10 - Fixed Luxi daemon socket permissions after master-failover
16 *(Released Thu, 07 Nov 2013)*
18 - DRBD: ensure peers are UpToDate for dual-primary
19 - Improve error message for replace-disks
20 - More dependency checks at configure time
21 - Placate warnings on ganeti.outils_unittest.py
27 *(Released Thu, 17 Oct 2013)*
29 - Correctly start/stop luxid during gnt-cluster master-failover
30 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
31 - Fix path for the job queue serial file
32 - Improved harep man page
33 - Minor documentation improvements
39 *(Released Mon, 30 Sep 2013)*
41 Incompatible/important changes
42 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
44 - Instance policy can contain multiple instance specs, as described in
45 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
46 <design-partitioned>`. As a consequence, it's not possible to partially change
47 or override instance specs. Bounding specs (min and max) can be specified as a
48 whole using the new option ``--ipolicy-bounds-specs``, while standard
49 specs use the new option ``--ipolicy-std-specs``.
50 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
51 gnt-instance is a valid YAML object.
52 - hail now honors network restrictions when allocating nodes. This led to an
53 update of the IAllocator protocol. See the IAllocator documentation for
55 - confd now only answers static configuration request over the network. luxid
56 was extracted, listens on the local LUXI socket and responds to live queries.
57 This allows finer grained permissions if using separate users.
62 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
63 to always require authentication, ``--require-authentication``. It can
64 be specified in ``$sysconfdir/default/ganeti``.
65 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
66 be used to manage the disk templates to be used by instances in the cluster.
67 Initially, it will be set to a list that includes plain, drbd, if they were
68 enabled by specifying a volume group name, and file and sharedfile, if those
69 were enabled at configure time. Additionally, it will include all disk
70 templates that are currently used by instances. The order of disk templates
71 will be based on Ganeti's history of supporting them. In the future, the
72 first entry of the list will be used as a default disk template on instance
74 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
75 configuration back to the previous stable version.
76 - Disk templates in group ipolicy can be restored to the default value.
77 - Initial support for diskless instances and virtual clusters in QA.
78 - More QA and unit tests for instance policies.
79 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
80 describing why the opcode itself was executed.
81 - The monitoring daemon is now available. It allows users to query the cluster
82 for obtaining information about the status of the system. The daemon is only
83 responsible for providing the information over the network: the actual data
84 gathering is performed by data collectors (currently, only the DRBD status
85 collector is available).
86 - In order to help developers work on Ganeti, a new script
87 (``devel/build_chroot``) is provided, for building a chroot that contains all
88 the required development libraries and tools for compiling Ganeti on a Debian
90 - A new tool, ``harep``, for performing self-repair and recreation of instances
91 in Ganeti has been added.
92 - Split queries are enabled for tags, network, exports, cluster info, groups,
94 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
95 It prints the command line to set the current policies, to ease
97 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
98 devices for KVM virtio-net interfaces will get created with VNET_HDR
99 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
100 virtio-net interfaces, which prevents host kernel tainting and log
101 flooding, when dealing with broken or malicious virtio-net drivers.
102 It's set to true by default.
103 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
105 - Support 'viridian' parameter in Xen HVM
106 - Support DSA SSH keys in bootstrap
107 - To simplify the work of packaging frameworks that want to add the needed users
108 and groups in a split-user setup themselves, at build time three files in
109 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
110 the groups to be generated, the ``users`` file contains, one per line, the
111 users to be generated, optionally followed by their primary group, where
112 important. The ``groupmemberships`` file contains, one per line, additional
113 user-group membership relations that need to be established. The syntax of
114 these files will remain stable in all future versions.
119 The following new dependencies have been added:
122 - The ``curl`` library is not optional anymore for compiling the Haskell code.
123 - ``snap-server`` library (if monitoring is enabled).
126 - The minimum Python version needed to run Ganeti is now 2.6.
127 - ``yaml`` library (only for running the QA).
131 - Perform proper cleanup on termination of Haskell daemons
132 - Fix corner-case in handling of remaining retry time
138 *(Released Tue, 17 Sep 2013)*
140 - To simplify the work of packaging frameworks that want to add the needed users
141 and groups in a split-user setup themselves, at build time three files in
142 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
143 the groups to be generated, the ``users`` file contains, one per line, the
144 users to be generated, optionally followed by their primary group, where
145 important. The ``groupmemberships`` file contains, one per line, additional
146 user-group membership relations that need to be established. The syntax of
147 these files will remain stable in all future versions.
148 - Add a default to file-driver when unspecified over RAPI (Issue 571)
149 - Mark the DSA host pubkey as optional, and remove it during config downgrade
151 - Some documentation fixes
157 *(Released Tue, 27 Aug 2013)*
159 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
161 - Support 'viridian' parameter in Xen HVM (Issue 233)
162 - Include VCS version in ``gnt-cluster version``
163 - Support DSA SSH keys in bootstrap (Issue 338)
164 - Fix batch creation of instances
165 - Use FQDN to check master node status (Issue 551)
166 - Make the DRBD collector more failure-resilient
172 *(Released Fri, 2 Aug 2013)*
174 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
176 - Fix upgrading/downgrading from 2.7
177 - Increase maximum RAPI message size
178 - Documentation updates
179 - Split ``confd`` between ``luxid`` and ``confd``
180 - Merge 2.7 series up to the 2.7.1 release
181 - Allow the ``modify_etc_hosts`` option to be changed
182 - Add better debugging for ``luxid`` queries
183 - Expose bulk parameter for GetJobs in RAPI client
184 - Expose missing ``network`` fields in RAPI
185 - Add some ``cluster verify`` tests
186 - Some unittest fixes
187 - Fix a malfunction in ``hspace``'s tiered allocation
188 - Fix query compatibility between haskell and python implementations
189 - Add the ``vnet_hdr`` HV parameter for KVM
190 - Add ``--cleanup`` to instance failover
191 - Change the connected groups format in ``gnt-network info`` output; it
192 was previously displayed as a raw list by mistake. (Merged from 2.7)
198 *(Released Mon, 24 Jun 2013)*
200 This was the first beta release of the 2.8 series. All important changes
201 are listed in the latest 2.8 entry.
207 *(Released Thu, 26 Sep 2013)*
209 - Change the connected groups format in ``gnt-network info`` output; it
210 was previously displayed as a raw list by mistake
211 - Check disk template in right dict when copying
212 - Support multi-instance allocs without iallocator
213 - Fix some errors in the documentation
214 - Fix formatting of tuple in an error message
220 *(Released Thu, 25 Jul 2013)*
222 - Add logrotate functionality in daemon-util
223 - Add logrotate example file
224 - Add missing fields to network queries over rapi
225 - Fix network object timestamps
226 - Add support for querying network timestamps
227 - Fix a typo in the example crontab
228 - Fix a documentation typo
234 *(Released Thu, 04 Jul 2013)*
236 Incompatible/important changes
237 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
239 - Instance policies for disk size were documented to be on a per-disk
240 basis, but hail applied them to the sum of all disks. This has been
242 - ``hbal`` will now exit with status 0 if, during job execution over
243 LUXI, early exit has been requested and all jobs are successful;
244 before, exit status 1 was used, which cannot be differentiated from
246 - Compatibility with newer versions of rbd has been fixed
247 - ``gnt-instance batch-create`` has been changed to use the bulk create
248 opcode from Ganeti. This lead to incompatible changes in the format of
249 the JSON file. It's now not a custom dict anymore but a dict
250 compatible with the ``OpInstanceCreate`` opcode.
251 - Parent directories for file storage need to be listed in
252 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
253 write the file automatically based on old configuration values, but it
254 can not distribute it across all nodes and the file contents should be
255 verified. Use ``gnt-cluster copyfile
256 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
257 upgraded. The reason for requiring this list of paths now is that
258 before it would have been possible to inject new paths via RPC,
259 allowing files to be created in arbitrary locations. The RPC protocol
260 is protected using SSL/X.509 certificates, but as a design principle
261 Ganeti does not permit arbitrary paths to be passed.
262 - The parsing of the variants file for OSes (see
263 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
264 lines and comment lines (starting with ``#``) are ignored for better
266 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
267 longer available. ``gnt-node add`` now invokes a new tool on the
268 destination node, named ``prepare-node-join``, to configure the SSH
269 daemon. Paramiko is no longer necessary to configure nodes' SSH
270 daemons via ``gnt-node add``.
271 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
272 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
273 the flag is not reset after a master failover.
274 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
275 is recommended. 2.6 will be mandatory from the 2.8 series.
281 - New network management functionality to support automatic allocation
282 of IP addresses and managing of network parameters. See
283 :manpage:`gnt-network(8)` for more details.
284 - New external storage backend, to allow managing arbitrary storage
285 systems external to the cluster. See
286 :manpage:`ganeti-extstorage-interface(7)`.
287 - New ``exclusive-storage`` node parameter added, restricted to
288 nodegroup level. When it's set to true, physical disks are assigned in
289 an exclusive fashion to instances, as documented in :doc:`Partitioned
290 Ganeti <design-partitioned>`. Currently, only instances using the
291 ``plain`` disk template are supported.
292 - The KVM hypervisor has been updated with many new hypervisor
293 parameters, including a generic one for passing arbitrary command line
294 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
295 compatible up to qemu 1.4.
296 - A new tool, called ``mon-collector``, is the stand-alone executor of
297 the data collectors for a monitoring system. As of this version, it
298 just includes the DRBD data collector, that can be executed by calling
299 ``mon-collector`` using the ``drbd`` parameter. See
300 :manpage:`mon-collector(7)`.
301 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
302 for RAPI users. It allows granting permissions to query for
303 information to a specific user without giving
304 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
305 - A new tool named ``node-cleanup`` has been added. It cleans remains of
306 a cluster from a machine by stopping all daemons, removing
307 certificates and ssconf files. Unless the ``--no-backup`` option is
308 given, copies of the certificates are made.
309 - Instance creations now support the use of opportunistic locking,
310 potentially speeding up the (parallel) creation of multiple instances.
311 This feature is currently only available via the :doc:`RAPI
312 <rapi>` interface and when an instance allocator is used. If the
313 ``opportunistic_locking`` parameter is set the opcode will try to
314 acquire as many locks as possible, but will not wait for any locks
315 held by other opcodes. If not enough resources can be found to
316 allocate the instance, the temporary error code
317 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
318 retried thereafter, with or without opportunistic locking.
319 - New experimental linux-ha resource scripts.
320 - Restricted-commands support: ganeti can now be asked (via command line
321 or rapi) to perform commands on a node. These are passed via ganeti
322 RPC rather than ssh. This functionality is restricted to commands
323 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
324 reasons. The file is not copied automatically.
330 - Diskless instances are now externally mirrored (Issue 237). This for
331 now has only been tested in conjunction with explicit target nodes for
333 - Queries not needing locks or RPC access to the node can now be
334 performed by the confd daemon, making them independent from jobs, and
335 thus faster to execute. This is selectable at configure time.
336 - The functionality for allocating multiple instances at once has been
337 overhauled and is now also available through :doc:`RAPI <rapi>`.
339 There are no significant changes from version 2.7.0~rc3.
345 *(Released Tue, 25 Jun 2013)*
347 - Fix permissions on the confd query socket (Issue 477)
348 - Fix permissions on the job archive dir (Issue 498)
349 - Fix handling of an internal exception in replace-disks (Issue 472)
350 - Fix gnt-node info handling of shortened names (Issue 497)
351 - Fix gnt-instance grow-disk when wiping is enabled
352 - Documentation improvements, and support for newer pandoc
353 - Fix hspace honoring ipolicy for disks (Issue 484)
354 - Improve handling of the ``kvm_extra`` HV parameter
360 *(Released Fri, 24 May 2013)*
362 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
364 - Disks added through ``gnt-instance modify`` or created through
365 ``gnt-instance recreate-disks`` are wiped, if the
366 ``prealloc_wipe_disks`` flag is set.
367 - If wiping newly created disks fails, the disks are removed. Also,
368 partial failures in creating disks through ``gnt-instance modify``
369 triggers a cleanup of the partially-created disks.
370 - Removing the master IP address doesn't fail if the address has been
372 - Fix ownership of the OS log dir
373 - Workaround missing SO_PEERCRED constant (Issue 191)
379 *(Released Fri, 3 May 2013)*
381 This was the first release candidate of the 2.7 series. Since beta3:
383 - Fix kvm compatibility with qemu 1.4 (Issue 389)
384 - Documentation updates (admin guide, upgrade notes, install
385 instructions) (Issue 372)
386 - Fix gnt-group list nodes and instances count (Issue 436)
387 - Fix compilation without non-mandatory libraries (Issue 441)
388 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
389 - Make confd logging more verbose at INFO level (Issue 435)
390 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
391 - Fix failure path for instance storage type conversion (Issue 229)
392 - Update htools text backend documentation
393 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
394 - Disable inter-cluster instance move for file-based instances, because
395 it is dependant on instance export, which is not supported for
396 file-based instances. (Issue 414)
397 - Fix gnt-job crashes on non-ascii characters (Issue 427)
398 - Fix volume group checks on non-vm-capable nodes (Issue 432)
404 *(Released Mon, 22 Apr 2013)*
406 This was the third beta release of the 2.7 series. Since beta2:
408 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
409 - Fix data loss on wrong usage of ``gnt-instance move``
410 - Properly export errors in confd-based job queries
411 - Add ``users-setup`` tool
412 - Fix iallocator protocol to report 0 as a disk size for diskless
413 instances. This avoids hail breaking when a diskless instance is
415 - Fix job queue directory permission problem that made confd job queries
416 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
417 for access to archived jobs (Issue 406).
418 - Limit the sizes of networks supported by ``gnt-network`` to something
419 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
420 - Fix bugs in instance disk template conversion
421 - Fix GHC 7 compatibility
422 - Fix ``burnin`` install path (Issue 426).
423 - Allow very small disk grows (Issue 347).
424 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
425 that noded doesn't import masterd code (Issue 419).
426 - Make sure the default metavg at cluster init is the same as the vg, if
427 unspecified (Issue 358).
428 - Fix cleanup of partially created disks (part of Issue 416)
434 *(Released Tue, 2 Apr 2013)*
436 This was the second beta release of the 2.7 series. Since beta1:
438 - Networks no longer have a "type" slot, since this information was
439 unused in Ganeti: instead of it tags should be used.
440 - The rapi client now has a ``target_node`` option to MigrateInstance.
441 - Fix early exit return code for hbal (Issue 386).
442 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
443 - Fix ``rbd showmapped`` output parsing (Issue 312).
444 - Networks are now referenced indexed by UUID, rather than name. This
445 will require running cfgupgrade, from 2.7.0beta1, if networks are in
447 - The OS environment now includes network information.
448 - Deleting of a network is now disallowed if any instance nic is using
449 it, to prevent dangling references.
450 - External storage is now documented in man pages.
451 - The exclusive_storage flag can now only be set at nodegroup level.
452 - Hbal can now submit an explicit priority with its jobs.
453 - Many network related locking fixes.
454 - Bump up the required pylint version to 0.25.1.
455 - Fix the ``no_remember`` option in RAPI client.
456 - Many ipolicy related tests, qa, and fixes.
457 - Many documentation improvements and fixes.
458 - Fix building with ``--disable-file-storage``.
459 - Fix ``-q`` option in htools, which was broken if passed more than
461 - Some haskell/python interaction improvements and fixes.
462 - Fix iallocator in case of missing LVM storage.
463 - Fix confd config load in case of ``--no-lvm-storage``.
464 - The confd/query functionality is now mentioned in the security
471 *(Released Wed, 6 Feb 2013)*
473 This was the first beta release of the 2.7 series. All important changes
474 are listed in the latest 2.7 entry.
480 *(Released Fri, 21 Dec 2012)*
482 Important behaviour change: hbal won't rebalance anymore instances which
483 have the ``auto_balance`` attribute set to false. This was the intention
484 all along, but until now it only skipped those from the N+1 memory
485 reservation (DRBD-specific).
487 A significant number of bug fixes in this release:
489 - Fixed disk adoption interaction with ipolicy checks.
490 - Fixed networking issues when instances are started, stopped or
491 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
492 - Fixed the warning in cluster verify for shared storage instances not
494 - Fixed removal of storage directory on shared file storage (issue 262).
495 - Fixed validation of LVM volume group name in OpClusterSetParams
496 (``gnt-cluster modify``) (issue 285).
497 - Fixed runtime memory increases (``gnt-instance modify -m``).
498 - Fixed live migration under Xen's ``xl`` mode.
499 - Fixed ``gnt-instance console`` with ``xl``.
500 - Fixed building with newer Haskell compiler/libraries.
501 - Fixed PID file writing in Haskell daemons (confd); this prevents
502 restart issues if confd was launched manually (outside of
503 ``daemon-util``) while another copy of it was running
504 - Fixed a type error when doing live migrations with KVM (issue 297) and
505 the error messages for failing migrations have been improved.
506 - Fixed opcode validation for the out-of-band commands (``gnt-node
508 - Fixed a type error when unsetting OS hypervisor parameters (issue
509 311); now it's possible to unset all OS-specific hypervisor
511 - Fixed the ``dry-run`` mode for many operations: verification of
512 results was over-zealous but didn't take into account the ``dry-run``
513 operation, resulting in "wrong" failures.
514 - Fixed bash completion in ``gnt-job list`` when the job queue has
515 hundreds of entries; especially with older ``bash`` versions, this
516 results in significant CPU usage.
518 And lastly, a few other improvements have been made:
520 - Added option to force master-failover without voting (issue 282).
521 - Clarified error message on lock conflict (issue 287).
522 - Logging of newly submitted jobs has been improved (issue 290).
523 - Hostname checks have been made uniform between instance rename and
525 - The ``--submit`` option is now supported by ``gnt-debug delay``.
526 - Shutting down the master daemon by sending SIGTERM now stops it from
527 processing jobs waiting for locks; instead, those jobs will be started
528 once again after the master daemon is started the next time (issue
530 - Support for Xen's ``xl`` program has been improved (besides the fixes
532 - Reduced logging noise in the Haskell confd daemon (only show one log
533 entry for each config reload, instead of two).
534 - Several man page updates and typo fixes.
540 *(Released Fri, 12 Oct 2012)*
542 A small bugfix release. Among the bugs fixed:
544 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
545 made the command unusable.
546 - Commands that issue many jobs don't fail anymore just because some jobs
547 take so long that other jobs are archived.
548 - Failures during ``gnt-instance reinstall`` are reflected by the exit
550 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
552 - When ``always_failover`` is set, ``--allow-failover`` is not required
553 in migrate commands anymore.
554 - ``bash_completion`` works even if extglob is disabled.
555 - Fixed bug with locks that made failover for RDB-based instances fail.
556 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
557 a random node instead of one based on the allocator metric.
558 - Support for newer versions of pylint and pep8.
559 - Hail doesn't fail anymore when trying to add an instance of type
560 ``file``, ``sharedfile`` or ``rbd``.
561 - Added new Makefile target to rebuild the whole distribution, so that
562 all files are included.
568 *(Released Fri, 27 Jul 2012)*
571 .. attention:: The ``LUXI`` protocol has been made more consistent
572 regarding its handling of command arguments. This, however, leads to
573 incompatibility issues with previous versions. Please ensure that you
574 restart Ganeti daemons soon after the upgrade, otherwise most
575 ``LUXI`` calls (job submission, setting/resetting the drain flag,
576 pausing/resuming the watcher, cancelling and archiving jobs, querying
577 the cluster configuration) will fail.
586 The current ``admin_up`` field, which used to denote whether an instance
587 should be running or not, has been removed. Instead, ``admin_state`` is
588 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
590 The rational behind this is that an instance being “down” can have
593 - it could be down during a reboot
594 - it could be temporarily be down for a reinstall
595 - or it could be down because it is deprecated and kept just for its
598 The previous Boolean state was making it difficult to do capacity
599 calculations: should Ganeti reserve memory for a down instance? Now, the
600 tri-state field makes it clear:
602 - in ``up`` and ``down`` state, all resources are reserved for the
603 instance, and it can be at any time brought up if it is down
604 - in ``offline`` state, only disk space is reserved for it, but not
607 The field can have an extra use: since the transition between ``up`` and
608 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
609 transition between ``offline`` and ``down`` is done via ``gnt-instance
610 modify``, it is possible to given different rights to users. For
611 example, owners of an instance could be allowed to start/stop it, but
612 not transition it out of the offline state.
614 Instance policies and specs
615 +++++++++++++++++++++++++++
617 In previous Ganeti versions, an instance creation request was not
618 limited on the minimum size and on the maximum size just by the cluster
619 resources. As such, any policy could be implemented only in third-party
620 clients (RAPI clients, or shell wrappers over ``gnt-*``
621 tools). Furthermore, calculating cluster capacity via ``hspace`` again
622 required external input with regards to instance sizes.
624 In order to improve these workflows and to allow for example better
625 per-node group differentiation, we introduced instance specs, which
628 - minimum instance disk size, disk count, memory size, cpu count
629 - maximum values for the above metrics
630 - and “standard” values (used in ``hspace`` to calculate the standard
633 The minimum/maximum values can be also customised at node-group level,
634 for example allowing more powerful hardware to support bigger instance
637 Beside the instance specs, there are a few other settings belonging to
638 the instance policy framework. It is possible now to customise, per
639 cluster and node-group:
641 - the list of allowed disk templates
642 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
643 - the maximum ratio of instance to spindles (see below for more
644 information) for local storage
646 All these together should allow all tools that talk to Ganeti to know
647 what are the ranges of allowed values for instances and the
648 over-subscription that is allowed.
650 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
651 instance configuration, and the physical CPU configuration from the
652 node. For the spindle ratios however, we didn't track before these
653 values, so new parameters have been added:
655 - a new node parameter ``spindle_count``, defaults to 1, customisable at
656 node group or node level
657 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
659 Note that spindles in this context doesn't need to mean actual
660 mechanical hard-drives; it's just a relative number for both the node
661 I/O capacity and instance I/O consumption.
663 Instance migration behaviour
664 ++++++++++++++++++++++++++++
666 While live-migration is in general desirable over failover, it is
667 possible that for some workloads it is actually worse, due to the
668 variable time of the “suspend” phase during live migration.
670 To allow the tools to work consistently over such instances (without
671 having to hard-code instance names), a new backend parameter
672 ``always_failover`` has been added to control the migration/failover
673 behaviour. When set to True, all migration requests for an instance will
674 instead fall-back to failover.
676 Instance memory ballooning
677 ++++++++++++++++++++++++++
679 Initial support for memory ballooning has been added. The memory for an
680 instance is no longer fixed (backend parameter ``memory``), but instead
681 can vary between minimum and maximum values (backend parameters
682 ``minmem`` and ``maxmem``). Currently we only change an instance's
685 - live migrating or failing over and instance and the target node
686 doesn't have enough memory
687 - user requests changing the memory via ``gnt-instance modify
693 In order to control the use of specific CPUs by instance, support for
694 controlling CPU pinning has been added for the Xen, HVM and LXC
695 hypervisors. This is controlled by a new hypervisor parameter
696 ``cpu_mask``; details about possible values for this are in the
697 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
698 VCPU-to-CPU mapping) form will work well only when all nodes in your
699 cluster have the same amount of CPUs.
704 Another area in which Ganeti was not customisable were the parameters
705 used for storage configuration, e.g. how many stripes to use for LVM,
706 DRBD resync configuration, etc.
708 To improve this area, we've added disks parameters, which are
709 customisable at cluster and node group level, and which allow to
710 specify various parameters for disks (DRBD has the most parameters
711 currently), for example:
713 - DRBD resync algorithm and parameters (e.g. speed)
714 - the default VG for meta-data volumes for DRBD
715 - number of stripes for LVM (plain disk template)
718 These parameters can be modified via ``gnt-cluster modify -D …`` and
719 ``gnt-group modify -D …``, and are used at either instance creation (in
720 case of LVM stripes, for example) or at disk “activation” time
723 Rados block device support
724 ++++++++++++++++++++++++++
726 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
727 denoted by the ``rbd`` disk template type. This is considered
728 experimental, feedback is welcome. For details on configuring it, see
729 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
734 The existing master IP functionality works well only in simple setups (a
735 single network shared by all nodes); however, if nodes belong to
736 different networks, then the ``/32`` setup and lack of routing
737 information is not enough.
739 To allow the master IP to function well in more complex cases, the
740 system was reworked as follows:
742 - a master IP netmask setting has been added
743 - the master IP activation/turn-down code was moved from the node daemon
745 - whether to run the Ganeti-supplied master IP script or a user-supplied
746 on is a ``gnt-cluster init`` setting
748 Details about the location of the standard and custom setup scripts are
749 in the man page :manpage:`gnt-cluster(8)`; for information about the
750 setup script protocol, look at the Ganeti-supplied script.
755 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
758 It is now possible to use TLS-protected connections, and when renewing
759 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
760 it is now possible to specify spice or spice CA certificates. Also, it
761 is possible to configure a password for SPICE sessions via the
762 hypervisor parameter ``spice_password_file``.
764 There are also new parameters to control the compression and streaming
765 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
766 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
767 for the spice parameters.
769 Lastly, it is now possible to see the SPICE connection information via
770 ``gnt-instance console``.
775 A new tool (``tools/ovfconverter``) has been added that supports
776 conversion between Ganeti and the `Open Virtualization Format
777 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
780 This relies on the ``qemu-img`` tool to convert the disk formats, so the
781 actual compatibility with other virtualization solutions depends on it.
786 The configuration query daemon (``ganeti-confd``) is now optional, and
787 has been rewritten in Haskell; whether to use the daemon at all, use the
788 Python (default) or the Haskell version is selectable at configure time
789 via the ``--enable-confd`` parameter, which can take one of the
790 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
791 daemon will result in a smaller footprint; for larger systems, we
792 welcome feedback on the Haskell version which might become the default
795 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
796 daemon running. The Python version doesn't implement the new call.
799 User interface changes
800 ~~~~~~~~~~~~~~~~~~~~~~
802 We have replaced the ``--disks`` option of ``gnt-instance
803 replace-disks`` with a more flexible ``--disk`` option, which allows
804 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
805 disk size and mode can be changed upon recreation (via ``gnt-instance
806 recreate-disks``, which accepts the same ``--disk`` option).
808 As many people are used to a ``show`` command, we have added that as an
809 alias to ``info`` on all ``gnt-*`` commands.
811 The ``gnt-instance grow-disk`` command has a new mode in which it can
812 accept the target size of the disk, instead of the delta; this can be
813 more safe since two runs in absolute mode will be idempotent, and
814 sometimes it's also easier to specify the desired size directly.
816 Also the handling of instances with regard to offline secondaries has
817 been improved. Instance operations should not fail because one of it's
818 secondary nodes is offline, even though it's safe to proceed.
820 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
821 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
822 minors to instance name.
827 RAPI coverage has improved, with (for example) new resources for
828 recreate-disks, node power-cycle, etc.
833 There is partial support for ``xl`` in the Xen hypervisor; feedback is
836 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
837 whether to still support Python 2.4 or move to Python 2.6 as minimum
840 Support for Fedora has been slightly improved; the provided example
841 init.d script should work better on it and the INSTALL file should
842 document the needed dependencies.
847 The deprecated ``QueryLocks`` LUXI request has been removed. Use
848 ``Query(what=QR_LOCK, ...)`` instead.
850 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
851 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
852 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
853 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
854 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
856 RAPI client: ``CertificateError`` now derives from
857 ``GanetiApiError``. This should make it more easy to handle Ganeti
860 Deprecation warnings due to PyCrypto/paramiko import in
861 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
862 make sure to run an up-to-date paramiko version, if you use this tool.
864 The QA scripts now depend on Python 2.5 or above (the main code base
865 still works with Python 2.4).
867 The configuration file (``config.data``) is now written without
868 indentation for performance reasons; if you want to edit it, it can be
869 re-formatted via ``tools/fmtjson``.
871 A number of bugs has been fixed in the cluster merge tool.
873 ``x509`` certification verification (used in import-export) has been
874 changed to allow the same clock skew as permitted by the cluster
875 verification. This will remove some rare but hard to diagnose errors in
882 *(Released Thu, 19 Jul 2012)*
884 Very few changes from rc4 to the final release, only bugfixes:
886 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
887 instance, fix CDROM booting for KVM instances)
888 - fixed node group modification of node parameters
889 - fixed issue in LUClusterVerifyGroup with multi-group clusters
890 - fixed generation of bash completion to ensure a stable ordering
897 *(Released Fri, 13 Jul 2012)*
899 Third release candidate for 2.6. The following changes were done from
902 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
903 - Fixed an inconsistency in the LUXI protocol with the provided
904 arguments (NOT backwards compatible)
905 - Fixed a bug with node groups ipolicy where ``min`` was greater than
906 the cluster ``std`` value
907 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
908 easier instance debugging on nodes (requires ``hconfd`` to work)
914 *(Released Tue, 03 Jul 2012)*
916 Second release candidate for 2.6. The following changes were done from
919 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
921 - Fixed a RAPI regression on missing beparams/memory
922 - Fixed redistribution of files on offline nodes
923 - Added possibility to run activate-disks even though secondaries are
924 offline. With this change it relaxes also the strictness on some other
925 commands which use activate disks internally:
926 * ``gnt-instance start|reboot|rename|backup|export``
927 - Made it possible to remove safely an instance if its secondaries are
929 - Made it possible to reinstall even though secondaries are offline
935 *(Released Mon, 25 Jun 2012)*
937 First release candidate for 2.6. The following changes were done from
940 - Fixed bugs with disk parameters and ``rbd`` templates as well as
942 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
943 behaviour. It supports now the modify operation
944 - ``hcheck`` implemented to analyze cluster health and possibility of
945 improving health by rebalance
946 - ``hbal`` has been improved in dealing with split instances
952 *(Released Mon, 11 Jun 2012)*
954 Second beta release of 2.6. The following changes were done from beta2
957 - Fixed ``daemon-util`` with non-root user models
958 - Fixed creation of plain instances with ``--no-wait-for-sync``
959 - Fix wrong iv_names when running ``cfgupgrade``
960 - Export more information in RAPI group queries
961 - Fixed bug when changing instance network interfaces
962 - Extended burnin to do NIC changes
963 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
964 - Changed default for DRBD barriers
965 - Fixed DRBD error reporting for syncer rate
966 - Verify the options on disk parameters
968 And of course various fixes to documentation and improved unittests and
975 *(Released Wed, 23 May 2012)*
977 First beta release of 2.6. The following changes were done from beta1 to
980 - integrated patch for distributions without ``start-stop-daemon``
981 - adapted example init.d script to work on Fedora
982 - fixed log handling in Haskell daemons
983 - adapted checks in the watcher for pycurl linked against libnss
984 - add partial support for ``xl`` instead of ``xm`` for Xen
985 - fixed a type issue in cluster verification
986 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
989 Plus integrated fixes from the 2.5 branch:
991 - fixed ``kvm-ifup`` to use ``/bin/bash``
992 - fixed parallel build failures
993 - KVM live migration when using a custom keymap
999 *(Released Tue, 24 Jul 2012)*
1001 A small bugfix release, with no new features:
1003 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1004 different default shell (e.g. Debian, Ubuntu)
1005 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1006 243 and Debian bug #650664)
1007 - fixed compatibility with KVM versions that don't support multiple boot
1008 devices (fixes Issue 230 and Debian bug #624256)
1010 Additionally, a few fixes were done to the build system (fixed parallel
1011 build failures) and to the unittests (fixed race condition in test for
1012 FileID functions, and the default enable/disable mode for QA test is now
1019 *(Released Fri, 11 May 2012)*
1021 A small bugfix release.
1023 The main issues solved are on the topic of compatibility with newer LVM
1026 - fixed parsing of ``lv_attr`` field
1027 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1028 the ``--force`` flag is needed
1030 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1031 to accept kernel 3.x too (was hardcoded to 2.6.*).
1033 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1034 of hook results and that also made display of other errors suboptimal
1035 was fixed; the code behaves now like 2.4 and earlier.
1037 Another change in 2.5, the cleanup of the OS scripts environment, is too
1038 aggressive: it removed even the ``PATH`` variable, which requires the OS
1039 scripts to *always* need to export it. Since this is a bit too strict,
1040 we now export a minimal PATH, the same that we export for hooks.
1042 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1043 integrated into this release.
1045 Finally, a few other miscellaneous changes were done (no new features,
1046 just small improvements):
1048 - Fix ``gnt-group --help`` display
1049 - Fix hardcoded Xen kernel path
1050 - Fix grow-disk handling of invalid units
1051 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1052 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1058 *(Released Thu, 12 Apr 2012)*
1060 Incompatible/important changes and bugfixes
1061 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1063 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1064 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1065 to match the underlying LUXI interface.
1066 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1067 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1068 not maintain backwards-compatibility as the underlying operation
1069 changed in an incompatible way. The RAPI client can talk to old
1070 servers, but it needs to be told so as the return value changed.
1071 - When creating file-based instances via RAPI, the ``file_driver``
1072 parameter no longer defaults to ``loop`` and must be specified.
1073 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1075 - Support for the undocumented and deprecated RAPI instance creation
1076 request format version 0 has been dropped. Use version 1, supported
1077 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1078 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1079 documentation <install>`.
1080 - The "cluster-verify" hooks are now executed per group by the
1081 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1082 if you just run ``gnt-cluster verify``, which generates one opcode per
1084 - The environment as passed to the OS scripts is cleared, and thus no
1085 environment variables defined in the node daemon's environment will be
1086 inherited by the scripts.
1087 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1089 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1090 support operations involving multiple node groups.
1091 - Offline nodes are ignored when failing over an instance.
1092 - Support for KVM version 1.0, which changed the version reporting format
1094 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1096 - ``Makefile`` is now compatible with Automake 1.11.2
1097 - Includes all bugfixes made in the 2.4 series
1102 - The ganeti-htools project has been merged into the ganeti-core source
1103 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1104 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1105 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1106 GPT-style partition tables (requires `parted
1107 <http://www.gnu.org/s/parted/>`_).
1108 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1109 - Allowed adding tags on instance creation.
1110 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1111 - Allow instances to be started in a paused state, enabling the user to
1112 see the complete console output on boot using the console.
1113 - Added new hypervisor flag to control default reboot behaviour
1114 (``reboot_behavior``).
1115 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1116 - Improved out-of-band management support:
1118 - Added ``gnt-node health`` command reporting the health status of
1120 - Added ``gnt-node power`` command to manage power status of nodes.
1121 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1123 - Instance migration can fall back to failover if instance is not
1125 - Filters can be used when listing nodes, instances, groups and locks;
1126 see :manpage:`ganeti(7)` manpage.
1127 - Added post-execution status as variables to :doc:`hooks <hooks>`
1129 - Instance tags are exported/imported together with the instance.
1130 - When given an explicit job ID, ``gnt-job info`` will work for archived
1132 - Jobs can define dependencies on other jobs (not yet supported via
1133 RAPI or command line, but used by internal commands and usable via
1136 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1139 - Instance failover is now available as a RAPI resource
1140 (``/2/instances/[instance_name]/failover``).
1141 - ``gnt-instance info`` defaults to static information if primary node
1143 - Opcodes have a new ``comment`` attribute.
1144 - Added basic SPICE support to KVM hypervisor.
1145 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1147 Node group improvements
1148 ~~~~~~~~~~~~~~~~~~~~~~~
1150 - ``gnt-cluster verify`` has been modified to check groups separately,
1151 thereby improving performance.
1152 - Node group support has been added to ``gnt-cluster verify-disks``,
1153 which now operates per node group.
1154 - Watcher has been changed to work better with node groups.
1156 - One process and state file per node group.
1157 - Slow watcher in one group doesn't block other group's watcher.
1159 - Added new command, ``gnt-group evacuate``, to move all instances in a
1160 node group to other groups.
1161 - Added ``gnt-instance change-group`` to move an instance to another
1163 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1164 per-group operations.
1165 - Node groups can be tagged.
1166 - Some operations switch from an exclusive to a shared lock as soon as
1168 - Instance's primary and secondary nodes' groups are now available as
1169 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1170 and ``snodes.group.uuid``).
1175 - Numerous updates to documentation and manpages.
1177 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1179 - Some opcode/job results are now also documented, see :doc:`RAPI
1182 - A lockset's internal lock is now also visible in lock monitor.
1183 - Log messages from job queue workers now contain information about the
1184 opcode they're processing.
1185 - ``gnt-instance console`` no longer requires the instance lock.
1186 - A short delay when waiting for job changes reduces the number of LUXI
1187 requests significantly.
1188 - DRBD metadata volumes are overwritten with zeros during disk creation.
1189 - Out-of-band commands no longer acquire the cluster lock in exclusive
1191 - ``devel/upload`` now uses correct permissions for directories.
1197 *(Released Fri, 23 Mar 2012)*
1199 This was the sixth release candidate of the 2.5 series.
1205 *(Released Mon, 9 Jan 2012)*
1207 This was the fifth release candidate of the 2.5 series.
1213 *(Released Thu, 27 Oct 2011)*
1215 This was the fourth release candidate of the 2.5 series.
1221 *(Released Wed, 26 Oct 2011)*
1223 This was the third release candidate of the 2.5 series.
1229 *(Released Tue, 18 Oct 2011)*
1231 This was the second release candidate of the 2.5 series.
1237 *(Released Tue, 4 Oct 2011)*
1239 This was the first release candidate of the 2.5 series.
1245 *(Released Wed, 31 Aug 2011)*
1247 This was the third beta release of the 2.5 series.
1253 *(Released Mon, 22 Aug 2011)*
1255 This was the second beta release of the 2.5 series.
1261 *(Released Fri, 12 Aug 2011)*
1263 This was the first beta release of the 2.5 series.
1269 *(Released Thu, 27 Oct 2011)*
1271 - Fixed bug when parsing command line parameter values ending in
1273 - Fixed assertion error after unclean master shutdown
1274 - Disable HTTP client pool for RPC, significantly reducing memory usage
1276 - Fixed queue archive creation with wrong permissions
1282 *(Released Tue, 23 Aug 2011)*
1286 - Fixed documentation for importing with ``--src-dir`` option
1287 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1288 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1294 *(Released Fri, 5 Aug 2011)*
1296 Many bug-fixes and a few small features:
1298 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1299 issues when you tried to add an instance with two MAC addresses in one
1301 - KVM: fixed per-instance stored UID value
1302 - KVM: configure bridged NICs at migration start
1303 - KVM: Fix a bug where instance will not start with never KVM versions
1305 - Added OS search path to ``gnt-cluster info``
1306 - Fixed an issue with ``file_storage_dir`` where you were forced to
1307 provide an absolute path, but the documentation states it is a
1308 relative path, the documentation was right
1309 - Added a new parameter to instance stop/start called ``--no-remember``
1310 that will make the state change to not be remembered
1311 - Implemented ``no_remember`` at RAPI level
1312 - Improved the documentation
1313 - Node evacuation: don't call IAllocator if node is already empty
1314 - Fixed bug in DRBD8 replace disks on current nodes
1315 - Fixed bug in recreate-disks for DRBD instances
1316 - Moved assertion checking locks in ``gnt-instance replace-disks``
1317 causing it to abort with not owning the right locks for some situation
1318 - Job queue: Fixed potential race condition when cancelling queued jobs
1319 - Fixed off-by-one bug in job serial generation
1320 - ``gnt-node volumes``: Fix instance names
1321 - Fixed aliases in bash completion
1322 - Fixed a bug in reopening log files after being sent a SIGHUP
1323 - Added a flag to burnin to allow specifying VCPU count
1324 - Bugfixes to non-root Ganeti configuration
1330 *(Released Thu, 12 May 2011)*
1332 Many bug-fixes and a few new small features:
1334 - Fixed a bug related to log opening failures
1335 - Fixed a bug in instance listing with orphan instances
1336 - Fixed a bug which prevented resetting the cluster-level node parameter
1337 ``oob_program`` to the default
1338 - Many fixes related to the ``cluster-merge`` tool
1339 - Fixed a race condition in the lock monitor, which caused failures
1340 during (at least) creation of many instances in parallel
1341 - Improved output for gnt-job info
1342 - Removed the quiet flag on some ssh calls which prevented debugging
1344 - Improved the N+1 failure messages in cluster verify by actually
1345 showing the memory values (needed and available)
1346 - Increased lock attempt timeouts so that when executing long operations
1347 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1348 too early and thus prevent the use of the 'fair' mechanism
1349 - Changed instance query data (``gnt-instance info``) to not acquire
1350 locks unless needed, thus allowing its use on locked instance if only
1351 static information is asked for
1352 - Improved behaviour with filesystems that do not support rename on an
1354 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1355 kept locks on all nodes during the wipe, which is unneeded
1356 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1357 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1358 10GiB) which caused the wipe to fail right at the end in some cases
1359 - Fixed master IP activation when doing master failover with no-voting
1360 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1361 the master node itself
1362 - Fixed potential data-loss in under disk full conditions, where Ganeti
1363 wouldn't check correctly the return code and would consider
1364 partially-written files 'correct'
1365 - Fixed bug related to multiple VGs and DRBD disk replacing
1366 - Added new disk parameter ``metavg`` that allows placement of the meta
1367 device for DRBD in a different volume group
1368 - Fixed error handling in the node daemon when the system libc doesn't
1369 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1370 - Fixed lock release during replace-disks, which kept cluster-wide locks
1371 when doing disk replaces with an iallocator script
1372 - Added check for missing bridges in cluster verify
1373 - Handle EPIPE errors while writing to the terminal better, so that
1374 piping the output to e.g. ``less`` doesn't cause a backtrace
1375 - Fixed rare case where a ^C during Luxi calls could have been
1376 interpreted as server errors, instead of simply terminating
1377 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1379 - Added a few more parameters to the KVM hypervisor, allowing a second
1380 CDROM, custom disk type for CDROMs and a floppy image
1381 - Removed redundant message in instance rename when the name is given
1383 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1384 disks on new nodes, allowing recreation when the original instance
1385 nodes are completely gone
1386 - Added option when converting disk templates to DRBD to skip waiting
1387 for the resync, in order to make the instance available sooner
1388 - Added two new variables to the OS scripts environment (containing the
1390 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1391 to allow use of ``pvgrub`` as bootloader
1392 - Changed the instance memory modifications to only check out-of-memory
1393 conditions on memory increases, and turned the secondary node warnings
1394 into errors (they can still be overridden via ``--force``)
1395 - Fixed the handling of a corner case when the Python installation gets
1396 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1397 to execute a command that doesn't exist
1398 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1399 node of the instance returned failures during instance shutdown; this
1400 adds the option ``--ignore-consistency`` to gnt-instance move
1402 And as usual, various improvements to the error messages, documentation
1409 *(Released Wed, 09 Mar 2011)*
1411 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1412 the RAPI users file if run twice (even with ``--dry-run``).
1414 The release fixes that bug (nothing else changed).
1420 *(Released Mon, 07 Mar 2011)*
1422 Final 2.4.0 release. Just a few small fixes:
1424 - Fixed RAPI node evacuate
1425 - Fixed the kvm-ifup script
1426 - Fixed internal error handling for special job cases
1427 - Updated man page to specify the escaping feature for options
1433 *(Released Mon, 28 Feb 2011)*
1435 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1436 that this feature wiped the disks of the wrong instance, leading to loss
1441 - Fixed title of query field containing instance name
1442 - Expanded the glossary in the documentation
1443 - Fixed one unittest (internal issue)
1449 *(Released Mon, 21 Feb 2011)*
1451 A number of bug fixes plus just a couple functionality changes.
1453 On the user-visible side, the ``gnt-* list`` command output has changed
1454 with respect to "special" field states. The current rc1 style of display
1455 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1456 the default output mode special fields states are displayed as follows:
1458 - Offline resource: ``*``
1459 - Unavailable/not applicable: ``-``
1460 - Data missing (RPC failure): ``?``
1461 - Unknown field: ``??``
1463 Another user-visible change is the addition of ``--force-join`` to
1468 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1469 - Fixed regression in RAPI/instance reinstall where all parameters were
1470 required (instead of optional)
1471 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1472 - Fixed iallocator usage (offline nodes were not considered offline)
1473 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1474 - Fixed hypervisor and OS parameter validation with respect to
1475 non-vm_capable nodes
1476 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1478 - Fixed ``tools/listrunner`` with respect to agent-based usage
1484 *(Released Fri, 4 Feb 2011)*
1486 Many changes and fixes since the beta1 release. While there were some
1487 internal changes, the code has been mostly stabilised for the RC
1490 Note: the dumb allocator was removed in this release, as it was not kept
1491 up-to-date with the IAllocator protocol changes. It is recommended to
1492 use the ``hail`` command from the ganeti-htools package.
1494 Note: the 2.4 and up versions of Ganeti are not compatible with the
1495 0.2.x branch of ganeti-htools. You need to upgrade to
1496 ganeti-htools-0.3.0 (or later).
1498 Regressions fixed from 2.3
1499 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1501 - Fixed the ``gnt-cluster verify-disks`` command
1502 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1504 - Fixed disk adoption breakage
1505 - Fixed wrong headers in instance listing for field aliases
1510 - Fixed corner case in KVM handling of NICs
1511 - Fixed many cases of wrong handling of non-vm_capable nodes
1512 - Fixed a bug where a missing instance symlink was not possible to
1513 recreate with any ``gnt-*`` command (now ``gnt-instance
1514 activate-disks`` does it)
1515 - Fixed the volume group name as reported by ``gnt-cluster
1517 - Increased timeouts for the import-export code, hopefully leading to
1518 fewer aborts due network or instance timeouts
1519 - Fixed bug in ``gnt-node list-storage``
1520 - Fixed bug where not all daemons were started on cluster
1521 initialisation, but only at the first watcher run
1522 - Fixed many bugs in the OOB implementation
1523 - Fixed watcher behaviour in presence of instances with offline
1525 - Fixed instance list output for instances running on the wrong node
1526 - a few fixes to the cluster-merge tool, but it still cannot merge
1527 multi-node groups (currently it is not recommended to use this tool)
1533 - Improved network configuration for the KVM hypervisor
1534 - Added e1000 as a supported NIC for Xen-HVM
1535 - Improved the lvmstrap tool to also be able to use partitions, as
1536 opposed to full disks
1537 - Improved speed of disk wiping (the cluster parameter
1538 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1539 of instance creations
1540 - Added documentation for the OS parameters
1541 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1542 hypervisor is not responding
1543 - Added display of blacklisted and hidden OS information in
1544 ``gnt-cluster info``
1545 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1546 NIC and node parameters, which might create problems with currently
1547 invalid (but undetected) configuration files, but prevents validation
1548 failures when unrelated parameters are modified
1549 - Changed cluster initialisation to wait for the master daemon to become
1551 - Expanded the RAPI interface:
1553 - Added config redistribution resource
1554 - Added activation/deactivation of instance disks
1555 - Added export of console information
1557 - Implemented log file reopening on SIGHUP, which allows using
1558 logrotate(8) for the Ganeti log files
1559 - Added a basic OOB helper script as an example
1565 *(Released Fri, 14 Jan 2011)*
1570 - Fixed timezone issues when formatting timestamps
1571 - Added support for node groups, available via ``gnt-group`` and other
1573 - Added out-of-band framework and management, see :doc:`design
1574 document <design-oob>`
1575 - Removed support for roman numbers from ``gnt-node list`` and
1576 ``gnt-instance list``.
1577 - Allowed modification of master network interface via ``gnt-cluster
1578 modify --master-netdev``
1579 - Accept offline secondaries while shutting down instance disks
1580 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1581 - Added support for multiple LVM volume groups
1582 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1584 - Added commands to list available fields:
1586 - ``gnt-node list-fields``
1587 - ``gnt-group list-fields``
1588 - ``gnt-instance list-fields``
1590 - Updated documentation and man pages
1595 - Moved ``rapi_users`` file into separate directory, now named
1596 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1598 - Added new tool for running commands on many machines,
1599 ``tools/ganeti-listrunner``
1600 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1601 improving the ``gnt-instance console`` output
1602 - Allowed customisation of disk index separator at ``configure`` time
1603 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1604 - Added support for non-partitioned md disks in ``lvmstrap``
1605 - Added script to gracefully power off KVM instances
1606 - Split ``utils`` module into smaller parts
1607 - Changed query operations to return more detailed information, e.g.
1608 whether an information is unavailable due to an offline node. To use
1609 this new functionality, the LUXI call ``Query`` must be used. Field
1610 information is now stored by the master daemon and can be retrieved
1611 using ``QueryFields``. Instances, nodes and groups can also be queried
1612 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1613 exposed via RAPI). The following commands make use of this
1614 infrastructure change:
1616 - ``gnt-group list``
1617 - ``gnt-group list-fields``
1619 - ``gnt-node list-fields``
1620 - ``gnt-instance list``
1621 - ``gnt-instance list-fields``
1622 - ``gnt-debug locks``
1627 - New RAPI resources (see :doc:`rapi`):
1631 - ``/2/groups/[group_name]``
1632 - ``/2/groups/[group_name]/assign-nodes``
1633 - ``/2/groups/[group_name]/modify``
1634 - ``/2/groups/[group_name]/rename``
1635 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1639 - Implemented ``no_install`` for instance creation
1640 - Implemented OS parameters for instance reinstallation, allowing
1641 use of special settings on reinstallation (e.g. for preserving data)
1646 - Added IPv6 support in import/export
1647 - Pause DRBD synchronization while wiping disks on instance creation
1648 - Updated unittests and QA scripts
1649 - Improved network parameters passed to KVM
1650 - Converted man pages from docbook to reStructuredText
1656 *(Released Mon, 20 Dec 2010)*
1658 Released version 2.3.1~rc1 without any changes.
1664 *(Released Wed, 1 Dec 2010)*
1666 - impexpd: Disable OpenSSL compression in socat if possible (backport
1667 from master, commit e90739d625b, see :doc:`installation guide
1668 <install-quick>` for details)
1669 - Changed unittest coverage report to exclude test scripts
1670 - Added script to check version format
1676 *(Released Wed, 1 Dec 2010)*
1678 Released version 2.3.0~rc1 without any changes.
1684 *(Released Fri, 19 Nov 2010)*
1686 A number of bugfixes and documentation updates:
1688 - Update ganeti-os-interface documentation
1689 - Fixed a bug related to duplicate MACs or similar items which should be
1691 - Fix breakage in OS state modify
1692 - Reinstall instance: disallow offline secondaries (fixes bug related to
1693 OS changing but reinstall failing)
1694 - plus all the other fixes between 2.2.1 and 2.2.2
1700 *(Released Tue, 2 Nov 2010)*
1702 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1703 - Fixed master failover race with watcher
1704 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1706 - Accept previously stopped instance for export with instance removal
1707 - Simplify and extend the environment variables for instance OS scripts
1708 - Added new node flags, ``master_capable`` and ``vm_capable``
1709 - Added optional instance disk wiping prior during allocation. This is a
1710 cluster-wide option and can be set/modified using
1711 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1712 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1713 :doc:`install-quick`
1714 - Added a new watcher option (``--ignore-pause``)
1715 - Added option to ignore offline node on instance start/stop
1716 (``--ignore-offline``)
1717 - Allow overriding OS parameters with ``gnt-instance reinstall``
1718 - Added ability to change node's secondary IP address using ``gnt-node
1720 - Implemented privilege separation for all daemons except
1721 ``ganeti-noded``, see ``configure`` options
1722 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1724 - Implemented job priorities (see ``ganeti(7)`` manpage)
1725 - Ignore failures while shutting down instances during failover from
1727 - Exit daemon's bootstrap process only once daemon is ready
1728 - Export more information via ``LUInstanceQuery``/remote API
1729 - Improved documentation, QA and unittests
1730 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1731 restart if the file was created or changed
1732 - Added LUXI protocol version sent with each request and response,
1733 allowing detection of server/client mismatches
1734 - Moved the Python scripts among gnt-* and ganeti-* into modules
1735 - Moved all code related to setting up SSH to an external script,
1737 - Infrastructure changes for node group support in future versions
1743 *(Released Fri, 19 Nov 2010)*
1745 A few small bugs fixed, and some improvements to the build system:
1747 - Fix documentation regarding conversion to drbd
1748 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1750 - Fix error handling in node modify with multiple changes
1751 - Allow remote imports without checked names
1757 *(Released Tue, 19 Oct 2010)*
1759 - Disable SSL session ID cache in RPC client
1765 *(Released Thu, 14 Oct 2010)*
1767 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1768 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1770 - Fix problems with interaction between Curl and Python's HTTP server,
1771 resulting in increased speed in many RPC calls
1772 - Improve our release script to prevent breakage with older aclocal and
1779 *(Released Thu, 7 Oct 2010)*
1781 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1782 value retrieved from master
1783 - Added support for blacklisted or hidden OS definitions
1784 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1785 - Added support for -mem-path in KVM hypervisor abstraction layer
1786 - Allow overriding instance parameters in tool for inter-cluster
1787 instance moves (``tools/move-instance``)
1788 - Improved opcode summaries (e.g. in ``gnt-job list``)
1789 - Improve consistency of OS listing by sorting it
1790 - Documentation updates
1796 *(Released Fri, 8 Oct 2010)*
1798 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1804 *(Released Mon, 4 Oct 2010)*
1806 - Fixed regression in ``gnt-instance rename``
1812 *(Released Wed, 22 Sep 2010)*
1814 - Fixed OS_VARIANT variable for OS scripts
1815 - Fixed cluster tag operations via RAPI
1816 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1817 - Disabled RAPI CA checks in watcher
1823 *(Released Mon, 23 Aug 2010)*
1825 - Support DRBD versions of the format "a.b.c.d"
1827 - Re-introduce support for usage from multiple threads in RAPI client
1828 - Instance renames and modify via RAPI
1829 - Work around race condition between processing and archival in job
1831 - Mark opcodes following failed one as failed, too
1832 - Job field ``lock_status`` was removed due to difficulties making it
1833 work with the changed job queue in Ganeti 2.2; a better way to monitor
1834 locks is expected for a later 2.2.x release
1835 - Fixed dry-run behaviour with many commands
1836 - Support ``ssh-agent`` again when adding nodes
1837 - Many additional bugfixes
1843 *(Released Fri, 30 Jul 2010)*
1845 Important change: the internal RPC mechanism between Ganeti nodes has
1846 changed from using a home-grown http library (based on the Python base
1847 libraries) to use the PycURL library. This requires that PycURL is
1848 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1849 against GnuTLS by default. cURL's support for GnuTLS had known issues
1850 before cURL 7.21.0 and we recommend using the latest cURL release or
1851 linking against OpenSSL. Most other distributions already link PycURL
1852 and cURL against OpenSSL. The command::
1854 python -c 'import pycurl; print pycurl.version'
1856 can be used to determine the libraries PycURL and cURL are linked
1859 Other significant changes:
1861 - Rewrote much of the internals of the job queue, in order to achieve
1862 better parallelism; this decouples job query operations from the job
1863 processing, and it should allow much nicer behaviour of the master
1864 daemon under load, and it also has uncovered some long-standing bugs
1865 related to the job serialisation (now fixed)
1866 - Added a default iallocator setting to the cluster parameters,
1867 eliminating the need to always pass nodes or an iallocator for
1868 operations that require selection of new node(s)
1869 - Added experimental support for the LXC virtualization method
1870 - Added support for OS parameters, which allows the installation of
1871 instances to pass parameter to OS scripts in order to customise the
1873 - Added a hypervisor parameter controlling the migration type (live or
1874 non-live), since hypervisors have various levels of reliability; this
1875 has renamed the 'live' parameter to 'mode'
1876 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1877 logical volumes, meaning that cluster verify will ignore them and not
1878 flag their presence as errors
1879 - The watcher will now reset the error count for failed instances after
1880 8 hours, thus allowing self-healing if the problem that caused the
1881 instances to be down/fail to start has cleared in the meantime
1882 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1883 check for, and warn, if the drbd module parameter ``usermode_helper``
1884 is not consistent with the cluster-wide setting; this is needed to
1885 make diagnose easier of failed drbd creations
1886 - Started adding base IPv6 support, but this is not yet
1887 enabled/available for use
1888 - Rename operations (cluster, instance) will now return the new name,
1889 which is especially useful if a short name was passed in
1890 - Added support for instance migration in RAPI
1891 - Added a tool to pre-configure nodes for the SSH setup, before joining
1892 them to the cluster; this will allow in the future a simplified model
1893 for node joining (but not yet fully enabled in 2.2); this needs the
1894 paramiko python library
1895 - Fixed handling of name-resolving errors
1896 - Fixed consistency of job results on the error path
1897 - Fixed master-failover race condition when executed multiple times in
1899 - Fixed many bugs related to the job queue (mostly introduced during the
1900 2.2 development cycle, so not all are impacting 2.1)
1901 - Fixed instance migration with missing disk symlinks
1902 - Fixed handling of unknown jobs in ``gnt-job archive``
1903 - And many other small fixes/improvements
1907 - Enhanced both the unittest and the QA coverage
1908 - Switched the opcode validation to a generic model, and extended the
1909 validation to all opcode parameters
1910 - Changed more parts of the code that write shell scripts to use the
1912 - Switched the master daemon to use the asyncore library for the Luxi
1919 *(Released Thu, 17 Jun 2010)*
1921 - Added tool (``move-instance``) and infrastructure to move instances
1922 between separate clusters (see :doc:`separate documentation
1923 <move-instance>` and :doc:`design document <design-2.2>`)
1924 - Added per-request RPC timeout
1925 - RAPI now requires a Content-Type header for requests with a body (e.g.
1926 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1927 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1928 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1930 - Implemented initial support for running Ganeti daemons as separate
1931 users, see configure-time flags ``--with-user-prefix`` and
1932 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1934 - Instances can be removed after export (``gnt-backup export
1935 --remove-instance``)
1936 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1937 key (instead of 1024 bit)
1938 - Added new cluster configuration file for cluster domain secret
1939 - Import/export now use SSL instead of SSH
1940 - Added support for showing estimated time when exporting an instance,
1941 see the ``ganeti-os-interface(7)`` manpage and look for
1948 *(Released Tue, 16 Nov 2010)*
1950 Some more bugfixes. Unless critical bugs occur, this will be the last
1953 - Fix case of MAC special-values
1954 - Fix mac checker regex
1955 - backend: Fix typo causing "out of range" error
1956 - Add missing --units in gnt-instance list man page
1962 *(Released Tue, 24 Aug 2010)*
1965 - Don't ignore secondary node silently on non-mirrored disk templates
1967 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1968 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1969 - Properly document the usb_mouse parameter
1970 - Fix path in ganeti-rapi(8) (issue 116)
1971 - Adjust error message when the ganeti user's .ssh directory is
1973 - Add same-node-check when changing the disk template to drbd
1979 *(Released Fri, 16 Jul 2010)*
1982 - Add an option to only select some reboot types during qa/burnin.
1983 (on some hypervisors consequent reboots are not supported)
1984 - Fix infrequent race condition in master failover. Sometimes the old
1985 master ip address would be still detected as up for a short time
1986 after it was removed, causing failover to fail.
1987 - Decrease mlockall warnings when the ctypes module is missing. On
1988 Python 2.4 we support running even if no ctypes module is installed,
1989 but we were too verbose about this issue.
1990 - Fix building on old distributions, on which man doesn't have a
1992 - Fix RAPI not to ignore the MAC address on instance creation
1993 - Implement the old instance creation format in the RAPI client.
1999 *(Released Thu, 01 Jul 2010)*
2001 A small bugfix release:
2002 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2003 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2004 a non-existing option
2005 - Fix instance create: the --force-variant option was ignored
2006 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2007 - Fix modify node storage with non-FQDN arguments
2008 - Fix RAPI client to authenticate under Python 2.6 when used
2009 for more than 5 requests needing authentication
2010 - Fix gnt-instance modify -t (storage) giving a wrong error message
2011 when converting a non-shutdown drbd instance to plain
2017 *(Released Fri, 18 Jun 2010)*
2019 A small bugfix release:
2021 - Fix live migration of KVM instances started with older Ganeti
2022 versions which had fewer hypervisor parameters
2023 - Fix gnt-instance grow-disk on down instances
2024 - Fix an error-reporting bug during instance migration
2025 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2026 silently ignoring broken ones
2027 - Fix an RPC error reporting bug affecting, for example, RAPI client
2029 - Fix bug triggered by different API version os-es on different nodes
2030 - Fix a bug in instance startup with custom hvparams: OS level
2031 parameters would fail to be applied.
2032 - Fix the RAPI client under Python 2.6 (but more work is needed to
2033 make it work completely well with OpenSSL)
2034 - Fix handling of errors when resolving names from DNS
2040 *(Released Thu, 3 Jun 2010)*
2042 A medium sized development cycle. Some new features, and some
2043 fixes/small improvements/cleanups.
2045 Significant features
2046 ~~~~~~~~~~~~~~~~~~~~
2048 The node deamon now tries to mlock itself into memory, unless the
2049 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2050 its logs, and falls back to console logging. This allows emergency
2051 features such as ``gnt-node powercycle`` to work even in the event of a
2052 broken node disk (tested offlining the disk hosting the node's
2053 filesystem and dropping its memory caches; don't try this at home)
2055 KVM: add vhost-net acceleration support. It can be tested with a new
2056 enough version of the kernel and of qemu-kvm.
2058 KVM: Add instance chrooting feature. If you use privilege dropping for
2059 your VMs you can also now force them to chroot to an empty directory,
2060 before starting the emulated guest.
2062 KVM: Add maximum migration bandwith and maximum downtime tweaking
2063 support (requires a new-enough version of qemu-kvm).
2065 Cluster verify will now warn if the master node doesn't have the master
2066 ip configured on it.
2068 Add a new (incompatible) instance creation request format to RAPI which
2069 supports all parameters (previously only a subset was supported, and it
2070 wasn't possible to extend the old format to accomodate all the new
2071 features. The old format is still supported, and a client can check for
2072 this feature, before using it, by checking for its presence in the
2073 ``features`` RAPI resource.
2075 Now with ancient latin support. Try it passing the ``--roman`` option to
2076 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2077 (requires the python-roman module to be installed, in order to work).
2082 As usual many internal code refactorings, documentation updates, and
2085 - Lots of improvements and cleanups to the experimental Remote API
2086 (RAPI) client library.
2087 - A new unit test suite for the core daemon libraries.
2088 - A fix to creating missing directories makes sure the umask is not
2089 applied anymore. This enforces the same directory permissions
2091 - Better handling terminating daemons with ctrl+c (used when running
2092 them in debugging mode).
2093 - Fix a race condition in live migrating a KVM instance, when stat()
2094 on the old proc status file returned EINVAL, which is an unexpected
2096 - Fixed manpage checking with newer man and utf-8 charachters. But now
2097 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2103 *(Released Fri, 7 May 2010)*
2105 Fix a bug which prevented untagged KVM instances from starting.
2111 *(Released Fri, 7 May 2010)*
2113 Another release with a long development cycle, during which many
2114 different features were added.
2116 Significant features
2117 ~~~~~~~~~~~~~~~~~~~~
2119 The KVM hypervisor now can run the individual instances as non-root, to
2120 reduce the impact of a VM being hijacked due to bugs in the
2121 hypervisor. It is possible to run all instances as a single (non-root)
2122 user, to manually specify a user for each instance, or to dynamically
2123 allocate a user out of a cluster-wide pool to each instance, with the
2124 guarantee that no two instances will run under the same user ID on any
2127 An experimental RAPI client library, that can be used standalone
2128 (without the other Ganeti libraries), is provided in the source tree as
2129 ``lib/rapi/client.py``. Note this client might change its interface in
2130 the future, as we iterate on its capabilities.
2132 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2133 replace the cluster's certificates and crypto keys. This might help in
2134 case they have been compromised, or have simply expired.
2136 A new disk option for instance creation has been added that allows one
2137 to "adopt" currently existing logical volumes, with data
2138 preservation. This should allow easier migration to Ganeti from
2139 unmanaged (or managed via other software) instances.
2141 Another disk improvement is the possibility to convert between redundant
2142 (DRBD) and plain (LVM) disk configuration for an instance. This should
2143 allow better scalability (starting with one node and growing the
2144 cluster, or shrinking a two-node cluster to one node).
2146 A new feature that could help with automated node failovers has been
2147 implemented: if a node sees itself as offline (by querying the master
2148 candidates), it will try to shutdown (hard) all instances and any active
2149 DRBD devices. This reduces the risk of duplicate instances if an
2150 external script automatically failovers the instances on such nodes. To
2151 enable this, the cluster parameter ``maintain_node_health`` should be
2152 enabled; in the future this option (per the name) will enable other
2153 automatic maintenance features.
2155 Instance export/import now will reuse the original instance
2156 specifications for all parameters; that means exporting an instance,
2157 deleting it and the importing it back should give an almost identical
2158 instance. Note that the default import behaviour has changed from
2159 before, where it created only one NIC; now it recreates the original
2162 Cluster verify has added a few new checks: SSL certificates validity,
2163 /etc/hosts consistency across the cluster, etc.
2168 As usual, many internal changes were done, documentation fixes,
2171 - Fixed cluster initialization with disabled cluster storage (regression
2172 introduced in 2.1.1)
2173 - File-based storage supports growing the disks
2174 - Fixed behaviour of node role changes
2175 - Fixed cluster verify for some corner cases, plus a general rewrite of
2176 cluster verify to allow future extension with more checks
2177 - Fixed log spamming by watcher and node daemon (regression introduced
2179 - Fixed possible validation issues when changing the list of enabled
2181 - Fixed cleanup of /etc/hosts during node removal
2182 - Fixed RAPI response for invalid methods
2183 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2184 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2185 from ide disks, etc.)
2186 - Allow OS changes without re-installation (to record a changed OS
2187 outside of Ganeti, or to allow OS renames)
2188 - Allow instance creation without OS installation (useful for example if
2189 the OS will be installed manually, or restored from a backup not in
2191 - Implemented option to make cluster ``copyfile`` use the replication
2193 - Added list of enabled hypervisors to ssconf (possibly useful for
2195 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2197 - A partial form of node re-IP is possible via node readd, which now
2198 allows changed node primary IP
2199 - Command line utilities now show an informational message if the job is
2201 - The logs of the master daemon now show the PID/UID/GID of the
2208 *(Released Fri, 12 Mar 2010)*
2210 During the 2.1.0 long release candidate cycle, a lot of improvements and
2211 changes have accumulated with were released later as 2.1.1.
2216 The node evacuate command (``gnt-node evacuate``) was significantly
2217 rewritten, and as such the IAllocator protocol was changed - a new
2218 request type has been added. This unfortunate change during a stable
2219 series is designed to improve performance of node evacuations; on
2220 clusters with more than about five nodes and which are well-balanced,
2221 evacuation should proceed in parallel for all instances of the node
2222 being evacuated. As such, any existing IAllocator scripts need to be
2223 updated, otherwise the above command will fail due to the unknown
2224 request. The provided "dumb" allocator has not been updated; but the
2225 ganeti-htools package supports the new protocol since version 0.2.4.
2227 Another important change is increased validation of node and instance
2228 names. This might create problems in special cases, if invalid host
2229 names are being used.
2231 Also, a new layer of hypervisor parameters has been added, that sits at
2232 OS level between the cluster defaults and the instance ones. This allows
2233 customisation of virtualization parameters depending on the installed
2234 OS. For example instances with OS 'X' may have a different KVM kernel
2235 (or any other parameter) than the cluster defaults. This is intended to
2236 help managing a multiple OSes on the same cluster, without manual
2237 modification of each instance's parameters.
2239 A tool for merging clusters, ``cluster-merge``, has been added in the
2240 tools sub-directory.
2245 - Improved the int/float conversions that should make the code more
2246 robust in face of errors from the node daemons
2247 - Fixed the remove node code in case of internal configuration errors
2248 - Fixed the node daemon behaviour in face of inconsistent queue
2249 directory (e.g. read-only file-system where we can't open the files
2251 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2252 now it either aborts cleanly or, if given the new "auto_promote"
2253 parameter, will automatically promote other nodes as needed
2254 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2255 completely prevent Ganeti from working
2256 - Fixed bug for instance export when not all disks were successfully
2258 - Fixed behaviour of node add when the new node is slow in starting up
2260 - Fixed handling of signals in the LUXI client, which should improve
2261 behaviour of command-line scripts
2262 - Added checks for invalid node/instance names in the configuration (now
2263 flagged during cluster verify)
2264 - Fixed watcher behaviour for disk activation errors
2265 - Fixed two potentially endless loops in http library, which led to the
2266 RAPI daemon hanging and consuming 100% CPU in some cases
2267 - Fixed bug in RAPI daemon related to hashed passwords
2268 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2270 - Enhanced compatibility with non-Debian OSes, but not using absolute
2271 path in some commands and allowing customisation of the ssh
2272 configuration directory
2273 - Fixed possible future issue with new Python versions by abiding to the
2274 proper use of ``__slots__`` attribute on classes
2275 - Added checks that should prevent directory traversal attacks
2276 - Many documentation fixes based on feedback from users
2281 - Added an "early_release" more for instance replace disks and node
2282 evacuate, where we release locks earlier and thus allow higher
2283 parallelism within the cluster
2284 - Added watcher hooks, intended to allow the watcher to restart other
2285 daemons (e.g. from the ganeti-nbma project), but they can be used of
2286 course for any other purpose
2287 - Added a compile-time disable for DRBD barriers, to increase
2288 performance if the administrator trusts the power supply or the
2289 storage system to not lose writes
2290 - Added the option of using syslog for logging instead of, or in
2291 addition to, Ganeti's own log files
2292 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2293 can indeed boot from a paravirtual NIC
2294 - Added a generic debug level for many operations; while this is not
2295 used widely yet, it allows one to pass the debug value all the way to
2297 - Enhanced the hooks environment for instance moves (failovers,
2298 migrations) where the primary/secondary nodes changed during the
2299 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2300 - Enhanced data validations for many user-supplied values; one important
2301 item is the restrictions imposed on instance and node names, which
2302 might reject some (invalid) host names
2303 - Add a configure-time option to disable file-based storage, if it's not
2304 needed; this allows greater security separation between the master
2305 node and the other nodes from the point of view of the inter-node RPC
2307 - Added user notification in interactive tools if job is waiting in the
2308 job queue or trying to acquire locks
2309 - Added log messages when a job is waiting for locks
2310 - Added filtering by node tags in instance operations which admit
2311 multiple instances (start, stop, reboot, reinstall)
2312 - Added a new tool for cluster mergers, ``cluster-merge``
2313 - Parameters from command line which are of the form ``a=b,c=d`` can now
2314 use backslash escapes to pass in values which contain commas,
2315 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2317 - For KVM, the instance name is the first parameter passed to KVM, so
2318 that it's more visible in the process list
2324 *(Released Tue, 2 Mar 2010)*
2326 Ganeti 2.1 brings many improvements with it. Major changes:
2328 - Added infrastructure to ease automated disk repairs
2329 - Added new daemon to export configuration data in a cheaper way than
2330 using the remote API
2331 - Instance NICs can now be routed instead of being associated with a
2333 - Improved job locking logic to reduce impact of jobs acquiring multiple
2334 locks waiting for other long-running jobs
2336 In-depth implementation details can be found in the Ganeti 2.1 design
2342 - Added chroot hypervisor
2343 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2345 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2346 ``bootloader_path`` and ``bootloader_args``)
2347 - Added the ``use_localtime`` option for the xen-hvm and kvm
2348 hypervisors, and the default value for this has changed to false (in
2349 2.0 xen-hvm always enabled it)
2350 - Added luxi call to submit multiple jobs in one go
2351 - Added cluster initialization option to not modify ``/etc/hosts``
2353 - Added network interface parameters
2354 - Added dry run mode to some LUs
2355 - Added RAPI resources:
2357 - ``/2/instances/[instance_name]/info``
2358 - ``/2/instances/[instance_name]/replace-disks``
2359 - ``/2/nodes/[node_name]/evacuate``
2360 - ``/2/nodes/[node_name]/migrate``
2361 - ``/2/nodes/[node_name]/role``
2362 - ``/2/nodes/[node_name]/storage``
2363 - ``/2/nodes/[node_name]/storage/modify``
2364 - ``/2/nodes/[node_name]/storage/repair``
2366 - Added OpCodes to evacuate or migrate all instances on a node
2367 - Added new command to list storage elements on nodes (``gnt-node
2368 list-storage``) and modify them (``gnt-node modify-storage``)
2369 - Added new ssconf files with master candidate IP address
2370 (``ssconf_master_candidates_ips``), node primary IP address
2371 (``ssconf_node_primary_ips``) and node secondary IP address
2372 (``ssconf_node_secondary_ips``)
2373 - Added ``ganeti-confd`` and a client library to query the Ganeti
2374 configuration via UDP
2375 - Added ability to run hooks after cluster initialization and before
2377 - Added automatic mode for disk replace (``gnt-instance replace-disks
2379 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2380 after catastrophic data-loss
2381 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2383 - Added ``gnt-instance move`` command to move instances
2384 - Added ``gnt-cluster watcher`` command to control watcher
2385 - Added ``gnt-node powercycle`` command to powercycle nodes
2386 - Added new job status field ``lock_status``
2387 - Added parseable error codes to cluster verification (``gnt-cluster
2388 verify --error-codes``) and made output less verbose (use
2389 ``--verbose`` to restore previous behaviour)
2390 - Added UUIDs to the main config entities (cluster, nodes, instances)
2391 - Added support for OS variants
2392 - Added support for hashed passwords in the Ganeti remote API users file
2394 - Added option to specify maximum timeout on instance shutdown
2395 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2396 - Added new helper script to start and stop Ganeti daemons
2397 (``daemon-util``), with the intent to reduce the work necessary to
2398 adjust Ganeti for non-Debian distributions and to start/stop daemons
2400 - Added more unittests
2401 - Fixed critical bug in ganeti-masterd startup
2402 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2403 now customisable at the cluster level for both the KVM and Xen
2404 hypervisors using the new ``migration_port`` parameter
2405 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2406 reinstalling an instance
2407 - Allowed ``@`` in tag names
2408 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2409 - Many documentation updates
2410 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2411 - ``gnt-instance reinstall`` can now reinstall multiple instances
2412 - Updated many command line parameters
2413 - Introduced new OS API version 15
2414 - No longer support a default hypervisor
2415 - Treat virtual LVs as inexistent
2416 - Improved job locking logic to reduce lock contention
2417 - Match instance and node names case insensitively
2418 - Reimplemented bash completion script to be more complete
2425 *(Released Thu, 4 Feb 2010)*
2427 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2428 - Fix a string formatting bug
2429 - Improve safety of the code in some error paths
2430 - Improve data validation in the master of values returned from nodes
2436 *(Released Thu, 17 Dec 2009)*
2438 - Fix security issue due to missing validation of iallocator names; this
2439 allows local and remote execution of arbitrary executables
2440 - Fix failure of gnt-node list during instance removal
2441 - Ship the RAPI documentation in the archive
2447 *(Released Wed, 30 Sep 2009)*
2449 - Fixed many wrong messages
2450 - Fixed a few bugs related to the locking library
2451 - Fixed MAC checking at instance creation time
2452 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2453 - Fixed a few issues related to signal handling in both daemons and
2455 - Fixed the example startup script provided
2456 - Fixed insserv dependencies in the example startup script (patch from
2458 - Fixed handling of drained nodes in the iallocator framework
2459 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2461 - Fixed error related to invalid job IDs in job polling
2462 - Fixed job/opcode persistence on unclean master shutdown
2463 - Fixed handling of partial job processing after unclean master
2465 - Fixed error reporting from LUs, previously all errors were converted
2466 into execution errors
2467 - Fixed error reporting from burnin
2468 - Decreased significantly the memory usage of the job queue
2469 - Optimised slightly multi-job submission
2470 - Optimised slightly opcode loading
2471 - Backported the multi-job submit framework from the development
2472 branch; multi-instance start and stop should be faster
2473 - Added script to clean archived jobs after 21 days; this will reduce
2474 the size of the queue directory
2475 - Added some extra checks in disk size tracking
2476 - Added an example ethers hook script
2477 - Added a cluster parameter that prevents Ganeti from modifying of
2479 - Added more node information to RAPI responses
2480 - Added a ``gnt-job watch`` command that allows following the ouput of a
2482 - Added a bind-address option to ganeti-rapi
2483 - Added more checks to the configuration verify
2484 - Enhanced the burnin script such that some operations can be retried
2486 - Converted instance reinstall to multi-instance model
2492 *(Released Fri, 7 Aug 2009)*
2494 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2495 to allow using the pre-2.0.2 behaviour in activation, if any existing
2496 instances have mismatched disk sizes in the configuration
2497 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2498 any configuration mismatches for disk sizes
2499 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2500 failover to work on two-node clusters
2501 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2503 - Fixed detection of OS script errors in ``gnt-backup export``
2504 - Fixed exit code of ``gnt-backup export``
2510 *(Released Fri, 17 Jul 2009)*
2512 - Added experimental support for stripped logical volumes; this should
2513 enhance performance but comes with a higher complexity in the block
2514 device handling; stripping is only enabled when passing
2515 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2516 affected even in the non-stripped mode
2517 - Improved resiliency against transient failures at the end of DRBD
2518 resyncs, and in general of DRBD resync checks
2519 - Fixed a couple of issues with exports and snapshot errors
2520 - Fixed a couple of issues in instance listing
2521 - Added display of the disk size in ``gnt-instance info``
2522 - Fixed checking for valid OSes in instance creation
2523 - Fixed handling of the "vcpus" parameter in instance listing and in
2524 general of invalid parameters
2525 - Fixed http server library, and thus RAPI, to handle invalid
2526 username/password combinations correctly; this means that now they
2527 report unauthorized for queries too, not only for modifications,
2528 allowing earlier detect of configuration problems
2529 - Added a new "role" node list field, equivalent to the master/master
2530 candidate/drained/offline flags combinations
2531 - Fixed cluster modify and changes of candidate pool size
2532 - Fixed cluster verify error messages for wrong files on regular nodes
2533 - Fixed a couple of issues with node demotion from master candidate role
2534 - Fixed node readd issues
2535 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2536 - Added a new ``--no-voting`` option for masterfailover to fix failover
2537 on two-nodes clusters when the former master node is unreachable
2538 - Added instance reinstall over RAPI
2544 *(Released Tue, 16 Jun 2009)*
2546 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2547 allow re-adding the start in single-user option (regression from 1.2)
2548 - the watcher writes the instance status to a file, to allow monitoring
2549 to report the instance status (from the master) based on cached
2550 results of the watcher's queries; while this can get stale if the
2551 watcher is being locked due to other work on the cluster, this is
2552 still an improvement
2553 - the watcher now also restarts the node daemon and the rapi daemon if
2555 - fixed the watcher to handle full and drained queue cases
2556 - hooks export more instance data in the environment, which helps if
2557 hook scripts need to take action based on the instance's properties
2558 (no longer need to query back into ganeti)
2559 - instance failovers when the instance is stopped do not check for free
2560 RAM, so that failing over a stopped instance is possible in low memory
2562 - rapi uses queries for tags instead of jobs (for less job traffic), and
2563 for cluster tags it won't talk to masterd at all but read them from
2565 - a couple of error handling fixes in RAPI
2566 - drbd handling: improved the error handling of inconsistent disks after
2567 resync to reduce the frequency of "there are some degraded disks for
2568 this instance" messages
2569 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2570 error handling path called a wrong function name)
2576 *(Released Wed, 27 May 2009)*
2578 - no changes from rc5
2584 *(Released Wed, 20 May 2009)*
2586 - fix a couple of bugs (validation, argument checks)
2587 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2588 - some small improvements to RAPI and IAllocator
2589 - make watcher automatically start the master daemon if down
2595 *(Released Mon, 27 Apr 2009)*
2597 - change the OS list to not require locks; this helps with big clusters
2598 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2599 volume group is broken
2600 - ``gnt-instance info``, without any arguments, doesn't run for all
2601 instances anymore; either pass ``--all`` or pass the desired
2602 instances; this helps against mistakes on big clusters where listing
2603 the information for all instances takes a long time
2604 - miscellaneous doc and man pages fixes
2610 *(Released Wed, 8 Apr 2009)*
2612 - Change the internal locking model of some ``gnt-node`` commands, in
2613 order to reduce contention (and blocking of master daemon) when
2614 batching many creation/reinstall jobs
2615 - Fixes to Xen soft reboot
2616 - No longer build documentation at build time, instead distribute it in
2617 the archive, in order to reduce the need for the whole docbook/rst
2624 *(Released Fri, 27 Mar 2009)*
2626 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2627 - Fix watcher startup sequence, improves the behaviour of busy clusters
2628 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2629 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2630 bind address and other places
2631 - Some documentation fixes and updates
2637 *(Released Mon, 2 Mar 2009)*
2639 - More documentation updates, now all docs should be more-or-less
2641 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2643 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2645 - Fix an issue related to $libdir/run/ganeti and cluster creation
2651 *(Released Thu, 19 Feb 2009)*
2653 - Xen PVM and KVM have switched the default value for the instance root
2654 disk to the first partition on the first drive, instead of the whole
2655 drive; this means that the OS installation scripts must be changed
2657 - Man pages have been updated
2658 - RAPI has been switched by default to HTTPS, and the exported functions
2659 should all work correctly
2660 - RAPI v1 has been removed
2661 - Many improvements to the KVM hypervisor
2662 - Block device errors are now better reported
2663 - Many other bugfixes and small improvements
2669 *(Released Mon, 26 Jan 2009)*
2671 - Version 2 is a general rewrite of the code and therefore the
2672 differences are too many to list, see the design document for 2.0 in
2673 the ``doc/`` subdirectory for more details
2674 - In this beta version there is not yet a migration path from 1.2 (there
2675 will be one in the final 2.0 release)
2676 - A few significant changes are:
2678 - all commands are executed by a daemon (``ganeti-masterd``) and the
2679 various ``gnt-*`` commands are just front-ends to it
2680 - all the commands are entered into, and executed from a job queue,
2681 see the ``gnt-job(8)`` manpage
2682 - the RAPI daemon supports read-write operations, secured by basic
2683 HTTP authentication on top of HTTPS
2684 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2685 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2686 to migrate to DRBD 8 first while still running Ganeti 1.2)
2687 - DRBD devices are using statically allocated minor numbers, which
2688 will be assigned to existing instances during the migration process
2689 - there is support for both Xen PVM and Xen HVM instances running on
2691 - KVM virtualization is supported too
2692 - file-based storage has been implemented, which means that it is
2693 possible to run the cluster without LVM and DRBD storage, for
2694 example using a shared filesystem exported from shared storage (and
2695 still have live migration)
2701 *(Released Tue, 13 Jan 2009)*
2703 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2704 - Reuse the old instance mac address by default on instance import, if
2705 the instance name is the same.
2706 - Handle situations in which the node info rpc returns incomplete
2708 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2709 - Improved version of batcher:
2711 - state file support
2712 - instance mac address support
2713 - support for HVM clusters/instances
2715 - Add an option to show the number of cpu sockets and nodes in
2717 - Support OSes that handle more than one version of the OS api (but do
2718 not change the current API in any other way)
2719 - Fix ``gnt-node migrate``
2720 - ``gnt-debug`` man page
2721 - Fixes various more typos and small issues
2722 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2728 *(Released Wed, 24 Sep 2008)*
2730 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2731 type of disk exported to fully virtualized instances.
2732 - provide access to the serial console of HVM instances
2733 - instance auto_balance flag, set by default. If turned off it will
2734 avoid warnings on cluster verify if there is not enough memory to fail
2735 over an instance. in the future it will prevent automatically failing
2736 it over when we will support that.
2737 - batcher tool for instance creation, see ``tools/README.batcher``
2738 - ``gnt-instance reinstall --select-os`` to interactively select a new
2739 operating system when reinstalling an instance.
2740 - when changing the memory amount on instance modify a check has been
2741 added that the instance will be able to start. also warnings are
2742 emitted if the instance will not be able to fail over, if auto_balance
2744 - documentation fixes
2745 - sync fields between ``gnt-instance list/modify/add/import``
2746 - fix a race condition in drbd when the sync speed was set after giving
2747 the device a remote peer.
2753 *(Released Tue, 22 Jul 2008)*
2755 - note: the allowed size and number of tags per object were reduced
2756 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2757 - fixed twisted 8.x compatibility
2758 - fixed ``gnt-instance replace-disks`` with iallocator
2759 - add TCP keepalives on twisted connections to detect restarted nodes
2760 - disk increase support, see ``gnt-instance grow-disk``
2761 - implement bulk node/instance query for RAPI
2762 - add tags in node/instance listing (optional)
2763 - experimental migration (and live migration) support, read the man page
2764 for ``gnt-instance migrate``
2765 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2766 has some small improvements in handling its state file
2772 *(Released Fri, 13 Jun 2008)*
2774 - Experimental readonly, REST-based remote API implementation;
2775 automatically started on master node, TCP port 5080, if enabled by
2776 ``--enable-rapi`` parameter to configure script.
2777 - Instance allocator support. Add and import instance accept a
2778 ``--iallocator`` parameter, and call that instance allocator to decide
2779 which node to use for the instance. The iallocator document describes
2780 what's expected from an allocator script.
2781 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2782 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2783 node is lost there is still enough memory to fail over the instances
2785 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2786 ``gnt-cluster verify``, to check for site-specific compliance. All the
2787 hooks will run, and their output, if any, will be displayed. Any
2788 failing hook will make the verification return an error value.
2789 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2790 primary and secondary interfaces
2791 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2792 that is still declared as part of the cluster and has failed.
2793 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2794 specifying output fields, that just adds the chosen fields to the
2796 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2797 export from the filesystem.
2798 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2799 have been added. Using them you can enable/disable acpi and pae
2800 support, and specify a path for a cd image to be exported to the
2801 instance. These parameters as the name suggest only work on HVM
2803 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2804 PAE support will be set to the previously hardcoded values, but the
2805 (previously hardcoded) path to the CDROM ISO image will be unset and
2806 if required, needs to be set manually with ``gnt-instance modify``
2808 - The address to which an instance's VNC console is bound is now
2809 selectable per-instance, rather than being cluster wide. Of course
2810 this only applies to instances controlled via VNC, so currently just
2811 applies to HVM clusters.
2817 *(Released Mon, 18 Feb 2008)*
2819 - more tweaks to the disk activation code (especially helpful for DRBD)
2820 - change the default ``gnt-instance list`` output format, now there is
2821 one combined status field (see the manpage for the exact values this
2823 - some more fixes for the mac export to hooks change
2824 - make Ganeti not break with DRBD 8.2.x (which changed the version
2825 format in ``/proc/drbd``) (issue 24)
2826 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2827 template, allowing migration from DRBD0.7+MD to DRBD8
2833 *(Released Wed, 30 Jan 2008)*
2835 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2837 - add command aliases infrastructure and a few aliases
2838 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2839 man pages and the ``--help`` option of ``gnt-node
2840 list``/``gnt-instance list``
2841 - fix ``gnt-backup list`` with down nodes (issue 21)
2842 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2843 - fix the dist archive and add a check for including svn/git files in
2845 - some developer-related changes: improve the burnin and the QA suite,
2846 add an upload script for testing during development
2852 *(Released Wed, 16 Jan 2008)*
2854 - experimental HVM support, read the install document, section
2855 "Initializing the cluster"
2856 - allow for the PVM hypervisor per-instance kernel and initrd paths
2857 - add a new command ``gnt-cluster verify-disks`` which uses a new
2858 algorithm to improve the reconnection of the DRBD pairs if the device
2859 on the secondary node has gone away
2860 - make logical volume code auto-activate LVs at disk activation time
2861 - slightly improve the speed of activating disks
2862 - allow specification of the MAC address at instance creation time, and
2863 changing it later via ``gnt-instance modify``
2864 - fix handling of external commands that generate lots of output on
2866 - update documentation with regard to minimum version of DRBD8 supported
2872 *(Released Tue, 4 Dec 2007)*
2874 - Log the ``xm create`` output to the node daemon log on failure (to
2875 help diagnosing the error)
2876 - In debug mode, log all external commands output if failed to the logs
2877 - Change parsing of lvm commands to ignore stderr
2883 *(Released Wed, 28 Nov 2007)*
2885 - Another round of updates to the DRBD 8 code to deal with more failures
2886 in the replace secondary node operation
2887 - Some more logging of failures in disk operations (lvm, drbd)
2888 - A few documentation updates
2895 *(Released Tue, 13 Nov 2007)*
2897 - Change configuration file format from Python's Pickle to JSON.
2898 Upgrading is possible using the cfgupgrade utility.
2899 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2900 faster replace disks and is more stable (DRBD 8 has many improvements
2901 compared to DRBD 0.7)
2902 - Added command line tags support (see man pages for ``gnt-instance``,
2903 ``gnt-node``, ``gnt-cluster``)
2904 - Added instance rename support
2905 - Added multi-instance startup/shutdown
2906 - Added cluster rename support
2907 - Added ``gnt-node evacuate`` to simplify some node operations
2908 - Added instance reboot operation that can speedup reboot as compared to
2910 - Soften the requirement that hostnames are in FQDN format
2911 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2913 - Removed dependency on debian's patched fping that uses the
2914 non-standard ``-S`` option
2915 - Now the OS definitions are searched for in multiple, configurable
2916 paths (easier for distros to package)
2917 - Some changes to the hooks infrastructure (especially the new
2918 post-configuration update hook)
2919 - Other small bugfixes
2921 .. vim: set textwidth=72 syntax=rst :