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 ``$sysconfdir/default/ganeti``.
14 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
15 be used to manage the disk templates to be used by instances in the cluster.
16 Initially, it will be set to a list that includes plain, drbd, if they were
17 enabled by specifying a volume group name, and file and sharedfile, if those
18 were enabled at configure time. Additionally, it will include all disk
19 templates that are currently used by instances. The order of disk templates
20 will be based on Ganeti's history of supporting them. In the future, the
21 first entry of the list will be used as a default disk template on instance
23 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
24 configuration back to the previous stable version.
30 *(Released Mon, 22 Apr 2013)*
32 Incompatible/important changes
33 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
35 - Instance policies for disk size were documented to be on a per-disk
36 basis, but hail applied them to the sum of all disks. This has been
38 - ``hbal`` will now exit with status 0 if, during job execution over
39 LUXI, early exit has been requested and all jobs are successful;
40 before, exit status 1 was used, which cannot be differentiated from
42 - Compatibility with newer versions of rbd has been fixed
43 - ``gnt-instance batch-create`` has been changed to use the bulk create
44 opcode from Ganeti. This lead to incompatible changes in the format of
45 the JSON file. It's now not a custom dict anymore but a dict
46 compatible with the ``OpInstanceCreate`` opcode.
47 - Parent directories for file storage need to be listed in
48 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
49 write the file automatically based on old configuration values, but it
50 can not distribute it across all nodes and the file contents should be
51 verified. Use ``gnt-cluster copyfile
52 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
53 upgraded. The reason for requiring this list of paths now is that
54 before it would have been possible to inject new paths via RPC,
55 allowing files to be created in arbitrary locations. The RPC protocol
56 is protected using SSL/X.509 certificates, but as a design principle
57 Ganeti does not permit arbitrary paths to be passed.
58 - The parsing of the variants file for OSes (see
59 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
60 lines and comment lines (starting with ``#``) are ignored for better
62 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
63 longer available. ``gnt-node add`` now invokes a new tool on the
64 destination node, named ``prepare-node-join``, to configure the SSH
65 daemon. Paramiko is no longer necessary to configure nodes' SSH
66 daemons via ``gnt-node add``.
67 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
68 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
69 the flag is not reset after a master failover.
70 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
71 is recommended. 2.6 will be mandatory from the 2.8 series.
77 - New network management functionality to support automatic allocation
78 of IP addresses and managing of network parameters. See
79 :manpage:`gnt-network(8)` for more details.
80 - New external storage backend, to allow managing arbitrary storage
81 systems external to the cluster. See
82 :manpage:`ganeti-extstorage-interface(7)`.
83 - New ``exclusive-storage`` node parameter added, restricted to
84 nodegroup level. When it's set to true, physical disks are assigned in
85 an exclusive fashion to instances, as documented in :doc:`Partitioned
86 Ganeti <design-partitioned>`. Currently, only instances using the
87 ``plain`` disk template are supported.
88 - The KVM hypervisor has been updated with many new hypervisor
89 parameters, including a generic one for passing arbitrary command line
90 values. See a complete list in :manpage:`gnt-instance(8)`.
91 - A new tool, called ``mon-collector``, is the stand-alone executor of
92 the data collectors for a monitoring system. As of this version, it
93 just includes the DRBD data collector, that can be executed by calling
94 ``mon-collector`` using the ``drbd`` parameter. See
95 :manpage:`mon-collector(7)`.
96 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
97 for RAPI users. It allows granting permissions to query for
98 information to a specific user without giving
99 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
100 - A new tool named ``node-cleanup`` has been added. It cleans remains of
101 a cluster from a machine by stopping all daemons, removing
102 certificates and ssconf files. Unless the ``--no-backup`` option is
103 given, copies of the certificates are made.
104 - Instance creations now support the use of opportunistic locking,
105 potentially speeding up the (parallel) creation of multiple instances.
106 This feature is currently only available via the :doc:`RAPI
107 <rapi>` interface and when an instance allocator is used. If the
108 ``opportunistic_locking`` parameter is set the opcode will try to
109 acquire as many locks as possible, but will not wait for any locks
110 held by other opcodes. If not enough resources can be found to
111 allocate the instance, the temporary error code
112 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
113 retried thereafter, with or without opportunistic locking.
114 - New experimental linux-ha resource scripts.
115 - Restricted-commands support: ganeti can now be asked (via command line
116 or rapi) to perform commands on a node. These are passed via ganeti
117 RPC rather than ssh. This functionality is restricted to commands
118 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
119 reasons. The file is not copied automatically.
125 - Diskless instances are now externally mirrored (Issue 237). This for
126 now has only been tested in conjunction with explicit target nodes for
128 - Queries not needing locks or RPC access to the node can now be
129 performed by the confd daemon, making them independent from jobs, and
130 thus faster to execute. This is selectable at configure time.
131 - The functionality for allocating multiple instances at once has been
132 overhauled and is now also available through :doc:`RAPI <rapi>`.
136 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
137 - Fix data loss on wrong usage of ``gnt-instance move``
138 - Properly export errors in confd-based job queries
139 - Add ``users-setup`` tool
140 - Fix iallocator protocol to report 0 as a disk size for diskless
141 instances. This avoids hail breaking when a diskless instance is
143 - Fix job queue directory permission problem that made confd job queries
144 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
145 for access to archived jobs (Issue 406).
146 - Limit the sizes of networks supported by ``gnt-network`` to something
147 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
148 - Fix bugs in instance disk template conversion
149 - Fix GHC 7 compatibility
150 - Fix ``burnin`` install path (Issue 426).
151 - Allow very small disk grows (Issue 347).
152 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
153 that noded doesn't import masterd code (Issue 419).
154 - Make sure the default metavg at cluster init is the same as the vg, if
155 unspecified (Issue 358).
156 - Fix cleanup of partially created disks (part of Issue 416)
162 *(Released Tue, 2 Apr 2013)*
164 This was the second beta release of the 2.7 series. Since beta1:
166 - Networks no longer have a "type" slot, since this information was
167 unused in Ganeti: instead of it tags should be used.
168 - The rapi client now has a ``target_node`` option to MigrateInstance.
169 - Fix early exit return code for hbal (Issue 386).
170 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
171 - Fix ``rbd showmapped`` output parsing (Issue 312).
172 - Networks are now referenced indexed by UUID, rather than name. This
173 will require running cfgupgrade, from 2.7.0beta1, if networks are in
175 - The OS environment now includes network information.
176 - Deleting of a network is now disallowed if any instance nic is using
177 it, to prevent dangling references.
178 - External storage is now documented in man pages.
179 - The exclusive_storage flag can now only be set at nodegroup level.
180 - Hbal can now submit an explicit priority with its jobs.
181 - Many network related locking fixes.
182 - Bump up the required pylint version to 0.25.1.
183 - Fix the ``no_remember`` option in RAPI client.
184 - Many ipolicy related tests, qa, and fixes.
185 - Many documentation improvements and fixes.
186 - Fix building with ``--disable-file-storage``.
187 - Fix ``-q`` option in htools, which was broken if passed more than
189 - Some haskell/python interaction improvements and fixes.
190 - Fix iallocator in case of missing LVM storage.
191 - Fix confd config load in case of ``--no-lvm-storage``.
192 - The confd/query functionality is now mentioned in the security
199 *(Released Wed, 6 Feb 2013)*
201 This was the first beta release of the 2.7 series. All important changes
202 are listed in the latest 2.7 entry.
208 *(Released Fri, 21 Dec 2012)*
210 Important behaviour change: hbal won't rebalance anymore instances which
211 have the ``auto_balance`` attribute set to false. This was the intention
212 all along, but until now it only skipped those from the N+1 memory
213 reservation (DRBD-specific).
215 A significant number of bug fixes in this release:
217 - Fixed disk adoption interaction with ipolicy checks.
218 - Fixed networking issues when instances are started, stopped or
219 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
220 - Fixed the warning in cluster verify for shared storage instances not
222 - Fixed removal of storage directory on shared file storage (issue 262).
223 - Fixed validation of LVM volume group name in OpClusterSetParams
224 (``gnt-cluster modify``) (issue 285).
225 - Fixed runtime memory increases (``gnt-instance modify -m``).
226 - Fixed live migration under Xen's ``xl`` mode.
227 - Fixed ``gnt-instance console`` with ``xl``.
228 - Fixed building with newer Haskell compiler/libraries.
229 - Fixed PID file writing in Haskell daemons (confd); this prevents
230 restart issues if confd was launched manually (outside of
231 ``daemon-util``) while another copy of it was running
232 - Fixed a type error when doing live migrations with KVM (issue 297) and
233 the error messages for failing migrations have been improved.
234 - Fixed opcode validation for the out-of-band commands (``gnt-node
236 - Fixed a type error when unsetting OS hypervisor parameters (issue
237 311); now it's possible to unset all OS-specific hypervisor
239 - Fixed the ``dry-run`` mode for many operations: verification of
240 results was over-zealous but didn't take into account the ``dry-run``
241 operation, resulting in "wrong" failures.
242 - Fixed bash completion in ``gnt-job list`` when the job queue has
243 hundreds of entries; especially with older ``bash`` versions, this
244 results in significant CPU usage.
246 And lastly, a few other improvements have been made:
248 - Added option to force master-failover without voting (issue 282).
249 - Clarified error message on lock conflict (issue 287).
250 - Logging of newly submitted jobs has been improved (issue 290).
251 - Hostname checks have been made uniform between instance rename and
253 - The ``--submit`` option is now supported by ``gnt-debug delay``.
254 - Shutting down the master daemon by sending SIGTERM now stops it from
255 processing jobs waiting for locks; instead, those jobs will be started
256 once again after the master daemon is started the next time (issue
258 - Support for Xen's ``xl`` program has been improved (besides the fixes
260 - Reduced logging noise in the Haskell confd daemon (only show one log
261 entry for each config reload, instead of two).
262 - Several man page updates and typo fixes.
268 *(Released Fri, 12 Oct 2012)*
270 A small bugfix release. Among the bugs fixed:
272 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
273 made the command unusable.
274 - Commands that issue many jobs don't fail anymore just because some jobs
275 take so long that other jobs are archived.
276 - Failures during ``gnt-instance reinstall`` are reflected by the exit
278 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
280 - When ``always_failover`` is set, ``--allow-failover`` is not required
281 in migrate commands anymore.
282 - ``bash_completion`` works even if extglob is disabled.
283 - Fixed bug with locks that made failover for RDB-based instances fail.
284 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
285 a random node instead of one based on the allocator metric.
286 - Support for newer versions of pylint and pep8.
287 - Hail doesn't fail anymore when trying to add an instance of type
288 ``file``, ``sharedfile`` or ``rbd``.
289 - Added new Makefile target to rebuild the whole distribution, so that
290 all files are included.
296 *(Released Fri, 27 Jul 2012)*
299 .. attention:: The ``LUXI`` protocol has been made more consistent
300 regarding its handling of command arguments. This, however, leads to
301 incompatibility issues with previous versions. Please ensure that you
302 restart Ganeti daemons soon after the upgrade, otherwise most
303 ``LUXI`` calls (job submission, setting/resetting the drain flag,
304 pausing/resuming the watcher, cancelling and archiving jobs, querying
305 the cluster configuration) will fail.
314 The current ``admin_up`` field, which used to denote whether an instance
315 should be running or not, has been removed. Instead, ``admin_state`` is
316 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
318 The rational behind this is that an instance being “down” can have
321 - it could be down during a reboot
322 - it could be temporarily be down for a reinstall
323 - or it could be down because it is deprecated and kept just for its
326 The previous Boolean state was making it difficult to do capacity
327 calculations: should Ganeti reserve memory for a down instance? Now, the
328 tri-state field makes it clear:
330 - in ``up`` and ``down`` state, all resources are reserved for the
331 instance, and it can be at any time brought up if it is down
332 - in ``offline`` state, only disk space is reserved for it, but not
335 The field can have an extra use: since the transition between ``up`` and
336 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
337 transition between ``offline`` and ``down`` is done via ``gnt-instance
338 modify``, it is possible to given different rights to users. For
339 example, owners of an instance could be allowed to start/stop it, but
340 not transition it out of the offline state.
342 Instance policies and specs
343 +++++++++++++++++++++++++++
345 In previous Ganeti versions, an instance creation request was not
346 limited on the minimum size and on the maximum size just by the cluster
347 resources. As such, any policy could be implemented only in third-party
348 clients (RAPI clients, or shell wrappers over ``gnt-*``
349 tools). Furthermore, calculating cluster capacity via ``hspace`` again
350 required external input with regards to instance sizes.
352 In order to improve these workflows and to allow for example better
353 per-node group differentiation, we introduced instance specs, which
356 - minimum instance disk size, disk count, memory size, cpu count
357 - maximum values for the above metrics
358 - and “standard” values (used in ``hspace`` to calculate the standard
361 The minimum/maximum values can be also customised at node-group level,
362 for example allowing more powerful hardware to support bigger instance
365 Beside the instance specs, there are a few other settings belonging to
366 the instance policy framework. It is possible now to customise, per
367 cluster and node-group:
369 - the list of allowed disk templates
370 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
371 - the maximum ratio of instance to spindles (see below for more
372 information) for local storage
374 All these together should allow all tools that talk to Ganeti to know
375 what are the ranges of allowed values for instances and the
376 over-subscription that is allowed.
378 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
379 instance configuration, and the physical CPU configuration from the
380 node. For the spindle ratios however, we didn't track before these
381 values, so new parameters have been added:
383 - a new node parameter ``spindle_count``, defaults to 1, customisable at
384 node group or node level
385 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
387 Note that spindles in this context doesn't need to mean actual
388 mechanical hard-drives; it's just a relative number for both the node
389 I/O capacity and instance I/O consumption.
391 Instance migration behaviour
392 ++++++++++++++++++++++++++++
394 While live-migration is in general desirable over failover, it is
395 possible that for some workloads it is actually worse, due to the
396 variable time of the “suspend” phase during live migration.
398 To allow the tools to work consistently over such instances (without
399 having to hard-code instance names), a new backend parameter
400 ``always_failover`` has been added to control the migration/failover
401 behaviour. When set to True, all migration requests for an instance will
402 instead fall-back to failover.
404 Instance memory ballooning
405 ++++++++++++++++++++++++++
407 Initial support for memory ballooning has been added. The memory for an
408 instance is no longer fixed (backend parameter ``memory``), but instead
409 can vary between minimum and maximum values (backend parameters
410 ``minmem`` and ``maxmem``). Currently we only change an instance's
413 - live migrating or failing over and instance and the target node
414 doesn't have enough memory
415 - user requests changing the memory via ``gnt-instance modify
421 In order to control the use of specific CPUs by instance, support for
422 controlling CPU pinning has been added for the Xen, HVM and LXC
423 hypervisors. This is controlled by a new hypervisor parameter
424 ``cpu_mask``; details about possible values for this are in the
425 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
426 VCPU-to-CPU mapping) form will work well only when all nodes in your
427 cluster have the same amount of CPUs.
432 Another area in which Ganeti was not customisable were the parameters
433 used for storage configuration, e.g. how many stripes to use for LVM,
434 DRBD resync configuration, etc.
436 To improve this area, we've added disks parameters, which are
437 customisable at cluster and node group level, and which allow to
438 specify various parameters for disks (DRBD has the most parameters
439 currently), for example:
441 - DRBD resync algorithm and parameters (e.g. speed)
442 - the default VG for meta-data volumes for DRBD
443 - number of stripes for LVM (plain disk template)
446 These parameters can be modified via ``gnt-cluster modify -D …`` and
447 ``gnt-group modify -D …``, and are used at either instance creation (in
448 case of LVM stripes, for example) or at disk “activation” time
451 Rados block device support
452 ++++++++++++++++++++++++++
454 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
455 denoted by the ``rbd`` disk template type. This is considered
456 experimental, feedback is welcome. For details on configuring it, see
457 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
462 The existing master IP functionality works well only in simple setups (a
463 single network shared by all nodes); however, if nodes belong to
464 different networks, then the ``/32`` setup and lack of routing
465 information is not enough.
467 To allow the master IP to function well in more complex cases, the
468 system was reworked as follows:
470 - a master IP netmask setting has been added
471 - the master IP activation/turn-down code was moved from the node daemon
473 - whether to run the Ganeti-supplied master IP script or a user-supplied
474 on is a ``gnt-cluster init`` setting
476 Details about the location of the standard and custom setup scripts are
477 in the man page :manpage:`gnt-cluster(8)`; for information about the
478 setup script protocol, look at the Ganeti-supplied script.
483 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
486 It is now possible to use TLS-protected connections, and when renewing
487 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
488 it is now possible to specify spice or spice CA certificates. Also, it
489 is possible to configure a password for SPICE sessions via the
490 hypervisor parameter ``spice_password_file``.
492 There are also new parameters to control the compression and streaming
493 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
494 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
495 for the spice parameters.
497 Lastly, it is now possible to see the SPICE connection information via
498 ``gnt-instance console``.
503 A new tool (``tools/ovfconverter``) has been added that supports
504 conversion between Ganeti and the `Open Virtualization Format
505 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
508 This relies on the ``qemu-img`` tool to convert the disk formats, so the
509 actual compatibility with other virtualization solutions depends on it.
514 The configuration query daemon (``ganeti-confd``) is now optional, and
515 has been rewritten in Haskell; whether to use the daemon at all, use the
516 Python (default) or the Haskell version is selectable at configure time
517 via the ``--enable-confd`` parameter, which can take one of the
518 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
519 daemon will result in a smaller footprint; for larger systems, we
520 welcome feedback on the Haskell version which might become the default
523 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
524 daemon running. The Python version doesn't implement the new call.
527 User interface changes
528 ~~~~~~~~~~~~~~~~~~~~~~
530 We have replaced the ``--disks`` option of ``gnt-instance
531 replace-disks`` with a more flexible ``--disk`` option, which allows
532 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
533 disk size and mode can be changed upon recreation (via ``gnt-instance
534 recreate-disks``, which accepts the same ``--disk`` option).
536 As many people are used to a ``show`` command, we have added that as an
537 alias to ``info`` on all ``gnt-*`` commands.
539 The ``gnt-instance grow-disk`` command has a new mode in which it can
540 accept the target size of the disk, instead of the delta; this can be
541 more safe since two runs in absolute mode will be idempotent, and
542 sometimes it's also easier to specify the desired size directly.
544 Also the handling of instances with regard to offline secondaries has
545 been improved. Instance operations should not fail because one of it's
546 secondary nodes is offline, even though it's safe to proceed.
548 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
549 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
550 minors to instance name.
555 RAPI coverage has improved, with (for example) new resources for
556 recreate-disks, node power-cycle, etc.
561 There is partial support for ``xl`` in the Xen hypervisor; feedback is
564 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
565 whether to still support Python 2.4 or move to Python 2.6 as minimum
568 Support for Fedora has been slightly improved; the provided example
569 init.d script should work better on it and the INSTALL file should
570 document the needed dependencies.
575 The deprecated ``QueryLocks`` LUXI request has been removed. Use
576 ``Query(what=QR_LOCK, ...)`` instead.
578 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
579 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
580 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
581 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
582 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
584 RAPI client: ``CertificateError`` now derives from
585 ``GanetiApiError``. This should make it more easy to handle Ganeti
588 Deprecation warnings due to PyCrypto/paramiko import in
589 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
590 make sure to run an up-to-date paramiko version, if you use this tool.
592 The QA scripts now depend on Python 2.5 or above (the main code base
593 still works with Python 2.4).
595 The configuration file (``config.data``) is now written without
596 indentation for performance reasons; if you want to edit it, it can be
597 re-formatted via ``tools/fmtjson``.
599 A number of bugs has been fixed in the cluster merge tool.
601 ``x509`` certification verification (used in import-export) has been
602 changed to allow the same clock skew as permitted by the cluster
603 verification. This will remove some rare but hard to diagnose errors in
610 *(Released Thu, 19 Jul 2012)*
612 Very few changes from rc4 to the final release, only bugfixes:
614 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
615 instance, fix CDROM booting for KVM instances)
616 - fixed node group modification of node parameters
617 - fixed issue in LUClusterVerifyGroup with multi-group clusters
618 - fixed generation of bash completion to ensure a stable ordering
625 *(Released Fri, 13 Jul 2012)*
627 Third release candidate for 2.6. The following changes were done from
630 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
631 - Fixed an inconsistency in the LUXI protocol with the provided
632 arguments (NOT backwards compatible)
633 - Fixed a bug with node groups ipolicy where ``min`` was greater than
634 the cluster ``std`` value
635 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
636 easier instance debugging on nodes (requires ``hconfd`` to work)
642 *(Released Tue, 03 Jul 2012)*
644 Second release candidate for 2.6. The following changes were done from
647 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
649 - Fixed a RAPI regression on missing beparams/memory
650 - Fixed redistribution of files on offline nodes
651 - Added possibility to run activate-disks even though secondaries are
652 offline. With this change it relaxes also the strictness on some other
653 commands which use activate disks internally:
654 * ``gnt-instance start|reboot|rename|backup|export``
655 - Made it possible to remove safely an instance if its secondaries are
657 - Made it possible to reinstall even though secondaries are offline
663 *(Released Mon, 25 Jun 2012)*
665 First release candidate for 2.6. The following changes were done from
668 - Fixed bugs with disk parameters and ``rbd`` templates as well as
670 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
671 behaviour. It supports now the modify operation
672 - ``hcheck`` implemented to analyze cluster health and possibility of
673 improving health by rebalance
674 - ``hbal`` has been improved in dealing with split instances
680 *(Released Mon, 11 Jun 2012)*
682 Second beta release of 2.6. The following changes were done from beta2
685 - Fixed ``daemon-util`` with non-root user models
686 - Fixed creation of plain instances with ``--no-wait-for-sync``
687 - Fix wrong iv_names when running ``cfgupgrade``
688 - Export more information in RAPI group queries
689 - Fixed bug when changing instance network interfaces
690 - Extended burnin to do NIC changes
691 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
692 - Changed default for DRBD barriers
693 - Fixed DRBD error reporting for syncer rate
694 - Verify the options on disk parameters
696 And of course various fixes to documentation and improved unittests and
703 *(Released Wed, 23 May 2012)*
705 First beta release of 2.6. The following changes were done from beta1 to
708 - integrated patch for distributions without ``start-stop-daemon``
709 - adapted example init.d script to work on Fedora
710 - fixed log handling in Haskell daemons
711 - adapted checks in the watcher for pycurl linked against libnss
712 - add partial support for ``xl`` instead of ``xm`` for Xen
713 - fixed a type issue in cluster verification
714 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
717 Plus integrated fixes from the 2.5 branch:
719 - fixed ``kvm-ifup`` to use ``/bin/bash``
720 - fixed parallel build failures
721 - KVM live migration when using a custom keymap
727 *(Released Tue, 24 Jul 2012)*
729 A small bugfix release, with no new features:
731 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
732 different default shell (e.g. Debian, Ubuntu)
733 - fixed KVM startup and live migration with a custom keymap (fixes Issue
734 243 and Debian bug #650664)
735 - fixed compatibility with KVM versions that don't support multiple boot
736 devices (fixes Issue 230 and Debian bug #624256)
738 Additionally, a few fixes were done to the build system (fixed parallel
739 build failures) and to the unittests (fixed race condition in test for
740 FileID functions, and the default enable/disable mode for QA test is now
747 *(Released Fri, 11 May 2012)*
749 A small bugfix release.
751 The main issues solved are on the topic of compatibility with newer LVM
754 - fixed parsing of ``lv_attr`` field
755 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
756 the ``--force`` flag is needed
758 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
759 to accept kernel 3.x too (was hardcoded to 2.6.*).
761 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
762 of hook results and that also made display of other errors suboptimal
763 was fixed; the code behaves now like 2.4 and earlier.
765 Another change in 2.5, the cleanup of the OS scripts environment, is too
766 aggressive: it removed even the ``PATH`` variable, which requires the OS
767 scripts to *always* need to export it. Since this is a bit too strict,
768 we now export a minimal PATH, the same that we export for hooks.
770 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
771 integrated into this release.
773 Finally, a few other miscellaneous changes were done (no new features,
774 just small improvements):
776 - Fix ``gnt-group --help`` display
777 - Fix hardcoded Xen kernel path
778 - Fix grow-disk handling of invalid units
779 - Update synopsis for ``gnt-cluster repair-disk-sizes``
780 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
786 *(Released Thu, 12 Apr 2012)*
788 Incompatible/important changes and bugfixes
789 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
791 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
792 resource's ``ip_check`` parameter changed from ``True`` to ``False``
793 to match the underlying LUXI interface.
794 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
795 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
796 not maintain backwards-compatibility as the underlying operation
797 changed in an incompatible way. The RAPI client can talk to old
798 servers, but it needs to be told so as the return value changed.
799 - When creating file-based instances via RAPI, the ``file_driver``
800 parameter no longer defaults to ``loop`` and must be specified.
801 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
803 - Support for the undocumented and deprecated RAPI instance creation
804 request format version 0 has been dropped. Use version 1, supported
805 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
806 - Pyparsing 1.4.6 or above is required, see :doc:`installation
807 documentation <install>`.
808 - The "cluster-verify" hooks are now executed per group by the
809 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
810 if you just run ``gnt-cluster verify``, which generates one opcode per
812 - The environment as passed to the OS scripts is cleared, and thus no
813 environment variables defined in the node daemon's environment will be
814 inherited by the scripts.
815 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
817 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
818 support operations involving multiple node groups.
819 - Offline nodes are ignored when failing over an instance.
820 - Support for KVM version 1.0, which changed the version reporting format
822 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
824 - ``Makefile`` is now compatible with Automake 1.11.2
825 - Includes all bugfixes made in the 2.4 series
830 - The ganeti-htools project has been merged into the ganeti-core source
831 tree and will be built as part of Ganeti (see :doc:`install-quick`).
832 - Implemented support for :doc:`shared storage <design-shared-storage>`.
833 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
834 GPT-style partition tables (requires `parted
835 <http://www.gnu.org/s/parted/>`_).
836 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
837 - Allowed adding tags on instance creation.
838 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
839 - Allow instances to be started in a paused state, enabling the user to
840 see the complete console output on boot using the console.
841 - Added new hypervisor flag to control default reboot behaviour
842 (``reboot_behavior``).
843 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
844 - Improved out-of-band management support:
846 - Added ``gnt-node health`` command reporting the health status of
848 - Added ``gnt-node power`` command to manage power status of nodes.
849 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
851 - Instance migration can fall back to failover if instance is not
853 - Filters can be used when listing nodes, instances, groups and locks;
854 see :manpage:`ganeti(7)` manpage.
855 - Added post-execution status as variables to :doc:`hooks <hooks>`
857 - Instance tags are exported/imported together with the instance.
858 - When given an explicit job ID, ``gnt-job info`` will work for archived
860 - Jobs can define dependencies on other jobs (not yet supported via
861 RAPI or command line, but used by internal commands and usable via
864 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
867 - Instance failover is now available as a RAPI resource
868 (``/2/instances/[instance_name]/failover``).
869 - ``gnt-instance info`` defaults to static information if primary node
871 - Opcodes have a new ``comment`` attribute.
872 - Added basic SPICE support to KVM hypervisor.
873 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
875 Node group improvements
876 ~~~~~~~~~~~~~~~~~~~~~~~
878 - ``gnt-cluster verify`` has been modified to check groups separately,
879 thereby improving performance.
880 - Node group support has been added to ``gnt-cluster verify-disks``,
881 which now operates per node group.
882 - Watcher has been changed to work better with node groups.
884 - One process and state file per node group.
885 - Slow watcher in one group doesn't block other group's watcher.
887 - Added new command, ``gnt-group evacuate``, to move all instances in a
888 node group to other groups.
889 - Added ``gnt-instance change-group`` to move an instance to another
891 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
892 per-group operations.
893 - Node groups can be tagged.
894 - Some operations switch from an exclusive to a shared lock as soon as
896 - Instance's primary and secondary nodes' groups are now available as
897 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
898 and ``snodes.group.uuid``).
903 - Numerous updates to documentation and manpages.
905 - :doc:`RAPI <rapi>` documentation now has detailed parameter
907 - Some opcode/job results are now also documented, see :doc:`RAPI
910 - A lockset's internal lock is now also visible in lock monitor.
911 - Log messages from job queue workers now contain information about the
912 opcode they're processing.
913 - ``gnt-instance console`` no longer requires the instance lock.
914 - A short delay when waiting for job changes reduces the number of LUXI
915 requests significantly.
916 - DRBD metadata volumes are overwritten with zeros during disk creation.
917 - Out-of-band commands no longer acquire the cluster lock in exclusive
919 - ``devel/upload`` now uses correct permissions for directories.
925 *(Released Fri, 23 Mar 2012)*
927 This was the sixth release candidate of the 2.5 series.
933 *(Released Mon, 9 Jan 2012)*
935 This was the fifth release candidate of the 2.5 series.
941 *(Released Thu, 27 Oct 2011)*
943 This was the fourth release candidate of the 2.5 series.
949 *(Released Wed, 26 Oct 2011)*
951 This was the third release candidate of the 2.5 series.
957 *(Released Tue, 18 Oct 2011)*
959 This was the second release candidate of the 2.5 series.
965 *(Released Tue, 4 Oct 2011)*
967 This was the first release candidate of the 2.5 series.
973 *(Released Wed, 31 Aug 2011)*
975 This was the third beta release of the 2.5 series.
981 *(Released Mon, 22 Aug 2011)*
983 This was the second beta release of the 2.5 series.
989 *(Released Fri, 12 Aug 2011)*
991 This was the first beta release of the 2.5 series.
997 *(Released Thu, 27 Oct 2011)*
999 - Fixed bug when parsing command line parameter values ending in
1001 - Fixed assertion error after unclean master shutdown
1002 - Disable HTTP client pool for RPC, significantly reducing memory usage
1004 - Fixed queue archive creation with wrong permissions
1010 *(Released Tue, 23 Aug 2011)*
1014 - Fixed documentation for importing with ``--src-dir`` option
1015 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1016 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1022 *(Released Fri, 5 Aug 2011)*
1024 Many bug-fixes and a few small features:
1026 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1027 issues when you tried to add an instance with two MAC addresses in one
1029 - KVM: fixed per-instance stored UID value
1030 - KVM: configure bridged NICs at migration start
1031 - KVM: Fix a bug where instance will not start with never KVM versions
1033 - Added OS search path to ``gnt-cluster info``
1034 - Fixed an issue with ``file_storage_dir`` where you were forced to
1035 provide an absolute path, but the documentation states it is a
1036 relative path, the documentation was right
1037 - Added a new parameter to instance stop/start called ``--no-remember``
1038 that will make the state change to not be remembered
1039 - Implemented ``no_remember`` at RAPI level
1040 - Improved the documentation
1041 - Node evacuation: don't call IAllocator if node is already empty
1042 - Fixed bug in DRBD8 replace disks on current nodes
1043 - Fixed bug in recreate-disks for DRBD instances
1044 - Moved assertion checking locks in ``gnt-instance replace-disks``
1045 causing it to abort with not owning the right locks for some situation
1046 - Job queue: Fixed potential race condition when cancelling queued jobs
1047 - Fixed off-by-one bug in job serial generation
1048 - ``gnt-node volumes``: Fix instance names
1049 - Fixed aliases in bash completion
1050 - Fixed a bug in reopening log files after being sent a SIGHUP
1051 - Added a flag to burnin to allow specifying VCPU count
1052 - Bugfixes to non-root Ganeti configuration
1058 *(Released Thu, 12 May 2011)*
1060 Many bug-fixes and a few new small features:
1062 - Fixed a bug related to log opening failures
1063 - Fixed a bug in instance listing with orphan instances
1064 - Fixed a bug which prevented resetting the cluster-level node parameter
1065 ``oob_program`` to the default
1066 - Many fixes related to the ``cluster-merge`` tool
1067 - Fixed a race condition in the lock monitor, which caused failures
1068 during (at least) creation of many instances in parallel
1069 - Improved output for gnt-job info
1070 - Removed the quiet flag on some ssh calls which prevented debugging
1072 - Improved the N+1 failure messages in cluster verify by actually
1073 showing the memory values (needed and available)
1074 - Increased lock attempt timeouts so that when executing long operations
1075 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1076 too early and thus prevent the use of the 'fair' mechanism
1077 - Changed instance query data (``gnt-instance info``) to not acquire
1078 locks unless needed, thus allowing its use on locked instance if only
1079 static information is asked for
1080 - Improved behaviour with filesystems that do not support rename on an
1082 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1083 kept locks on all nodes during the wipe, which is unneeded
1084 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1085 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1086 10GiB) which caused the wipe to fail right at the end in some cases
1087 - Fixed master IP activation when doing master failover with no-voting
1088 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1089 the master node itself
1090 - Fixed potential data-loss in under disk full conditions, where Ganeti
1091 wouldn't check correctly the return code and would consider
1092 partially-written files 'correct'
1093 - Fixed bug related to multiple VGs and DRBD disk replacing
1094 - Added new disk parameter ``metavg`` that allows placement of the meta
1095 device for DRBD in a different volume group
1096 - Fixed error handling in the node daemon when the system libc doesn't
1097 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1098 - Fixed lock release during replace-disks, which kept cluster-wide locks
1099 when doing disk replaces with an iallocator script
1100 - Added check for missing bridges in cluster verify
1101 - Handle EPIPE errors while writing to the terminal better, so that
1102 piping the output to e.g. ``less`` doesn't cause a backtrace
1103 - Fixed rare case where a ^C during Luxi calls could have been
1104 interpreted as server errors, instead of simply terminating
1105 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1107 - Added a few more parameters to the KVM hypervisor, allowing a second
1108 CDROM, custom disk type for CDROMs and a floppy image
1109 - Removed redundant message in instance rename when the name is given
1111 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1112 disks on new nodes, allowing recreation when the original instance
1113 nodes are completely gone
1114 - Added option when converting disk templates to DRBD to skip waiting
1115 for the resync, in order to make the instance available sooner
1116 - Added two new variables to the OS scripts environment (containing the
1118 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1119 to allow use of ``pvgrub`` as bootloader
1120 - Changed the instance memory modifications to only check out-of-memory
1121 conditions on memory increases, and turned the secondary node warnings
1122 into errors (they can still be overridden via ``--force``)
1123 - Fixed the handling of a corner case when the Python installation gets
1124 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1125 to execute a command that doesn't exist
1126 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1127 node of the instance returned failures during instance shutdown; this
1128 adds the option ``--ignore-consistency`` to gnt-instance move
1130 And as usual, various improvements to the error messages, documentation
1137 *(Released Wed, 09 Mar 2011)*
1139 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1140 the RAPI users file if run twice (even with ``--dry-run``).
1142 The release fixes that bug (nothing else changed).
1148 *(Released Mon, 07 Mar 2011)*
1150 Final 2.4.0 release. Just a few small fixes:
1152 - Fixed RAPI node evacuate
1153 - Fixed the kvm-ifup script
1154 - Fixed internal error handling for special job cases
1155 - Updated man page to specify the escaping feature for options
1161 *(Released Mon, 28 Feb 2011)*
1163 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1164 that this feature wiped the disks of the wrong instance, leading to loss
1169 - Fixed title of query field containing instance name
1170 - Expanded the glossary in the documentation
1171 - Fixed one unittest (internal issue)
1177 *(Released Mon, 21 Feb 2011)*
1179 A number of bug fixes plus just a couple functionality changes.
1181 On the user-visible side, the ``gnt-* list`` command output has changed
1182 with respect to "special" field states. The current rc1 style of display
1183 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1184 the default output mode special fields states are displayed as follows:
1186 - Offline resource: ``*``
1187 - Unavailable/not applicable: ``-``
1188 - Data missing (RPC failure): ``?``
1189 - Unknown field: ``??``
1191 Another user-visible change is the addition of ``--force-join`` to
1196 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1197 - Fixed regression in RAPI/instance reinstall where all parameters were
1198 required (instead of optional)
1199 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1200 - Fixed iallocator usage (offline nodes were not considered offline)
1201 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1202 - Fixed hypervisor and OS parameter validation with respect to
1203 non-vm_capable nodes
1204 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1206 - Fixed ``tools/listrunner`` with respect to agent-based usage
1212 *(Released Fri, 4 Feb 2011)*
1214 Many changes and fixes since the beta1 release. While there were some
1215 internal changes, the code has been mostly stabilised for the RC
1218 Note: the dumb allocator was removed in this release, as it was not kept
1219 up-to-date with the IAllocator protocol changes. It is recommended to
1220 use the ``hail`` command from the ganeti-htools package.
1222 Note: the 2.4 and up versions of Ganeti are not compatible with the
1223 0.2.x branch of ganeti-htools. You need to upgrade to
1224 ganeti-htools-0.3.0 (or later).
1226 Regressions fixed from 2.3
1227 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1229 - Fixed the ``gnt-cluster verify-disks`` command
1230 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1232 - Fixed disk adoption breakage
1233 - Fixed wrong headers in instance listing for field aliases
1238 - Fixed corner case in KVM handling of NICs
1239 - Fixed many cases of wrong handling of non-vm_capable nodes
1240 - Fixed a bug where a missing instance symlink was not possible to
1241 recreate with any ``gnt-*`` command (now ``gnt-instance
1242 activate-disks`` does it)
1243 - Fixed the volume group name as reported by ``gnt-cluster
1245 - Increased timeouts for the import-export code, hopefully leading to
1246 fewer aborts due network or instance timeouts
1247 - Fixed bug in ``gnt-node list-storage``
1248 - Fixed bug where not all daemons were started on cluster
1249 initialisation, but only at the first watcher run
1250 - Fixed many bugs in the OOB implementation
1251 - Fixed watcher behaviour in presence of instances with offline
1253 - Fixed instance list output for instances running on the wrong node
1254 - a few fixes to the cluster-merge tool, but it still cannot merge
1255 multi-node groups (currently it is not recommended to use this tool)
1261 - Improved network configuration for the KVM hypervisor
1262 - Added e1000 as a supported NIC for Xen-HVM
1263 - Improved the lvmstrap tool to also be able to use partitions, as
1264 opposed to full disks
1265 - Improved speed of disk wiping (the cluster parameter
1266 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1267 of instance creations
1268 - Added documentation for the OS parameters
1269 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1270 hypervisor is not responding
1271 - Added display of blacklisted and hidden OS information in
1272 ``gnt-cluster info``
1273 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1274 NIC and node parameters, which might create problems with currently
1275 invalid (but undetected) configuration files, but prevents validation
1276 failures when unrelated parameters are modified
1277 - Changed cluster initialisation to wait for the master daemon to become
1279 - Expanded the RAPI interface:
1281 - Added config redistribution resource
1282 - Added activation/deactivation of instance disks
1283 - Added export of console information
1285 - Implemented log file reopening on SIGHUP, which allows using
1286 logrotate(8) for the Ganeti log files
1287 - Added a basic OOB helper script as an example
1293 *(Released Fri, 14 Jan 2011)*
1298 - Fixed timezone issues when formatting timestamps
1299 - Added support for node groups, available via ``gnt-group`` and other
1301 - Added out-of-band framework and management, see :doc:`design
1302 document <design-oob>`
1303 - Removed support for roman numbers from ``gnt-node list`` and
1304 ``gnt-instance list``.
1305 - Allowed modification of master network interface via ``gnt-cluster
1306 modify --master-netdev``
1307 - Accept offline secondaries while shutting down instance disks
1308 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1309 - Added support for multiple LVM volume groups
1310 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1312 - Added commands to list available fields:
1314 - ``gnt-node list-fields``
1315 - ``gnt-group list-fields``
1316 - ``gnt-instance list-fields``
1318 - Updated documentation and man pages
1323 - Moved ``rapi_users`` file into separate directory, now named
1324 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1326 - Added new tool for running commands on many machines,
1327 ``tools/ganeti-listrunner``
1328 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1329 improving the ``gnt-instance console`` output
1330 - Allowed customisation of disk index separator at ``configure`` time
1331 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1332 - Added support for non-partitioned md disks in ``lvmstrap``
1333 - Added script to gracefully power off KVM instances
1334 - Split ``utils`` module into smaller parts
1335 - Changed query operations to return more detailed information, e.g.
1336 whether an information is unavailable due to an offline node. To use
1337 this new functionality, the LUXI call ``Query`` must be used. Field
1338 information is now stored by the master daemon and can be retrieved
1339 using ``QueryFields``. Instances, nodes and groups can also be queried
1340 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1341 exposed via RAPI). The following commands make use of this
1342 infrastructure change:
1344 - ``gnt-group list``
1345 - ``gnt-group list-fields``
1347 - ``gnt-node list-fields``
1348 - ``gnt-instance list``
1349 - ``gnt-instance list-fields``
1350 - ``gnt-debug locks``
1355 - New RAPI resources (see :doc:`rapi`):
1359 - ``/2/groups/[group_name]``
1360 - ``/2/groups/[group_name]/assign-nodes``
1361 - ``/2/groups/[group_name]/modify``
1362 - ``/2/groups/[group_name]/rename``
1363 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1367 - Implemented ``no_install`` for instance creation
1368 - Implemented OS parameters for instance reinstallation, allowing
1369 use of special settings on reinstallation (e.g. for preserving data)
1374 - Added IPv6 support in import/export
1375 - Pause DRBD synchronization while wiping disks on instance creation
1376 - Updated unittests and QA scripts
1377 - Improved network parameters passed to KVM
1378 - Converted man pages from docbook to reStructuredText
1384 *(Released Mon, 20 Dec 2010)*
1386 Released version 2.3.1~rc1 without any changes.
1392 *(Released Wed, 1 Dec 2010)*
1394 - impexpd: Disable OpenSSL compression in socat if possible (backport
1395 from master, commit e90739d625b, see :doc:`installation guide
1396 <install-quick>` for details)
1397 - Changed unittest coverage report to exclude test scripts
1398 - Added script to check version format
1404 *(Released Wed, 1 Dec 2010)*
1406 Released version 2.3.0~rc1 without any changes.
1412 *(Released Fri, 19 Nov 2010)*
1414 A number of bugfixes and documentation updates:
1416 - Update ganeti-os-interface documentation
1417 - Fixed a bug related to duplicate MACs or similar items which should be
1419 - Fix breakage in OS state modify
1420 - Reinstall instance: disallow offline secondaries (fixes bug related to
1421 OS changing but reinstall failing)
1422 - plus all the other fixes between 2.2.1 and 2.2.2
1428 *(Released Tue, 2 Nov 2010)*
1430 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1431 - Fixed master failover race with watcher
1432 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1434 - Accept previously stopped instance for export with instance removal
1435 - Simplify and extend the environment variables for instance OS scripts
1436 - Added new node flags, ``master_capable`` and ``vm_capable``
1437 - Added optional instance disk wiping prior during allocation. This is a
1438 cluster-wide option and can be set/modified using
1439 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1440 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1441 :doc:`install-quick`
1442 - Added a new watcher option (``--ignore-pause``)
1443 - Added option to ignore offline node on instance start/stop
1444 (``--ignore-offline``)
1445 - Allow overriding OS parameters with ``gnt-instance reinstall``
1446 - Added ability to change node's secondary IP address using ``gnt-node
1448 - Implemented privilege separation for all daemons except
1449 ``ganeti-noded``, see ``configure`` options
1450 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1452 - Implemented job priorities (see ``ganeti(7)`` manpage)
1453 - Ignore failures while shutting down instances during failover from
1455 - Exit daemon's bootstrap process only once daemon is ready
1456 - Export more information via ``LUInstanceQuery``/remote API
1457 - Improved documentation, QA and unittests
1458 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1459 restart if the file was created or changed
1460 - Added LUXI protocol version sent with each request and response,
1461 allowing detection of server/client mismatches
1462 - Moved the Python scripts among gnt-* and ganeti-* into modules
1463 - Moved all code related to setting up SSH to an external script,
1465 - Infrastructure changes for node group support in future versions
1471 *(Released Fri, 19 Nov 2010)*
1473 A few small bugs fixed, and some improvements to the build system:
1475 - Fix documentation regarding conversion to drbd
1476 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1478 - Fix error handling in node modify with multiple changes
1479 - Allow remote imports without checked names
1485 *(Released Tue, 19 Oct 2010)*
1487 - Disable SSL session ID cache in RPC client
1493 *(Released Thu, 14 Oct 2010)*
1495 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1496 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1498 - Fix problems with interaction between Curl and Python's HTTP server,
1499 resulting in increased speed in many RPC calls
1500 - Improve our release script to prevent breakage with older aclocal and
1507 *(Released Thu, 7 Oct 2010)*
1509 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1510 value retrieved from master
1511 - Added support for blacklisted or hidden OS definitions
1512 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1513 - Added support for -mem-path in KVM hypervisor abstraction layer
1514 - Allow overriding instance parameters in tool for inter-cluster
1515 instance moves (``tools/move-instance``)
1516 - Improved opcode summaries (e.g. in ``gnt-job list``)
1517 - Improve consistency of OS listing by sorting it
1518 - Documentation updates
1524 *(Released Fri, 8 Oct 2010)*
1526 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1532 *(Released Mon, 4 Oct 2010)*
1534 - Fixed regression in ``gnt-instance rename``
1540 *(Released Wed, 22 Sep 2010)*
1542 - Fixed OS_VARIANT variable for OS scripts
1543 - Fixed cluster tag operations via RAPI
1544 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1545 - Disabled RAPI CA checks in watcher
1551 *(Released Mon, 23 Aug 2010)*
1553 - Support DRBD versions of the format "a.b.c.d"
1555 - Re-introduce support for usage from multiple threads in RAPI client
1556 - Instance renames and modify via RAPI
1557 - Work around race condition between processing and archival in job
1559 - Mark opcodes following failed one as failed, too
1560 - Job field ``lock_status`` was removed due to difficulties making it
1561 work with the changed job queue in Ganeti 2.2; a better way to monitor
1562 locks is expected for a later 2.2.x release
1563 - Fixed dry-run behaviour with many commands
1564 - Support ``ssh-agent`` again when adding nodes
1565 - Many additional bugfixes
1571 *(Released Fri, 30 Jul 2010)*
1573 Important change: the internal RPC mechanism between Ganeti nodes has
1574 changed from using a home-grown http library (based on the Python base
1575 libraries) to use the PycURL library. This requires that PycURL is
1576 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1577 against GnuTLS by default. cURL's support for GnuTLS had known issues
1578 before cURL 7.21.0 and we recommend using the latest cURL release or
1579 linking against OpenSSL. Most other distributions already link PycURL
1580 and cURL against OpenSSL. The command::
1582 python -c 'import pycurl; print pycurl.version'
1584 can be used to determine the libraries PycURL and cURL are linked
1587 Other significant changes:
1589 - Rewrote much of the internals of the job queue, in order to achieve
1590 better parallelism; this decouples job query operations from the job
1591 processing, and it should allow much nicer behaviour of the master
1592 daemon under load, and it also has uncovered some long-standing bugs
1593 related to the job serialisation (now fixed)
1594 - Added a default iallocator setting to the cluster parameters,
1595 eliminating the need to always pass nodes or an iallocator for
1596 operations that require selection of new node(s)
1597 - Added experimental support for the LXC virtualization method
1598 - Added support for OS parameters, which allows the installation of
1599 instances to pass parameter to OS scripts in order to customise the
1601 - Added a hypervisor parameter controlling the migration type (live or
1602 non-live), since hypervisors have various levels of reliability; this
1603 has renamed the 'live' parameter to 'mode'
1604 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1605 logical volumes, meaning that cluster verify will ignore them and not
1606 flag their presence as errors
1607 - The watcher will now reset the error count for failed instances after
1608 8 hours, thus allowing self-healing if the problem that caused the
1609 instances to be down/fail to start has cleared in the meantime
1610 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1611 check for, and warn, if the drbd module parameter ``usermode_helper``
1612 is not consistent with the cluster-wide setting; this is needed to
1613 make diagnose easier of failed drbd creations
1614 - Started adding base IPv6 support, but this is not yet
1615 enabled/available for use
1616 - Rename operations (cluster, instance) will now return the new name,
1617 which is especially useful if a short name was passed in
1618 - Added support for instance migration in RAPI
1619 - Added a tool to pre-configure nodes for the SSH setup, before joining
1620 them to the cluster; this will allow in the future a simplified model
1621 for node joining (but not yet fully enabled in 2.2); this needs the
1622 paramiko python library
1623 - Fixed handling of name-resolving errors
1624 - Fixed consistency of job results on the error path
1625 - Fixed master-failover race condition when executed multiple times in
1627 - Fixed many bugs related to the job queue (mostly introduced during the
1628 2.2 development cycle, so not all are impacting 2.1)
1629 - Fixed instance migration with missing disk symlinks
1630 - Fixed handling of unknown jobs in ``gnt-job archive``
1631 - And many other small fixes/improvements
1635 - Enhanced both the unittest and the QA coverage
1636 - Switched the opcode validation to a generic model, and extended the
1637 validation to all opcode parameters
1638 - Changed more parts of the code that write shell scripts to use the
1640 - Switched the master daemon to use the asyncore library for the Luxi
1647 *(Released Thu, 17 Jun 2010)*
1649 - Added tool (``move-instance``) and infrastructure to move instances
1650 between separate clusters (see :doc:`separate documentation
1651 <move-instance>` and :doc:`design document <design-2.2>`)
1652 - Added per-request RPC timeout
1653 - RAPI now requires a Content-Type header for requests with a body (e.g.
1654 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1655 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1656 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1658 - Implemented initial support for running Ganeti daemons as separate
1659 users, see configure-time flags ``--with-user-prefix`` and
1660 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1662 - Instances can be removed after export (``gnt-backup export
1663 --remove-instance``)
1664 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1665 key (instead of 1024 bit)
1666 - Added new cluster configuration file for cluster domain secret
1667 - Import/export now use SSL instead of SSH
1668 - Added support for showing estimated time when exporting an instance,
1669 see the ``ganeti-os-interface(7)`` manpage and look for
1676 *(Released Tue, 16 Nov 2010)*
1678 Some more bugfixes. Unless critical bugs occur, this will be the last
1681 - Fix case of MAC special-values
1682 - Fix mac checker regex
1683 - backend: Fix typo causing "out of range" error
1684 - Add missing --units in gnt-instance list man page
1690 *(Released Tue, 24 Aug 2010)*
1693 - Don't ignore secondary node silently on non-mirrored disk templates
1695 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1696 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1697 - Properly document the usb_mouse parameter
1698 - Fix path in ganeti-rapi(8) (issue 116)
1699 - Adjust error message when the ganeti user's .ssh directory is
1701 - Add same-node-check when changing the disk template to drbd
1707 *(Released Fri, 16 Jul 2010)*
1710 - Add an option to only select some reboot types during qa/burnin.
1711 (on some hypervisors consequent reboots are not supported)
1712 - Fix infrequent race condition in master failover. Sometimes the old
1713 master ip address would be still detected as up for a short time
1714 after it was removed, causing failover to fail.
1715 - Decrease mlockall warnings when the ctypes module is missing. On
1716 Python 2.4 we support running even if no ctypes module is installed,
1717 but we were too verbose about this issue.
1718 - Fix building on old distributions, on which man doesn't have a
1720 - Fix RAPI not to ignore the MAC address on instance creation
1721 - Implement the old instance creation format in the RAPI client.
1727 *(Released Thu, 01 Jul 2010)*
1729 A small bugfix release:
1730 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1731 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1732 a non-existing option
1733 - Fix instance create: the --force-variant option was ignored
1734 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1735 - Fix modify node storage with non-FQDN arguments
1736 - Fix RAPI client to authenticate under Python 2.6 when used
1737 for more than 5 requests needing authentication
1738 - Fix gnt-instance modify -t (storage) giving a wrong error message
1739 when converting a non-shutdown drbd instance to plain
1745 *(Released Fri, 18 Jun 2010)*
1747 A small bugfix release:
1749 - Fix live migration of KVM instances started with older Ganeti
1750 versions which had fewer hypervisor parameters
1751 - Fix gnt-instance grow-disk on down instances
1752 - Fix an error-reporting bug during instance migration
1753 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1754 silently ignoring broken ones
1755 - Fix an RPC error reporting bug affecting, for example, RAPI client
1757 - Fix bug triggered by different API version os-es on different nodes
1758 - Fix a bug in instance startup with custom hvparams: OS level
1759 parameters would fail to be applied.
1760 - Fix the RAPI client under Python 2.6 (but more work is needed to
1761 make it work completely well with OpenSSL)
1762 - Fix handling of errors when resolving names from DNS
1768 *(Released Thu, 3 Jun 2010)*
1770 A medium sized development cycle. Some new features, and some
1771 fixes/small improvements/cleanups.
1773 Significant features
1774 ~~~~~~~~~~~~~~~~~~~~
1776 The node deamon now tries to mlock itself into memory, unless the
1777 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1778 its logs, and falls back to console logging. This allows emergency
1779 features such as ``gnt-node powercycle`` to work even in the event of a
1780 broken node disk (tested offlining the disk hosting the node's
1781 filesystem and dropping its memory caches; don't try this at home)
1783 KVM: add vhost-net acceleration support. It can be tested with a new
1784 enough version of the kernel and of qemu-kvm.
1786 KVM: Add instance chrooting feature. If you use privilege dropping for
1787 your VMs you can also now force them to chroot to an empty directory,
1788 before starting the emulated guest.
1790 KVM: Add maximum migration bandwith and maximum downtime tweaking
1791 support (requires a new-enough version of qemu-kvm).
1793 Cluster verify will now warn if the master node doesn't have the master
1794 ip configured on it.
1796 Add a new (incompatible) instance creation request format to RAPI which
1797 supports all parameters (previously only a subset was supported, and it
1798 wasn't possible to extend the old format to accomodate all the new
1799 features. The old format is still supported, and a client can check for
1800 this feature, before using it, by checking for its presence in the
1801 ``features`` RAPI resource.
1803 Now with ancient latin support. Try it passing the ``--roman`` option to
1804 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1805 (requires the python-roman module to be installed, in order to work).
1810 As usual many internal code refactorings, documentation updates, and
1813 - Lots of improvements and cleanups to the experimental Remote API
1814 (RAPI) client library.
1815 - A new unit test suite for the core daemon libraries.
1816 - A fix to creating missing directories makes sure the umask is not
1817 applied anymore. This enforces the same directory permissions
1819 - Better handling terminating daemons with ctrl+c (used when running
1820 them in debugging mode).
1821 - Fix a race condition in live migrating a KVM instance, when stat()
1822 on the old proc status file returned EINVAL, which is an unexpected
1824 - Fixed manpage checking with newer man and utf-8 charachters. But now
1825 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1831 *(Released Fri, 7 May 2010)*
1833 Fix a bug which prevented untagged KVM instances from starting.
1839 *(Released Fri, 7 May 2010)*
1841 Another release with a long development cycle, during which many
1842 different features were added.
1844 Significant features
1845 ~~~~~~~~~~~~~~~~~~~~
1847 The KVM hypervisor now can run the individual instances as non-root, to
1848 reduce the impact of a VM being hijacked due to bugs in the
1849 hypervisor. It is possible to run all instances as a single (non-root)
1850 user, to manually specify a user for each instance, or to dynamically
1851 allocate a user out of a cluster-wide pool to each instance, with the
1852 guarantee that no two instances will run under the same user ID on any
1855 An experimental RAPI client library, that can be used standalone
1856 (without the other Ganeti libraries), is provided in the source tree as
1857 ``lib/rapi/client.py``. Note this client might change its interface in
1858 the future, as we iterate on its capabilities.
1860 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1861 replace the cluster's certificates and crypto keys. This might help in
1862 case they have been compromised, or have simply expired.
1864 A new disk option for instance creation has been added that allows one
1865 to "adopt" currently existing logical volumes, with data
1866 preservation. This should allow easier migration to Ganeti from
1867 unmanaged (or managed via other software) instances.
1869 Another disk improvement is the possibility to convert between redundant
1870 (DRBD) and plain (LVM) disk configuration for an instance. This should
1871 allow better scalability (starting with one node and growing the
1872 cluster, or shrinking a two-node cluster to one node).
1874 A new feature that could help with automated node failovers has been
1875 implemented: if a node sees itself as offline (by querying the master
1876 candidates), it will try to shutdown (hard) all instances and any active
1877 DRBD devices. This reduces the risk of duplicate instances if an
1878 external script automatically failovers the instances on such nodes. To
1879 enable this, the cluster parameter ``maintain_node_health`` should be
1880 enabled; in the future this option (per the name) will enable other
1881 automatic maintenance features.
1883 Instance export/import now will reuse the original instance
1884 specifications for all parameters; that means exporting an instance,
1885 deleting it and the importing it back should give an almost identical
1886 instance. Note that the default import behaviour has changed from
1887 before, where it created only one NIC; now it recreates the original
1890 Cluster verify has added a few new checks: SSL certificates validity,
1891 /etc/hosts consistency across the cluster, etc.
1896 As usual, many internal changes were done, documentation fixes,
1899 - Fixed cluster initialization with disabled cluster storage (regression
1900 introduced in 2.1.1)
1901 - File-based storage supports growing the disks
1902 - Fixed behaviour of node role changes
1903 - Fixed cluster verify for some corner cases, plus a general rewrite of
1904 cluster verify to allow future extension with more checks
1905 - Fixed log spamming by watcher and node daemon (regression introduced
1907 - Fixed possible validation issues when changing the list of enabled
1909 - Fixed cleanup of /etc/hosts during node removal
1910 - Fixed RAPI response for invalid methods
1911 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1912 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1913 from ide disks, etc.)
1914 - Allow OS changes without re-installation (to record a changed OS
1915 outside of Ganeti, or to allow OS renames)
1916 - Allow instance creation without OS installation (useful for example if
1917 the OS will be installed manually, or restored from a backup not in
1919 - Implemented option to make cluster ``copyfile`` use the replication
1921 - Added list of enabled hypervisors to ssconf (possibly useful for
1923 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1925 - A partial form of node re-IP is possible via node readd, which now
1926 allows changed node primary IP
1927 - Command line utilities now show an informational message if the job is
1929 - The logs of the master daemon now show the PID/UID/GID of the
1936 *(Released Fri, 12 Mar 2010)*
1938 During the 2.1.0 long release candidate cycle, a lot of improvements and
1939 changes have accumulated with were released later as 2.1.1.
1944 The node evacuate command (``gnt-node evacuate``) was significantly
1945 rewritten, and as such the IAllocator protocol was changed - a new
1946 request type has been added. This unfortunate change during a stable
1947 series is designed to improve performance of node evacuations; on
1948 clusters with more than about five nodes and which are well-balanced,
1949 evacuation should proceed in parallel for all instances of the node
1950 being evacuated. As such, any existing IAllocator scripts need to be
1951 updated, otherwise the above command will fail due to the unknown
1952 request. The provided "dumb" allocator has not been updated; but the
1953 ganeti-htools package supports the new protocol since version 0.2.4.
1955 Another important change is increased validation of node and instance
1956 names. This might create problems in special cases, if invalid host
1957 names are being used.
1959 Also, a new layer of hypervisor parameters has been added, that sits at
1960 OS level between the cluster defaults and the instance ones. This allows
1961 customisation of virtualization parameters depending on the installed
1962 OS. For example instances with OS 'X' may have a different KVM kernel
1963 (or any other parameter) than the cluster defaults. This is intended to
1964 help managing a multiple OSes on the same cluster, without manual
1965 modification of each instance's parameters.
1967 A tool for merging clusters, ``cluster-merge``, has been added in the
1968 tools sub-directory.
1973 - Improved the int/float conversions that should make the code more
1974 robust in face of errors from the node daemons
1975 - Fixed the remove node code in case of internal configuration errors
1976 - Fixed the node daemon behaviour in face of inconsistent queue
1977 directory (e.g. read-only file-system where we can't open the files
1979 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1980 now it either aborts cleanly or, if given the new "auto_promote"
1981 parameter, will automatically promote other nodes as needed
1982 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1983 completely prevent Ganeti from working
1984 - Fixed bug for instance export when not all disks were successfully
1986 - Fixed behaviour of node add when the new node is slow in starting up
1988 - Fixed handling of signals in the LUXI client, which should improve
1989 behaviour of command-line scripts
1990 - Added checks for invalid node/instance names in the configuration (now
1991 flagged during cluster verify)
1992 - Fixed watcher behaviour for disk activation errors
1993 - Fixed two potentially endless loops in http library, which led to the
1994 RAPI daemon hanging and consuming 100% CPU in some cases
1995 - Fixed bug in RAPI daemon related to hashed passwords
1996 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1998 - Enhanced compatibility with non-Debian OSes, but not using absolute
1999 path in some commands and allowing customisation of the ssh
2000 configuration directory
2001 - Fixed possible future issue with new Python versions by abiding to the
2002 proper use of ``__slots__`` attribute on classes
2003 - Added checks that should prevent directory traversal attacks
2004 - Many documentation fixes based on feedback from users
2009 - Added an "early_release" more for instance replace disks and node
2010 evacuate, where we release locks earlier and thus allow higher
2011 parallelism within the cluster
2012 - Added watcher hooks, intended to allow the watcher to restart other
2013 daemons (e.g. from the ganeti-nbma project), but they can be used of
2014 course for any other purpose
2015 - Added a compile-time disable for DRBD barriers, to increase
2016 performance if the administrator trusts the power supply or the
2017 storage system to not lose writes
2018 - Added the option of using syslog for logging instead of, or in
2019 addition to, Ganeti's own log files
2020 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2021 can indeed boot from a paravirtual NIC
2022 - Added a generic debug level for many operations; while this is not
2023 used widely yet, it allows one to pass the debug value all the way to
2025 - Enhanced the hooks environment for instance moves (failovers,
2026 migrations) where the primary/secondary nodes changed during the
2027 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2028 - Enhanced data validations for many user-supplied values; one important
2029 item is the restrictions imposed on instance and node names, which
2030 might reject some (invalid) host names
2031 - Add a configure-time option to disable file-based storage, if it's not
2032 needed; this allows greater security separation between the master
2033 node and the other nodes from the point of view of the inter-node RPC
2035 - Added user notification in interactive tools if job is waiting in the
2036 job queue or trying to acquire locks
2037 - Added log messages when a job is waiting for locks
2038 - Added filtering by node tags in instance operations which admit
2039 multiple instances (start, stop, reboot, reinstall)
2040 - Added a new tool for cluster mergers, ``cluster-merge``
2041 - Parameters from command line which are of the form ``a=b,c=d`` can now
2042 use backslash escapes to pass in values which contain commas,
2043 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2045 - For KVM, the instance name is the first parameter passed to KVM, so
2046 that it's more visible in the process list
2052 *(Released Tue, 2 Mar 2010)*
2054 Ganeti 2.1 brings many improvements with it. Major changes:
2056 - Added infrastructure to ease automated disk repairs
2057 - Added new daemon to export configuration data in a cheaper way than
2058 using the remote API
2059 - Instance NICs can now be routed instead of being associated with a
2061 - Improved job locking logic to reduce impact of jobs acquiring multiple
2062 locks waiting for other long-running jobs
2064 In-depth implementation details can be found in the Ganeti 2.1 design
2070 - Added chroot hypervisor
2071 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2073 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2074 ``bootloader_path`` and ``bootloader_args``)
2075 - Added the ``use_localtime`` option for the xen-hvm and kvm
2076 hypervisors, and the default value for this has changed to false (in
2077 2.0 xen-hvm always enabled it)
2078 - Added luxi call to submit multiple jobs in one go
2079 - Added cluster initialization option to not modify ``/etc/hosts``
2081 - Added network interface parameters
2082 - Added dry run mode to some LUs
2083 - Added RAPI resources:
2085 - ``/2/instances/[instance_name]/info``
2086 - ``/2/instances/[instance_name]/replace-disks``
2087 - ``/2/nodes/[node_name]/evacuate``
2088 - ``/2/nodes/[node_name]/migrate``
2089 - ``/2/nodes/[node_name]/role``
2090 - ``/2/nodes/[node_name]/storage``
2091 - ``/2/nodes/[node_name]/storage/modify``
2092 - ``/2/nodes/[node_name]/storage/repair``
2094 - Added OpCodes to evacuate or migrate all instances on a node
2095 - Added new command to list storage elements on nodes (``gnt-node
2096 list-storage``) and modify them (``gnt-node modify-storage``)
2097 - Added new ssconf files with master candidate IP address
2098 (``ssconf_master_candidates_ips``), node primary IP address
2099 (``ssconf_node_primary_ips``) and node secondary IP address
2100 (``ssconf_node_secondary_ips``)
2101 - Added ``ganeti-confd`` and a client library to query the Ganeti
2102 configuration via UDP
2103 - Added ability to run hooks after cluster initialization and before
2105 - Added automatic mode for disk replace (``gnt-instance replace-disks
2107 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2108 after catastrophic data-loss
2109 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2111 - Added ``gnt-instance move`` command to move instances
2112 - Added ``gnt-cluster watcher`` command to control watcher
2113 - Added ``gnt-node powercycle`` command to powercycle nodes
2114 - Added new job status field ``lock_status``
2115 - Added parseable error codes to cluster verification (``gnt-cluster
2116 verify --error-codes``) and made output less verbose (use
2117 ``--verbose`` to restore previous behaviour)
2118 - Added UUIDs to the main config entities (cluster, nodes, instances)
2119 - Added support for OS variants
2120 - Added support for hashed passwords in the Ganeti remote API users file
2122 - Added option to specify maximum timeout on instance shutdown
2123 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2124 - Added new helper script to start and stop Ganeti daemons
2125 (``daemon-util``), with the intent to reduce the work necessary to
2126 adjust Ganeti for non-Debian distributions and to start/stop daemons
2128 - Added more unittests
2129 - Fixed critical bug in ganeti-masterd startup
2130 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2131 now customisable at the cluster level for both the KVM and Xen
2132 hypervisors using the new ``migration_port`` parameter
2133 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2134 reinstalling an instance
2135 - Allowed ``@`` in tag names
2136 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2137 - Many documentation updates
2138 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2139 - ``gnt-instance reinstall`` can now reinstall multiple instances
2140 - Updated many command line parameters
2141 - Introduced new OS API version 15
2142 - No longer support a default hypervisor
2143 - Treat virtual LVs as inexistent
2144 - Improved job locking logic to reduce lock contention
2145 - Match instance and node names case insensitively
2146 - Reimplemented bash completion script to be more complete
2153 *(Released Thu, 4 Feb 2010)*
2155 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2156 - Fix a string formatting bug
2157 - Improve safety of the code in some error paths
2158 - Improve data validation in the master of values returned from nodes
2164 *(Released Thu, 17 Dec 2009)*
2166 - Fix security issue due to missing validation of iallocator names; this
2167 allows local and remote execution of arbitrary executables
2168 - Fix failure of gnt-node list during instance removal
2169 - Ship the RAPI documentation in the archive
2175 *(Released Wed, 30 Sep 2009)*
2177 - Fixed many wrong messages
2178 - Fixed a few bugs related to the locking library
2179 - Fixed MAC checking at instance creation time
2180 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2181 - Fixed a few issues related to signal handling in both daemons and
2183 - Fixed the example startup script provided
2184 - Fixed insserv dependencies in the example startup script (patch from
2186 - Fixed handling of drained nodes in the iallocator framework
2187 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2189 - Fixed error related to invalid job IDs in job polling
2190 - Fixed job/opcode persistence on unclean master shutdown
2191 - Fixed handling of partial job processing after unclean master
2193 - Fixed error reporting from LUs, previously all errors were converted
2194 into execution errors
2195 - Fixed error reporting from burnin
2196 - Decreased significantly the memory usage of the job queue
2197 - Optimised slightly multi-job submission
2198 - Optimised slightly opcode loading
2199 - Backported the multi-job submit framework from the development
2200 branch; multi-instance start and stop should be faster
2201 - Added script to clean archived jobs after 21 days; this will reduce
2202 the size of the queue directory
2203 - Added some extra checks in disk size tracking
2204 - Added an example ethers hook script
2205 - Added a cluster parameter that prevents Ganeti from modifying of
2207 - Added more node information to RAPI responses
2208 - Added a ``gnt-job watch`` command that allows following the ouput of a
2210 - Added a bind-address option to ganeti-rapi
2211 - Added more checks to the configuration verify
2212 - Enhanced the burnin script such that some operations can be retried
2214 - Converted instance reinstall to multi-instance model
2220 *(Released Fri, 7 Aug 2009)*
2222 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2223 to allow using the pre-2.0.2 behaviour in activation, if any existing
2224 instances have mismatched disk sizes in the configuration
2225 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2226 any configuration mismatches for disk sizes
2227 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2228 failover to work on two-node clusters
2229 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2231 - Fixed detection of OS script errors in ``gnt-backup export``
2232 - Fixed exit code of ``gnt-backup export``
2238 *(Released Fri, 17 Jul 2009)*
2240 - Added experimental support for stripped logical volumes; this should
2241 enhance performance but comes with a higher complexity in the block
2242 device handling; stripping is only enabled when passing
2243 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2244 affected even in the non-stripped mode
2245 - Improved resiliency against transient failures at the end of DRBD
2246 resyncs, and in general of DRBD resync checks
2247 - Fixed a couple of issues with exports and snapshot errors
2248 - Fixed a couple of issues in instance listing
2249 - Added display of the disk size in ``gnt-instance info``
2250 - Fixed checking for valid OSes in instance creation
2251 - Fixed handling of the "vcpus" parameter in instance listing and in
2252 general of invalid parameters
2253 - Fixed http server library, and thus RAPI, to handle invalid
2254 username/password combinations correctly; this means that now they
2255 report unauthorized for queries too, not only for modifications,
2256 allowing earlier detect of configuration problems
2257 - Added a new "role" node list field, equivalent to the master/master
2258 candidate/drained/offline flags combinations
2259 - Fixed cluster modify and changes of candidate pool size
2260 - Fixed cluster verify error messages for wrong files on regular nodes
2261 - Fixed a couple of issues with node demotion from master candidate role
2262 - Fixed node readd issues
2263 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2264 - Added a new ``--no-voting`` option for masterfailover to fix failover
2265 on two-nodes clusters when the former master node is unreachable
2266 - Added instance reinstall over RAPI
2272 *(Released Tue, 16 Jun 2009)*
2274 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2275 allow re-adding the start in single-user option (regression from 1.2)
2276 - the watcher writes the instance status to a file, to allow monitoring
2277 to report the instance status (from the master) based on cached
2278 results of the watcher's queries; while this can get stale if the
2279 watcher is being locked due to other work on the cluster, this is
2280 still an improvement
2281 - the watcher now also restarts the node daemon and the rapi daemon if
2283 - fixed the watcher to handle full and drained queue cases
2284 - hooks export more instance data in the environment, which helps if
2285 hook scripts need to take action based on the instance's properties
2286 (no longer need to query back into ganeti)
2287 - instance failovers when the instance is stopped do not check for free
2288 RAM, so that failing over a stopped instance is possible in low memory
2290 - rapi uses queries for tags instead of jobs (for less job traffic), and
2291 for cluster tags it won't talk to masterd at all but read them from
2293 - a couple of error handling fixes in RAPI
2294 - drbd handling: improved the error handling of inconsistent disks after
2295 resync to reduce the frequency of "there are some degraded disks for
2296 this instance" messages
2297 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2298 error handling path called a wrong function name)
2304 *(Released Wed, 27 May 2009)*
2306 - no changes from rc5
2312 *(Released Wed, 20 May 2009)*
2314 - fix a couple of bugs (validation, argument checks)
2315 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2316 - some small improvements to RAPI and IAllocator
2317 - make watcher automatically start the master daemon if down
2323 *(Released Mon, 27 Apr 2009)*
2325 - change the OS list to not require locks; this helps with big clusters
2326 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2327 volume group is broken
2328 - ``gnt-instance info``, without any arguments, doesn't run for all
2329 instances anymore; either pass ``--all`` or pass the desired
2330 instances; this helps against mistakes on big clusters where listing
2331 the information for all instances takes a long time
2332 - miscellaneous doc and man pages fixes
2338 *(Released Wed, 8 Apr 2009)*
2340 - Change the internal locking model of some ``gnt-node`` commands, in
2341 order to reduce contention (and blocking of master daemon) when
2342 batching many creation/reinstall jobs
2343 - Fixes to Xen soft reboot
2344 - No longer build documentation at build time, instead distribute it in
2345 the archive, in order to reduce the need for the whole docbook/rst
2352 *(Released Fri, 27 Mar 2009)*
2354 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2355 - Fix watcher startup sequence, improves the behaviour of busy clusters
2356 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2357 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2358 bind address and other places
2359 - Some documentation fixes and updates
2365 *(Released Mon, 2 Mar 2009)*
2367 - More documentation updates, now all docs should be more-or-less
2369 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2371 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2373 - Fix an issue related to $libdir/run/ganeti and cluster creation
2379 *(Released Thu, 19 Feb 2009)*
2381 - Xen PVM and KVM have switched the default value for the instance root
2382 disk to the first partition on the first drive, instead of the whole
2383 drive; this means that the OS installation scripts must be changed
2385 - Man pages have been updated
2386 - RAPI has been switched by default to HTTPS, and the exported functions
2387 should all work correctly
2388 - RAPI v1 has been removed
2389 - Many improvements to the KVM hypervisor
2390 - Block device errors are now better reported
2391 - Many other bugfixes and small improvements
2397 *(Released Mon, 26 Jan 2009)*
2399 - Version 2 is a general rewrite of the code and therefore the
2400 differences are too many to list, see the design document for 2.0 in
2401 the ``doc/`` subdirectory for more details
2402 - In this beta version there is not yet a migration path from 1.2 (there
2403 will be one in the final 2.0 release)
2404 - A few significant changes are:
2406 - all commands are executed by a daemon (``ganeti-masterd``) and the
2407 various ``gnt-*`` commands are just front-ends to it
2408 - all the commands are entered into, and executed from a job queue,
2409 see the ``gnt-job(8)`` manpage
2410 - the RAPI daemon supports read-write operations, secured by basic
2411 HTTP authentication on top of HTTPS
2412 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2413 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2414 to migrate to DRBD 8 first while still running Ganeti 1.2)
2415 - DRBD devices are using statically allocated minor numbers, which
2416 will be assigned to existing instances during the migration process
2417 - there is support for both Xen PVM and Xen HVM instances running on
2419 - KVM virtualization is supported too
2420 - file-based storage has been implemented, which means that it is
2421 possible to run the cluster without LVM and DRBD storage, for
2422 example using a shared filesystem exported from shared storage (and
2423 still have live migration)
2429 *(Released Tue, 13 Jan 2009)*
2431 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2432 - Reuse the old instance mac address by default on instance import, if
2433 the instance name is the same.
2434 - Handle situations in which the node info rpc returns incomplete
2436 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2437 - Improved version of batcher:
2439 - state file support
2440 - instance mac address support
2441 - support for HVM clusters/instances
2443 - Add an option to show the number of cpu sockets and nodes in
2445 - Support OSes that handle more than one version of the OS api (but do
2446 not change the current API in any other way)
2447 - Fix ``gnt-node migrate``
2448 - ``gnt-debug`` man page
2449 - Fixes various more typos and small issues
2450 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2456 *(Released Wed, 24 Sep 2008)*
2458 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2459 type of disk exported to fully virtualized instances.
2460 - provide access to the serial console of HVM instances
2461 - instance auto_balance flag, set by default. If turned off it will
2462 avoid warnings on cluster verify if there is not enough memory to fail
2463 over an instance. in the future it will prevent automatically failing
2464 it over when we will support that.
2465 - batcher tool for instance creation, see ``tools/README.batcher``
2466 - ``gnt-instance reinstall --select-os`` to interactively select a new
2467 operating system when reinstalling an instance.
2468 - when changing the memory amount on instance modify a check has been
2469 added that the instance will be able to start. also warnings are
2470 emitted if the instance will not be able to fail over, if auto_balance
2472 - documentation fixes
2473 - sync fields between ``gnt-instance list/modify/add/import``
2474 - fix a race condition in drbd when the sync speed was set after giving
2475 the device a remote peer.
2481 *(Released Tue, 22 Jul 2008)*
2483 - note: the allowed size and number of tags per object were reduced
2484 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2485 - fixed twisted 8.x compatibility
2486 - fixed ``gnt-instance replace-disks`` with iallocator
2487 - add TCP keepalives on twisted connections to detect restarted nodes
2488 - disk increase support, see ``gnt-instance grow-disk``
2489 - implement bulk node/instance query for RAPI
2490 - add tags in node/instance listing (optional)
2491 - experimental migration (and live migration) support, read the man page
2492 for ``gnt-instance migrate``
2493 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2494 has some small improvements in handling its state file
2500 *(Released Fri, 13 Jun 2008)*
2502 - Experimental readonly, REST-based remote API implementation;
2503 automatically started on master node, TCP port 5080, if enabled by
2504 ``--enable-rapi`` parameter to configure script.
2505 - Instance allocator support. Add and import instance accept a
2506 ``--iallocator`` parameter, and call that instance allocator to decide
2507 which node to use for the instance. The iallocator document describes
2508 what's expected from an allocator script.
2509 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2510 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2511 node is lost there is still enough memory to fail over the instances
2513 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2514 ``gnt-cluster verify``, to check for site-specific compliance. All the
2515 hooks will run, and their output, if any, will be displayed. Any
2516 failing hook will make the verification return an error value.
2517 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2518 primary and secondary interfaces
2519 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2520 that is still declared as part of the cluster and has failed.
2521 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2522 specifying output fields, that just adds the chosen fields to the
2524 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2525 export from the filesystem.
2526 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2527 have been added. Using them you can enable/disable acpi and pae
2528 support, and specify a path for a cd image to be exported to the
2529 instance. These parameters as the name suggest only work on HVM
2531 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2532 PAE support will be set to the previously hardcoded values, but the
2533 (previously hardcoded) path to the CDROM ISO image will be unset and
2534 if required, needs to be set manually with ``gnt-instance modify``
2536 - The address to which an instance's VNC console is bound is now
2537 selectable per-instance, rather than being cluster wide. Of course
2538 this only applies to instances controlled via VNC, so currently just
2539 applies to HVM clusters.
2545 *(Released Mon, 18 Feb 2008)*
2547 - more tweaks to the disk activation code (especially helpful for DRBD)
2548 - change the default ``gnt-instance list`` output format, now there is
2549 one combined status field (see the manpage for the exact values this
2551 - some more fixes for the mac export to hooks change
2552 - make Ganeti not break with DRBD 8.2.x (which changed the version
2553 format in ``/proc/drbd``) (issue 24)
2554 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2555 template, allowing migration from DRBD0.7+MD to DRBD8
2561 *(Released Wed, 30 Jan 2008)*
2563 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2565 - add command aliases infrastructure and a few aliases
2566 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2567 man pages and the ``--help`` option of ``gnt-node
2568 list``/``gnt-instance list``
2569 - fix ``gnt-backup list`` with down nodes (issue 21)
2570 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2571 - fix the dist archive and add a check for including svn/git files in
2573 - some developer-related changes: improve the burnin and the QA suite,
2574 add an upload script for testing during development
2580 *(Released Wed, 16 Jan 2008)*
2582 - experimental HVM support, read the install document, section
2583 "Initializing the cluster"
2584 - allow for the PVM hypervisor per-instance kernel and initrd paths
2585 - add a new command ``gnt-cluster verify-disks`` which uses a new
2586 algorithm to improve the reconnection of the DRBD pairs if the device
2587 on the secondary node has gone away
2588 - make logical volume code auto-activate LVs at disk activation time
2589 - slightly improve the speed of activating disks
2590 - allow specification of the MAC address at instance creation time, and
2591 changing it later via ``gnt-instance modify``
2592 - fix handling of external commands that generate lots of output on
2594 - update documentation with regard to minimum version of DRBD8 supported
2600 *(Released Tue, 4 Dec 2007)*
2602 - Log the ``xm create`` output to the node daemon log on failure (to
2603 help diagnosing the error)
2604 - In debug mode, log all external commands output if failed to the logs
2605 - Change parsing of lvm commands to ignore stderr
2611 *(Released Wed, 28 Nov 2007)*
2613 - Another round of updates to the DRBD 8 code to deal with more failures
2614 in the replace secondary node operation
2615 - Some more logging of failures in disk operations (lvm, drbd)
2616 - A few documentation updates
2623 *(Released Tue, 13 Nov 2007)*
2625 - Change configuration file format from Python's Pickle to JSON.
2626 Upgrading is possible using the cfgupgrade utility.
2627 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2628 faster replace disks and is more stable (DRBD 8 has many improvements
2629 compared to DRBD 0.7)
2630 - Added command line tags support (see man pages for ``gnt-instance``,
2631 ``gnt-node``, ``gnt-cluster``)
2632 - Added instance rename support
2633 - Added multi-instance startup/shutdown
2634 - Added cluster rename support
2635 - Added ``gnt-node evacuate`` to simplify some node operations
2636 - Added instance reboot operation that can speedup reboot as compared to
2638 - Soften the requirement that hostnames are in FQDN format
2639 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2641 - Removed dependency on debian's patched fping that uses the
2642 non-standard ``-S`` option
2643 - Now the OS definitions are searched for in multiple, configurable
2644 paths (easier for distros to package)
2645 - Some changes to the hooks infrastructure (especially the new
2646 post-configuration update hook)
2647 - Other small bugfixes
2649 .. vim: set textwidth=72 syntax=rst :