10 Incompatible/important changes
11 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
13 - Adding disks with 'gnt-instance modify' now waits for the disks to sync per
14 default. Specify --no-wait-for-sync to override this behavior.
15 - The Ganeti python code now adheres to a private-module layout. In particular,
16 the module 'ganeti' is no longer in the python search path.
20 - KVM hypervisors can now access RBD storage directly without having to
21 go through a block device.
27 *(Released Tue, 1 Oct 2013)*
29 Incompatible/important changes
30 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
32 - hroller now also plans for capacity to move non-redundant instances off
33 any node to be rebooted; the old behavior of completely ignoring any
34 non-redundant instances can be restored by adding the --ignore-non-redundant
36 - The cluster option '--no-lvm-storage' was removed in favor of the new option
37 '--enabled-disk-templates'.
38 - On instance creation, disk templates no longer need to be specified
39 with '-t'. The default disk template will be taken from the list of
40 enabled disk templates.
41 - The monitoring daemon is now running as root, in order to be able to collect
42 information only available to root (such as the state of Xen instances).
43 - The ConfD client is now IPv6 compatible.
44 - File and shared file storage is no longer dis/enabled at configure time,
45 but using the option '--enabled-disk-templates' at cluster initialization and
47 - The default directories for file and shared file storage are not anymore
48 specified at configure time, but taken from the cluster's configuration.
49 They can be set at cluster initialization and modification with
50 '--file-storage-dir' and '--shared-file-storage-dir'.
51 - Cluster verification now includes stricter checks regarding the
52 default file and shared file storage directories. It now checks that
53 the directories are explicitely allowed in the 'file-storage-paths' file and
54 that the directories exist on all nodes.
55 - The list of allowed disk templates in the instance policy and the list
56 of cluster-wide enabled disk templates is now checked for consistency
57 on cluster or group modification. On cluster initialization, the ipolicy
58 disk templates are ensured to be a subset of the cluster-wide enabled
64 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
65 the correct command syntax. It is possible to use different DRBD versions
66 on different nodes as long as they are compatible to each other. This
67 enables rolling upgrades of DRBD with no downtime. As permanent operation
68 of different DRBD versions within a node group is discouraged,
69 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
70 - New "inst-status-xen" data collector for the monitoring daemon, providing
71 information about the state of the xen instances on the nodes.
72 - New "lv" data collector for the monitoring daemon, collecting data about the
73 logical volumes on the nodes, and pairing them with the name of the instances
75 - New "diskstats" data collector, collecting the data from /proc/diskstats and
76 presenting them over the monitoring daemon interface.
77 - The ConfD client is now IPv6 compatible.
81 The following new dependencies have been added.
85 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
86 for the unit tests (and only used for testing).
90 - ``hslogger`` (http://software.complete.org/hslogger) is now always
91 required, even if confd is not enabled.
97 - update of the documentation, in particular installation instructions
98 - merging of LD_* constants into DT_* constants
99 - python style changes to be compatible with newer versions of pylint
105 *(Released Thu, 29 Aug 2013)*
107 This was the first beta release of the 2.9 series. All important changes
108 are listed in the latest 2.9 entry.
114 *(Released Mon, 30 Sep 2013)*
116 Incompatible/important changes
117 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
119 - Instance policy can contain multiple instance specs, as described in
120 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
121 <design-partitioned>`. As a consequence, it's not possible to partially change
122 or override instance specs. Bounding specs (min and max) can be specified as a
123 whole using the new option ``--ipolicy-bounds-specs``, while standard
124 specs use the new option ``--ipolicy-std-specs``.
125 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
126 gnt-instance is a valid YAML object.
127 - hail now honors network restrictions when allocating nodes. This led to an
128 update of the IAllocator protocol. See the IAllocator documentation for
130 - confd now only answers static configuration request over the network. luxid
131 was extracted, listens on the local LUXI socket and responds to live queries.
132 This allows finer grained permissions if using separate users.
137 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
138 to always require authentication, ``--require-authentication``. It can
139 be specified in ``$sysconfdir/default/ganeti``.
140 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
141 be used to manage the disk templates to be used by instances in the cluster.
142 Initially, it will be set to a list that includes plain, drbd, if they were
143 enabled by specifying a volume group name, and file and sharedfile, if those
144 were enabled at configure time. Additionally, it will include all disk
145 templates that are currently used by instances. The order of disk templates
146 will be based on Ganeti's history of supporting them. In the future, the
147 first entry of the list will be used as a default disk template on instance
149 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
150 configuration back to the previous stable version.
151 - Disk templates in group ipolicy can be restored to the default value.
152 - Initial support for diskless instances and virtual clusters in QA.
153 - More QA and unit tests for instance policies.
154 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
155 describing why the opcode itself was executed.
156 - The monitoring daemon is now available. It allows users to query the cluster
157 for obtaining information about the status of the system. The daemon is only
158 responsible for providing the information over the network: the actual data
159 gathering is performed by data collectors (currently, only the DRBD status
160 collector is available).
161 - In order to help developers work on Ganeti, a new script
162 (``devel/build_chroot``) is provided, for building a chroot that contains all
163 the required development libraries and tools for compiling Ganeti on a Debian
165 - A new tool, ``harep``, for performing self-repair and recreation of instances
166 in Ganeti has been added.
167 - Split queries are enabled for tags, network, exports, cluster info, groups,
169 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
170 It prints the command line to set the current policies, to ease
172 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
173 devices for KVM virtio-net interfaces will get created with VNET_HDR
174 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
175 virtio-net interfaces, which prevents host kernel tainting and log
176 flooding, when dealing with broken or malicious virtio-net drivers.
177 It's set to true by default.
178 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
180 - Support 'viridian' parameter in Xen HVM
181 - Support DSA SSH keys in bootstrap
182 - To simplify the work of packaging frameworks that want to add the needed users
183 and groups in a split-user setup themselves, at build time three files in
184 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
185 the groups to be generated, the ``users`` file contains, one per line, the
186 users to be generated, optionally followed by their primary group, where
187 important. The ``groupmemberships`` file contains, one per line, additional
188 user-group membership relations that need to be established. The syntax of
189 these files will remain stable in all future versions.
194 The following new dependencies have been added:
197 - The ``curl`` library is not optional anymore for compiling the Haskell code.
198 - ``snap-server`` library (if monitoring is enabled).
201 - The minimum Python version needed to run Ganeti is now 2.6.
202 - ``yaml`` library (only for running the QA).
206 - Perform proper cleanup on termination of Haskell daemons
207 - Fix corner-case in handling of remaining retry time
213 *(Released Tue, 17 Sep 2013)*
215 - To simplify the work of packaging frameworks that want to add the needed users
216 and groups in a split-user setup themselves, at build time three files in
217 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
218 the groups to be generated, the ``users`` file contains, one per line, the
219 users to be generated, optionally followed by their primary group, where
220 important. The ``groupmemberships`` file contains, one per line, additional
221 user-group membership relations that need to be established. The syntax of
222 these files will remain stable in all future versions.
223 - Add a default to file-driver when unspecified over RAPI (Issue 571)
224 - Mark the DSA host pubkey as optional, and remove it during config downgrade
226 - Some documentation fixes
232 *(Released Tue, 27 Aug 2013)*
234 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
236 - Support 'viridian' parameter in Xen HVM (Issue 233)
237 - Include VCS version in ``gnt-cluster version``
238 - Support DSA SSH keys in bootstrap (Issue 338)
239 - Fix batch creation of instances
240 - Use FQDN to check master node status (Issue 551)
241 - Make the DRBD collector more failure-resilient
247 *(Released Fri, 2 Aug 2013)*
249 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
251 - Fix upgrading/downgrading from 2.7
252 - Increase maximum RAPI message size
253 - Documentation updates
254 - Split ``confd`` between ``luxid`` and ``confd``
255 - Merge 2.7 series up to the 2.7.1 release
256 - Allow the ``modify_etc_hosts`` option to be changed
257 - Add better debugging for ``luxid`` queries
258 - Expose bulk parameter for GetJobs in RAPI client
259 - Expose missing ``network`` fields in RAPI
260 - Add some ``cluster verify`` tests
261 - Some unittest fixes
262 - Fix a malfunction in ``hspace``'s tiered allocation
263 - Fix query compatibility between haskell and python implementations
264 - Add the ``vnet_hdr`` HV parameter for KVM
265 - Add ``--cleanup`` to instance failover
266 - Change the connected groups format in ``gnt-network info`` output; it
267 was previously displayed as a raw list by mistake. (Merged from 2.7)
273 *(Released Mon, 24 Jun 2013)*
275 This was the first beta release of the 2.8 series. All important changes
276 are listed in the latest 2.8 entry.
282 *(Released Thu, 26 Sep 2013)*
284 - Change the connected groups format in ``gnt-network info`` output; it
285 was previously displayed as a raw list by mistake
286 - Check disk template in right dict when copying
287 - Support multi-instance allocs without iallocator
288 - Fix some errors in the documentation
289 - Fix formatting of tuple in an error message
295 *(Released Thu, 25 Jul 2013)*
297 - Add logrotate functionality in daemon-util
298 - Add logrotate example file
299 - Add missing fields to network queries over rapi
300 - Fix network object timestamps
301 - Add support for querying network timestamps
302 - Fix a typo in the example crontab
303 - Fix a documentation typo
309 *(Released Thu, 04 Jul 2013)*
311 Incompatible/important changes
312 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
314 - Instance policies for disk size were documented to be on a per-disk
315 basis, but hail applied them to the sum of all disks. This has been
317 - ``hbal`` will now exit with status 0 if, during job execution over
318 LUXI, early exit has been requested and all jobs are successful;
319 before, exit status 1 was used, which cannot be differentiated from
321 - Compatibility with newer versions of rbd has been fixed
322 - ``gnt-instance batch-create`` has been changed to use the bulk create
323 opcode from Ganeti. This lead to incompatible changes in the format of
324 the JSON file. It's now not a custom dict anymore but a dict
325 compatible with the ``OpInstanceCreate`` opcode.
326 - Parent directories for file storage need to be listed in
327 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
328 write the file automatically based on old configuration values, but it
329 can not distribute it across all nodes and the file contents should be
330 verified. Use ``gnt-cluster copyfile
331 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
332 upgraded. The reason for requiring this list of paths now is that
333 before it would have been possible to inject new paths via RPC,
334 allowing files to be created in arbitrary locations. The RPC protocol
335 is protected using SSL/X.509 certificates, but as a design principle
336 Ganeti does not permit arbitrary paths to be passed.
337 - The parsing of the variants file for OSes (see
338 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
339 lines and comment lines (starting with ``#``) are ignored for better
341 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
342 longer available. ``gnt-node add`` now invokes a new tool on the
343 destination node, named ``prepare-node-join``, to configure the SSH
344 daemon. Paramiko is no longer necessary to configure nodes' SSH
345 daemons via ``gnt-node add``.
346 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
347 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
348 the flag is not reset after a master failover.
349 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
350 is recommended. 2.6 will be mandatory from the 2.8 series.
356 - New network management functionality to support automatic allocation
357 of IP addresses and managing of network parameters. See
358 :manpage:`gnt-network(8)` for more details.
359 - New external storage backend, to allow managing arbitrary storage
360 systems external to the cluster. See
361 :manpage:`ganeti-extstorage-interface(7)`.
362 - New ``exclusive-storage`` node parameter added, restricted to
363 nodegroup level. When it's set to true, physical disks are assigned in
364 an exclusive fashion to instances, as documented in :doc:`Partitioned
365 Ganeti <design-partitioned>`. Currently, only instances using the
366 ``plain`` disk template are supported.
367 - The KVM hypervisor has been updated with many new hypervisor
368 parameters, including a generic one for passing arbitrary command line
369 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
370 compatible up to qemu 1.4.
371 - A new tool, called ``mon-collector``, is the stand-alone executor of
372 the data collectors for a monitoring system. As of this version, it
373 just includes the DRBD data collector, that can be executed by calling
374 ``mon-collector`` using the ``drbd`` parameter. See
375 :manpage:`mon-collector(7)`.
376 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
377 for RAPI users. It allows granting permissions to query for
378 information to a specific user without giving
379 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
380 - A new tool named ``node-cleanup`` has been added. It cleans remains of
381 a cluster from a machine by stopping all daemons, removing
382 certificates and ssconf files. Unless the ``--no-backup`` option is
383 given, copies of the certificates are made.
384 - Instance creations now support the use of opportunistic locking,
385 potentially speeding up the (parallel) creation of multiple instances.
386 This feature is currently only available via the :doc:`RAPI
387 <rapi>` interface and when an instance allocator is used. If the
388 ``opportunistic_locking`` parameter is set the opcode will try to
389 acquire as many locks as possible, but will not wait for any locks
390 held by other opcodes. If not enough resources can be found to
391 allocate the instance, the temporary error code
392 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
393 retried thereafter, with or without opportunistic locking.
394 - New experimental linux-ha resource scripts.
395 - Restricted-commands support: ganeti can now be asked (via command line
396 or rapi) to perform commands on a node. These are passed via ganeti
397 RPC rather than ssh. This functionality is restricted to commands
398 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
399 reasons. The file is not copied automatically.
405 - Diskless instances are now externally mirrored (Issue 237). This for
406 now has only been tested in conjunction with explicit target nodes for
408 - Queries not needing locks or RPC access to the node can now be
409 performed by the confd daemon, making them independent from jobs, and
410 thus faster to execute. This is selectable at configure time.
411 - The functionality for allocating multiple instances at once has been
412 overhauled and is now also available through :doc:`RAPI <rapi>`.
414 There are no significant changes from version 2.7.0~rc3.
420 *(Released Tue, 25 Jun 2013)*
422 - Fix permissions on the confd query socket (Issue 477)
423 - Fix permissions on the job archive dir (Issue 498)
424 - Fix handling of an internal exception in replace-disks (Issue 472)
425 - Fix gnt-node info handling of shortened names (Issue 497)
426 - Fix gnt-instance grow-disk when wiping is enabled
427 - Documentation improvements, and support for newer pandoc
428 - Fix hspace honoring ipolicy for disks (Issue 484)
429 - Improve handling of the ``kvm_extra`` HV parameter
435 *(Released Fri, 24 May 2013)*
437 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
439 - Disks added through ``gnt-instance modify`` or created through
440 ``gnt-instance recreate-disks`` are wiped, if the
441 ``prealloc_wipe_disks`` flag is set.
442 - If wiping newly created disks fails, the disks are removed. Also,
443 partial failures in creating disks through ``gnt-instance modify``
444 triggers a cleanup of the partially-created disks.
445 - Removing the master IP address doesn't fail if the address has been
447 - Fix ownership of the OS log dir
448 - Workaround missing SO_PEERCRED constant (Issue 191)
454 *(Released Fri, 3 May 2013)*
456 This was the first release candidate of the 2.7 series. Since beta3:
458 - Fix kvm compatibility with qemu 1.4 (Issue 389)
459 - Documentation updates (admin guide, upgrade notes, install
460 instructions) (Issue 372)
461 - Fix gnt-group list nodes and instances count (Issue 436)
462 - Fix compilation without non-mandatory libraries (Issue 441)
463 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
464 - Make confd logging more verbose at INFO level (Issue 435)
465 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
466 - Fix failure path for instance storage type conversion (Issue 229)
467 - Update htools text backend documentation
468 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
469 - Disable inter-cluster instance move for file-based instances, because
470 it is dependant on instance export, which is not supported for
471 file-based instances. (Issue 414)
472 - Fix gnt-job crashes on non-ascii characters (Issue 427)
473 - Fix volume group checks on non-vm-capable nodes (Issue 432)
479 *(Released Mon, 22 Apr 2013)*
481 This was the third beta release of the 2.7 series. Since beta2:
483 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
484 - Fix data loss on wrong usage of ``gnt-instance move``
485 - Properly export errors in confd-based job queries
486 - Add ``users-setup`` tool
487 - Fix iallocator protocol to report 0 as a disk size for diskless
488 instances. This avoids hail breaking when a diskless instance is
490 - Fix job queue directory permission problem that made confd job queries
491 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
492 for access to archived jobs (Issue 406).
493 - Limit the sizes of networks supported by ``gnt-network`` to something
494 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
495 - Fix bugs in instance disk template conversion
496 - Fix GHC 7 compatibility
497 - Fix ``burnin`` install path (Issue 426).
498 - Allow very small disk grows (Issue 347).
499 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
500 that noded doesn't import masterd code (Issue 419).
501 - Make sure the default metavg at cluster init is the same as the vg, if
502 unspecified (Issue 358).
503 - Fix cleanup of partially created disks (part of Issue 416)
509 *(Released Tue, 2 Apr 2013)*
511 This was the second beta release of the 2.7 series. Since beta1:
513 - Networks no longer have a "type" slot, since this information was
514 unused in Ganeti: instead of it tags should be used.
515 - The rapi client now has a ``target_node`` option to MigrateInstance.
516 - Fix early exit return code for hbal (Issue 386).
517 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
518 - Fix ``rbd showmapped`` output parsing (Issue 312).
519 - Networks are now referenced indexed by UUID, rather than name. This
520 will require running cfgupgrade, from 2.7.0beta1, if networks are in
522 - The OS environment now includes network information.
523 - Deleting of a network is now disallowed if any instance nic is using
524 it, to prevent dangling references.
525 - External storage is now documented in man pages.
526 - The exclusive_storage flag can now only be set at nodegroup level.
527 - Hbal can now submit an explicit priority with its jobs.
528 - Many network related locking fixes.
529 - Bump up the required pylint version to 0.25.1.
530 - Fix the ``no_remember`` option in RAPI client.
531 - Many ipolicy related tests, qa, and fixes.
532 - Many documentation improvements and fixes.
533 - Fix building with ``--disable-file-storage``.
534 - Fix ``-q`` option in htools, which was broken if passed more than
536 - Some haskell/python interaction improvements and fixes.
537 - Fix iallocator in case of missing LVM storage.
538 - Fix confd config load in case of ``--no-lvm-storage``.
539 - The confd/query functionality is now mentioned in the security
546 *(Released Wed, 6 Feb 2013)*
548 This was the first beta release of the 2.7 series. All important changes
549 are listed in the latest 2.7 entry.
555 *(Released Fri, 21 Dec 2012)*
557 Important behaviour change: hbal won't rebalance anymore instances which
558 have the ``auto_balance`` attribute set to false. This was the intention
559 all along, but until now it only skipped those from the N+1 memory
560 reservation (DRBD-specific).
562 A significant number of bug fixes in this release:
564 - Fixed disk adoption interaction with ipolicy checks.
565 - Fixed networking issues when instances are started, stopped or
566 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
567 - Fixed the warning in cluster verify for shared storage instances not
569 - Fixed removal of storage directory on shared file storage (issue 262).
570 - Fixed validation of LVM volume group name in OpClusterSetParams
571 (``gnt-cluster modify``) (issue 285).
572 - Fixed runtime memory increases (``gnt-instance modify -m``).
573 - Fixed live migration under Xen's ``xl`` mode.
574 - Fixed ``gnt-instance console`` with ``xl``.
575 - Fixed building with newer Haskell compiler/libraries.
576 - Fixed PID file writing in Haskell daemons (confd); this prevents
577 restart issues if confd was launched manually (outside of
578 ``daemon-util``) while another copy of it was running
579 - Fixed a type error when doing live migrations with KVM (issue 297) and
580 the error messages for failing migrations have been improved.
581 - Fixed opcode validation for the out-of-band commands (``gnt-node
583 - Fixed a type error when unsetting OS hypervisor parameters (issue
584 311); now it's possible to unset all OS-specific hypervisor
586 - Fixed the ``dry-run`` mode for many operations: verification of
587 results was over-zealous but didn't take into account the ``dry-run``
588 operation, resulting in "wrong" failures.
589 - Fixed bash completion in ``gnt-job list`` when the job queue has
590 hundreds of entries; especially with older ``bash`` versions, this
591 results in significant CPU usage.
593 And lastly, a few other improvements have been made:
595 - Added option to force master-failover without voting (issue 282).
596 - Clarified error message on lock conflict (issue 287).
597 - Logging of newly submitted jobs has been improved (issue 290).
598 - Hostname checks have been made uniform between instance rename and
600 - The ``--submit`` option is now supported by ``gnt-debug delay``.
601 - Shutting down the master daemon by sending SIGTERM now stops it from
602 processing jobs waiting for locks; instead, those jobs will be started
603 once again after the master daemon is started the next time (issue
605 - Support for Xen's ``xl`` program has been improved (besides the fixes
607 - Reduced logging noise in the Haskell confd daemon (only show one log
608 entry for each config reload, instead of two).
609 - Several man page updates and typo fixes.
615 *(Released Fri, 12 Oct 2012)*
617 A small bugfix release. Among the bugs fixed:
619 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
620 made the command unusable.
621 - Commands that issue many jobs don't fail anymore just because some jobs
622 take so long that other jobs are archived.
623 - Failures during ``gnt-instance reinstall`` are reflected by the exit
625 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
627 - When ``always_failover`` is set, ``--allow-failover`` is not required
628 in migrate commands anymore.
629 - ``bash_completion`` works even if extglob is disabled.
630 - Fixed bug with locks that made failover for RDB-based instances fail.
631 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
632 a random node instead of one based on the allocator metric.
633 - Support for newer versions of pylint and pep8.
634 - Hail doesn't fail anymore when trying to add an instance of type
635 ``file``, ``sharedfile`` or ``rbd``.
636 - Added new Makefile target to rebuild the whole distribution, so that
637 all files are included.
643 *(Released Fri, 27 Jul 2012)*
646 .. attention:: The ``LUXI`` protocol has been made more consistent
647 regarding its handling of command arguments. This, however, leads to
648 incompatibility issues with previous versions. Please ensure that you
649 restart Ganeti daemons soon after the upgrade, otherwise most
650 ``LUXI`` calls (job submission, setting/resetting the drain flag,
651 pausing/resuming the watcher, cancelling and archiving jobs, querying
652 the cluster configuration) will fail.
661 The current ``admin_up`` field, which used to denote whether an instance
662 should be running or not, has been removed. Instead, ``admin_state`` is
663 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
665 The rational behind this is that an instance being “down” can have
668 - it could be down during a reboot
669 - it could be temporarily be down for a reinstall
670 - or it could be down because it is deprecated and kept just for its
673 The previous Boolean state was making it difficult to do capacity
674 calculations: should Ganeti reserve memory for a down instance? Now, the
675 tri-state field makes it clear:
677 - in ``up`` and ``down`` state, all resources are reserved for the
678 instance, and it can be at any time brought up if it is down
679 - in ``offline`` state, only disk space is reserved for it, but not
682 The field can have an extra use: since the transition between ``up`` and
683 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
684 transition between ``offline`` and ``down`` is done via ``gnt-instance
685 modify``, it is possible to given different rights to users. For
686 example, owners of an instance could be allowed to start/stop it, but
687 not transition it out of the offline state.
689 Instance policies and specs
690 +++++++++++++++++++++++++++
692 In previous Ganeti versions, an instance creation request was not
693 limited on the minimum size and on the maximum size just by the cluster
694 resources. As such, any policy could be implemented only in third-party
695 clients (RAPI clients, or shell wrappers over ``gnt-*``
696 tools). Furthermore, calculating cluster capacity via ``hspace`` again
697 required external input with regards to instance sizes.
699 In order to improve these workflows and to allow for example better
700 per-node group differentiation, we introduced instance specs, which
703 - minimum instance disk size, disk count, memory size, cpu count
704 - maximum values for the above metrics
705 - and “standard” values (used in ``hspace`` to calculate the standard
708 The minimum/maximum values can be also customised at node-group level,
709 for example allowing more powerful hardware to support bigger instance
712 Beside the instance specs, there are a few other settings belonging to
713 the instance policy framework. It is possible now to customise, per
714 cluster and node-group:
716 - the list of allowed disk templates
717 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
718 - the maximum ratio of instance to spindles (see below for more
719 information) for local storage
721 All these together should allow all tools that talk to Ganeti to know
722 what are the ranges of allowed values for instances and the
723 over-subscription that is allowed.
725 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
726 instance configuration, and the physical CPU configuration from the
727 node. For the spindle ratios however, we didn't track before these
728 values, so new parameters have been added:
730 - a new node parameter ``spindle_count``, defaults to 1, customisable at
731 node group or node level
732 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
734 Note that spindles in this context doesn't need to mean actual
735 mechanical hard-drives; it's just a relative number for both the node
736 I/O capacity and instance I/O consumption.
738 Instance migration behaviour
739 ++++++++++++++++++++++++++++
741 While live-migration is in general desirable over failover, it is
742 possible that for some workloads it is actually worse, due to the
743 variable time of the “suspend” phase during live migration.
745 To allow the tools to work consistently over such instances (without
746 having to hard-code instance names), a new backend parameter
747 ``always_failover`` has been added to control the migration/failover
748 behaviour. When set to True, all migration requests for an instance will
749 instead fall-back to failover.
751 Instance memory ballooning
752 ++++++++++++++++++++++++++
754 Initial support for memory ballooning has been added. The memory for an
755 instance is no longer fixed (backend parameter ``memory``), but instead
756 can vary between minimum and maximum values (backend parameters
757 ``minmem`` and ``maxmem``). Currently we only change an instance's
760 - live migrating or failing over and instance and the target node
761 doesn't have enough memory
762 - user requests changing the memory via ``gnt-instance modify
768 In order to control the use of specific CPUs by instance, support for
769 controlling CPU pinning has been added for the Xen, HVM and LXC
770 hypervisors. This is controlled by a new hypervisor parameter
771 ``cpu_mask``; details about possible values for this are in the
772 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
773 VCPU-to-CPU mapping) form will work well only when all nodes in your
774 cluster have the same amount of CPUs.
779 Another area in which Ganeti was not customisable were the parameters
780 used for storage configuration, e.g. how many stripes to use for LVM,
781 DRBD resync configuration, etc.
783 To improve this area, we've added disks parameters, which are
784 customisable at cluster and node group level, and which allow to
785 specify various parameters for disks (DRBD has the most parameters
786 currently), for example:
788 - DRBD resync algorithm and parameters (e.g. speed)
789 - the default VG for meta-data volumes for DRBD
790 - number of stripes for LVM (plain disk template)
793 These parameters can be modified via ``gnt-cluster modify -D …`` and
794 ``gnt-group modify -D …``, and are used at either instance creation (in
795 case of LVM stripes, for example) or at disk “activation” time
798 Rados block device support
799 ++++++++++++++++++++++++++
801 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
802 denoted by the ``rbd`` disk template type. This is considered
803 experimental, feedback is welcome. For details on configuring it, see
804 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
809 The existing master IP functionality works well only in simple setups (a
810 single network shared by all nodes); however, if nodes belong to
811 different networks, then the ``/32`` setup and lack of routing
812 information is not enough.
814 To allow the master IP to function well in more complex cases, the
815 system was reworked as follows:
817 - a master IP netmask setting has been added
818 - the master IP activation/turn-down code was moved from the node daemon
820 - whether to run the Ganeti-supplied master IP script or a user-supplied
821 on is a ``gnt-cluster init`` setting
823 Details about the location of the standard and custom setup scripts are
824 in the man page :manpage:`gnt-cluster(8)`; for information about the
825 setup script protocol, look at the Ganeti-supplied script.
830 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
833 It is now possible to use TLS-protected connections, and when renewing
834 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
835 it is now possible to specify spice or spice CA certificates. Also, it
836 is possible to configure a password for SPICE sessions via the
837 hypervisor parameter ``spice_password_file``.
839 There are also new parameters to control the compression and streaming
840 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
841 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
842 for the spice parameters.
844 Lastly, it is now possible to see the SPICE connection information via
845 ``gnt-instance console``.
850 A new tool (``tools/ovfconverter``) has been added that supports
851 conversion between Ganeti and the `Open Virtualization Format
852 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
855 This relies on the ``qemu-img`` tool to convert the disk formats, so the
856 actual compatibility with other virtualization solutions depends on it.
861 The configuration query daemon (``ganeti-confd``) is now optional, and
862 has been rewritten in Haskell; whether to use the daemon at all, use the
863 Python (default) or the Haskell version is selectable at configure time
864 via the ``--enable-confd`` parameter, which can take one of the
865 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
866 daemon will result in a smaller footprint; for larger systems, we
867 welcome feedback on the Haskell version which might become the default
870 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
871 daemon running. The Python version doesn't implement the new call.
874 User interface changes
875 ~~~~~~~~~~~~~~~~~~~~~~
877 We have replaced the ``--disks`` option of ``gnt-instance
878 replace-disks`` with a more flexible ``--disk`` option, which allows
879 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
880 disk size and mode can be changed upon recreation (via ``gnt-instance
881 recreate-disks``, which accepts the same ``--disk`` option).
883 As many people are used to a ``show`` command, we have added that as an
884 alias to ``info`` on all ``gnt-*`` commands.
886 The ``gnt-instance grow-disk`` command has a new mode in which it can
887 accept the target size of the disk, instead of the delta; this can be
888 more safe since two runs in absolute mode will be idempotent, and
889 sometimes it's also easier to specify the desired size directly.
891 Also the handling of instances with regard to offline secondaries has
892 been improved. Instance operations should not fail because one of it's
893 secondary nodes is offline, even though it's safe to proceed.
895 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
896 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
897 minors to instance name.
902 RAPI coverage has improved, with (for example) new resources for
903 recreate-disks, node power-cycle, etc.
908 There is partial support for ``xl`` in the Xen hypervisor; feedback is
911 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
912 whether to still support Python 2.4 or move to Python 2.6 as minimum
915 Support for Fedora has been slightly improved; the provided example
916 init.d script should work better on it and the INSTALL file should
917 document the needed dependencies.
922 The deprecated ``QueryLocks`` LUXI request has been removed. Use
923 ``Query(what=QR_LOCK, ...)`` instead.
925 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
926 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
927 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
928 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
929 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
931 RAPI client: ``CertificateError`` now derives from
932 ``GanetiApiError``. This should make it more easy to handle Ganeti
935 Deprecation warnings due to PyCrypto/paramiko import in
936 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
937 make sure to run an up-to-date paramiko version, if you use this tool.
939 The QA scripts now depend on Python 2.5 or above (the main code base
940 still works with Python 2.4).
942 The configuration file (``config.data``) is now written without
943 indentation for performance reasons; if you want to edit it, it can be
944 re-formatted via ``tools/fmtjson``.
946 A number of bugs has been fixed in the cluster merge tool.
948 ``x509`` certification verification (used in import-export) has been
949 changed to allow the same clock skew as permitted by the cluster
950 verification. This will remove some rare but hard to diagnose errors in
957 *(Released Thu, 19 Jul 2012)*
959 Very few changes from rc4 to the final release, only bugfixes:
961 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
962 instance, fix CDROM booting for KVM instances)
963 - fixed node group modification of node parameters
964 - fixed issue in LUClusterVerifyGroup with multi-group clusters
965 - fixed generation of bash completion to ensure a stable ordering
972 *(Released Fri, 13 Jul 2012)*
974 Third release candidate for 2.6. The following changes were done from
977 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
978 - Fixed an inconsistency in the LUXI protocol with the provided
979 arguments (NOT backwards compatible)
980 - Fixed a bug with node groups ipolicy where ``min`` was greater than
981 the cluster ``std`` value
982 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
983 easier instance debugging on nodes (requires ``hconfd`` to work)
989 *(Released Tue, 03 Jul 2012)*
991 Second release candidate for 2.6. The following changes were done from
994 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
996 - Fixed a RAPI regression on missing beparams/memory
997 - Fixed redistribution of files on offline nodes
998 - Added possibility to run activate-disks even though secondaries are
999 offline. With this change it relaxes also the strictness on some other
1000 commands which use activate disks internally:
1001 * ``gnt-instance start|reboot|rename|backup|export``
1002 - Made it possible to remove safely an instance if its secondaries are
1004 - Made it possible to reinstall even though secondaries are offline
1010 *(Released Mon, 25 Jun 2012)*
1012 First release candidate for 2.6. The following changes were done from
1015 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1017 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1018 behaviour. It supports now the modify operation
1019 - ``hcheck`` implemented to analyze cluster health and possibility of
1020 improving health by rebalance
1021 - ``hbal`` has been improved in dealing with split instances
1027 *(Released Mon, 11 Jun 2012)*
1029 Second beta release of 2.6. The following changes were done from beta2
1032 - Fixed ``daemon-util`` with non-root user models
1033 - Fixed creation of plain instances with ``--no-wait-for-sync``
1034 - Fix wrong iv_names when running ``cfgupgrade``
1035 - Export more information in RAPI group queries
1036 - Fixed bug when changing instance network interfaces
1037 - Extended burnin to do NIC changes
1038 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1039 - Changed default for DRBD barriers
1040 - Fixed DRBD error reporting for syncer rate
1041 - Verify the options on disk parameters
1043 And of course various fixes to documentation and improved unittests and
1050 *(Released Wed, 23 May 2012)*
1052 First beta release of 2.6. The following changes were done from beta1 to
1055 - integrated patch for distributions without ``start-stop-daemon``
1056 - adapted example init.d script to work on Fedora
1057 - fixed log handling in Haskell daemons
1058 - adapted checks in the watcher for pycurl linked against libnss
1059 - add partial support for ``xl`` instead of ``xm`` for Xen
1060 - fixed a type issue in cluster verification
1061 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1064 Plus integrated fixes from the 2.5 branch:
1066 - fixed ``kvm-ifup`` to use ``/bin/bash``
1067 - fixed parallel build failures
1068 - KVM live migration when using a custom keymap
1074 *(Released Tue, 24 Jul 2012)*
1076 A small bugfix release, with no new features:
1078 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1079 different default shell (e.g. Debian, Ubuntu)
1080 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1081 243 and Debian bug #650664)
1082 - fixed compatibility with KVM versions that don't support multiple boot
1083 devices (fixes Issue 230 and Debian bug #624256)
1085 Additionally, a few fixes were done to the build system (fixed parallel
1086 build failures) and to the unittests (fixed race condition in test for
1087 FileID functions, and the default enable/disable mode for QA test is now
1094 *(Released Fri, 11 May 2012)*
1096 A small bugfix release.
1098 The main issues solved are on the topic of compatibility with newer LVM
1101 - fixed parsing of ``lv_attr`` field
1102 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1103 the ``--force`` flag is needed
1105 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1106 to accept kernel 3.x too (was hardcoded to 2.6.*).
1108 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1109 of hook results and that also made display of other errors suboptimal
1110 was fixed; the code behaves now like 2.4 and earlier.
1112 Another change in 2.5, the cleanup of the OS scripts environment, is too
1113 aggressive: it removed even the ``PATH`` variable, which requires the OS
1114 scripts to *always* need to export it. Since this is a bit too strict,
1115 we now export a minimal PATH, the same that we export for hooks.
1117 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1118 integrated into this release.
1120 Finally, a few other miscellaneous changes were done (no new features,
1121 just small improvements):
1123 - Fix ``gnt-group --help`` display
1124 - Fix hardcoded Xen kernel path
1125 - Fix grow-disk handling of invalid units
1126 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1127 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1133 *(Released Thu, 12 Apr 2012)*
1135 Incompatible/important changes and bugfixes
1136 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1138 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1139 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1140 to match the underlying LUXI interface.
1141 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1142 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1143 not maintain backwards-compatibility as the underlying operation
1144 changed in an incompatible way. The RAPI client can talk to old
1145 servers, but it needs to be told so as the return value changed.
1146 - When creating file-based instances via RAPI, the ``file_driver``
1147 parameter no longer defaults to ``loop`` and must be specified.
1148 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1150 - Support for the undocumented and deprecated RAPI instance creation
1151 request format version 0 has been dropped. Use version 1, supported
1152 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1153 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1154 documentation <install>`.
1155 - The "cluster-verify" hooks are now executed per group by the
1156 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1157 if you just run ``gnt-cluster verify``, which generates one opcode per
1159 - The environment as passed to the OS scripts is cleared, and thus no
1160 environment variables defined in the node daemon's environment will be
1161 inherited by the scripts.
1162 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1164 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1165 support operations involving multiple node groups.
1166 - Offline nodes are ignored when failing over an instance.
1167 - Support for KVM version 1.0, which changed the version reporting format
1169 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1171 - ``Makefile`` is now compatible with Automake 1.11.2
1172 - Includes all bugfixes made in the 2.4 series
1177 - The ganeti-htools project has been merged into the ganeti-core source
1178 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1179 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1180 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1181 GPT-style partition tables (requires `parted
1182 <http://www.gnu.org/s/parted/>`_).
1183 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1184 - Allowed adding tags on instance creation.
1185 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1186 - Allow instances to be started in a paused state, enabling the user to
1187 see the complete console output on boot using the console.
1188 - Added new hypervisor flag to control default reboot behaviour
1189 (``reboot_behavior``).
1190 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1191 - Improved out-of-band management support:
1193 - Added ``gnt-node health`` command reporting the health status of
1195 - Added ``gnt-node power`` command to manage power status of nodes.
1196 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1198 - Instance migration can fall back to failover if instance is not
1200 - Filters can be used when listing nodes, instances, groups and locks;
1201 see :manpage:`ganeti(7)` manpage.
1202 - Added post-execution status as variables to :doc:`hooks <hooks>`
1204 - Instance tags are exported/imported together with the instance.
1205 - When given an explicit job ID, ``gnt-job info`` will work for archived
1207 - Jobs can define dependencies on other jobs (not yet supported via
1208 RAPI or command line, but used by internal commands and usable via
1211 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1214 - Instance failover is now available as a RAPI resource
1215 (``/2/instances/[instance_name]/failover``).
1216 - ``gnt-instance info`` defaults to static information if primary node
1218 - Opcodes have a new ``comment`` attribute.
1219 - Added basic SPICE support to KVM hypervisor.
1220 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1222 Node group improvements
1223 ~~~~~~~~~~~~~~~~~~~~~~~
1225 - ``gnt-cluster verify`` has been modified to check groups separately,
1226 thereby improving performance.
1227 - Node group support has been added to ``gnt-cluster verify-disks``,
1228 which now operates per node group.
1229 - Watcher has been changed to work better with node groups.
1231 - One process and state file per node group.
1232 - Slow watcher in one group doesn't block other group's watcher.
1234 - Added new command, ``gnt-group evacuate``, to move all instances in a
1235 node group to other groups.
1236 - Added ``gnt-instance change-group`` to move an instance to another
1238 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1239 per-group operations.
1240 - Node groups can be tagged.
1241 - Some operations switch from an exclusive to a shared lock as soon as
1243 - Instance's primary and secondary nodes' groups are now available as
1244 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1245 and ``snodes.group.uuid``).
1250 - Numerous updates to documentation and manpages.
1252 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1254 - Some opcode/job results are now also documented, see :doc:`RAPI
1257 - A lockset's internal lock is now also visible in lock monitor.
1258 - Log messages from job queue workers now contain information about the
1259 opcode they're processing.
1260 - ``gnt-instance console`` no longer requires the instance lock.
1261 - A short delay when waiting for job changes reduces the number of LUXI
1262 requests significantly.
1263 - DRBD metadata volumes are overwritten with zeros during disk creation.
1264 - Out-of-band commands no longer acquire the cluster lock in exclusive
1266 - ``devel/upload`` now uses correct permissions for directories.
1272 *(Released Fri, 23 Mar 2012)*
1274 This was the sixth release candidate of the 2.5 series.
1280 *(Released Mon, 9 Jan 2012)*
1282 This was the fifth release candidate of the 2.5 series.
1288 *(Released Thu, 27 Oct 2011)*
1290 This was the fourth release candidate of the 2.5 series.
1296 *(Released Wed, 26 Oct 2011)*
1298 This was the third release candidate of the 2.5 series.
1304 *(Released Tue, 18 Oct 2011)*
1306 This was the second release candidate of the 2.5 series.
1312 *(Released Tue, 4 Oct 2011)*
1314 This was the first release candidate of the 2.5 series.
1320 *(Released Wed, 31 Aug 2011)*
1322 This was the third beta release of the 2.5 series.
1328 *(Released Mon, 22 Aug 2011)*
1330 This was the second beta release of the 2.5 series.
1336 *(Released Fri, 12 Aug 2011)*
1338 This was the first beta release of the 2.5 series.
1344 *(Released Thu, 27 Oct 2011)*
1346 - Fixed bug when parsing command line parameter values ending in
1348 - Fixed assertion error after unclean master shutdown
1349 - Disable HTTP client pool for RPC, significantly reducing memory usage
1351 - Fixed queue archive creation with wrong permissions
1357 *(Released Tue, 23 Aug 2011)*
1361 - Fixed documentation for importing with ``--src-dir`` option
1362 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1363 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1369 *(Released Fri, 5 Aug 2011)*
1371 Many bug-fixes and a few small features:
1373 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1374 issues when you tried to add an instance with two MAC addresses in one
1376 - KVM: fixed per-instance stored UID value
1377 - KVM: configure bridged NICs at migration start
1378 - KVM: Fix a bug where instance will not start with never KVM versions
1380 - Added OS search path to ``gnt-cluster info``
1381 - Fixed an issue with ``file_storage_dir`` where you were forced to
1382 provide an absolute path, but the documentation states it is a
1383 relative path, the documentation was right
1384 - Added a new parameter to instance stop/start called ``--no-remember``
1385 that will make the state change to not be remembered
1386 - Implemented ``no_remember`` at RAPI level
1387 - Improved the documentation
1388 - Node evacuation: don't call IAllocator if node is already empty
1389 - Fixed bug in DRBD8 replace disks on current nodes
1390 - Fixed bug in recreate-disks for DRBD instances
1391 - Moved assertion checking locks in ``gnt-instance replace-disks``
1392 causing it to abort with not owning the right locks for some situation
1393 - Job queue: Fixed potential race condition when cancelling queued jobs
1394 - Fixed off-by-one bug in job serial generation
1395 - ``gnt-node volumes``: Fix instance names
1396 - Fixed aliases in bash completion
1397 - Fixed a bug in reopening log files after being sent a SIGHUP
1398 - Added a flag to burnin to allow specifying VCPU count
1399 - Bugfixes to non-root Ganeti configuration
1405 *(Released Thu, 12 May 2011)*
1407 Many bug-fixes and a few new small features:
1409 - Fixed a bug related to log opening failures
1410 - Fixed a bug in instance listing with orphan instances
1411 - Fixed a bug which prevented resetting the cluster-level node parameter
1412 ``oob_program`` to the default
1413 - Many fixes related to the ``cluster-merge`` tool
1414 - Fixed a race condition in the lock monitor, which caused failures
1415 during (at least) creation of many instances in parallel
1416 - Improved output for gnt-job info
1417 - Removed the quiet flag on some ssh calls which prevented debugging
1419 - Improved the N+1 failure messages in cluster verify by actually
1420 showing the memory values (needed and available)
1421 - Increased lock attempt timeouts so that when executing long operations
1422 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1423 too early and thus prevent the use of the 'fair' mechanism
1424 - Changed instance query data (``gnt-instance info``) to not acquire
1425 locks unless needed, thus allowing its use on locked instance if only
1426 static information is asked for
1427 - Improved behaviour with filesystems that do not support rename on an
1429 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1430 kept locks on all nodes during the wipe, which is unneeded
1431 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1432 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1433 10GiB) which caused the wipe to fail right at the end in some cases
1434 - Fixed master IP activation when doing master failover with no-voting
1435 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1436 the master node itself
1437 - Fixed potential data-loss in under disk full conditions, where Ganeti
1438 wouldn't check correctly the return code and would consider
1439 partially-written files 'correct'
1440 - Fixed bug related to multiple VGs and DRBD disk replacing
1441 - Added new disk parameter ``metavg`` that allows placement of the meta
1442 device for DRBD in a different volume group
1443 - Fixed error handling in the node daemon when the system libc doesn't
1444 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1445 - Fixed lock release during replace-disks, which kept cluster-wide locks
1446 when doing disk replaces with an iallocator script
1447 - Added check for missing bridges in cluster verify
1448 - Handle EPIPE errors while writing to the terminal better, so that
1449 piping the output to e.g. ``less`` doesn't cause a backtrace
1450 - Fixed rare case where a ^C during Luxi calls could have been
1451 interpreted as server errors, instead of simply terminating
1452 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1454 - Added a few more parameters to the KVM hypervisor, allowing a second
1455 CDROM, custom disk type for CDROMs and a floppy image
1456 - Removed redundant message in instance rename when the name is given
1458 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1459 disks on new nodes, allowing recreation when the original instance
1460 nodes are completely gone
1461 - Added option when converting disk templates to DRBD to skip waiting
1462 for the resync, in order to make the instance available sooner
1463 - Added two new variables to the OS scripts environment (containing the
1465 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1466 to allow use of ``pvgrub`` as bootloader
1467 - Changed the instance memory modifications to only check out-of-memory
1468 conditions on memory increases, and turned the secondary node warnings
1469 into errors (they can still be overridden via ``--force``)
1470 - Fixed the handling of a corner case when the Python installation gets
1471 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1472 to execute a command that doesn't exist
1473 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1474 node of the instance returned failures during instance shutdown; this
1475 adds the option ``--ignore-consistency`` to gnt-instance move
1477 And as usual, various improvements to the error messages, documentation
1484 *(Released Wed, 09 Mar 2011)*
1486 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1487 the RAPI users file if run twice (even with ``--dry-run``).
1489 The release fixes that bug (nothing else changed).
1495 *(Released Mon, 07 Mar 2011)*
1497 Final 2.4.0 release. Just a few small fixes:
1499 - Fixed RAPI node evacuate
1500 - Fixed the kvm-ifup script
1501 - Fixed internal error handling for special job cases
1502 - Updated man page to specify the escaping feature for options
1508 *(Released Mon, 28 Feb 2011)*
1510 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1511 that this feature wiped the disks of the wrong instance, leading to loss
1516 - Fixed title of query field containing instance name
1517 - Expanded the glossary in the documentation
1518 - Fixed one unittest (internal issue)
1524 *(Released Mon, 21 Feb 2011)*
1526 A number of bug fixes plus just a couple functionality changes.
1528 On the user-visible side, the ``gnt-* list`` command output has changed
1529 with respect to "special" field states. The current rc1 style of display
1530 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1531 the default output mode special fields states are displayed as follows:
1533 - Offline resource: ``*``
1534 - Unavailable/not applicable: ``-``
1535 - Data missing (RPC failure): ``?``
1536 - Unknown field: ``??``
1538 Another user-visible change is the addition of ``--force-join`` to
1543 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1544 - Fixed regression in RAPI/instance reinstall where all parameters were
1545 required (instead of optional)
1546 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1547 - Fixed iallocator usage (offline nodes were not considered offline)
1548 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1549 - Fixed hypervisor and OS parameter validation with respect to
1550 non-vm_capable nodes
1551 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1553 - Fixed ``tools/listrunner`` with respect to agent-based usage
1559 *(Released Fri, 4 Feb 2011)*
1561 Many changes and fixes since the beta1 release. While there were some
1562 internal changes, the code has been mostly stabilised for the RC
1565 Note: the dumb allocator was removed in this release, as it was not kept
1566 up-to-date with the IAllocator protocol changes. It is recommended to
1567 use the ``hail`` command from the ganeti-htools package.
1569 Note: the 2.4 and up versions of Ganeti are not compatible with the
1570 0.2.x branch of ganeti-htools. You need to upgrade to
1571 ganeti-htools-0.3.0 (or later).
1573 Regressions fixed from 2.3
1574 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1576 - Fixed the ``gnt-cluster verify-disks`` command
1577 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1579 - Fixed disk adoption breakage
1580 - Fixed wrong headers in instance listing for field aliases
1585 - Fixed corner case in KVM handling of NICs
1586 - Fixed many cases of wrong handling of non-vm_capable nodes
1587 - Fixed a bug where a missing instance symlink was not possible to
1588 recreate with any ``gnt-*`` command (now ``gnt-instance
1589 activate-disks`` does it)
1590 - Fixed the volume group name as reported by ``gnt-cluster
1592 - Increased timeouts for the import-export code, hopefully leading to
1593 fewer aborts due network or instance timeouts
1594 - Fixed bug in ``gnt-node list-storage``
1595 - Fixed bug where not all daemons were started on cluster
1596 initialisation, but only at the first watcher run
1597 - Fixed many bugs in the OOB implementation
1598 - Fixed watcher behaviour in presence of instances with offline
1600 - Fixed instance list output for instances running on the wrong node
1601 - a few fixes to the cluster-merge tool, but it still cannot merge
1602 multi-node groups (currently it is not recommended to use this tool)
1608 - Improved network configuration for the KVM hypervisor
1609 - Added e1000 as a supported NIC for Xen-HVM
1610 - Improved the lvmstrap tool to also be able to use partitions, as
1611 opposed to full disks
1612 - Improved speed of disk wiping (the cluster parameter
1613 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1614 of instance creations
1615 - Added documentation for the OS parameters
1616 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1617 hypervisor is not responding
1618 - Added display of blacklisted and hidden OS information in
1619 ``gnt-cluster info``
1620 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1621 NIC and node parameters, which might create problems with currently
1622 invalid (but undetected) configuration files, but prevents validation
1623 failures when unrelated parameters are modified
1624 - Changed cluster initialisation to wait for the master daemon to become
1626 - Expanded the RAPI interface:
1628 - Added config redistribution resource
1629 - Added activation/deactivation of instance disks
1630 - Added export of console information
1632 - Implemented log file reopening on SIGHUP, which allows using
1633 logrotate(8) for the Ganeti log files
1634 - Added a basic OOB helper script as an example
1640 *(Released Fri, 14 Jan 2011)*
1645 - Fixed timezone issues when formatting timestamps
1646 - Added support for node groups, available via ``gnt-group`` and other
1648 - Added out-of-band framework and management, see :doc:`design
1649 document <design-oob>`
1650 - Removed support for roman numbers from ``gnt-node list`` and
1651 ``gnt-instance list``.
1652 - Allowed modification of master network interface via ``gnt-cluster
1653 modify --master-netdev``
1654 - Accept offline secondaries while shutting down instance disks
1655 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1656 - Added support for multiple LVM volume groups
1657 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1659 - Added commands to list available fields:
1661 - ``gnt-node list-fields``
1662 - ``gnt-group list-fields``
1663 - ``gnt-instance list-fields``
1665 - Updated documentation and man pages
1670 - Moved ``rapi_users`` file into separate directory, now named
1671 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1673 - Added new tool for running commands on many machines,
1674 ``tools/ganeti-listrunner``
1675 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1676 improving the ``gnt-instance console`` output
1677 - Allowed customisation of disk index separator at ``configure`` time
1678 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1679 - Added support for non-partitioned md disks in ``lvmstrap``
1680 - Added script to gracefully power off KVM instances
1681 - Split ``utils`` module into smaller parts
1682 - Changed query operations to return more detailed information, e.g.
1683 whether an information is unavailable due to an offline node. To use
1684 this new functionality, the LUXI call ``Query`` must be used. Field
1685 information is now stored by the master daemon and can be retrieved
1686 using ``QueryFields``. Instances, nodes and groups can also be queried
1687 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1688 exposed via RAPI). The following commands make use of this
1689 infrastructure change:
1691 - ``gnt-group list``
1692 - ``gnt-group list-fields``
1694 - ``gnt-node list-fields``
1695 - ``gnt-instance list``
1696 - ``gnt-instance list-fields``
1697 - ``gnt-debug locks``
1702 - New RAPI resources (see :doc:`rapi`):
1706 - ``/2/groups/[group_name]``
1707 - ``/2/groups/[group_name]/assign-nodes``
1708 - ``/2/groups/[group_name]/modify``
1709 - ``/2/groups/[group_name]/rename``
1710 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1714 - Implemented ``no_install`` for instance creation
1715 - Implemented OS parameters for instance reinstallation, allowing
1716 use of special settings on reinstallation (e.g. for preserving data)
1721 - Added IPv6 support in import/export
1722 - Pause DRBD synchronization while wiping disks on instance creation
1723 - Updated unittests and QA scripts
1724 - Improved network parameters passed to KVM
1725 - Converted man pages from docbook to reStructuredText
1731 *(Released Mon, 20 Dec 2010)*
1733 Released version 2.3.1~rc1 without any changes.
1739 *(Released Wed, 1 Dec 2010)*
1741 - impexpd: Disable OpenSSL compression in socat if possible (backport
1742 from master, commit e90739d625b, see :doc:`installation guide
1743 <install-quick>` for details)
1744 - Changed unittest coverage report to exclude test scripts
1745 - Added script to check version format
1751 *(Released Wed, 1 Dec 2010)*
1753 Released version 2.3.0~rc1 without any changes.
1759 *(Released Fri, 19 Nov 2010)*
1761 A number of bugfixes and documentation updates:
1763 - Update ganeti-os-interface documentation
1764 - Fixed a bug related to duplicate MACs or similar items which should be
1766 - Fix breakage in OS state modify
1767 - Reinstall instance: disallow offline secondaries (fixes bug related to
1768 OS changing but reinstall failing)
1769 - plus all the other fixes between 2.2.1 and 2.2.2
1775 *(Released Tue, 2 Nov 2010)*
1777 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1778 - Fixed master failover race with watcher
1779 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1781 - Accept previously stopped instance for export with instance removal
1782 - Simplify and extend the environment variables for instance OS scripts
1783 - Added new node flags, ``master_capable`` and ``vm_capable``
1784 - Added optional instance disk wiping prior during allocation. This is a
1785 cluster-wide option and can be set/modified using
1786 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1787 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1788 :doc:`install-quick`
1789 - Added a new watcher option (``--ignore-pause``)
1790 - Added option to ignore offline node on instance start/stop
1791 (``--ignore-offline``)
1792 - Allow overriding OS parameters with ``gnt-instance reinstall``
1793 - Added ability to change node's secondary IP address using ``gnt-node
1795 - Implemented privilege separation for all daemons except
1796 ``ganeti-noded``, see ``configure`` options
1797 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1799 - Implemented job priorities (see ``ganeti(7)`` manpage)
1800 - Ignore failures while shutting down instances during failover from
1802 - Exit daemon's bootstrap process only once daemon is ready
1803 - Export more information via ``LUInstanceQuery``/remote API
1804 - Improved documentation, QA and unittests
1805 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1806 restart if the file was created or changed
1807 - Added LUXI protocol version sent with each request and response,
1808 allowing detection of server/client mismatches
1809 - Moved the Python scripts among gnt-* and ganeti-* into modules
1810 - Moved all code related to setting up SSH to an external script,
1812 - Infrastructure changes for node group support in future versions
1818 *(Released Fri, 19 Nov 2010)*
1820 A few small bugs fixed, and some improvements to the build system:
1822 - Fix documentation regarding conversion to drbd
1823 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1825 - Fix error handling in node modify with multiple changes
1826 - Allow remote imports without checked names
1832 *(Released Tue, 19 Oct 2010)*
1834 - Disable SSL session ID cache in RPC client
1840 *(Released Thu, 14 Oct 2010)*
1842 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1843 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1845 - Fix problems with interaction between Curl and Python's HTTP server,
1846 resulting in increased speed in many RPC calls
1847 - Improve our release script to prevent breakage with older aclocal and
1854 *(Released Thu, 7 Oct 2010)*
1856 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1857 value retrieved from master
1858 - Added support for blacklisted or hidden OS definitions
1859 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1860 - Added support for -mem-path in KVM hypervisor abstraction layer
1861 - Allow overriding instance parameters in tool for inter-cluster
1862 instance moves (``tools/move-instance``)
1863 - Improved opcode summaries (e.g. in ``gnt-job list``)
1864 - Improve consistency of OS listing by sorting it
1865 - Documentation updates
1871 *(Released Fri, 8 Oct 2010)*
1873 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1879 *(Released Mon, 4 Oct 2010)*
1881 - Fixed regression in ``gnt-instance rename``
1887 *(Released Wed, 22 Sep 2010)*
1889 - Fixed OS_VARIANT variable for OS scripts
1890 - Fixed cluster tag operations via RAPI
1891 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1892 - Disabled RAPI CA checks in watcher
1898 *(Released Mon, 23 Aug 2010)*
1900 - Support DRBD versions of the format "a.b.c.d"
1902 - Re-introduce support for usage from multiple threads in RAPI client
1903 - Instance renames and modify via RAPI
1904 - Work around race condition between processing and archival in job
1906 - Mark opcodes following failed one as failed, too
1907 - Job field ``lock_status`` was removed due to difficulties making it
1908 work with the changed job queue in Ganeti 2.2; a better way to monitor
1909 locks is expected for a later 2.2.x release
1910 - Fixed dry-run behaviour with many commands
1911 - Support ``ssh-agent`` again when adding nodes
1912 - Many additional bugfixes
1918 *(Released Fri, 30 Jul 2010)*
1920 Important change: the internal RPC mechanism between Ganeti nodes has
1921 changed from using a home-grown http library (based on the Python base
1922 libraries) to use the PycURL library. This requires that PycURL is
1923 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1924 against GnuTLS by default. cURL's support for GnuTLS had known issues
1925 before cURL 7.21.0 and we recommend using the latest cURL release or
1926 linking against OpenSSL. Most other distributions already link PycURL
1927 and cURL against OpenSSL. The command::
1929 python -c 'import pycurl; print pycurl.version'
1931 can be used to determine the libraries PycURL and cURL are linked
1934 Other significant changes:
1936 - Rewrote much of the internals of the job queue, in order to achieve
1937 better parallelism; this decouples job query operations from the job
1938 processing, and it should allow much nicer behaviour of the master
1939 daemon under load, and it also has uncovered some long-standing bugs
1940 related to the job serialisation (now fixed)
1941 - Added a default iallocator setting to the cluster parameters,
1942 eliminating the need to always pass nodes or an iallocator for
1943 operations that require selection of new node(s)
1944 - Added experimental support for the LXC virtualization method
1945 - Added support for OS parameters, which allows the installation of
1946 instances to pass parameter to OS scripts in order to customise the
1948 - Added a hypervisor parameter controlling the migration type (live or
1949 non-live), since hypervisors have various levels of reliability; this
1950 has renamed the 'live' parameter to 'mode'
1951 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1952 logical volumes, meaning that cluster verify will ignore them and not
1953 flag their presence as errors
1954 - The watcher will now reset the error count for failed instances after
1955 8 hours, thus allowing self-healing if the problem that caused the
1956 instances to be down/fail to start has cleared in the meantime
1957 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1958 check for, and warn, if the drbd module parameter ``usermode_helper``
1959 is not consistent with the cluster-wide setting; this is needed to
1960 make diagnose easier of failed drbd creations
1961 - Started adding base IPv6 support, but this is not yet
1962 enabled/available for use
1963 - Rename operations (cluster, instance) will now return the new name,
1964 which is especially useful if a short name was passed in
1965 - Added support for instance migration in RAPI
1966 - Added a tool to pre-configure nodes for the SSH setup, before joining
1967 them to the cluster; this will allow in the future a simplified model
1968 for node joining (but not yet fully enabled in 2.2); this needs the
1969 paramiko python library
1970 - Fixed handling of name-resolving errors
1971 - Fixed consistency of job results on the error path
1972 - Fixed master-failover race condition when executed multiple times in
1974 - Fixed many bugs related to the job queue (mostly introduced during the
1975 2.2 development cycle, so not all are impacting 2.1)
1976 - Fixed instance migration with missing disk symlinks
1977 - Fixed handling of unknown jobs in ``gnt-job archive``
1978 - And many other small fixes/improvements
1982 - Enhanced both the unittest and the QA coverage
1983 - Switched the opcode validation to a generic model, and extended the
1984 validation to all opcode parameters
1985 - Changed more parts of the code that write shell scripts to use the
1987 - Switched the master daemon to use the asyncore library for the Luxi
1994 *(Released Thu, 17 Jun 2010)*
1996 - Added tool (``move-instance``) and infrastructure to move instances
1997 between separate clusters (see :doc:`separate documentation
1998 <move-instance>` and :doc:`design document <design-2.2>`)
1999 - Added per-request RPC timeout
2000 - RAPI now requires a Content-Type header for requests with a body (e.g.
2001 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2002 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2003 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2005 - Implemented initial support for running Ganeti daemons as separate
2006 users, see configure-time flags ``--with-user-prefix`` and
2007 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2009 - Instances can be removed after export (``gnt-backup export
2010 --remove-instance``)
2011 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2012 key (instead of 1024 bit)
2013 - Added new cluster configuration file for cluster domain secret
2014 - Import/export now use SSL instead of SSH
2015 - Added support for showing estimated time when exporting an instance,
2016 see the ``ganeti-os-interface(7)`` manpage and look for
2023 *(Released Tue, 16 Nov 2010)*
2025 Some more bugfixes. Unless critical bugs occur, this will be the last
2028 - Fix case of MAC special-values
2029 - Fix mac checker regex
2030 - backend: Fix typo causing "out of range" error
2031 - Add missing --units in gnt-instance list man page
2037 *(Released Tue, 24 Aug 2010)*
2040 - Don't ignore secondary node silently on non-mirrored disk templates
2042 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2043 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2044 - Properly document the usb_mouse parameter
2045 - Fix path in ganeti-rapi(8) (issue 116)
2046 - Adjust error message when the ganeti user's .ssh directory is
2048 - Add same-node-check when changing the disk template to drbd
2054 *(Released Fri, 16 Jul 2010)*
2057 - Add an option to only select some reboot types during qa/burnin.
2058 (on some hypervisors consequent reboots are not supported)
2059 - Fix infrequent race condition in master failover. Sometimes the old
2060 master ip address would be still detected as up for a short time
2061 after it was removed, causing failover to fail.
2062 - Decrease mlockall warnings when the ctypes module is missing. On
2063 Python 2.4 we support running even if no ctypes module is installed,
2064 but we were too verbose about this issue.
2065 - Fix building on old distributions, on which man doesn't have a
2067 - Fix RAPI not to ignore the MAC address on instance creation
2068 - Implement the old instance creation format in the RAPI client.
2074 *(Released Thu, 01 Jul 2010)*
2076 A small bugfix release:
2077 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2078 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2079 a non-existing option
2080 - Fix instance create: the --force-variant option was ignored
2081 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2082 - Fix modify node storage with non-FQDN arguments
2083 - Fix RAPI client to authenticate under Python 2.6 when used
2084 for more than 5 requests needing authentication
2085 - Fix gnt-instance modify -t (storage) giving a wrong error message
2086 when converting a non-shutdown drbd instance to plain
2092 *(Released Fri, 18 Jun 2010)*
2094 A small bugfix release:
2096 - Fix live migration of KVM instances started with older Ganeti
2097 versions which had fewer hypervisor parameters
2098 - Fix gnt-instance grow-disk on down instances
2099 - Fix an error-reporting bug during instance migration
2100 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2101 silently ignoring broken ones
2102 - Fix an RPC error reporting bug affecting, for example, RAPI client
2104 - Fix bug triggered by different API version os-es on different nodes
2105 - Fix a bug in instance startup with custom hvparams: OS level
2106 parameters would fail to be applied.
2107 - Fix the RAPI client under Python 2.6 (but more work is needed to
2108 make it work completely well with OpenSSL)
2109 - Fix handling of errors when resolving names from DNS
2115 *(Released Thu, 3 Jun 2010)*
2117 A medium sized development cycle. Some new features, and some
2118 fixes/small improvements/cleanups.
2120 Significant features
2121 ~~~~~~~~~~~~~~~~~~~~
2123 The node deamon now tries to mlock itself into memory, unless the
2124 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2125 its logs, and falls back to console logging. This allows emergency
2126 features such as ``gnt-node powercycle`` to work even in the event of a
2127 broken node disk (tested offlining the disk hosting the node's
2128 filesystem and dropping its memory caches; don't try this at home)
2130 KVM: add vhost-net acceleration support. It can be tested with a new
2131 enough version of the kernel and of qemu-kvm.
2133 KVM: Add instance chrooting feature. If you use privilege dropping for
2134 your VMs you can also now force them to chroot to an empty directory,
2135 before starting the emulated guest.
2137 KVM: Add maximum migration bandwith and maximum downtime tweaking
2138 support (requires a new-enough version of qemu-kvm).
2140 Cluster verify will now warn if the master node doesn't have the master
2141 ip configured on it.
2143 Add a new (incompatible) instance creation request format to RAPI which
2144 supports all parameters (previously only a subset was supported, and it
2145 wasn't possible to extend the old format to accomodate all the new
2146 features. The old format is still supported, and a client can check for
2147 this feature, before using it, by checking for its presence in the
2148 ``features`` RAPI resource.
2150 Now with ancient latin support. Try it passing the ``--roman`` option to
2151 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2152 (requires the python-roman module to be installed, in order to work).
2157 As usual many internal code refactorings, documentation updates, and
2160 - Lots of improvements and cleanups to the experimental Remote API
2161 (RAPI) client library.
2162 - A new unit test suite for the core daemon libraries.
2163 - A fix to creating missing directories makes sure the umask is not
2164 applied anymore. This enforces the same directory permissions
2166 - Better handling terminating daemons with ctrl+c (used when running
2167 them in debugging mode).
2168 - Fix a race condition in live migrating a KVM instance, when stat()
2169 on the old proc status file returned EINVAL, which is an unexpected
2171 - Fixed manpage checking with newer man and utf-8 charachters. But now
2172 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2178 *(Released Fri, 7 May 2010)*
2180 Fix a bug which prevented untagged KVM instances from starting.
2186 *(Released Fri, 7 May 2010)*
2188 Another release with a long development cycle, during which many
2189 different features were added.
2191 Significant features
2192 ~~~~~~~~~~~~~~~~~~~~
2194 The KVM hypervisor now can run the individual instances as non-root, to
2195 reduce the impact of a VM being hijacked due to bugs in the
2196 hypervisor. It is possible to run all instances as a single (non-root)
2197 user, to manually specify a user for each instance, or to dynamically
2198 allocate a user out of a cluster-wide pool to each instance, with the
2199 guarantee that no two instances will run under the same user ID on any
2202 An experimental RAPI client library, that can be used standalone
2203 (without the other Ganeti libraries), is provided in the source tree as
2204 ``lib/rapi/client.py``. Note this client might change its interface in
2205 the future, as we iterate on its capabilities.
2207 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2208 replace the cluster's certificates and crypto keys. This might help in
2209 case they have been compromised, or have simply expired.
2211 A new disk option for instance creation has been added that allows one
2212 to "adopt" currently existing logical volumes, with data
2213 preservation. This should allow easier migration to Ganeti from
2214 unmanaged (or managed via other software) instances.
2216 Another disk improvement is the possibility to convert between redundant
2217 (DRBD) and plain (LVM) disk configuration for an instance. This should
2218 allow better scalability (starting with one node and growing the
2219 cluster, or shrinking a two-node cluster to one node).
2221 A new feature that could help with automated node failovers has been
2222 implemented: if a node sees itself as offline (by querying the master
2223 candidates), it will try to shutdown (hard) all instances and any active
2224 DRBD devices. This reduces the risk of duplicate instances if an
2225 external script automatically failovers the instances on such nodes. To
2226 enable this, the cluster parameter ``maintain_node_health`` should be
2227 enabled; in the future this option (per the name) will enable other
2228 automatic maintenance features.
2230 Instance export/import now will reuse the original instance
2231 specifications for all parameters; that means exporting an instance,
2232 deleting it and the importing it back should give an almost identical
2233 instance. Note that the default import behaviour has changed from
2234 before, where it created only one NIC; now it recreates the original
2237 Cluster verify has added a few new checks: SSL certificates validity,
2238 /etc/hosts consistency across the cluster, etc.
2243 As usual, many internal changes were done, documentation fixes,
2246 - Fixed cluster initialization with disabled cluster storage (regression
2247 introduced in 2.1.1)
2248 - File-based storage supports growing the disks
2249 - Fixed behaviour of node role changes
2250 - Fixed cluster verify for some corner cases, plus a general rewrite of
2251 cluster verify to allow future extension with more checks
2252 - Fixed log spamming by watcher and node daemon (regression introduced
2254 - Fixed possible validation issues when changing the list of enabled
2256 - Fixed cleanup of /etc/hosts during node removal
2257 - Fixed RAPI response for invalid methods
2258 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2259 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2260 from ide disks, etc.)
2261 - Allow OS changes without re-installation (to record a changed OS
2262 outside of Ganeti, or to allow OS renames)
2263 - Allow instance creation without OS installation (useful for example if
2264 the OS will be installed manually, or restored from a backup not in
2266 - Implemented option to make cluster ``copyfile`` use the replication
2268 - Added list of enabled hypervisors to ssconf (possibly useful for
2270 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2272 - A partial form of node re-IP is possible via node readd, which now
2273 allows changed node primary IP
2274 - Command line utilities now show an informational message if the job is
2276 - The logs of the master daemon now show the PID/UID/GID of the
2283 *(Released Fri, 12 Mar 2010)*
2285 During the 2.1.0 long release candidate cycle, a lot of improvements and
2286 changes have accumulated with were released later as 2.1.1.
2291 The node evacuate command (``gnt-node evacuate``) was significantly
2292 rewritten, and as such the IAllocator protocol was changed - a new
2293 request type has been added. This unfortunate change during a stable
2294 series is designed to improve performance of node evacuations; on
2295 clusters with more than about five nodes and which are well-balanced,
2296 evacuation should proceed in parallel for all instances of the node
2297 being evacuated. As such, any existing IAllocator scripts need to be
2298 updated, otherwise the above command will fail due to the unknown
2299 request. The provided "dumb" allocator has not been updated; but the
2300 ganeti-htools package supports the new protocol since version 0.2.4.
2302 Another important change is increased validation of node and instance
2303 names. This might create problems in special cases, if invalid host
2304 names are being used.
2306 Also, a new layer of hypervisor parameters has been added, that sits at
2307 OS level between the cluster defaults and the instance ones. This allows
2308 customisation of virtualization parameters depending on the installed
2309 OS. For example instances with OS 'X' may have a different KVM kernel
2310 (or any other parameter) than the cluster defaults. This is intended to
2311 help managing a multiple OSes on the same cluster, without manual
2312 modification of each instance's parameters.
2314 A tool for merging clusters, ``cluster-merge``, has been added in the
2315 tools sub-directory.
2320 - Improved the int/float conversions that should make the code more
2321 robust in face of errors from the node daemons
2322 - Fixed the remove node code in case of internal configuration errors
2323 - Fixed the node daemon behaviour in face of inconsistent queue
2324 directory (e.g. read-only file-system where we can't open the files
2326 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2327 now it either aborts cleanly or, if given the new "auto_promote"
2328 parameter, will automatically promote other nodes as needed
2329 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2330 completely prevent Ganeti from working
2331 - Fixed bug for instance export when not all disks were successfully
2333 - Fixed behaviour of node add when the new node is slow in starting up
2335 - Fixed handling of signals in the LUXI client, which should improve
2336 behaviour of command-line scripts
2337 - Added checks for invalid node/instance names in the configuration (now
2338 flagged during cluster verify)
2339 - Fixed watcher behaviour for disk activation errors
2340 - Fixed two potentially endless loops in http library, which led to the
2341 RAPI daemon hanging and consuming 100% CPU in some cases
2342 - Fixed bug in RAPI daemon related to hashed passwords
2343 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2345 - Enhanced compatibility with non-Debian OSes, but not using absolute
2346 path in some commands and allowing customisation of the ssh
2347 configuration directory
2348 - Fixed possible future issue with new Python versions by abiding to the
2349 proper use of ``__slots__`` attribute on classes
2350 - Added checks that should prevent directory traversal attacks
2351 - Many documentation fixes based on feedback from users
2356 - Added an "early_release" more for instance replace disks and node
2357 evacuate, where we release locks earlier and thus allow higher
2358 parallelism within the cluster
2359 - Added watcher hooks, intended to allow the watcher to restart other
2360 daemons (e.g. from the ganeti-nbma project), but they can be used of
2361 course for any other purpose
2362 - Added a compile-time disable for DRBD barriers, to increase
2363 performance if the administrator trusts the power supply or the
2364 storage system to not lose writes
2365 - Added the option of using syslog for logging instead of, or in
2366 addition to, Ganeti's own log files
2367 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2368 can indeed boot from a paravirtual NIC
2369 - Added a generic debug level for many operations; while this is not
2370 used widely yet, it allows one to pass the debug value all the way to
2372 - Enhanced the hooks environment for instance moves (failovers,
2373 migrations) where the primary/secondary nodes changed during the
2374 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2375 - Enhanced data validations for many user-supplied values; one important
2376 item is the restrictions imposed on instance and node names, which
2377 might reject some (invalid) host names
2378 - Add a configure-time option to disable file-based storage, if it's not
2379 needed; this allows greater security separation between the master
2380 node and the other nodes from the point of view of the inter-node RPC
2382 - Added user notification in interactive tools if job is waiting in the
2383 job queue or trying to acquire locks
2384 - Added log messages when a job is waiting for locks
2385 - Added filtering by node tags in instance operations which admit
2386 multiple instances (start, stop, reboot, reinstall)
2387 - Added a new tool for cluster mergers, ``cluster-merge``
2388 - Parameters from command line which are of the form ``a=b,c=d`` can now
2389 use backslash escapes to pass in values which contain commas,
2390 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2392 - For KVM, the instance name is the first parameter passed to KVM, so
2393 that it's more visible in the process list
2399 *(Released Tue, 2 Mar 2010)*
2401 Ganeti 2.1 brings many improvements with it. Major changes:
2403 - Added infrastructure to ease automated disk repairs
2404 - Added new daemon to export configuration data in a cheaper way than
2405 using the remote API
2406 - Instance NICs can now be routed instead of being associated with a
2408 - Improved job locking logic to reduce impact of jobs acquiring multiple
2409 locks waiting for other long-running jobs
2411 In-depth implementation details can be found in the Ganeti 2.1 design
2417 - Added chroot hypervisor
2418 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2420 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2421 ``bootloader_path`` and ``bootloader_args``)
2422 - Added the ``use_localtime`` option for the xen-hvm and kvm
2423 hypervisors, and the default value for this has changed to false (in
2424 2.0 xen-hvm always enabled it)
2425 - Added luxi call to submit multiple jobs in one go
2426 - Added cluster initialization option to not modify ``/etc/hosts``
2428 - Added network interface parameters
2429 - Added dry run mode to some LUs
2430 - Added RAPI resources:
2432 - ``/2/instances/[instance_name]/info``
2433 - ``/2/instances/[instance_name]/replace-disks``
2434 - ``/2/nodes/[node_name]/evacuate``
2435 - ``/2/nodes/[node_name]/migrate``
2436 - ``/2/nodes/[node_name]/role``
2437 - ``/2/nodes/[node_name]/storage``
2438 - ``/2/nodes/[node_name]/storage/modify``
2439 - ``/2/nodes/[node_name]/storage/repair``
2441 - Added OpCodes to evacuate or migrate all instances on a node
2442 - Added new command to list storage elements on nodes (``gnt-node
2443 list-storage``) and modify them (``gnt-node modify-storage``)
2444 - Added new ssconf files with master candidate IP address
2445 (``ssconf_master_candidates_ips``), node primary IP address
2446 (``ssconf_node_primary_ips``) and node secondary IP address
2447 (``ssconf_node_secondary_ips``)
2448 - Added ``ganeti-confd`` and a client library to query the Ganeti
2449 configuration via UDP
2450 - Added ability to run hooks after cluster initialization and before
2452 - Added automatic mode for disk replace (``gnt-instance replace-disks
2454 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2455 after catastrophic data-loss
2456 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2458 - Added ``gnt-instance move`` command to move instances
2459 - Added ``gnt-cluster watcher`` command to control watcher
2460 - Added ``gnt-node powercycle`` command to powercycle nodes
2461 - Added new job status field ``lock_status``
2462 - Added parseable error codes to cluster verification (``gnt-cluster
2463 verify --error-codes``) and made output less verbose (use
2464 ``--verbose`` to restore previous behaviour)
2465 - Added UUIDs to the main config entities (cluster, nodes, instances)
2466 - Added support for OS variants
2467 - Added support for hashed passwords in the Ganeti remote API users file
2469 - Added option to specify maximum timeout on instance shutdown
2470 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2471 - Added new helper script to start and stop Ganeti daemons
2472 (``daemon-util``), with the intent to reduce the work necessary to
2473 adjust Ganeti for non-Debian distributions and to start/stop daemons
2475 - Added more unittests
2476 - Fixed critical bug in ganeti-masterd startup
2477 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2478 now customisable at the cluster level for both the KVM and Xen
2479 hypervisors using the new ``migration_port`` parameter
2480 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2481 reinstalling an instance
2482 - Allowed ``@`` in tag names
2483 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2484 - Many documentation updates
2485 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2486 - ``gnt-instance reinstall`` can now reinstall multiple instances
2487 - Updated many command line parameters
2488 - Introduced new OS API version 15
2489 - No longer support a default hypervisor
2490 - Treat virtual LVs as inexistent
2491 - Improved job locking logic to reduce lock contention
2492 - Match instance and node names case insensitively
2493 - Reimplemented bash completion script to be more complete
2500 *(Released Thu, 4 Feb 2010)*
2502 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2503 - Fix a string formatting bug
2504 - Improve safety of the code in some error paths
2505 - Improve data validation in the master of values returned from nodes
2511 *(Released Thu, 17 Dec 2009)*
2513 - Fix security issue due to missing validation of iallocator names; this
2514 allows local and remote execution of arbitrary executables
2515 - Fix failure of gnt-node list during instance removal
2516 - Ship the RAPI documentation in the archive
2522 *(Released Wed, 30 Sep 2009)*
2524 - Fixed many wrong messages
2525 - Fixed a few bugs related to the locking library
2526 - Fixed MAC checking at instance creation time
2527 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2528 - Fixed a few issues related to signal handling in both daemons and
2530 - Fixed the example startup script provided
2531 - Fixed insserv dependencies in the example startup script (patch from
2533 - Fixed handling of drained nodes in the iallocator framework
2534 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2536 - Fixed error related to invalid job IDs in job polling
2537 - Fixed job/opcode persistence on unclean master shutdown
2538 - Fixed handling of partial job processing after unclean master
2540 - Fixed error reporting from LUs, previously all errors were converted
2541 into execution errors
2542 - Fixed error reporting from burnin
2543 - Decreased significantly the memory usage of the job queue
2544 - Optimised slightly multi-job submission
2545 - Optimised slightly opcode loading
2546 - Backported the multi-job submit framework from the development
2547 branch; multi-instance start and stop should be faster
2548 - Added script to clean archived jobs after 21 days; this will reduce
2549 the size of the queue directory
2550 - Added some extra checks in disk size tracking
2551 - Added an example ethers hook script
2552 - Added a cluster parameter that prevents Ganeti from modifying of
2554 - Added more node information to RAPI responses
2555 - Added a ``gnt-job watch`` command that allows following the ouput of a
2557 - Added a bind-address option to ganeti-rapi
2558 - Added more checks to the configuration verify
2559 - Enhanced the burnin script such that some operations can be retried
2561 - Converted instance reinstall to multi-instance model
2567 *(Released Fri, 7 Aug 2009)*
2569 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2570 to allow using the pre-2.0.2 behaviour in activation, if any existing
2571 instances have mismatched disk sizes in the configuration
2572 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2573 any configuration mismatches for disk sizes
2574 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2575 failover to work on two-node clusters
2576 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2578 - Fixed detection of OS script errors in ``gnt-backup export``
2579 - Fixed exit code of ``gnt-backup export``
2585 *(Released Fri, 17 Jul 2009)*
2587 - Added experimental support for stripped logical volumes; this should
2588 enhance performance but comes with a higher complexity in the block
2589 device handling; stripping is only enabled when passing
2590 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2591 affected even in the non-stripped mode
2592 - Improved resiliency against transient failures at the end of DRBD
2593 resyncs, and in general of DRBD resync checks
2594 - Fixed a couple of issues with exports and snapshot errors
2595 - Fixed a couple of issues in instance listing
2596 - Added display of the disk size in ``gnt-instance info``
2597 - Fixed checking for valid OSes in instance creation
2598 - Fixed handling of the "vcpus" parameter in instance listing and in
2599 general of invalid parameters
2600 - Fixed http server library, and thus RAPI, to handle invalid
2601 username/password combinations correctly; this means that now they
2602 report unauthorized for queries too, not only for modifications,
2603 allowing earlier detect of configuration problems
2604 - Added a new "role" node list field, equivalent to the master/master
2605 candidate/drained/offline flags combinations
2606 - Fixed cluster modify and changes of candidate pool size
2607 - Fixed cluster verify error messages for wrong files on regular nodes
2608 - Fixed a couple of issues with node demotion from master candidate role
2609 - Fixed node readd issues
2610 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2611 - Added a new ``--no-voting`` option for masterfailover to fix failover
2612 on two-nodes clusters when the former master node is unreachable
2613 - Added instance reinstall over RAPI
2619 *(Released Tue, 16 Jun 2009)*
2621 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2622 allow re-adding the start in single-user option (regression from 1.2)
2623 - the watcher writes the instance status to a file, to allow monitoring
2624 to report the instance status (from the master) based on cached
2625 results of the watcher's queries; while this can get stale if the
2626 watcher is being locked due to other work on the cluster, this is
2627 still an improvement
2628 - the watcher now also restarts the node daemon and the rapi daemon if
2630 - fixed the watcher to handle full and drained queue cases
2631 - hooks export more instance data in the environment, which helps if
2632 hook scripts need to take action based on the instance's properties
2633 (no longer need to query back into ganeti)
2634 - instance failovers when the instance is stopped do not check for free
2635 RAM, so that failing over a stopped instance is possible in low memory
2637 - rapi uses queries for tags instead of jobs (for less job traffic), and
2638 for cluster tags it won't talk to masterd at all but read them from
2640 - a couple of error handling fixes in RAPI
2641 - drbd handling: improved the error handling of inconsistent disks after
2642 resync to reduce the frequency of "there are some degraded disks for
2643 this instance" messages
2644 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2645 error handling path called a wrong function name)
2651 *(Released Wed, 27 May 2009)*
2653 - no changes from rc5
2659 *(Released Wed, 20 May 2009)*
2661 - fix a couple of bugs (validation, argument checks)
2662 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2663 - some small improvements to RAPI and IAllocator
2664 - make watcher automatically start the master daemon if down
2670 *(Released Mon, 27 Apr 2009)*
2672 - change the OS list to not require locks; this helps with big clusters
2673 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2674 volume group is broken
2675 - ``gnt-instance info``, without any arguments, doesn't run for all
2676 instances anymore; either pass ``--all`` or pass the desired
2677 instances; this helps against mistakes on big clusters where listing
2678 the information for all instances takes a long time
2679 - miscellaneous doc and man pages fixes
2685 *(Released Wed, 8 Apr 2009)*
2687 - Change the internal locking model of some ``gnt-node`` commands, in
2688 order to reduce contention (and blocking of master daemon) when
2689 batching many creation/reinstall jobs
2690 - Fixes to Xen soft reboot
2691 - No longer build documentation at build time, instead distribute it in
2692 the archive, in order to reduce the need for the whole docbook/rst
2699 *(Released Fri, 27 Mar 2009)*
2701 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2702 - Fix watcher startup sequence, improves the behaviour of busy clusters
2703 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2704 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2705 bind address and other places
2706 - Some documentation fixes and updates
2712 *(Released Mon, 2 Mar 2009)*
2714 - More documentation updates, now all docs should be more-or-less
2716 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2718 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2720 - Fix an issue related to $libdir/run/ganeti and cluster creation
2726 *(Released Thu, 19 Feb 2009)*
2728 - Xen PVM and KVM have switched the default value for the instance root
2729 disk to the first partition on the first drive, instead of the whole
2730 drive; this means that the OS installation scripts must be changed
2732 - Man pages have been updated
2733 - RAPI has been switched by default to HTTPS, and the exported functions
2734 should all work correctly
2735 - RAPI v1 has been removed
2736 - Many improvements to the KVM hypervisor
2737 - Block device errors are now better reported
2738 - Many other bugfixes and small improvements
2744 *(Released Mon, 26 Jan 2009)*
2746 - Version 2 is a general rewrite of the code and therefore the
2747 differences are too many to list, see the design document for 2.0 in
2748 the ``doc/`` subdirectory for more details
2749 - In this beta version there is not yet a migration path from 1.2 (there
2750 will be one in the final 2.0 release)
2751 - A few significant changes are:
2753 - all commands are executed by a daemon (``ganeti-masterd``) and the
2754 various ``gnt-*`` commands are just front-ends to it
2755 - all the commands are entered into, and executed from a job queue,
2756 see the ``gnt-job(8)`` manpage
2757 - the RAPI daemon supports read-write operations, secured by basic
2758 HTTP authentication on top of HTTPS
2759 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2760 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2761 to migrate to DRBD 8 first while still running Ganeti 1.2)
2762 - DRBD devices are using statically allocated minor numbers, which
2763 will be assigned to existing instances during the migration process
2764 - there is support for both Xen PVM and Xen HVM instances running on
2766 - KVM virtualization is supported too
2767 - file-based storage has been implemented, which means that it is
2768 possible to run the cluster without LVM and DRBD storage, for
2769 example using a shared filesystem exported from shared storage (and
2770 still have live migration)
2776 *(Released Tue, 13 Jan 2009)*
2778 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2779 - Reuse the old instance mac address by default on instance import, if
2780 the instance name is the same.
2781 - Handle situations in which the node info rpc returns incomplete
2783 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2784 - Improved version of batcher:
2786 - state file support
2787 - instance mac address support
2788 - support for HVM clusters/instances
2790 - Add an option to show the number of cpu sockets and nodes in
2792 - Support OSes that handle more than one version of the OS api (but do
2793 not change the current API in any other way)
2794 - Fix ``gnt-node migrate``
2795 - ``gnt-debug`` man page
2796 - Fixes various more typos and small issues
2797 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2803 *(Released Wed, 24 Sep 2008)*
2805 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2806 type of disk exported to fully virtualized instances.
2807 - provide access to the serial console of HVM instances
2808 - instance auto_balance flag, set by default. If turned off it will
2809 avoid warnings on cluster verify if there is not enough memory to fail
2810 over an instance. in the future it will prevent automatically failing
2811 it over when we will support that.
2812 - batcher tool for instance creation, see ``tools/README.batcher``
2813 - ``gnt-instance reinstall --select-os`` to interactively select a new
2814 operating system when reinstalling an instance.
2815 - when changing the memory amount on instance modify a check has been
2816 added that the instance will be able to start. also warnings are
2817 emitted if the instance will not be able to fail over, if auto_balance
2819 - documentation fixes
2820 - sync fields between ``gnt-instance list/modify/add/import``
2821 - fix a race condition in drbd when the sync speed was set after giving
2822 the device a remote peer.
2828 *(Released Tue, 22 Jul 2008)*
2830 - note: the allowed size and number of tags per object were reduced
2831 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2832 - fixed twisted 8.x compatibility
2833 - fixed ``gnt-instance replace-disks`` with iallocator
2834 - add TCP keepalives on twisted connections to detect restarted nodes
2835 - disk increase support, see ``gnt-instance grow-disk``
2836 - implement bulk node/instance query for RAPI
2837 - add tags in node/instance listing (optional)
2838 - experimental migration (and live migration) support, read the man page
2839 for ``gnt-instance migrate``
2840 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2841 has some small improvements in handling its state file
2847 *(Released Fri, 13 Jun 2008)*
2849 - Experimental readonly, REST-based remote API implementation;
2850 automatically started on master node, TCP port 5080, if enabled by
2851 ``--enable-rapi`` parameter to configure script.
2852 - Instance allocator support. Add and import instance accept a
2853 ``--iallocator`` parameter, and call that instance allocator to decide
2854 which node to use for the instance. The iallocator document describes
2855 what's expected from an allocator script.
2856 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2857 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2858 node is lost there is still enough memory to fail over the instances
2860 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2861 ``gnt-cluster verify``, to check for site-specific compliance. All the
2862 hooks will run, and their output, if any, will be displayed. Any
2863 failing hook will make the verification return an error value.
2864 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2865 primary and secondary interfaces
2866 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2867 that is still declared as part of the cluster and has failed.
2868 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2869 specifying output fields, that just adds the chosen fields to the
2871 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2872 export from the filesystem.
2873 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2874 have been added. Using them you can enable/disable acpi and pae
2875 support, and specify a path for a cd image to be exported to the
2876 instance. These parameters as the name suggest only work on HVM
2878 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2879 PAE support will be set to the previously hardcoded values, but the
2880 (previously hardcoded) path to the CDROM ISO image will be unset and
2881 if required, needs to be set manually with ``gnt-instance modify``
2883 - The address to which an instance's VNC console is bound is now
2884 selectable per-instance, rather than being cluster wide. Of course
2885 this only applies to instances controlled via VNC, so currently just
2886 applies to HVM clusters.
2892 *(Released Mon, 18 Feb 2008)*
2894 - more tweaks to the disk activation code (especially helpful for DRBD)
2895 - change the default ``gnt-instance list`` output format, now there is
2896 one combined status field (see the manpage for the exact values this
2898 - some more fixes for the mac export to hooks change
2899 - make Ganeti not break with DRBD 8.2.x (which changed the version
2900 format in ``/proc/drbd``) (issue 24)
2901 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2902 template, allowing migration from DRBD0.7+MD to DRBD8
2908 *(Released Wed, 30 Jan 2008)*
2910 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2912 - add command aliases infrastructure and a few aliases
2913 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2914 man pages and the ``--help`` option of ``gnt-node
2915 list``/``gnt-instance list``
2916 - fix ``gnt-backup list`` with down nodes (issue 21)
2917 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2918 - fix the dist archive and add a check for including svn/git files in
2920 - some developer-related changes: improve the burnin and the QA suite,
2921 add an upload script for testing during development
2927 *(Released Wed, 16 Jan 2008)*
2929 - experimental HVM support, read the install document, section
2930 "Initializing the cluster"
2931 - allow for the PVM hypervisor per-instance kernel and initrd paths
2932 - add a new command ``gnt-cluster verify-disks`` which uses a new
2933 algorithm to improve the reconnection of the DRBD pairs if the device
2934 on the secondary node has gone away
2935 - make logical volume code auto-activate LVs at disk activation time
2936 - slightly improve the speed of activating disks
2937 - allow specification of the MAC address at instance creation time, and
2938 changing it later via ``gnt-instance modify``
2939 - fix handling of external commands that generate lots of output on
2941 - update documentation with regard to minimum version of DRBD8 supported
2947 *(Released Tue, 4 Dec 2007)*
2949 - Log the ``xm create`` output to the node daemon log on failure (to
2950 help diagnosing the error)
2951 - In debug mode, log all external commands output if failed to the logs
2952 - Change parsing of lvm commands to ignore stderr
2958 *(Released Wed, 28 Nov 2007)*
2960 - Another round of updates to the DRBD 8 code to deal with more failures
2961 in the replace secondary node operation
2962 - Some more logging of failures in disk operations (lvm, drbd)
2963 - A few documentation updates
2970 *(Released Tue, 13 Nov 2007)*
2972 - Change configuration file format from Python's Pickle to JSON.
2973 Upgrading is possible using the cfgupgrade utility.
2974 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2975 faster replace disks and is more stable (DRBD 8 has many improvements
2976 compared to DRBD 0.7)
2977 - Added command line tags support (see man pages for ``gnt-instance``,
2978 ``gnt-node``, ``gnt-cluster``)
2979 - Added instance rename support
2980 - Added multi-instance startup/shutdown
2981 - Added cluster rename support
2982 - Added ``gnt-node evacuate`` to simplify some node operations
2983 - Added instance reboot operation that can speedup reboot as compared to
2985 - Soften the requirement that hostnames are in FQDN format
2986 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2988 - Removed dependency on debian's patched fping that uses the
2989 non-standard ``-S`` option
2990 - Now the OS definitions are searched for in multiple, configurable
2991 paths (easier for distros to package)
2992 - Some changes to the hooks infrastructure (especially the new
2993 post-configuration update hook)
2994 - Other small bugfixes
2996 .. vim: set textwidth=72 syntax=rst :