10 - The minimum Python version needed to run Ganeti is now 2.6.
11 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
12 to always require authentication, ``--require-authentication``. It can
13 be specified in ``$sysconfdir/default/ganeti``.
14 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
15 be used to manage the disk templates to be used by instances in the cluster.
16 Initially, it will be set to a list that includes plain, drbd, if they were
17 enabled by specifying a volume group name, and file and sharedfile, if those
18 were enabled at configure time. Additionally, it will include all disk
19 templates that are currently used by instances. The order of disk templates
20 will be based on Ganeti's history of supporting them. In the future, the
21 first entry of the list will be used as a default disk template on instance
23 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
24 configuration back to the previous stable version.
25 - The cluster option '--no-lvm-storage' was removed in favor of the new option
26 '--enabled-disk-templates'.
34 Incompatible/important changes
35 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
37 - Instance policies for disk size were documented to be on a per-disk
38 basis, but hail applied them to the sum of all disks. This has been
40 - ``hbal`` will now exit with status 0 if, during job execution over
41 LUXI, early exit has been requested and all jobs are successful;
42 before, exit status 1 was used, which cannot be differentiated from
44 - Compatibility with newer versions of rbd has been fixed
45 - ``gnt-instance batch-create`` has been changed to use the bulk create
46 opcode from Ganeti. This lead to incompatible changes in the format of
47 the JSON file. It's now not a custom dict anymore but a dict
48 compatible with the ``OpInstanceCreate`` opcode.
49 - Parent directories for file storage need to be listed in
50 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
51 write the file automatically based on old configuration values, but it
52 can not distribute it across all nodes and the file contents should be
53 verified. Use ``gnt-cluster copyfile
54 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
55 upgraded. The reason for requiring this list of paths now is that
56 before it would have been possible to inject new paths via RPC,
57 allowing files to be created in arbitrary locations. The RPC protocol
58 is protected using SSL/X.509 certificates, but as a design principle
59 Ganeti does not permit arbitrary paths to be passed.
60 - The parsing of the variants file for OSes (see
61 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
62 lines and comment lines (starting with ``#``) are ignored for better
64 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
65 longer available. ``gnt-node add`` now invokes a new tool on the
66 destination node, named ``prepare-node-join``, to configure the SSH
67 daemon. Paramiko is no longer necessary to configure nodes' SSH
68 daemons via ``gnt-node add``.
69 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
70 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
71 the flag is not reset after a master failover.
72 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
73 is recommended. 2.6 will be mandatory from the 2.8 series.
79 - New network management functionality to support automatic allocation
80 of IP addresses and managing of network parameters. See
81 :manpage:`gnt-network(8)` for more details.
82 - New external storage backend, to allow managing arbitrary storage
83 systems external to the cluster. See
84 :manpage:`ganeti-extstorage-interface(7)`.
85 - New ``exclusive-storage`` node parameter added, restricted to
86 nodegroup level. When it's set to true, physical disks are assigned in
87 an exclusive fashion to instances, as documented in :doc:`Partitioned
88 Ganeti <design-partitioned>`. Currently, only instances using the
89 ``plain`` disk template are supported.
90 - The KVM hypervisor has been updated with many new hypervisor
91 parameters, including a generic one for passing arbitrary command line
92 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
93 compatible up to qemu 1.4.
94 - A new tool, called ``mon-collector``, is the stand-alone executor of
95 the data collectors for a monitoring system. As of this version, it
96 just includes the DRBD data collector, that can be executed by calling
97 ``mon-collector`` using the ``drbd`` parameter. See
98 :manpage:`mon-collector(7)`.
99 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
100 for RAPI users. It allows granting permissions to query for
101 information to a specific user without giving
102 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
103 - A new tool named ``node-cleanup`` has been added. It cleans remains of
104 a cluster from a machine by stopping all daemons, removing
105 certificates and ssconf files. Unless the ``--no-backup`` option is
106 given, copies of the certificates are made.
107 - Instance creations now support the use of opportunistic locking,
108 potentially speeding up the (parallel) creation of multiple instances.
109 This feature is currently only available via the :doc:`RAPI
110 <rapi>` interface and when an instance allocator is used. If the
111 ``opportunistic_locking`` parameter is set the opcode will try to
112 acquire as many locks as possible, but will not wait for any locks
113 held by other opcodes. If not enough resources can be found to
114 allocate the instance, the temporary error code
115 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
116 retried thereafter, with or without opportunistic locking.
117 - New experimental linux-ha resource scripts.
118 - Restricted-commands support: ganeti can now be asked (via command line
119 or rapi) to perform commands on a node. These are passed via ganeti
120 RPC rather than ssh. This functionality is restricted to commands
121 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
122 reasons. The file is not copied automatically.
128 - Diskless instances are now externally mirrored (Issue 237). This for
129 now has only been tested in conjunction with explicit target nodes for
131 - Queries not needing locks or RPC access to the node can now be
132 performed by the confd daemon, making them independent from jobs, and
133 thus faster to execute. This is selectable at configure time.
134 - The functionality for allocating multiple instances at once has been
135 overhauled and is now also available through :doc:`RAPI <rapi>`.
140 - Fix kvm compatibility with qemu 1.4 (Issue 389)
141 - Documentation updates (admin guide, upgrade notes, install
142 instructions) (Issue 372)
143 - Fix gnt-group list nodes and instances count (Issue 436)
144 - Fix compilation without non-mandatory libraries (Issue 441)
145 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
146 - Make confd logging more verbose at INFO level (Issue 435)
147 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
148 - Fix failure path for instance storage type conversion (Issue 229)
149 - Update htools text backend documentation
150 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
151 - Disable inter-cluster instance move for file-based instances, because
152 it is dependant on instance export, which is not supported for
153 file-based instances. (Issue 414)
159 *(Released Mon, 22 Apr 2013)*
161 This was the third beta release of the 2.7 series. Since beta2:
163 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
164 - Fix data loss on wrong usage of ``gnt-instance move``
165 - Properly export errors in confd-based job queries
166 - Add ``users-setup`` tool
167 - Fix iallocator protocol to report 0 as a disk size for diskless
168 instances. This avoids hail breaking when a diskless instance is
170 - Fix job queue directory permission problem that made confd job queries
171 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
172 for access to archived jobs (Issue 406).
173 - Limit the sizes of networks supported by ``gnt-network`` to something
174 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
175 - Fix bugs in instance disk template conversion
176 - Fix GHC 7 compatibility
177 - Fix ``burnin`` install path (Issue 426).
178 - Allow very small disk grows (Issue 347).
179 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
180 that noded doesn't import masterd code (Issue 419).
181 - Make sure the default metavg at cluster init is the same as the vg, if
182 unspecified (Issue 358).
183 - Fix cleanup of partially created disks (part of Issue 416)
189 *(Released Tue, 2 Apr 2013)*
191 This was the second beta release of the 2.7 series. Since beta1:
193 - Networks no longer have a "type" slot, since this information was
194 unused in Ganeti: instead of it tags should be used.
195 - The rapi client now has a ``target_node`` option to MigrateInstance.
196 - Fix early exit return code for hbal (Issue 386).
197 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
198 - Fix ``rbd showmapped`` output parsing (Issue 312).
199 - Networks are now referenced indexed by UUID, rather than name. This
200 will require running cfgupgrade, from 2.7.0beta1, if networks are in
202 - The OS environment now includes network information.
203 - Deleting of a network is now disallowed if any instance nic is using
204 it, to prevent dangling references.
205 - External storage is now documented in man pages.
206 - The exclusive_storage flag can now only be set at nodegroup level.
207 - Hbal can now submit an explicit priority with its jobs.
208 - Many network related locking fixes.
209 - Bump up the required pylint version to 0.25.1.
210 - Fix the ``no_remember`` option in RAPI client.
211 - Many ipolicy related tests, qa, and fixes.
212 - Many documentation improvements and fixes.
213 - Fix building with ``--disable-file-storage``.
214 - Fix ``-q`` option in htools, which was broken if passed more than
216 - Some haskell/python interaction improvements and fixes.
217 - Fix iallocator in case of missing LVM storage.
218 - Fix confd config load in case of ``--no-lvm-storage``.
219 - The confd/query functionality is now mentioned in the security
226 *(Released Wed, 6 Feb 2013)*
228 This was the first beta release of the 2.7 series. All important changes
229 are listed in the latest 2.7 entry.
235 *(Released Fri, 21 Dec 2012)*
237 Important behaviour change: hbal won't rebalance anymore instances which
238 have the ``auto_balance`` attribute set to false. This was the intention
239 all along, but until now it only skipped those from the N+1 memory
240 reservation (DRBD-specific).
242 A significant number of bug fixes in this release:
244 - Fixed disk adoption interaction with ipolicy checks.
245 - Fixed networking issues when instances are started, stopped or
246 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
247 - Fixed the warning in cluster verify for shared storage instances not
249 - Fixed removal of storage directory on shared file storage (issue 262).
250 - Fixed validation of LVM volume group name in OpClusterSetParams
251 (``gnt-cluster modify``) (issue 285).
252 - Fixed runtime memory increases (``gnt-instance modify -m``).
253 - Fixed live migration under Xen's ``xl`` mode.
254 - Fixed ``gnt-instance console`` with ``xl``.
255 - Fixed building with newer Haskell compiler/libraries.
256 - Fixed PID file writing in Haskell daemons (confd); this prevents
257 restart issues if confd was launched manually (outside of
258 ``daemon-util``) while another copy of it was running
259 - Fixed a type error when doing live migrations with KVM (issue 297) and
260 the error messages for failing migrations have been improved.
261 - Fixed opcode validation for the out-of-band commands (``gnt-node
263 - Fixed a type error when unsetting OS hypervisor parameters (issue
264 311); now it's possible to unset all OS-specific hypervisor
266 - Fixed the ``dry-run`` mode for many operations: verification of
267 results was over-zealous but didn't take into account the ``dry-run``
268 operation, resulting in "wrong" failures.
269 - Fixed bash completion in ``gnt-job list`` when the job queue has
270 hundreds of entries; especially with older ``bash`` versions, this
271 results in significant CPU usage.
273 And lastly, a few other improvements have been made:
275 - Added option to force master-failover without voting (issue 282).
276 - Clarified error message on lock conflict (issue 287).
277 - Logging of newly submitted jobs has been improved (issue 290).
278 - Hostname checks have been made uniform between instance rename and
280 - The ``--submit`` option is now supported by ``gnt-debug delay``.
281 - Shutting down the master daemon by sending SIGTERM now stops it from
282 processing jobs waiting for locks; instead, those jobs will be started
283 once again after the master daemon is started the next time (issue
285 - Support for Xen's ``xl`` program has been improved (besides the fixes
287 - Reduced logging noise in the Haskell confd daemon (only show one log
288 entry for each config reload, instead of two).
289 - Several man page updates and typo fixes.
295 *(Released Fri, 12 Oct 2012)*
297 A small bugfix release. Among the bugs fixed:
299 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
300 made the command unusable.
301 - Commands that issue many jobs don't fail anymore just because some jobs
302 take so long that other jobs are archived.
303 - Failures during ``gnt-instance reinstall`` are reflected by the exit
305 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
307 - When ``always_failover`` is set, ``--allow-failover`` is not required
308 in migrate commands anymore.
309 - ``bash_completion`` works even if extglob is disabled.
310 - Fixed bug with locks that made failover for RDB-based instances fail.
311 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
312 a random node instead of one based on the allocator metric.
313 - Support for newer versions of pylint and pep8.
314 - Hail doesn't fail anymore when trying to add an instance of type
315 ``file``, ``sharedfile`` or ``rbd``.
316 - Added new Makefile target to rebuild the whole distribution, so that
317 all files are included.
323 *(Released Fri, 27 Jul 2012)*
326 .. attention:: The ``LUXI`` protocol has been made more consistent
327 regarding its handling of command arguments. This, however, leads to
328 incompatibility issues with previous versions. Please ensure that you
329 restart Ganeti daemons soon after the upgrade, otherwise most
330 ``LUXI`` calls (job submission, setting/resetting the drain flag,
331 pausing/resuming the watcher, cancelling and archiving jobs, querying
332 the cluster configuration) will fail.
341 The current ``admin_up`` field, which used to denote whether an instance
342 should be running or not, has been removed. Instead, ``admin_state`` is
343 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
345 The rational behind this is that an instance being “down” can have
348 - it could be down during a reboot
349 - it could be temporarily be down for a reinstall
350 - or it could be down because it is deprecated and kept just for its
353 The previous Boolean state was making it difficult to do capacity
354 calculations: should Ganeti reserve memory for a down instance? Now, the
355 tri-state field makes it clear:
357 - in ``up`` and ``down`` state, all resources are reserved for the
358 instance, and it can be at any time brought up if it is down
359 - in ``offline`` state, only disk space is reserved for it, but not
362 The field can have an extra use: since the transition between ``up`` and
363 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
364 transition between ``offline`` and ``down`` is done via ``gnt-instance
365 modify``, it is possible to given different rights to users. For
366 example, owners of an instance could be allowed to start/stop it, but
367 not transition it out of the offline state.
369 Instance policies and specs
370 +++++++++++++++++++++++++++
372 In previous Ganeti versions, an instance creation request was not
373 limited on the minimum size and on the maximum size just by the cluster
374 resources. As such, any policy could be implemented only in third-party
375 clients (RAPI clients, or shell wrappers over ``gnt-*``
376 tools). Furthermore, calculating cluster capacity via ``hspace`` again
377 required external input with regards to instance sizes.
379 In order to improve these workflows and to allow for example better
380 per-node group differentiation, we introduced instance specs, which
383 - minimum instance disk size, disk count, memory size, cpu count
384 - maximum values for the above metrics
385 - and “standard” values (used in ``hspace`` to calculate the standard
388 The minimum/maximum values can be also customised at node-group level,
389 for example allowing more powerful hardware to support bigger instance
392 Beside the instance specs, there are a few other settings belonging to
393 the instance policy framework. It is possible now to customise, per
394 cluster and node-group:
396 - the list of allowed disk templates
397 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
398 - the maximum ratio of instance to spindles (see below for more
399 information) for local storage
401 All these together should allow all tools that talk to Ganeti to know
402 what are the ranges of allowed values for instances and the
403 over-subscription that is allowed.
405 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
406 instance configuration, and the physical CPU configuration from the
407 node. For the spindle ratios however, we didn't track before these
408 values, so new parameters have been added:
410 - a new node parameter ``spindle_count``, defaults to 1, customisable at
411 node group or node level
412 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
414 Note that spindles in this context doesn't need to mean actual
415 mechanical hard-drives; it's just a relative number for both the node
416 I/O capacity and instance I/O consumption.
418 Instance migration behaviour
419 ++++++++++++++++++++++++++++
421 While live-migration is in general desirable over failover, it is
422 possible that for some workloads it is actually worse, due to the
423 variable time of the “suspend” phase during live migration.
425 To allow the tools to work consistently over such instances (without
426 having to hard-code instance names), a new backend parameter
427 ``always_failover`` has been added to control the migration/failover
428 behaviour. When set to True, all migration requests for an instance will
429 instead fall-back to failover.
431 Instance memory ballooning
432 ++++++++++++++++++++++++++
434 Initial support for memory ballooning has been added. The memory for an
435 instance is no longer fixed (backend parameter ``memory``), but instead
436 can vary between minimum and maximum values (backend parameters
437 ``minmem`` and ``maxmem``). Currently we only change an instance's
440 - live migrating or failing over and instance and the target node
441 doesn't have enough memory
442 - user requests changing the memory via ``gnt-instance modify
448 In order to control the use of specific CPUs by instance, support for
449 controlling CPU pinning has been added for the Xen, HVM and LXC
450 hypervisors. This is controlled by a new hypervisor parameter
451 ``cpu_mask``; details about possible values for this are in the
452 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
453 VCPU-to-CPU mapping) form will work well only when all nodes in your
454 cluster have the same amount of CPUs.
459 Another area in which Ganeti was not customisable were the parameters
460 used for storage configuration, e.g. how many stripes to use for LVM,
461 DRBD resync configuration, etc.
463 To improve this area, we've added disks parameters, which are
464 customisable at cluster and node group level, and which allow to
465 specify various parameters for disks (DRBD has the most parameters
466 currently), for example:
468 - DRBD resync algorithm and parameters (e.g. speed)
469 - the default VG for meta-data volumes for DRBD
470 - number of stripes for LVM (plain disk template)
473 These parameters can be modified via ``gnt-cluster modify -D …`` and
474 ``gnt-group modify -D …``, and are used at either instance creation (in
475 case of LVM stripes, for example) or at disk “activation” time
478 Rados block device support
479 ++++++++++++++++++++++++++
481 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
482 denoted by the ``rbd`` disk template type. This is considered
483 experimental, feedback is welcome. For details on configuring it, see
484 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
489 The existing master IP functionality works well only in simple setups (a
490 single network shared by all nodes); however, if nodes belong to
491 different networks, then the ``/32`` setup and lack of routing
492 information is not enough.
494 To allow the master IP to function well in more complex cases, the
495 system was reworked as follows:
497 - a master IP netmask setting has been added
498 - the master IP activation/turn-down code was moved from the node daemon
500 - whether to run the Ganeti-supplied master IP script or a user-supplied
501 on is a ``gnt-cluster init`` setting
503 Details about the location of the standard and custom setup scripts are
504 in the man page :manpage:`gnt-cluster(8)`; for information about the
505 setup script protocol, look at the Ganeti-supplied script.
510 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
513 It is now possible to use TLS-protected connections, and when renewing
514 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
515 it is now possible to specify spice or spice CA certificates. Also, it
516 is possible to configure a password for SPICE sessions via the
517 hypervisor parameter ``spice_password_file``.
519 There are also new parameters to control the compression and streaming
520 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
521 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
522 for the spice parameters.
524 Lastly, it is now possible to see the SPICE connection information via
525 ``gnt-instance console``.
530 A new tool (``tools/ovfconverter``) has been added that supports
531 conversion between Ganeti and the `Open Virtualization Format
532 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
535 This relies on the ``qemu-img`` tool to convert the disk formats, so the
536 actual compatibility with other virtualization solutions depends on it.
541 The configuration query daemon (``ganeti-confd``) is now optional, and
542 has been rewritten in Haskell; whether to use the daemon at all, use the
543 Python (default) or the Haskell version is selectable at configure time
544 via the ``--enable-confd`` parameter, which can take one of the
545 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
546 daemon will result in a smaller footprint; for larger systems, we
547 welcome feedback on the Haskell version which might become the default
550 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
551 daemon running. The Python version doesn't implement the new call.
554 User interface changes
555 ~~~~~~~~~~~~~~~~~~~~~~
557 We have replaced the ``--disks`` option of ``gnt-instance
558 replace-disks`` with a more flexible ``--disk`` option, which allows
559 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
560 disk size and mode can be changed upon recreation (via ``gnt-instance
561 recreate-disks``, which accepts the same ``--disk`` option).
563 As many people are used to a ``show`` command, we have added that as an
564 alias to ``info`` on all ``gnt-*`` commands.
566 The ``gnt-instance grow-disk`` command has a new mode in which it can
567 accept the target size of the disk, instead of the delta; this can be
568 more safe since two runs in absolute mode will be idempotent, and
569 sometimes it's also easier to specify the desired size directly.
571 Also the handling of instances with regard to offline secondaries has
572 been improved. Instance operations should not fail because one of it's
573 secondary nodes is offline, even though it's safe to proceed.
575 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
576 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
577 minors to instance name.
582 RAPI coverage has improved, with (for example) new resources for
583 recreate-disks, node power-cycle, etc.
588 There is partial support for ``xl`` in the Xen hypervisor; feedback is
591 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
592 whether to still support Python 2.4 or move to Python 2.6 as minimum
595 Support for Fedora has been slightly improved; the provided example
596 init.d script should work better on it and the INSTALL file should
597 document the needed dependencies.
602 The deprecated ``QueryLocks`` LUXI request has been removed. Use
603 ``Query(what=QR_LOCK, ...)`` instead.
605 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
606 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
607 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
608 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
609 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
611 RAPI client: ``CertificateError`` now derives from
612 ``GanetiApiError``. This should make it more easy to handle Ganeti
615 Deprecation warnings due to PyCrypto/paramiko import in
616 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
617 make sure to run an up-to-date paramiko version, if you use this tool.
619 The QA scripts now depend on Python 2.5 or above (the main code base
620 still works with Python 2.4).
622 The configuration file (``config.data``) is now written without
623 indentation for performance reasons; if you want to edit it, it can be
624 re-formatted via ``tools/fmtjson``.
626 A number of bugs has been fixed in the cluster merge tool.
628 ``x509`` certification verification (used in import-export) has been
629 changed to allow the same clock skew as permitted by the cluster
630 verification. This will remove some rare but hard to diagnose errors in
637 *(Released Thu, 19 Jul 2012)*
639 Very few changes from rc4 to the final release, only bugfixes:
641 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
642 instance, fix CDROM booting for KVM instances)
643 - fixed node group modification of node parameters
644 - fixed issue in LUClusterVerifyGroup with multi-group clusters
645 - fixed generation of bash completion to ensure a stable ordering
652 *(Released Fri, 13 Jul 2012)*
654 Third release candidate for 2.6. The following changes were done from
657 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
658 - Fixed an inconsistency in the LUXI protocol with the provided
659 arguments (NOT backwards compatible)
660 - Fixed a bug with node groups ipolicy where ``min`` was greater than
661 the cluster ``std`` value
662 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
663 easier instance debugging on nodes (requires ``hconfd`` to work)
669 *(Released Tue, 03 Jul 2012)*
671 Second release candidate for 2.6. The following changes were done from
674 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
676 - Fixed a RAPI regression on missing beparams/memory
677 - Fixed redistribution of files on offline nodes
678 - Added possibility to run activate-disks even though secondaries are
679 offline. With this change it relaxes also the strictness on some other
680 commands which use activate disks internally:
681 * ``gnt-instance start|reboot|rename|backup|export``
682 - Made it possible to remove safely an instance if its secondaries are
684 - Made it possible to reinstall even though secondaries are offline
690 *(Released Mon, 25 Jun 2012)*
692 First release candidate for 2.6. The following changes were done from
695 - Fixed bugs with disk parameters and ``rbd`` templates as well as
697 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
698 behaviour. It supports now the modify operation
699 - ``hcheck`` implemented to analyze cluster health and possibility of
700 improving health by rebalance
701 - ``hbal`` has been improved in dealing with split instances
707 *(Released Mon, 11 Jun 2012)*
709 Second beta release of 2.6. The following changes were done from beta2
712 - Fixed ``daemon-util`` with non-root user models
713 - Fixed creation of plain instances with ``--no-wait-for-sync``
714 - Fix wrong iv_names when running ``cfgupgrade``
715 - Export more information in RAPI group queries
716 - Fixed bug when changing instance network interfaces
717 - Extended burnin to do NIC changes
718 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
719 - Changed default for DRBD barriers
720 - Fixed DRBD error reporting for syncer rate
721 - Verify the options on disk parameters
723 And of course various fixes to documentation and improved unittests and
730 *(Released Wed, 23 May 2012)*
732 First beta release of 2.6. The following changes were done from beta1 to
735 - integrated patch for distributions without ``start-stop-daemon``
736 - adapted example init.d script to work on Fedora
737 - fixed log handling in Haskell daemons
738 - adapted checks in the watcher for pycurl linked against libnss
739 - add partial support for ``xl`` instead of ``xm`` for Xen
740 - fixed a type issue in cluster verification
741 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
744 Plus integrated fixes from the 2.5 branch:
746 - fixed ``kvm-ifup`` to use ``/bin/bash``
747 - fixed parallel build failures
748 - KVM live migration when using a custom keymap
754 *(Released Tue, 24 Jul 2012)*
756 A small bugfix release, with no new features:
758 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
759 different default shell (e.g. Debian, Ubuntu)
760 - fixed KVM startup and live migration with a custom keymap (fixes Issue
761 243 and Debian bug #650664)
762 - fixed compatibility with KVM versions that don't support multiple boot
763 devices (fixes Issue 230 and Debian bug #624256)
765 Additionally, a few fixes were done to the build system (fixed parallel
766 build failures) and to the unittests (fixed race condition in test for
767 FileID functions, and the default enable/disable mode for QA test is now
774 *(Released Fri, 11 May 2012)*
776 A small bugfix release.
778 The main issues solved are on the topic of compatibility with newer LVM
781 - fixed parsing of ``lv_attr`` field
782 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
783 the ``--force`` flag is needed
785 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
786 to accept kernel 3.x too (was hardcoded to 2.6.*).
788 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
789 of hook results and that also made display of other errors suboptimal
790 was fixed; the code behaves now like 2.4 and earlier.
792 Another change in 2.5, the cleanup of the OS scripts environment, is too
793 aggressive: it removed even the ``PATH`` variable, which requires the OS
794 scripts to *always* need to export it. Since this is a bit too strict,
795 we now export a minimal PATH, the same that we export for hooks.
797 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
798 integrated into this release.
800 Finally, a few other miscellaneous changes were done (no new features,
801 just small improvements):
803 - Fix ``gnt-group --help`` display
804 - Fix hardcoded Xen kernel path
805 - Fix grow-disk handling of invalid units
806 - Update synopsis for ``gnt-cluster repair-disk-sizes``
807 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
813 *(Released Thu, 12 Apr 2012)*
815 Incompatible/important changes and bugfixes
816 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
818 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
819 resource's ``ip_check`` parameter changed from ``True`` to ``False``
820 to match the underlying LUXI interface.
821 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
822 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
823 not maintain backwards-compatibility as the underlying operation
824 changed in an incompatible way. The RAPI client can talk to old
825 servers, but it needs to be told so as the return value changed.
826 - When creating file-based instances via RAPI, the ``file_driver``
827 parameter no longer defaults to ``loop`` and must be specified.
828 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
830 - Support for the undocumented and deprecated RAPI instance creation
831 request format version 0 has been dropped. Use version 1, supported
832 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
833 - Pyparsing 1.4.6 or above is required, see :doc:`installation
834 documentation <install>`.
835 - The "cluster-verify" hooks are now executed per group by the
836 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
837 if you just run ``gnt-cluster verify``, which generates one opcode per
839 - The environment as passed to the OS scripts is cleared, and thus no
840 environment variables defined in the node daemon's environment will be
841 inherited by the scripts.
842 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
844 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
845 support operations involving multiple node groups.
846 - Offline nodes are ignored when failing over an instance.
847 - Support for KVM version 1.0, which changed the version reporting format
849 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
851 - ``Makefile`` is now compatible with Automake 1.11.2
852 - Includes all bugfixes made in the 2.4 series
857 - The ganeti-htools project has been merged into the ganeti-core source
858 tree and will be built as part of Ganeti (see :doc:`install-quick`).
859 - Implemented support for :doc:`shared storage <design-shared-storage>`.
860 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
861 GPT-style partition tables (requires `parted
862 <http://www.gnu.org/s/parted/>`_).
863 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
864 - Allowed adding tags on instance creation.
865 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
866 - Allow instances to be started in a paused state, enabling the user to
867 see the complete console output on boot using the console.
868 - Added new hypervisor flag to control default reboot behaviour
869 (``reboot_behavior``).
870 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
871 - Improved out-of-band management support:
873 - Added ``gnt-node health`` command reporting the health status of
875 - Added ``gnt-node power`` command to manage power status of nodes.
876 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
878 - Instance migration can fall back to failover if instance is not
880 - Filters can be used when listing nodes, instances, groups and locks;
881 see :manpage:`ganeti(7)` manpage.
882 - Added post-execution status as variables to :doc:`hooks <hooks>`
884 - Instance tags are exported/imported together with the instance.
885 - When given an explicit job ID, ``gnt-job info`` will work for archived
887 - Jobs can define dependencies on other jobs (not yet supported via
888 RAPI or command line, but used by internal commands and usable via
891 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
894 - Instance failover is now available as a RAPI resource
895 (``/2/instances/[instance_name]/failover``).
896 - ``gnt-instance info`` defaults to static information if primary node
898 - Opcodes have a new ``comment`` attribute.
899 - Added basic SPICE support to KVM hypervisor.
900 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
902 Node group improvements
903 ~~~~~~~~~~~~~~~~~~~~~~~
905 - ``gnt-cluster verify`` has been modified to check groups separately,
906 thereby improving performance.
907 - Node group support has been added to ``gnt-cluster verify-disks``,
908 which now operates per node group.
909 - Watcher has been changed to work better with node groups.
911 - One process and state file per node group.
912 - Slow watcher in one group doesn't block other group's watcher.
914 - Added new command, ``gnt-group evacuate``, to move all instances in a
915 node group to other groups.
916 - Added ``gnt-instance change-group`` to move an instance to another
918 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
919 per-group operations.
920 - Node groups can be tagged.
921 - Some operations switch from an exclusive to a shared lock as soon as
923 - Instance's primary and secondary nodes' groups are now available as
924 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
925 and ``snodes.group.uuid``).
930 - Numerous updates to documentation and manpages.
932 - :doc:`RAPI <rapi>` documentation now has detailed parameter
934 - Some opcode/job results are now also documented, see :doc:`RAPI
937 - A lockset's internal lock is now also visible in lock monitor.
938 - Log messages from job queue workers now contain information about the
939 opcode they're processing.
940 - ``gnt-instance console`` no longer requires the instance lock.
941 - A short delay when waiting for job changes reduces the number of LUXI
942 requests significantly.
943 - DRBD metadata volumes are overwritten with zeros during disk creation.
944 - Out-of-band commands no longer acquire the cluster lock in exclusive
946 - ``devel/upload`` now uses correct permissions for directories.
952 *(Released Fri, 23 Mar 2012)*
954 This was the sixth release candidate of the 2.5 series.
960 *(Released Mon, 9 Jan 2012)*
962 This was the fifth release candidate of the 2.5 series.
968 *(Released Thu, 27 Oct 2011)*
970 This was the fourth release candidate of the 2.5 series.
976 *(Released Wed, 26 Oct 2011)*
978 This was the third release candidate of the 2.5 series.
984 *(Released Tue, 18 Oct 2011)*
986 This was the second release candidate of the 2.5 series.
992 *(Released Tue, 4 Oct 2011)*
994 This was the first release candidate of the 2.5 series.
1000 *(Released Wed, 31 Aug 2011)*
1002 This was the third beta release of the 2.5 series.
1008 *(Released Mon, 22 Aug 2011)*
1010 This was the second beta release of the 2.5 series.
1016 *(Released Fri, 12 Aug 2011)*
1018 This was the first beta release of the 2.5 series.
1024 *(Released Thu, 27 Oct 2011)*
1026 - Fixed bug when parsing command line parameter values ending in
1028 - Fixed assertion error after unclean master shutdown
1029 - Disable HTTP client pool for RPC, significantly reducing memory usage
1031 - Fixed queue archive creation with wrong permissions
1037 *(Released Tue, 23 Aug 2011)*
1041 - Fixed documentation for importing with ``--src-dir`` option
1042 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1043 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1049 *(Released Fri, 5 Aug 2011)*
1051 Many bug-fixes and a few small features:
1053 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1054 issues when you tried to add an instance with two MAC addresses in one
1056 - KVM: fixed per-instance stored UID value
1057 - KVM: configure bridged NICs at migration start
1058 - KVM: Fix a bug where instance will not start with never KVM versions
1060 - Added OS search path to ``gnt-cluster info``
1061 - Fixed an issue with ``file_storage_dir`` where you were forced to
1062 provide an absolute path, but the documentation states it is a
1063 relative path, the documentation was right
1064 - Added a new parameter to instance stop/start called ``--no-remember``
1065 that will make the state change to not be remembered
1066 - Implemented ``no_remember`` at RAPI level
1067 - Improved the documentation
1068 - Node evacuation: don't call IAllocator if node is already empty
1069 - Fixed bug in DRBD8 replace disks on current nodes
1070 - Fixed bug in recreate-disks for DRBD instances
1071 - Moved assertion checking locks in ``gnt-instance replace-disks``
1072 causing it to abort with not owning the right locks for some situation
1073 - Job queue: Fixed potential race condition when cancelling queued jobs
1074 - Fixed off-by-one bug in job serial generation
1075 - ``gnt-node volumes``: Fix instance names
1076 - Fixed aliases in bash completion
1077 - Fixed a bug in reopening log files after being sent a SIGHUP
1078 - Added a flag to burnin to allow specifying VCPU count
1079 - Bugfixes to non-root Ganeti configuration
1085 *(Released Thu, 12 May 2011)*
1087 Many bug-fixes and a few new small features:
1089 - Fixed a bug related to log opening failures
1090 - Fixed a bug in instance listing with orphan instances
1091 - Fixed a bug which prevented resetting the cluster-level node parameter
1092 ``oob_program`` to the default
1093 - Many fixes related to the ``cluster-merge`` tool
1094 - Fixed a race condition in the lock monitor, which caused failures
1095 during (at least) creation of many instances in parallel
1096 - Improved output for gnt-job info
1097 - Removed the quiet flag on some ssh calls which prevented debugging
1099 - Improved the N+1 failure messages in cluster verify by actually
1100 showing the memory values (needed and available)
1101 - Increased lock attempt timeouts so that when executing long operations
1102 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1103 too early and thus prevent the use of the 'fair' mechanism
1104 - Changed instance query data (``gnt-instance info``) to not acquire
1105 locks unless needed, thus allowing its use on locked instance if only
1106 static information is asked for
1107 - Improved behaviour with filesystems that do not support rename on an
1109 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1110 kept locks on all nodes during the wipe, which is unneeded
1111 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1112 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1113 10GiB) which caused the wipe to fail right at the end in some cases
1114 - Fixed master IP activation when doing master failover with no-voting
1115 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1116 the master node itself
1117 - Fixed potential data-loss in under disk full conditions, where Ganeti
1118 wouldn't check correctly the return code and would consider
1119 partially-written files 'correct'
1120 - Fixed bug related to multiple VGs and DRBD disk replacing
1121 - Added new disk parameter ``metavg`` that allows placement of the meta
1122 device for DRBD in a different volume group
1123 - Fixed error handling in the node daemon when the system libc doesn't
1124 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1125 - Fixed lock release during replace-disks, which kept cluster-wide locks
1126 when doing disk replaces with an iallocator script
1127 - Added check for missing bridges in cluster verify
1128 - Handle EPIPE errors while writing to the terminal better, so that
1129 piping the output to e.g. ``less`` doesn't cause a backtrace
1130 - Fixed rare case where a ^C during Luxi calls could have been
1131 interpreted as server errors, instead of simply terminating
1132 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1134 - Added a few more parameters to the KVM hypervisor, allowing a second
1135 CDROM, custom disk type for CDROMs and a floppy image
1136 - Removed redundant message in instance rename when the name is given
1138 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1139 disks on new nodes, allowing recreation when the original instance
1140 nodes are completely gone
1141 - Added option when converting disk templates to DRBD to skip waiting
1142 for the resync, in order to make the instance available sooner
1143 - Added two new variables to the OS scripts environment (containing the
1145 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1146 to allow use of ``pvgrub`` as bootloader
1147 - Changed the instance memory modifications to only check out-of-memory
1148 conditions on memory increases, and turned the secondary node warnings
1149 into errors (they can still be overridden via ``--force``)
1150 - Fixed the handling of a corner case when the Python installation gets
1151 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1152 to execute a command that doesn't exist
1153 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1154 node of the instance returned failures during instance shutdown; this
1155 adds the option ``--ignore-consistency`` to gnt-instance move
1157 And as usual, various improvements to the error messages, documentation
1164 *(Released Wed, 09 Mar 2011)*
1166 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1167 the RAPI users file if run twice (even with ``--dry-run``).
1169 The release fixes that bug (nothing else changed).
1175 *(Released Mon, 07 Mar 2011)*
1177 Final 2.4.0 release. Just a few small fixes:
1179 - Fixed RAPI node evacuate
1180 - Fixed the kvm-ifup script
1181 - Fixed internal error handling for special job cases
1182 - Updated man page to specify the escaping feature for options
1188 *(Released Mon, 28 Feb 2011)*
1190 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1191 that this feature wiped the disks of the wrong instance, leading to loss
1196 - Fixed title of query field containing instance name
1197 - Expanded the glossary in the documentation
1198 - Fixed one unittest (internal issue)
1204 *(Released Mon, 21 Feb 2011)*
1206 A number of bug fixes plus just a couple functionality changes.
1208 On the user-visible side, the ``gnt-* list`` command output has changed
1209 with respect to "special" field states. The current rc1 style of display
1210 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1211 the default output mode special fields states are displayed as follows:
1213 - Offline resource: ``*``
1214 - Unavailable/not applicable: ``-``
1215 - Data missing (RPC failure): ``?``
1216 - Unknown field: ``??``
1218 Another user-visible change is the addition of ``--force-join`` to
1223 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1224 - Fixed regression in RAPI/instance reinstall where all parameters were
1225 required (instead of optional)
1226 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1227 - Fixed iallocator usage (offline nodes were not considered offline)
1228 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1229 - Fixed hypervisor and OS parameter validation with respect to
1230 non-vm_capable nodes
1231 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1233 - Fixed ``tools/listrunner`` with respect to agent-based usage
1239 *(Released Fri, 4 Feb 2011)*
1241 Many changes and fixes since the beta1 release. While there were some
1242 internal changes, the code has been mostly stabilised for the RC
1245 Note: the dumb allocator was removed in this release, as it was not kept
1246 up-to-date with the IAllocator protocol changes. It is recommended to
1247 use the ``hail`` command from the ganeti-htools package.
1249 Note: the 2.4 and up versions of Ganeti are not compatible with the
1250 0.2.x branch of ganeti-htools. You need to upgrade to
1251 ganeti-htools-0.3.0 (or later).
1253 Regressions fixed from 2.3
1254 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1256 - Fixed the ``gnt-cluster verify-disks`` command
1257 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1259 - Fixed disk adoption breakage
1260 - Fixed wrong headers in instance listing for field aliases
1265 - Fixed corner case in KVM handling of NICs
1266 - Fixed many cases of wrong handling of non-vm_capable nodes
1267 - Fixed a bug where a missing instance symlink was not possible to
1268 recreate with any ``gnt-*`` command (now ``gnt-instance
1269 activate-disks`` does it)
1270 - Fixed the volume group name as reported by ``gnt-cluster
1272 - Increased timeouts for the import-export code, hopefully leading to
1273 fewer aborts due network or instance timeouts
1274 - Fixed bug in ``gnt-node list-storage``
1275 - Fixed bug where not all daemons were started on cluster
1276 initialisation, but only at the first watcher run
1277 - Fixed many bugs in the OOB implementation
1278 - Fixed watcher behaviour in presence of instances with offline
1280 - Fixed instance list output for instances running on the wrong node
1281 - a few fixes to the cluster-merge tool, but it still cannot merge
1282 multi-node groups (currently it is not recommended to use this tool)
1288 - Improved network configuration for the KVM hypervisor
1289 - Added e1000 as a supported NIC for Xen-HVM
1290 - Improved the lvmstrap tool to also be able to use partitions, as
1291 opposed to full disks
1292 - Improved speed of disk wiping (the cluster parameter
1293 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1294 of instance creations
1295 - Added documentation for the OS parameters
1296 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1297 hypervisor is not responding
1298 - Added display of blacklisted and hidden OS information in
1299 ``gnt-cluster info``
1300 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1301 NIC and node parameters, which might create problems with currently
1302 invalid (but undetected) configuration files, but prevents validation
1303 failures when unrelated parameters are modified
1304 - Changed cluster initialisation to wait for the master daemon to become
1306 - Expanded the RAPI interface:
1308 - Added config redistribution resource
1309 - Added activation/deactivation of instance disks
1310 - Added export of console information
1312 - Implemented log file reopening on SIGHUP, which allows using
1313 logrotate(8) for the Ganeti log files
1314 - Added a basic OOB helper script as an example
1320 *(Released Fri, 14 Jan 2011)*
1325 - Fixed timezone issues when formatting timestamps
1326 - Added support for node groups, available via ``gnt-group`` and other
1328 - Added out-of-band framework and management, see :doc:`design
1329 document <design-oob>`
1330 - Removed support for roman numbers from ``gnt-node list`` and
1331 ``gnt-instance list``.
1332 - Allowed modification of master network interface via ``gnt-cluster
1333 modify --master-netdev``
1334 - Accept offline secondaries while shutting down instance disks
1335 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1336 - Added support for multiple LVM volume groups
1337 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1339 - Added commands to list available fields:
1341 - ``gnt-node list-fields``
1342 - ``gnt-group list-fields``
1343 - ``gnt-instance list-fields``
1345 - Updated documentation and man pages
1350 - Moved ``rapi_users`` file into separate directory, now named
1351 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1353 - Added new tool for running commands on many machines,
1354 ``tools/ganeti-listrunner``
1355 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1356 improving the ``gnt-instance console`` output
1357 - Allowed customisation of disk index separator at ``configure`` time
1358 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1359 - Added support for non-partitioned md disks in ``lvmstrap``
1360 - Added script to gracefully power off KVM instances
1361 - Split ``utils`` module into smaller parts
1362 - Changed query operations to return more detailed information, e.g.
1363 whether an information is unavailable due to an offline node. To use
1364 this new functionality, the LUXI call ``Query`` must be used. Field
1365 information is now stored by the master daemon and can be retrieved
1366 using ``QueryFields``. Instances, nodes and groups can also be queried
1367 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1368 exposed via RAPI). The following commands make use of this
1369 infrastructure change:
1371 - ``gnt-group list``
1372 - ``gnt-group list-fields``
1374 - ``gnt-node list-fields``
1375 - ``gnt-instance list``
1376 - ``gnt-instance list-fields``
1377 - ``gnt-debug locks``
1382 - New RAPI resources (see :doc:`rapi`):
1386 - ``/2/groups/[group_name]``
1387 - ``/2/groups/[group_name]/assign-nodes``
1388 - ``/2/groups/[group_name]/modify``
1389 - ``/2/groups/[group_name]/rename``
1390 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1394 - Implemented ``no_install`` for instance creation
1395 - Implemented OS parameters for instance reinstallation, allowing
1396 use of special settings on reinstallation (e.g. for preserving data)
1401 - Added IPv6 support in import/export
1402 - Pause DRBD synchronization while wiping disks on instance creation
1403 - Updated unittests and QA scripts
1404 - Improved network parameters passed to KVM
1405 - Converted man pages from docbook to reStructuredText
1411 *(Released Mon, 20 Dec 2010)*
1413 Released version 2.3.1~rc1 without any changes.
1419 *(Released Wed, 1 Dec 2010)*
1421 - impexpd: Disable OpenSSL compression in socat if possible (backport
1422 from master, commit e90739d625b, see :doc:`installation guide
1423 <install-quick>` for details)
1424 - Changed unittest coverage report to exclude test scripts
1425 - Added script to check version format
1431 *(Released Wed, 1 Dec 2010)*
1433 Released version 2.3.0~rc1 without any changes.
1439 *(Released Fri, 19 Nov 2010)*
1441 A number of bugfixes and documentation updates:
1443 - Update ganeti-os-interface documentation
1444 - Fixed a bug related to duplicate MACs or similar items which should be
1446 - Fix breakage in OS state modify
1447 - Reinstall instance: disallow offline secondaries (fixes bug related to
1448 OS changing but reinstall failing)
1449 - plus all the other fixes between 2.2.1 and 2.2.2
1455 *(Released Tue, 2 Nov 2010)*
1457 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1458 - Fixed master failover race with watcher
1459 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1461 - Accept previously stopped instance for export with instance removal
1462 - Simplify and extend the environment variables for instance OS scripts
1463 - Added new node flags, ``master_capable`` and ``vm_capable``
1464 - Added optional instance disk wiping prior during allocation. This is a
1465 cluster-wide option and can be set/modified using
1466 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1467 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1468 :doc:`install-quick`
1469 - Added a new watcher option (``--ignore-pause``)
1470 - Added option to ignore offline node on instance start/stop
1471 (``--ignore-offline``)
1472 - Allow overriding OS parameters with ``gnt-instance reinstall``
1473 - Added ability to change node's secondary IP address using ``gnt-node
1475 - Implemented privilege separation for all daemons except
1476 ``ganeti-noded``, see ``configure`` options
1477 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1479 - Implemented job priorities (see ``ganeti(7)`` manpage)
1480 - Ignore failures while shutting down instances during failover from
1482 - Exit daemon's bootstrap process only once daemon is ready
1483 - Export more information via ``LUInstanceQuery``/remote API
1484 - Improved documentation, QA and unittests
1485 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1486 restart if the file was created or changed
1487 - Added LUXI protocol version sent with each request and response,
1488 allowing detection of server/client mismatches
1489 - Moved the Python scripts among gnt-* and ganeti-* into modules
1490 - Moved all code related to setting up SSH to an external script,
1492 - Infrastructure changes for node group support in future versions
1498 *(Released Fri, 19 Nov 2010)*
1500 A few small bugs fixed, and some improvements to the build system:
1502 - Fix documentation regarding conversion to drbd
1503 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1505 - Fix error handling in node modify with multiple changes
1506 - Allow remote imports without checked names
1512 *(Released Tue, 19 Oct 2010)*
1514 - Disable SSL session ID cache in RPC client
1520 *(Released Thu, 14 Oct 2010)*
1522 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1523 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1525 - Fix problems with interaction between Curl and Python's HTTP server,
1526 resulting in increased speed in many RPC calls
1527 - Improve our release script to prevent breakage with older aclocal and
1534 *(Released Thu, 7 Oct 2010)*
1536 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1537 value retrieved from master
1538 - Added support for blacklisted or hidden OS definitions
1539 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1540 - Added support for -mem-path in KVM hypervisor abstraction layer
1541 - Allow overriding instance parameters in tool for inter-cluster
1542 instance moves (``tools/move-instance``)
1543 - Improved opcode summaries (e.g. in ``gnt-job list``)
1544 - Improve consistency of OS listing by sorting it
1545 - Documentation updates
1551 *(Released Fri, 8 Oct 2010)*
1553 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1559 *(Released Mon, 4 Oct 2010)*
1561 - Fixed regression in ``gnt-instance rename``
1567 *(Released Wed, 22 Sep 2010)*
1569 - Fixed OS_VARIANT variable for OS scripts
1570 - Fixed cluster tag operations via RAPI
1571 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1572 - Disabled RAPI CA checks in watcher
1578 *(Released Mon, 23 Aug 2010)*
1580 - Support DRBD versions of the format "a.b.c.d"
1582 - Re-introduce support for usage from multiple threads in RAPI client
1583 - Instance renames and modify via RAPI
1584 - Work around race condition between processing and archival in job
1586 - Mark opcodes following failed one as failed, too
1587 - Job field ``lock_status`` was removed due to difficulties making it
1588 work with the changed job queue in Ganeti 2.2; a better way to monitor
1589 locks is expected for a later 2.2.x release
1590 - Fixed dry-run behaviour with many commands
1591 - Support ``ssh-agent`` again when adding nodes
1592 - Many additional bugfixes
1598 *(Released Fri, 30 Jul 2010)*
1600 Important change: the internal RPC mechanism between Ganeti nodes has
1601 changed from using a home-grown http library (based on the Python base
1602 libraries) to use the PycURL library. This requires that PycURL is
1603 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1604 against GnuTLS by default. cURL's support for GnuTLS had known issues
1605 before cURL 7.21.0 and we recommend using the latest cURL release or
1606 linking against OpenSSL. Most other distributions already link PycURL
1607 and cURL against OpenSSL. The command::
1609 python -c 'import pycurl; print pycurl.version'
1611 can be used to determine the libraries PycURL and cURL are linked
1614 Other significant changes:
1616 - Rewrote much of the internals of the job queue, in order to achieve
1617 better parallelism; this decouples job query operations from the job
1618 processing, and it should allow much nicer behaviour of the master
1619 daemon under load, and it also has uncovered some long-standing bugs
1620 related to the job serialisation (now fixed)
1621 - Added a default iallocator setting to the cluster parameters,
1622 eliminating the need to always pass nodes or an iallocator for
1623 operations that require selection of new node(s)
1624 - Added experimental support for the LXC virtualization method
1625 - Added support for OS parameters, which allows the installation of
1626 instances to pass parameter to OS scripts in order to customise the
1628 - Added a hypervisor parameter controlling the migration type (live or
1629 non-live), since hypervisors have various levels of reliability; this
1630 has renamed the 'live' parameter to 'mode'
1631 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1632 logical volumes, meaning that cluster verify will ignore them and not
1633 flag their presence as errors
1634 - The watcher will now reset the error count for failed instances after
1635 8 hours, thus allowing self-healing if the problem that caused the
1636 instances to be down/fail to start has cleared in the meantime
1637 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1638 check for, and warn, if the drbd module parameter ``usermode_helper``
1639 is not consistent with the cluster-wide setting; this is needed to
1640 make diagnose easier of failed drbd creations
1641 - Started adding base IPv6 support, but this is not yet
1642 enabled/available for use
1643 - Rename operations (cluster, instance) will now return the new name,
1644 which is especially useful if a short name was passed in
1645 - Added support for instance migration in RAPI
1646 - Added a tool to pre-configure nodes for the SSH setup, before joining
1647 them to the cluster; this will allow in the future a simplified model
1648 for node joining (but not yet fully enabled in 2.2); this needs the
1649 paramiko python library
1650 - Fixed handling of name-resolving errors
1651 - Fixed consistency of job results on the error path
1652 - Fixed master-failover race condition when executed multiple times in
1654 - Fixed many bugs related to the job queue (mostly introduced during the
1655 2.2 development cycle, so not all are impacting 2.1)
1656 - Fixed instance migration with missing disk symlinks
1657 - Fixed handling of unknown jobs in ``gnt-job archive``
1658 - And many other small fixes/improvements
1662 - Enhanced both the unittest and the QA coverage
1663 - Switched the opcode validation to a generic model, and extended the
1664 validation to all opcode parameters
1665 - Changed more parts of the code that write shell scripts to use the
1667 - Switched the master daemon to use the asyncore library for the Luxi
1674 *(Released Thu, 17 Jun 2010)*
1676 - Added tool (``move-instance``) and infrastructure to move instances
1677 between separate clusters (see :doc:`separate documentation
1678 <move-instance>` and :doc:`design document <design-2.2>`)
1679 - Added per-request RPC timeout
1680 - RAPI now requires a Content-Type header for requests with a body (e.g.
1681 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1682 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1683 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1685 - Implemented initial support for running Ganeti daemons as separate
1686 users, see configure-time flags ``--with-user-prefix`` and
1687 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1689 - Instances can be removed after export (``gnt-backup export
1690 --remove-instance``)
1691 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1692 key (instead of 1024 bit)
1693 - Added new cluster configuration file for cluster domain secret
1694 - Import/export now use SSL instead of SSH
1695 - Added support for showing estimated time when exporting an instance,
1696 see the ``ganeti-os-interface(7)`` manpage and look for
1703 *(Released Tue, 16 Nov 2010)*
1705 Some more bugfixes. Unless critical bugs occur, this will be the last
1708 - Fix case of MAC special-values
1709 - Fix mac checker regex
1710 - backend: Fix typo causing "out of range" error
1711 - Add missing --units in gnt-instance list man page
1717 *(Released Tue, 24 Aug 2010)*
1720 - Don't ignore secondary node silently on non-mirrored disk templates
1722 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1723 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1724 - Properly document the usb_mouse parameter
1725 - Fix path in ganeti-rapi(8) (issue 116)
1726 - Adjust error message when the ganeti user's .ssh directory is
1728 - Add same-node-check when changing the disk template to drbd
1734 *(Released Fri, 16 Jul 2010)*
1737 - Add an option to only select some reboot types during qa/burnin.
1738 (on some hypervisors consequent reboots are not supported)
1739 - Fix infrequent race condition in master failover. Sometimes the old
1740 master ip address would be still detected as up for a short time
1741 after it was removed, causing failover to fail.
1742 - Decrease mlockall warnings when the ctypes module is missing. On
1743 Python 2.4 we support running even if no ctypes module is installed,
1744 but we were too verbose about this issue.
1745 - Fix building on old distributions, on which man doesn't have a
1747 - Fix RAPI not to ignore the MAC address on instance creation
1748 - Implement the old instance creation format in the RAPI client.
1754 *(Released Thu, 01 Jul 2010)*
1756 A small bugfix release:
1757 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1758 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1759 a non-existing option
1760 - Fix instance create: the --force-variant option was ignored
1761 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1762 - Fix modify node storage with non-FQDN arguments
1763 - Fix RAPI client to authenticate under Python 2.6 when used
1764 for more than 5 requests needing authentication
1765 - Fix gnt-instance modify -t (storage) giving a wrong error message
1766 when converting a non-shutdown drbd instance to plain
1772 *(Released Fri, 18 Jun 2010)*
1774 A small bugfix release:
1776 - Fix live migration of KVM instances started with older Ganeti
1777 versions which had fewer hypervisor parameters
1778 - Fix gnt-instance grow-disk on down instances
1779 - Fix an error-reporting bug during instance migration
1780 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1781 silently ignoring broken ones
1782 - Fix an RPC error reporting bug affecting, for example, RAPI client
1784 - Fix bug triggered by different API version os-es on different nodes
1785 - Fix a bug in instance startup with custom hvparams: OS level
1786 parameters would fail to be applied.
1787 - Fix the RAPI client under Python 2.6 (but more work is needed to
1788 make it work completely well with OpenSSL)
1789 - Fix handling of errors when resolving names from DNS
1795 *(Released Thu, 3 Jun 2010)*
1797 A medium sized development cycle. Some new features, and some
1798 fixes/small improvements/cleanups.
1800 Significant features
1801 ~~~~~~~~~~~~~~~~~~~~
1803 The node deamon now tries to mlock itself into memory, unless the
1804 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1805 its logs, and falls back to console logging. This allows emergency
1806 features such as ``gnt-node powercycle`` to work even in the event of a
1807 broken node disk (tested offlining the disk hosting the node's
1808 filesystem and dropping its memory caches; don't try this at home)
1810 KVM: add vhost-net acceleration support. It can be tested with a new
1811 enough version of the kernel and of qemu-kvm.
1813 KVM: Add instance chrooting feature. If you use privilege dropping for
1814 your VMs you can also now force them to chroot to an empty directory,
1815 before starting the emulated guest.
1817 KVM: Add maximum migration bandwith and maximum downtime tweaking
1818 support (requires a new-enough version of qemu-kvm).
1820 Cluster verify will now warn if the master node doesn't have the master
1821 ip configured on it.
1823 Add a new (incompatible) instance creation request format to RAPI which
1824 supports all parameters (previously only a subset was supported, and it
1825 wasn't possible to extend the old format to accomodate all the new
1826 features. The old format is still supported, and a client can check for
1827 this feature, before using it, by checking for its presence in the
1828 ``features`` RAPI resource.
1830 Now with ancient latin support. Try it passing the ``--roman`` option to
1831 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1832 (requires the python-roman module to be installed, in order to work).
1837 As usual many internal code refactorings, documentation updates, and
1840 - Lots of improvements and cleanups to the experimental Remote API
1841 (RAPI) client library.
1842 - A new unit test suite for the core daemon libraries.
1843 - A fix to creating missing directories makes sure the umask is not
1844 applied anymore. This enforces the same directory permissions
1846 - Better handling terminating daemons with ctrl+c (used when running
1847 them in debugging mode).
1848 - Fix a race condition in live migrating a KVM instance, when stat()
1849 on the old proc status file returned EINVAL, which is an unexpected
1851 - Fixed manpage checking with newer man and utf-8 charachters. But now
1852 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1858 *(Released Fri, 7 May 2010)*
1860 Fix a bug which prevented untagged KVM instances from starting.
1866 *(Released Fri, 7 May 2010)*
1868 Another release with a long development cycle, during which many
1869 different features were added.
1871 Significant features
1872 ~~~~~~~~~~~~~~~~~~~~
1874 The KVM hypervisor now can run the individual instances as non-root, to
1875 reduce the impact of a VM being hijacked due to bugs in the
1876 hypervisor. It is possible to run all instances as a single (non-root)
1877 user, to manually specify a user for each instance, or to dynamically
1878 allocate a user out of a cluster-wide pool to each instance, with the
1879 guarantee that no two instances will run under the same user ID on any
1882 An experimental RAPI client library, that can be used standalone
1883 (without the other Ganeti libraries), is provided in the source tree as
1884 ``lib/rapi/client.py``. Note this client might change its interface in
1885 the future, as we iterate on its capabilities.
1887 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1888 replace the cluster's certificates and crypto keys. This might help in
1889 case they have been compromised, or have simply expired.
1891 A new disk option for instance creation has been added that allows one
1892 to "adopt" currently existing logical volumes, with data
1893 preservation. This should allow easier migration to Ganeti from
1894 unmanaged (or managed via other software) instances.
1896 Another disk improvement is the possibility to convert between redundant
1897 (DRBD) and plain (LVM) disk configuration for an instance. This should
1898 allow better scalability (starting with one node and growing the
1899 cluster, or shrinking a two-node cluster to one node).
1901 A new feature that could help with automated node failovers has been
1902 implemented: if a node sees itself as offline (by querying the master
1903 candidates), it will try to shutdown (hard) all instances and any active
1904 DRBD devices. This reduces the risk of duplicate instances if an
1905 external script automatically failovers the instances on such nodes. To
1906 enable this, the cluster parameter ``maintain_node_health`` should be
1907 enabled; in the future this option (per the name) will enable other
1908 automatic maintenance features.
1910 Instance export/import now will reuse the original instance
1911 specifications for all parameters; that means exporting an instance,
1912 deleting it and the importing it back should give an almost identical
1913 instance. Note that the default import behaviour has changed from
1914 before, where it created only one NIC; now it recreates the original
1917 Cluster verify has added a few new checks: SSL certificates validity,
1918 /etc/hosts consistency across the cluster, etc.
1923 As usual, many internal changes were done, documentation fixes,
1926 - Fixed cluster initialization with disabled cluster storage (regression
1927 introduced in 2.1.1)
1928 - File-based storage supports growing the disks
1929 - Fixed behaviour of node role changes
1930 - Fixed cluster verify for some corner cases, plus a general rewrite of
1931 cluster verify to allow future extension with more checks
1932 - Fixed log spamming by watcher and node daemon (regression introduced
1934 - Fixed possible validation issues when changing the list of enabled
1936 - Fixed cleanup of /etc/hosts during node removal
1937 - Fixed RAPI response for invalid methods
1938 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1939 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1940 from ide disks, etc.)
1941 - Allow OS changes without re-installation (to record a changed OS
1942 outside of Ganeti, or to allow OS renames)
1943 - Allow instance creation without OS installation (useful for example if
1944 the OS will be installed manually, or restored from a backup not in
1946 - Implemented option to make cluster ``copyfile`` use the replication
1948 - Added list of enabled hypervisors to ssconf (possibly useful for
1950 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1952 - A partial form of node re-IP is possible via node readd, which now
1953 allows changed node primary IP
1954 - Command line utilities now show an informational message if the job is
1956 - The logs of the master daemon now show the PID/UID/GID of the
1963 *(Released Fri, 12 Mar 2010)*
1965 During the 2.1.0 long release candidate cycle, a lot of improvements and
1966 changes have accumulated with were released later as 2.1.1.
1971 The node evacuate command (``gnt-node evacuate``) was significantly
1972 rewritten, and as such the IAllocator protocol was changed - a new
1973 request type has been added. This unfortunate change during a stable
1974 series is designed to improve performance of node evacuations; on
1975 clusters with more than about five nodes and which are well-balanced,
1976 evacuation should proceed in parallel for all instances of the node
1977 being evacuated. As such, any existing IAllocator scripts need to be
1978 updated, otherwise the above command will fail due to the unknown
1979 request. The provided "dumb" allocator has not been updated; but the
1980 ganeti-htools package supports the new protocol since version 0.2.4.
1982 Another important change is increased validation of node and instance
1983 names. This might create problems in special cases, if invalid host
1984 names are being used.
1986 Also, a new layer of hypervisor parameters has been added, that sits at
1987 OS level between the cluster defaults and the instance ones. This allows
1988 customisation of virtualization parameters depending on the installed
1989 OS. For example instances with OS 'X' may have a different KVM kernel
1990 (or any other parameter) than the cluster defaults. This is intended to
1991 help managing a multiple OSes on the same cluster, without manual
1992 modification of each instance's parameters.
1994 A tool for merging clusters, ``cluster-merge``, has been added in the
1995 tools sub-directory.
2000 - Improved the int/float conversions that should make the code more
2001 robust in face of errors from the node daemons
2002 - Fixed the remove node code in case of internal configuration errors
2003 - Fixed the node daemon behaviour in face of inconsistent queue
2004 directory (e.g. read-only file-system where we can't open the files
2006 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2007 now it either aborts cleanly or, if given the new "auto_promote"
2008 parameter, will automatically promote other nodes as needed
2009 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2010 completely prevent Ganeti from working
2011 - Fixed bug for instance export when not all disks were successfully
2013 - Fixed behaviour of node add when the new node is slow in starting up
2015 - Fixed handling of signals in the LUXI client, which should improve
2016 behaviour of command-line scripts
2017 - Added checks for invalid node/instance names in the configuration (now
2018 flagged during cluster verify)
2019 - Fixed watcher behaviour for disk activation errors
2020 - Fixed two potentially endless loops in http library, which led to the
2021 RAPI daemon hanging and consuming 100% CPU in some cases
2022 - Fixed bug in RAPI daemon related to hashed passwords
2023 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2025 - Enhanced compatibility with non-Debian OSes, but not using absolute
2026 path in some commands and allowing customisation of the ssh
2027 configuration directory
2028 - Fixed possible future issue with new Python versions by abiding to the
2029 proper use of ``__slots__`` attribute on classes
2030 - Added checks that should prevent directory traversal attacks
2031 - Many documentation fixes based on feedback from users
2036 - Added an "early_release" more for instance replace disks and node
2037 evacuate, where we release locks earlier and thus allow higher
2038 parallelism within the cluster
2039 - Added watcher hooks, intended to allow the watcher to restart other
2040 daemons (e.g. from the ganeti-nbma project), but they can be used of
2041 course for any other purpose
2042 - Added a compile-time disable for DRBD barriers, to increase
2043 performance if the administrator trusts the power supply or the
2044 storage system to not lose writes
2045 - Added the option of using syslog for logging instead of, or in
2046 addition to, Ganeti's own log files
2047 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2048 can indeed boot from a paravirtual NIC
2049 - Added a generic debug level for many operations; while this is not
2050 used widely yet, it allows one to pass the debug value all the way to
2052 - Enhanced the hooks environment for instance moves (failovers,
2053 migrations) where the primary/secondary nodes changed during the
2054 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2055 - Enhanced data validations for many user-supplied values; one important
2056 item is the restrictions imposed on instance and node names, which
2057 might reject some (invalid) host names
2058 - Add a configure-time option to disable file-based storage, if it's not
2059 needed; this allows greater security separation between the master
2060 node and the other nodes from the point of view of the inter-node RPC
2062 - Added user notification in interactive tools if job is waiting in the
2063 job queue or trying to acquire locks
2064 - Added log messages when a job is waiting for locks
2065 - Added filtering by node tags in instance operations which admit
2066 multiple instances (start, stop, reboot, reinstall)
2067 - Added a new tool for cluster mergers, ``cluster-merge``
2068 - Parameters from command line which are of the form ``a=b,c=d`` can now
2069 use backslash escapes to pass in values which contain commas,
2070 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2072 - For KVM, the instance name is the first parameter passed to KVM, so
2073 that it's more visible in the process list
2079 *(Released Tue, 2 Mar 2010)*
2081 Ganeti 2.1 brings many improvements with it. Major changes:
2083 - Added infrastructure to ease automated disk repairs
2084 - Added new daemon to export configuration data in a cheaper way than
2085 using the remote API
2086 - Instance NICs can now be routed instead of being associated with a
2088 - Improved job locking logic to reduce impact of jobs acquiring multiple
2089 locks waiting for other long-running jobs
2091 In-depth implementation details can be found in the Ganeti 2.1 design
2097 - Added chroot hypervisor
2098 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2100 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2101 ``bootloader_path`` and ``bootloader_args``)
2102 - Added the ``use_localtime`` option for the xen-hvm and kvm
2103 hypervisors, and the default value for this has changed to false (in
2104 2.0 xen-hvm always enabled it)
2105 - Added luxi call to submit multiple jobs in one go
2106 - Added cluster initialization option to not modify ``/etc/hosts``
2108 - Added network interface parameters
2109 - Added dry run mode to some LUs
2110 - Added RAPI resources:
2112 - ``/2/instances/[instance_name]/info``
2113 - ``/2/instances/[instance_name]/replace-disks``
2114 - ``/2/nodes/[node_name]/evacuate``
2115 - ``/2/nodes/[node_name]/migrate``
2116 - ``/2/nodes/[node_name]/role``
2117 - ``/2/nodes/[node_name]/storage``
2118 - ``/2/nodes/[node_name]/storage/modify``
2119 - ``/2/nodes/[node_name]/storage/repair``
2121 - Added OpCodes to evacuate or migrate all instances on a node
2122 - Added new command to list storage elements on nodes (``gnt-node
2123 list-storage``) and modify them (``gnt-node modify-storage``)
2124 - Added new ssconf files with master candidate IP address
2125 (``ssconf_master_candidates_ips``), node primary IP address
2126 (``ssconf_node_primary_ips``) and node secondary IP address
2127 (``ssconf_node_secondary_ips``)
2128 - Added ``ganeti-confd`` and a client library to query the Ganeti
2129 configuration via UDP
2130 - Added ability to run hooks after cluster initialization and before
2132 - Added automatic mode for disk replace (``gnt-instance replace-disks
2134 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2135 after catastrophic data-loss
2136 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2138 - Added ``gnt-instance move`` command to move instances
2139 - Added ``gnt-cluster watcher`` command to control watcher
2140 - Added ``gnt-node powercycle`` command to powercycle nodes
2141 - Added new job status field ``lock_status``
2142 - Added parseable error codes to cluster verification (``gnt-cluster
2143 verify --error-codes``) and made output less verbose (use
2144 ``--verbose`` to restore previous behaviour)
2145 - Added UUIDs to the main config entities (cluster, nodes, instances)
2146 - Added support for OS variants
2147 - Added support for hashed passwords in the Ganeti remote API users file
2149 - Added option to specify maximum timeout on instance shutdown
2150 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2151 - Added new helper script to start and stop Ganeti daemons
2152 (``daemon-util``), with the intent to reduce the work necessary to
2153 adjust Ganeti for non-Debian distributions and to start/stop daemons
2155 - Added more unittests
2156 - Fixed critical bug in ganeti-masterd startup
2157 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2158 now customisable at the cluster level for both the KVM and Xen
2159 hypervisors using the new ``migration_port`` parameter
2160 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2161 reinstalling an instance
2162 - Allowed ``@`` in tag names
2163 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2164 - Many documentation updates
2165 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2166 - ``gnt-instance reinstall`` can now reinstall multiple instances
2167 - Updated many command line parameters
2168 - Introduced new OS API version 15
2169 - No longer support a default hypervisor
2170 - Treat virtual LVs as inexistent
2171 - Improved job locking logic to reduce lock contention
2172 - Match instance and node names case insensitively
2173 - Reimplemented bash completion script to be more complete
2180 *(Released Thu, 4 Feb 2010)*
2182 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2183 - Fix a string formatting bug
2184 - Improve safety of the code in some error paths
2185 - Improve data validation in the master of values returned from nodes
2191 *(Released Thu, 17 Dec 2009)*
2193 - Fix security issue due to missing validation of iallocator names; this
2194 allows local and remote execution of arbitrary executables
2195 - Fix failure of gnt-node list during instance removal
2196 - Ship the RAPI documentation in the archive
2202 *(Released Wed, 30 Sep 2009)*
2204 - Fixed many wrong messages
2205 - Fixed a few bugs related to the locking library
2206 - Fixed MAC checking at instance creation time
2207 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2208 - Fixed a few issues related to signal handling in both daemons and
2210 - Fixed the example startup script provided
2211 - Fixed insserv dependencies in the example startup script (patch from
2213 - Fixed handling of drained nodes in the iallocator framework
2214 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2216 - Fixed error related to invalid job IDs in job polling
2217 - Fixed job/opcode persistence on unclean master shutdown
2218 - Fixed handling of partial job processing after unclean master
2220 - Fixed error reporting from LUs, previously all errors were converted
2221 into execution errors
2222 - Fixed error reporting from burnin
2223 - Decreased significantly the memory usage of the job queue
2224 - Optimised slightly multi-job submission
2225 - Optimised slightly opcode loading
2226 - Backported the multi-job submit framework from the development
2227 branch; multi-instance start and stop should be faster
2228 - Added script to clean archived jobs after 21 days; this will reduce
2229 the size of the queue directory
2230 - Added some extra checks in disk size tracking
2231 - Added an example ethers hook script
2232 - Added a cluster parameter that prevents Ganeti from modifying of
2234 - Added more node information to RAPI responses
2235 - Added a ``gnt-job watch`` command that allows following the ouput of a
2237 - Added a bind-address option to ganeti-rapi
2238 - Added more checks to the configuration verify
2239 - Enhanced the burnin script such that some operations can be retried
2241 - Converted instance reinstall to multi-instance model
2247 *(Released Fri, 7 Aug 2009)*
2249 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2250 to allow using the pre-2.0.2 behaviour in activation, if any existing
2251 instances have mismatched disk sizes in the configuration
2252 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2253 any configuration mismatches for disk sizes
2254 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2255 failover to work on two-node clusters
2256 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2258 - Fixed detection of OS script errors in ``gnt-backup export``
2259 - Fixed exit code of ``gnt-backup export``
2265 *(Released Fri, 17 Jul 2009)*
2267 - Added experimental support for stripped logical volumes; this should
2268 enhance performance but comes with a higher complexity in the block
2269 device handling; stripping is only enabled when passing
2270 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2271 affected even in the non-stripped mode
2272 - Improved resiliency against transient failures at the end of DRBD
2273 resyncs, and in general of DRBD resync checks
2274 - Fixed a couple of issues with exports and snapshot errors
2275 - Fixed a couple of issues in instance listing
2276 - Added display of the disk size in ``gnt-instance info``
2277 - Fixed checking for valid OSes in instance creation
2278 - Fixed handling of the "vcpus" parameter in instance listing and in
2279 general of invalid parameters
2280 - Fixed http server library, and thus RAPI, to handle invalid
2281 username/password combinations correctly; this means that now they
2282 report unauthorized for queries too, not only for modifications,
2283 allowing earlier detect of configuration problems
2284 - Added a new "role" node list field, equivalent to the master/master
2285 candidate/drained/offline flags combinations
2286 - Fixed cluster modify and changes of candidate pool size
2287 - Fixed cluster verify error messages for wrong files on regular nodes
2288 - Fixed a couple of issues with node demotion from master candidate role
2289 - Fixed node readd issues
2290 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2291 - Added a new ``--no-voting`` option for masterfailover to fix failover
2292 on two-nodes clusters when the former master node is unreachable
2293 - Added instance reinstall over RAPI
2299 *(Released Tue, 16 Jun 2009)*
2301 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2302 allow re-adding the start in single-user option (regression from 1.2)
2303 - the watcher writes the instance status to a file, to allow monitoring
2304 to report the instance status (from the master) based on cached
2305 results of the watcher's queries; while this can get stale if the
2306 watcher is being locked due to other work on the cluster, this is
2307 still an improvement
2308 - the watcher now also restarts the node daemon and the rapi daemon if
2310 - fixed the watcher to handle full and drained queue cases
2311 - hooks export more instance data in the environment, which helps if
2312 hook scripts need to take action based on the instance's properties
2313 (no longer need to query back into ganeti)
2314 - instance failovers when the instance is stopped do not check for free
2315 RAM, so that failing over a stopped instance is possible in low memory
2317 - rapi uses queries for tags instead of jobs (for less job traffic), and
2318 for cluster tags it won't talk to masterd at all but read them from
2320 - a couple of error handling fixes in RAPI
2321 - drbd handling: improved the error handling of inconsistent disks after
2322 resync to reduce the frequency of "there are some degraded disks for
2323 this instance" messages
2324 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2325 error handling path called a wrong function name)
2331 *(Released Wed, 27 May 2009)*
2333 - no changes from rc5
2339 *(Released Wed, 20 May 2009)*
2341 - fix a couple of bugs (validation, argument checks)
2342 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2343 - some small improvements to RAPI and IAllocator
2344 - make watcher automatically start the master daemon if down
2350 *(Released Mon, 27 Apr 2009)*
2352 - change the OS list to not require locks; this helps with big clusters
2353 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2354 volume group is broken
2355 - ``gnt-instance info``, without any arguments, doesn't run for all
2356 instances anymore; either pass ``--all`` or pass the desired
2357 instances; this helps against mistakes on big clusters where listing
2358 the information for all instances takes a long time
2359 - miscellaneous doc and man pages fixes
2365 *(Released Wed, 8 Apr 2009)*
2367 - Change the internal locking model of some ``gnt-node`` commands, in
2368 order to reduce contention (and blocking of master daemon) when
2369 batching many creation/reinstall jobs
2370 - Fixes to Xen soft reboot
2371 - No longer build documentation at build time, instead distribute it in
2372 the archive, in order to reduce the need for the whole docbook/rst
2379 *(Released Fri, 27 Mar 2009)*
2381 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2382 - Fix watcher startup sequence, improves the behaviour of busy clusters
2383 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2384 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2385 bind address and other places
2386 - Some documentation fixes and updates
2392 *(Released Mon, 2 Mar 2009)*
2394 - More documentation updates, now all docs should be more-or-less
2396 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2398 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2400 - Fix an issue related to $libdir/run/ganeti and cluster creation
2406 *(Released Thu, 19 Feb 2009)*
2408 - Xen PVM and KVM have switched the default value for the instance root
2409 disk to the first partition on the first drive, instead of the whole
2410 drive; this means that the OS installation scripts must be changed
2412 - Man pages have been updated
2413 - RAPI has been switched by default to HTTPS, and the exported functions
2414 should all work correctly
2415 - RAPI v1 has been removed
2416 - Many improvements to the KVM hypervisor
2417 - Block device errors are now better reported
2418 - Many other bugfixes and small improvements
2424 *(Released Mon, 26 Jan 2009)*
2426 - Version 2 is a general rewrite of the code and therefore the
2427 differences are too many to list, see the design document for 2.0 in
2428 the ``doc/`` subdirectory for more details
2429 - In this beta version there is not yet a migration path from 1.2 (there
2430 will be one in the final 2.0 release)
2431 - A few significant changes are:
2433 - all commands are executed by a daemon (``ganeti-masterd``) and the
2434 various ``gnt-*`` commands are just front-ends to it
2435 - all the commands are entered into, and executed from a job queue,
2436 see the ``gnt-job(8)`` manpage
2437 - the RAPI daemon supports read-write operations, secured by basic
2438 HTTP authentication on top of HTTPS
2439 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2440 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2441 to migrate to DRBD 8 first while still running Ganeti 1.2)
2442 - DRBD devices are using statically allocated minor numbers, which
2443 will be assigned to existing instances during the migration process
2444 - there is support for both Xen PVM and Xen HVM instances running on
2446 - KVM virtualization is supported too
2447 - file-based storage has been implemented, which means that it is
2448 possible to run the cluster without LVM and DRBD storage, for
2449 example using a shared filesystem exported from shared storage (and
2450 still have live migration)
2456 *(Released Tue, 13 Jan 2009)*
2458 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2459 - Reuse the old instance mac address by default on instance import, if
2460 the instance name is the same.
2461 - Handle situations in which the node info rpc returns incomplete
2463 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2464 - Improved version of batcher:
2466 - state file support
2467 - instance mac address support
2468 - support for HVM clusters/instances
2470 - Add an option to show the number of cpu sockets and nodes in
2472 - Support OSes that handle more than one version of the OS api (but do
2473 not change the current API in any other way)
2474 - Fix ``gnt-node migrate``
2475 - ``gnt-debug`` man page
2476 - Fixes various more typos and small issues
2477 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2483 *(Released Wed, 24 Sep 2008)*
2485 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2486 type of disk exported to fully virtualized instances.
2487 - provide access to the serial console of HVM instances
2488 - instance auto_balance flag, set by default. If turned off it will
2489 avoid warnings on cluster verify if there is not enough memory to fail
2490 over an instance. in the future it will prevent automatically failing
2491 it over when we will support that.
2492 - batcher tool for instance creation, see ``tools/README.batcher``
2493 - ``gnt-instance reinstall --select-os`` to interactively select a new
2494 operating system when reinstalling an instance.
2495 - when changing the memory amount on instance modify a check has been
2496 added that the instance will be able to start. also warnings are
2497 emitted if the instance will not be able to fail over, if auto_balance
2499 - documentation fixes
2500 - sync fields between ``gnt-instance list/modify/add/import``
2501 - fix a race condition in drbd when the sync speed was set after giving
2502 the device a remote peer.
2508 *(Released Tue, 22 Jul 2008)*
2510 - note: the allowed size and number of tags per object were reduced
2511 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2512 - fixed twisted 8.x compatibility
2513 - fixed ``gnt-instance replace-disks`` with iallocator
2514 - add TCP keepalives on twisted connections to detect restarted nodes
2515 - disk increase support, see ``gnt-instance grow-disk``
2516 - implement bulk node/instance query for RAPI
2517 - add tags in node/instance listing (optional)
2518 - experimental migration (and live migration) support, read the man page
2519 for ``gnt-instance migrate``
2520 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2521 has some small improvements in handling its state file
2527 *(Released Fri, 13 Jun 2008)*
2529 - Experimental readonly, REST-based remote API implementation;
2530 automatically started on master node, TCP port 5080, if enabled by
2531 ``--enable-rapi`` parameter to configure script.
2532 - Instance allocator support. Add and import instance accept a
2533 ``--iallocator`` parameter, and call that instance allocator to decide
2534 which node to use for the instance. The iallocator document describes
2535 what's expected from an allocator script.
2536 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2537 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2538 node is lost there is still enough memory to fail over the instances
2540 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2541 ``gnt-cluster verify``, to check for site-specific compliance. All the
2542 hooks will run, and their output, if any, will be displayed. Any
2543 failing hook will make the verification return an error value.
2544 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2545 primary and secondary interfaces
2546 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2547 that is still declared as part of the cluster and has failed.
2548 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2549 specifying output fields, that just adds the chosen fields to the
2551 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2552 export from the filesystem.
2553 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2554 have been added. Using them you can enable/disable acpi and pae
2555 support, and specify a path for a cd image to be exported to the
2556 instance. These parameters as the name suggest only work on HVM
2558 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2559 PAE support will be set to the previously hardcoded values, but the
2560 (previously hardcoded) path to the CDROM ISO image will be unset and
2561 if required, needs to be set manually with ``gnt-instance modify``
2563 - The address to which an instance's VNC console is bound is now
2564 selectable per-instance, rather than being cluster wide. Of course
2565 this only applies to instances controlled via VNC, so currently just
2566 applies to HVM clusters.
2572 *(Released Mon, 18 Feb 2008)*
2574 - more tweaks to the disk activation code (especially helpful for DRBD)
2575 - change the default ``gnt-instance list`` output format, now there is
2576 one combined status field (see the manpage for the exact values this
2578 - some more fixes for the mac export to hooks change
2579 - make Ganeti not break with DRBD 8.2.x (which changed the version
2580 format in ``/proc/drbd``) (issue 24)
2581 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2582 template, allowing migration from DRBD0.7+MD to DRBD8
2588 *(Released Wed, 30 Jan 2008)*
2590 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2592 - add command aliases infrastructure and a few aliases
2593 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2594 man pages and the ``--help`` option of ``gnt-node
2595 list``/``gnt-instance list``
2596 - fix ``gnt-backup list`` with down nodes (issue 21)
2597 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2598 - fix the dist archive and add a check for including svn/git files in
2600 - some developer-related changes: improve the burnin and the QA suite,
2601 add an upload script for testing during development
2607 *(Released Wed, 16 Jan 2008)*
2609 - experimental HVM support, read the install document, section
2610 "Initializing the cluster"
2611 - allow for the PVM hypervisor per-instance kernel and initrd paths
2612 - add a new command ``gnt-cluster verify-disks`` which uses a new
2613 algorithm to improve the reconnection of the DRBD pairs if the device
2614 on the secondary node has gone away
2615 - make logical volume code auto-activate LVs at disk activation time
2616 - slightly improve the speed of activating disks
2617 - allow specification of the MAC address at instance creation time, and
2618 changing it later via ``gnt-instance modify``
2619 - fix handling of external commands that generate lots of output on
2621 - update documentation with regard to minimum version of DRBD8 supported
2627 *(Released Tue, 4 Dec 2007)*
2629 - Log the ``xm create`` output to the node daemon log on failure (to
2630 help diagnosing the error)
2631 - In debug mode, log all external commands output if failed to the logs
2632 - Change parsing of lvm commands to ignore stderr
2638 *(Released Wed, 28 Nov 2007)*
2640 - Another round of updates to the DRBD 8 code to deal with more failures
2641 in the replace secondary node operation
2642 - Some more logging of failures in disk operations (lvm, drbd)
2643 - A few documentation updates
2650 *(Released Tue, 13 Nov 2007)*
2652 - Change configuration file format from Python's Pickle to JSON.
2653 Upgrading is possible using the cfgupgrade utility.
2654 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2655 faster replace disks and is more stable (DRBD 8 has many improvements
2656 compared to DRBD 0.7)
2657 - Added command line tags support (see man pages for ``gnt-instance``,
2658 ``gnt-node``, ``gnt-cluster``)
2659 - Added instance rename support
2660 - Added multi-instance startup/shutdown
2661 - Added cluster rename support
2662 - Added ``gnt-node evacuate`` to simplify some node operations
2663 - Added instance reboot operation that can speedup reboot as compared to
2665 - Soften the requirement that hostnames are in FQDN format
2666 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2668 - Removed dependency on debian's patched fping that uses the
2669 non-standard ``-S`` option
2670 - Now the OS definitions are searched for in multiple, configurable
2671 paths (easier for distros to package)
2672 - Some changes to the hooks infrastructure (especially the new
2673 post-configuration update hook)
2674 - Other small bugfixes
2676 .. vim: set textwidth=72 syntax=rst :