk size tracking
2545 - Added an example ethers hook script
2546 - Added a cluster parameter that prevents Ganeti from modifying of
2548 - Added more node information to RAPI responses
2549 - Added a ``gnt-job watch`` command that allows following the ouput of a
2551 - Added a bind-address option to ganeti-rapi
2552 - Added more checks to the configuration verify
2553 - Enhanced the burnin script such that some operations can be retried
2555 - Converted instance reinstall to multi-instance model
2561 *(Released Fri, 7 Aug 2009)*
2563 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2564 to allow using the pre-2.0.2 behaviour in activation, if any existing
2565 instances have mismatched disk sizes in the configuration
2566 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2567 any configuration mismatches for disk sizes
2568 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2569 failover to work on two-node clusters
2570 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2572 - Fixed detection of OS script errors in ``gnt-backup export``
2573 - Fixed exit code of ``gnt-backup export``
2579 *(Released Fri, 17 Jul 2009)*
2581 - Added experimental support for stripped logical volumes; this should
2582 enhance performance but comes with a higher complexity in the block
2583 device handling; stripping is only enabled when passing
2584 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2585 affected even in the non-stripped mode
2586 - Improved resiliency against transient failures at the end of DRBD
2587 resyncs, and in general of DRBD resync checks
2588 - Fixed a couple of issues with exports and snapshot errors
2589 - Fixed a couple of issues in instance listing
2590 - Added display of the disk size in ``gnt-instance info``
2591 - Fixed checking for valid OSes in instance creation
2592 - Fixed handling of the "vcpus" parameter in instance listing and in
2593 general of invalid parameters
2594 - Fixed http server library, and thus RAPI, to handle invalid
2595 username/password combinations correctly; this means that now they
2596 report unauthorized for queries too, not only for modifications,
2597 allowing earlier detect of configuration problems
2598 - Added a new "role" node list field, equivalent to the master/master
2599 candidate/drained/offline flags combinations
2600 - Fixed cluster modify and changes of candidate pool size
2601 - Fixed cluster verify error messages for wrong files on regular nodes
2602 - Fixed a couple of issues with node demotion from master candidate role
2603 - Fixed node readd issues
2604 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2605 - Added a new ``--no-voting`` option for masterfailover to fix failover
2606 on two-nodes clusters when the former master node is unreachable
2607 - Added instance reinstall over RAPI
2613 *(Released Tue, 16 Jun 2009)*
2615 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2616 allow re-adding the start in single-user option (regression from 1.2)
2617 - the watcher writes the instance status to a file, to allow monitoring
2618 to report the instance status (from the master) based on cached
2619 results of the watcher's queries; while this can get stale if the
2620 watcher is being locked due to other work on the cluster, this is
2621 still an improvement
2622 - the watcher now also restarts the node daemon and the rapi daemon if
2624 - fixed the watcher to handle full and drained queue cases
2625 - hooks export more instance data in the environment, which helps if
2626 hook scripts need to take action based on the instance's properties
2627 (no longer need to query back into ganeti)
2628 - instance failovers when the instance is stopped do not check for free
2629 RAM, so that failing over a stopped instance is possible in low memory
2631 - rapi uses queries for tags instead of jobs (for less job traffic), and
2632 for cluster tags it won't talk to masterd at all but read them from
2634 - a couple of error handling fixes in RAPI
2635 - drbd handling: improved the error handling of inconsistent disks after
2636 resync to reduce the frequency of "there are some degraded disks for
2637 this instance" messages
2638 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2639 error handling path called a wrong function name)
2645 *(Released Wed, 27 May 2009)*
2647 - no changes from rc5
2653 *(Released Wed, 20 May 2009)*
2655 - fix a couple of bugs (validation, argument checks)
2656 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2657 - some small improvements to RAPI and IAllocator
2658 - make watcher automatically start the master daemon if down
2664 *(Released Mon, 27 Apr 2009)*
2666 - change the OS list to not require locks; this helps with big clusters
2667 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2668 volume group is broken
2669 - ``gnt-instance info``, without any arguments, doesn't run for all
2670 instances anymore; either pass ``--all`` or pass the desired
2671 instances; this helps against mistakes on big clusters where listing
2672 the information for all instances takes a long time
2673 - miscellaneous doc and man pages fixes
2679 *(Released Wed, 8 Apr 2009)*
2681 - Change the internal locking model of some ``gnt-node`` commands, in
2682 order to reduce contention (and blocking of master daemon) when
2683 batching many creation/reinstall jobs
2684 - Fixes to Xen soft reboot
2685 - No longer build documentation at build time, instead distribute it in
2686 the archive, in order to reduce the need for the whole docbook/rst
2693 *(Released Fri, 27 Mar 2009)*
2695 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2696 - Fix watcher startup sequence, improves the behaviour of busy clusters
2697 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2698 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2699 bind address and other places
2700 - Some documentation fixes and updates
2706 *(Released Mon, 2 Mar 2009)*
2708 - More documentation updates, now all docs should be more-or-less
2710 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2712 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2714 - Fix an issue related to $libdir/run/ganeti and cluster creation
2720 *(Released Thu, 19 Feb 2009)*
2722 - Xen PVM and KVM have switched the default value for the instance root
2723 disk to the first partition on the first drive, instead of the whole
2724 drive; this means that the OS installation scripts must be changed
2726 - Man pages have been updated
2727 - RAPI has been switched by default to HTTPS, and the exported functions
2728 should all work correctly
2729 - RAPI v1 has been removed
2730 - Many improvements to the KVM hypervisor
2731 - Block device errors are now better reported
2732 - Many other bugfixes and small improvements
2738 *(Released Mon, 26 Jan 2009)*
2740 - Version 2 is a general rewrite of the code and therefore the
2741 differences are too many to list, see the design document for 2.0 in
2742 the ``doc/`` subdirectory for more details
2743 - In this beta version there is not yet a migration path from 1.2 (there
2744 will be one in the final 2.0 release)
2745 - A few significant changes are:
2747 - all commands are executed by a daemon (``ganeti-masterd``) and the
2748 various ``gnt-*`` commands are just front-ends to it
2749 - all the commands are entered into, and executed from a job queue,
2750 see the ``gnt-job(8)`` manpage
2751 - the RAPI daemon supports read-write operations, secured by basic
2752 HTTP authentication on top of HTTPS
2753 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2754 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2755 to migrate to DRBD 8 first while still running Ganeti 1.2)
2756 - DRBD devices are using statically allocated minor numbers, which
2757 will be assigned to existing instances during the migration process
2758 - there is support for both Xen PVM and Xen HVM instances running on
2760 - KVM virtualization is supported too
2761 - file-based storage has been implemented, which means that it is
2762 possible to run the cluster without LVM and DRBD storage, for
2763 example using a shared filesystem exported from shared storage (and
2764 still have live migration)
2770 *(Released Tue, 13 Jan 2009)*
2772 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2773 - Reuse the old instance mac address by default on instance import, if
2774 the instance name is the same.
2775 - Handle situations in which the node info rpc returns incomplete
2777 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2778 - Improved version of batcher:
2780 - state file support
2781 - instance mac address support
2782 - support for HVM clusters/instances
2784 - Add an option to show the number of cpu sockets and nodes in
2786 - Support OSes that handle more than one version of the OS api (but do
2787 not change the current API in any other way)
2788 - Fix ``gnt-node migrate``
2789 - ``gnt-debug`` man page
2790 - Fixes various more typos and small issues
2791 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2797 *(Released Wed, 24 Sep 2008)*
2799 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2800 type of disk exported to fully virtualized instances.
2801 - provide access to the serial console of HVM instances
2802 - instance auto_balance flag, set by default. If turned off it will
2803 avoid warnings on cluster verify if there is not enough memory to fail
2804 over an instance. in the future it will prevent automatically failing
2805 it over when we will support that.
2806 - batcher tool for instance creation, see ``tools/README.batcher``
2807 - ``gnt-instance reinstall --select-os`` to interactively select a new
2808 operating system when reinstalling an instance.
2809 - when changing the memory amount on instance modify a check has been
2810 added that the instance will be able to start. also warnings are
2811 emitted if the instance will not be able to fail over, if auto_balance
2813 - documentation fixes
2814 - sync fields between ``gnt-instance list/modify/add/import``
2815 - fix a race condition in drbd when the sync speed was set after giving
2816 the device a remote peer.
2822 *(Released Tue, 22 Jul 2008)*
2824 - note: the allowed size and number of tags per object were reduced
2825 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2826 - fixed twisted 8.x compatibility
2827 - fixed ``gnt-instance replace-disks`` with iallocator
2828 - add TCP keepalives on twisted connections to detect restarted nodes
2829 - disk increase support, see ``gnt-instance grow-disk``
2830 - implement bulk node/instance query for RAPI
2831 - add tags in node/instance listing (optional)
2832 - experimental migration (and live migration) support, read the man page
2833 for ``gnt-instance migrate``
2834 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2835 has some small improvements in handling its state file
2841 *(Released Fri, 13 Jun 2008)*
2843 - Experimental readonly, REST-based remote API implementation;
2844 automatically started on master node, TCP port 5080, if enabled by
2845 ``--enable-rapi`` parameter to configure script.
2846 - Instance allocator support. Add and import instance accept a
2847 ``--iallocator`` parameter, and call that instance allocator to decide
2848 which node to use for the instance. The iallocator document describes
2849 what's expected from an allocator script.
2850 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2851 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2852 node is lost there is still enough memory to fail over the instances
2854 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2855 ``gnt-cluster verify``, to check for site-specific compliance. All the
2856 hooks will run, and their output, if any, will be displayed. Any
2857 failing hook will make the verification return an error value.
2858 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2859 primary and secondary interfaces
2860 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2861 that is still declared as part of the cluster and has failed.
2862 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2863 specifying output fields, that just adds the chosen fields to the
2865 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2866 export from the filesystem.
2867 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2868 have been added. Using them you can enable/disable acpi and pae
2869 support, and specify a path for a cd image to be exported to the
2870 instance. These parameters as the name suggest only work on HVM
2872 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2873 PAE support will be set to the previously hardcoded values, but the
2874 (previously hardcoded) path to the CDROM ISO image will be unset and
2875 if required, needs to be set manually with ``gnt-instance modify``
2877 - The address to which an instance's VNC console is bound is now
2878 selectable per-instance, rather than being cluster wide. Of course
2879 this only applies to instances controlled via VNC, so currently just
2880 applies to HVM clusters.
2886 *(Released Mon, 18 Feb 2008)*
2888 - more tweaks to the disk activation code (especially helpful for DRBD)
2889 - change the default ``gnt-instance list`` output format, now there is
2890 one combined status field (see the manpage for the exact values this
2892 - some more fixes for the mac export to hooks change
2893 - make Ganeti not break with DRBD 8.2.x (which changed the version
2894 format in ``/proc/drbd``) (issue 24)
2895 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2896 template, allowing migration from DRBD0.7+MD to DRBD8
2902 *(Released Wed, 30 Jan 2008)*
2904 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2906 - add command aliases infrastructure and a few aliases
2907 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2908 man pages and the ``--help`` option of ``gnt-node
2909 list``/``gnt-instance list``
2910 - fix ``gnt-backup list`` with down nodes (issue 21)
2911 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2912 - fix the dist archive and add a check for including svn/git files in
2914 - some developer-related changes: improve the burnin and the QA suite,
2915 add an upload script for testing during development
2921 *(Released Wed, 16 Jan 2008)*
2923 - experimental HVM support, read the install document, section
2924 "Initializing the cluster"
2925 - allow for the PVM hypervisor per-instance kernel and initrd paths
2926 - add a new command ``gnt-cluster verify-disks`` which uses a new
2927 algorithm to improve the reconnection of the DRBD pairs if the device
2928 on the secondary node has gone away
2929 - make logical volume code auto-activate LVs at disk activation time
2930 - slightly improve the speed of activating disks
2931 - allow specification of the MAC address at instance creation time, and
2932 changing it later via ``gnt-instance modify``
2933 - fix handling of external commands that generate lots of output on
2935 - update documentation with regard to minimum version of DRBD8 supported
2941 *(Released Tue, 4 Dec 2007)*
2943 - Log the ``xm create`` output to the node daemon log on failure (to
2944 help diagnosing the error)
2945 - In debug mode, log all external commands output if failed to the logs
2946 - Change parsing of lvm commands to ignore stderr
2952 *(Released Wed, 28 Nov 2007)*
2954 - Another round of updates to the DRBD 8 code to deal with more failures
2955 in the replace secondary node operation
2956 - Some more logging of failures in disk operations (lvm, drbd)
2957 - A few documentation updates
2964 *(Released Tue, 13 Nov 2007)*
2966 - Change configuration file format from Python's Pickle to JSON.
2967 Upgrading is possible using the cfgupgrade utility.
2968 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2969 faster replace disks and is more stable (DRBD 8 has many improvements
2970 compared to DRBD 0.7)
2971 - Added command line tags support (see man pages for ``gnt-instance``,
2972 ``gnt-node``, ``gnt-cluster``)
2973 - Added instance rename support
2974 - Added multi-instance startup/shutdown
2975 - Added cluster rename support
2976 - Added ``gnt-node evacuate`` to simplify some node operations
2977 - Added instance reboot operation that can speedup reboot as compared to
2979 - Soften the requirement that hostnames are in FQDN format
2980 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2982 - Removed dependency on debian's patched fping that uses the
2983 non-standard ``-S`` option
2984 - Now the OS definitions are searched for in multiple, configurable
2985 paths (easier for distros to package)
2986 - Some changes to the hooks infrastructure (especially the new
2987 post-configuration update hook)
2988 - Other small bugfixes
2990 .. vim: set textwidth=72 syntax=rst :