8 *(Released Thu, 25 Jul 2013)*
10 - Add logrotate functionality in daemon-util
11 - Add logrotate example file
12 - Add missing fields to network queries over rapi
13 - Fix network object timestamps
14 - Add support for querying network timestamps
15 - Fix a typo in the example crontab
16 - Fix a documentation typo
22 *(Released Thu, 04 Jul 2013)*
24 Incompatible/important changes
25 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
27 - Instance policies for disk size were documented to be on a per-disk
28 basis, but hail applied them to the sum of all disks. This has been
30 - ``hbal`` will now exit with status 0 if, during job execution over
31 LUXI, early exit has been requested and all jobs are successful;
32 before, exit status 1 was used, which cannot be differentiated from
34 - Compatibility with newer versions of rbd has been fixed
35 - ``gnt-instance batch-create`` has been changed to use the bulk create
36 opcode from Ganeti. This lead to incompatible changes in the format of
37 the JSON file. It's now not a custom dict anymore but a dict
38 compatible with the ``OpInstanceCreate`` opcode.
39 - Parent directories for file storage need to be listed in
40 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
41 write the file automatically based on old configuration values, but it
42 can not distribute it across all nodes and the file contents should be
43 verified. Use ``gnt-cluster copyfile
44 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
45 upgraded. The reason for requiring this list of paths now is that
46 before it would have been possible to inject new paths via RPC,
47 allowing files to be created in arbitrary locations. The RPC protocol
48 is protected using SSL/X.509 certificates, but as a design principle
49 Ganeti does not permit arbitrary paths to be passed.
50 - The parsing of the variants file for OSes (see
51 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
52 lines and comment lines (starting with ``#``) are ignored for better
54 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
55 longer available. ``gnt-node add`` now invokes a new tool on the
56 destination node, named ``prepare-node-join``, to configure the SSH
57 daemon. Paramiko is no longer necessary to configure nodes' SSH
58 daemons via ``gnt-node add``.
59 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
60 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
61 the flag is not reset after a master failover.
62 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
63 is recommended. 2.6 will be mandatory from the 2.8 series.
69 - New network management functionality to support automatic allocation
70 of IP addresses and managing of network parameters. See
71 :manpage:`gnt-network(8)` for more details.
72 - New external storage backend, to allow managing arbitrary storage
73 systems external to the cluster. See
74 :manpage:`ganeti-extstorage-interface(7)`.
75 - New ``exclusive-storage`` node parameter added, restricted to
76 nodegroup level. When it's set to true, physical disks are assigned in
77 an exclusive fashion to instances, as documented in :doc:`Partitioned
78 Ganeti <design-partitioned>`. Currently, only instances using the
79 ``plain`` disk template are supported.
80 - The KVM hypervisor has been updated with many new hypervisor
81 parameters, including a generic one for passing arbitrary command line
82 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
83 compatible up to qemu 1.4.
84 - A new tool, called ``mon-collector``, is the stand-alone executor of
85 the data collectors for a monitoring system. As of this version, it
86 just includes the DRBD data collector, that can be executed by calling
87 ``mon-collector`` using the ``drbd`` parameter. See
88 :manpage:`mon-collector(7)`.
89 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
90 for RAPI users. It allows granting permissions to query for
91 information to a specific user without giving
92 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
93 - A new tool named ``node-cleanup`` has been added. It cleans remains of
94 a cluster from a machine by stopping all daemons, removing
95 certificates and ssconf files. Unless the ``--no-backup`` option is
96 given, copies of the certificates are made.
97 - Instance creations now support the use of opportunistic locking,
98 potentially speeding up the (parallel) creation of multiple instances.
99 This feature is currently only available via the :doc:`RAPI
100 <rapi>` interface and when an instance allocator is used. If the
101 ``opportunistic_locking`` parameter is set the opcode will try to
102 acquire as many locks as possible, but will not wait for any locks
103 held by other opcodes. If not enough resources can be found to
104 allocate the instance, the temporary error code
105 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
106 retried thereafter, with or without opportunistic locking.
107 - New experimental linux-ha resource scripts.
108 - Restricted-commands support: ganeti can now be asked (via command line
109 or rapi) to perform commands on a node. These are passed via ganeti
110 RPC rather than ssh. This functionality is restricted to commands
111 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
112 reasons. The file is not copied automatically.
118 - Diskless instances are now externally mirrored (Issue 237). This for
119 now has only been tested in conjunction with explicit target nodes for
121 - Queries not needing locks or RPC access to the node can now be
122 performed by the confd daemon, making them independent from jobs, and
123 thus faster to execute. This is selectable at configure time.
124 - The functionality for allocating multiple instances at once has been
125 overhauled and is now also available through :doc:`RAPI <rapi>`.
127 There are no significant changes from version 2.7.0~rc3.
133 *(Released Tue, 25 Jun 2013)*
135 - Fix permissions on the confd query socket (Issue 477)
136 - Fix permissions on the job archive dir (Issue 498)
137 - Fix handling of an internal exception in replace-disks (Issue 472)
138 - Fix gnt-node info handling of shortened names (Issue 497)
139 - Fix gnt-instance grow-disk when wiping is enabled
140 - Documentation improvements, and support for newer pandoc
141 - Fix hspace honoring ipolicy for disks (Issue 484)
142 - Improve handling of the ``kvm_extra`` HV parameter
148 *(Released Fri, 24 May 2013)*
150 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
152 - Disks added through ``gnt-instance modify`` or created through
153 ``gnt-instance recreate-disks`` are wiped, if the
154 ``prealloc_wipe_disks`` flag is set.
155 - If wiping newly created disks fails, the disks are removed. Also,
156 partial failures in creating disks through ``gnt-instance modify``
157 triggers a cleanup of the partially-created disks.
158 - Removing the master IP address doesn't fail if the address has been
160 - Fix ownership of the OS log dir
161 - Workaround missing SO_PEERCRED constant (Issue 191)
167 *(Released Fri, 3 May 2013)*
169 This was the first release candidate of the 2.7 series. Since beta3:
171 - Fix kvm compatibility with qemu 1.4 (Issue 389)
172 - Documentation updates (admin guide, upgrade notes, install
173 instructions) (Issue 372)
174 - Fix gnt-group list nodes and instances count (Issue 436)
175 - Fix compilation without non-mandatory libraries (Issue 441)
176 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
177 - Make confd logging more verbose at INFO level (Issue 435)
178 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
179 - Fix failure path for instance storage type conversion (Issue 229)
180 - Update htools text backend documentation
181 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
182 - Disable inter-cluster instance move for file-based instances, because
183 it is dependant on instance export, which is not supported for
184 file-based instances. (Issue 414)
185 - Fix gnt-job crashes on non-ascii characters (Issue 427)
186 - Fix volume group checks on non-vm-capable nodes (Issue 432)
192 *(Released Mon, 22 Apr 2013)*
194 This was the third beta release of the 2.7 series. Since beta2:
196 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
197 - Fix data loss on wrong usage of ``gnt-instance move``
198 - Properly export errors in confd-based job queries
199 - Add ``users-setup`` tool
200 - Fix iallocator protocol to report 0 as a disk size for diskless
201 instances. This avoids hail breaking when a diskless instance is
203 - Fix job queue directory permission problem that made confd job queries
204 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
205 for access to archived jobs (Issue 406).
206 - Limit the sizes of networks supported by ``gnt-network`` to something
207 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
208 - Fix bugs in instance disk template conversion
209 - Fix GHC 7 compatibility
210 - Fix ``burnin`` install path (Issue 426).
211 - Allow very small disk grows (Issue 347).
212 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
213 that noded doesn't import masterd code (Issue 419).
214 - Make sure the default metavg at cluster init is the same as the vg, if
215 unspecified (Issue 358).
216 - Fix cleanup of partially created disks (part of Issue 416)
222 *(Released Tue, 2 Apr 2013)*
224 This was the second beta release of the 2.7 series. Since beta1:
226 - Networks no longer have a "type" slot, since this information was
227 unused in Ganeti: instead of it tags should be used.
228 - The rapi client now has a ``target_node`` option to MigrateInstance.
229 - Fix early exit return code for hbal (Issue 386).
230 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
231 - Fix ``rbd showmapped`` output parsing (Issue 312).
232 - Networks are now referenced indexed by UUID, rather than name. This
233 will require running cfgupgrade, from 2.7.0beta1, if networks are in
235 - The OS environment now includes network information.
236 - Deleting of a network is now disallowed if any instance nic is using
237 it, to prevent dangling references.
238 - External storage is now documented in man pages.
239 - The exclusive_storage flag can now only be set at nodegroup level.
240 - Hbal can now submit an explicit priority with its jobs.
241 - Many network related locking fixes.
242 - Bump up the required pylint version to 0.25.1.
243 - Fix the ``no_remember`` option in RAPI client.
244 - Many ipolicy related tests, qa, and fixes.
245 - Many documentation improvements and fixes.
246 - Fix building with ``--disable-file-storage``.
247 - Fix ``-q`` option in htools, which was broken if passed more than
249 - Some haskell/python interaction improvements and fixes.
250 - Fix iallocator in case of missing LVM storage.
251 - Fix confd config load in case of ``--no-lvm-storage``.
252 - The confd/query functionality is now mentioned in the security
259 *(Released Wed, 6 Feb 2013)*
261 This was the first beta release of the 2.7 series. All important changes
262 are listed in the latest 2.7 entry.
268 *(Released Fri, 21 Dec 2012)*
270 Important behaviour change: hbal won't rebalance anymore instances which
271 have the ``auto_balance`` attribute set to false. This was the intention
272 all along, but until now it only skipped those from the N+1 memory
273 reservation (DRBD-specific).
275 A significant number of bug fixes in this release:
277 - Fixed disk adoption interaction with ipolicy checks.
278 - Fixed networking issues when instances are started, stopped or
279 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
280 - Fixed the warning in cluster verify for shared storage instances not
282 - Fixed removal of storage directory on shared file storage (issue 262).
283 - Fixed validation of LVM volume group name in OpClusterSetParams
284 (``gnt-cluster modify``) (issue 285).
285 - Fixed runtime memory increases (``gnt-instance modify -m``).
286 - Fixed live migration under Xen's ``xl`` mode.
287 - Fixed ``gnt-instance console`` with ``xl``.
288 - Fixed building with newer Haskell compiler/libraries.
289 - Fixed PID file writing in Haskell daemons (confd); this prevents
290 restart issues if confd was launched manually (outside of
291 ``daemon-util``) while another copy of it was running
292 - Fixed a type error when doing live migrations with KVM (issue 297) and
293 the error messages for failing migrations have been improved.
294 - Fixed opcode validation for the out-of-band commands (``gnt-node
296 - Fixed a type error when unsetting OS hypervisor parameters (issue
297 311); now it's possible to unset all OS-specific hypervisor
299 - Fixed the ``dry-run`` mode for many operations: verification of
300 results was over-zealous but didn't take into account the ``dry-run``
301 operation, resulting in "wrong" failures.
302 - Fixed bash completion in ``gnt-job list`` when the job queue has
303 hundreds of entries; especially with older ``bash`` versions, this
304 results in significant CPU usage.
306 And lastly, a few other improvements have been made:
308 - Added option to force master-failover without voting (issue 282).
309 - Clarified error message on lock conflict (issue 287).
310 - Logging of newly submitted jobs has been improved (issue 290).
311 - Hostname checks have been made uniform between instance rename and
313 - The ``--submit`` option is now supported by ``gnt-debug delay``.
314 - Shutting down the master daemon by sending SIGTERM now stops it from
315 processing jobs waiting for locks; instead, those jobs will be started
316 once again after the master daemon is started the next time (issue
318 - Support for Xen's ``xl`` program has been improved (besides the fixes
320 - Reduced logging noise in the Haskell confd daemon (only show one log
321 entry for each config reload, instead of two).
322 - Several man page updates and typo fixes.
328 *(Released Fri, 12 Oct 2012)*
330 A small bugfix release. Among the bugs fixed:
332 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
333 made the command unusable.
334 - Commands that issue many jobs don't fail anymore just because some jobs
335 take so long that other jobs are archived.
336 - Failures during ``gnt-instance reinstall`` are reflected by the exit
338 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
340 - When ``always_failover`` is set, ``--allow-failover`` is not required
341 in migrate commands anymore.
342 - ``bash_completion`` works even if extglob is disabled.
343 - Fixed bug with locks that made failover for RDB-based instances fail.
344 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
345 a random node instead of one based on the allocator metric.
346 - Support for newer versions of pylint and pep8.
347 - Hail doesn't fail anymore when trying to add an instance of type
348 ``file``, ``sharedfile`` or ``rbd``.
349 - Added new Makefile target to rebuild the whole distribution, so that
350 all files are included.
356 *(Released Fri, 27 Jul 2012)*
359 .. attention:: The ``LUXI`` protocol has been made more consistent
360 regarding its handling of command arguments. This, however, leads to
361 incompatibility issues with previous versions. Please ensure that you
362 restart Ganeti daemons soon after the upgrade, otherwise most
363 ``LUXI`` calls (job submission, setting/resetting the drain flag,
364 pausing/resuming the watcher, cancelling and archiving jobs, querying
365 the cluster configuration) will fail.
374 The current ``admin_up`` field, which used to denote whether an instance
375 should be running or not, has been removed. Instead, ``admin_state`` is
376 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
378 The rational behind this is that an instance being “down” can have
381 - it could be down during a reboot
382 - it could be temporarily be down for a reinstall
383 - or it could be down because it is deprecated and kept just for its
386 The previous Boolean state was making it difficult to do capacity
387 calculations: should Ganeti reserve memory for a down instance? Now, the
388 tri-state field makes it clear:
390 - in ``up`` and ``down`` state, all resources are reserved for the
391 instance, and it can be at any time brought up if it is down
392 - in ``offline`` state, only disk space is reserved for it, but not
395 The field can have an extra use: since the transition between ``up`` and
396 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
397 transition between ``offline`` and ``down`` is done via ``gnt-instance
398 modify``, it is possible to given different rights to users. For
399 example, owners of an instance could be allowed to start/stop it, but
400 not transition it out of the offline state.
402 Instance policies and specs
403 +++++++++++++++++++++++++++
405 In previous Ganeti versions, an instance creation request was not
406 limited on the minimum size and on the maximum size just by the cluster
407 resources. As such, any policy could be implemented only in third-party
408 clients (RAPI clients, or shell wrappers over ``gnt-*``
409 tools). Furthermore, calculating cluster capacity via ``hspace`` again
410 required external input with regards to instance sizes.
412 In order to improve these workflows and to allow for example better
413 per-node group differentiation, we introduced instance specs, which
416 - minimum instance disk size, disk count, memory size, cpu count
417 - maximum values for the above metrics
418 - and “standard” values (used in ``hspace`` to calculate the standard
421 The minimum/maximum values can be also customised at node-group level,
422 for example allowing more powerful hardware to support bigger instance
425 Beside the instance specs, there are a few other settings belonging to
426 the instance policy framework. It is possible now to customise, per
427 cluster and node-group:
429 - the list of allowed disk templates
430 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
431 - the maximum ratio of instance to spindles (see below for more
432 information) for local storage
434 All these together should allow all tools that talk to Ganeti to know
435 what are the ranges of allowed values for instances and the
436 over-subscription that is allowed.
438 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
439 instance configuration, and the physical CPU configuration from the
440 node. For the spindle ratios however, we didn't track before these
441 values, so new parameters have been added:
443 - a new node parameter ``spindle_count``, defaults to 1, customisable at
444 node group or node level
445 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
447 Note that spindles in this context doesn't need to mean actual
448 mechanical hard-drives; it's just a relative number for both the node
449 I/O capacity and instance I/O consumption.
451 Instance migration behaviour
452 ++++++++++++++++++++++++++++
454 While live-migration is in general desirable over failover, it is
455 possible that for some workloads it is actually worse, due to the
456 variable time of the “suspend” phase during live migration.
458 To allow the tools to work consistently over such instances (without
459 having to hard-code instance names), a new backend parameter
460 ``always_failover`` has been added to control the migration/failover
461 behaviour. When set to True, all migration requests for an instance will
462 instead fall-back to failover.
464 Instance memory ballooning
465 ++++++++++++++++++++++++++
467 Initial support for memory ballooning has been added. The memory for an
468 instance is no longer fixed (backend parameter ``memory``), but instead
469 can vary between minimum and maximum values (backend parameters
470 ``minmem`` and ``maxmem``). Currently we only change an instance's
473 - live migrating or failing over and instance and the target node
474 doesn't have enough memory
475 - user requests changing the memory via ``gnt-instance modify
481 In order to control the use of specific CPUs by instance, support for
482 controlling CPU pinning has been added for the Xen, HVM and LXC
483 hypervisors. This is controlled by a new hypervisor parameter
484 ``cpu_mask``; details about possible values for this are in the
485 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
486 VCPU-to-CPU mapping) form will work well only when all nodes in your
487 cluster have the same amount of CPUs.
492 Another area in which Ganeti was not customisable were the parameters
493 used for storage configuration, e.g. how many stripes to use for LVM,
494 DRBD resync configuration, etc.
496 To improve this area, we've added disks parameters, which are
497 customisable at cluster and node group level, and which allow to
498 specify various parameters for disks (DRBD has the most parameters
499 currently), for example:
501 - DRBD resync algorithm and parameters (e.g. speed)
502 - the default VG for meta-data volumes for DRBD
503 - number of stripes for LVM (plain disk template)
506 These parameters can be modified via ``gnt-cluster modify -D …`` and
507 ``gnt-group modify -D …``, and are used at either instance creation (in
508 case of LVM stripes, for example) or at disk “activation” time
511 Rados block device support
512 ++++++++++++++++++++++++++
514 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
515 denoted by the ``rbd`` disk template type. This is considered
516 experimental, feedback is welcome. For details on configuring it, see
517 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
522 The existing master IP functionality works well only in simple setups (a
523 single network shared by all nodes); however, if nodes belong to
524 different networks, then the ``/32`` setup and lack of routing
525 information is not enough.
527 To allow the master IP to function well in more complex cases, the
528 system was reworked as follows:
530 - a master IP netmask setting has been added
531 - the master IP activation/turn-down code was moved from the node daemon
533 - whether to run the Ganeti-supplied master IP script or a user-supplied
534 on is a ``gnt-cluster init`` setting
536 Details about the location of the standard and custom setup scripts are
537 in the man page :manpage:`gnt-cluster(8)`; for information about the
538 setup script protocol, look at the Ganeti-supplied script.
543 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
546 It is now possible to use TLS-protected connections, and when renewing
547 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
548 it is now possible to specify spice or spice CA certificates. Also, it
549 is possible to configure a password for SPICE sessions via the
550 hypervisor parameter ``spice_password_file``.
552 There are also new parameters to control the compression and streaming
553 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
554 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
555 for the spice parameters.
557 Lastly, it is now possible to see the SPICE connection information via
558 ``gnt-instance console``.
563 A new tool (``tools/ovfconverter``) has been added that supports
564 conversion between Ganeti and the `Open Virtualization Format
565 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
568 This relies on the ``qemu-img`` tool to convert the disk formats, so the
569 actual compatibility with other virtualization solutions depends on it.
574 The configuration query daemon (``ganeti-confd``) is now optional, and
575 has been rewritten in Haskell; whether to use the daemon at all, use the
576 Python (default) or the Haskell version is selectable at configure time
577 via the ``--enable-confd`` parameter, which can take one of the
578 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
579 daemon will result in a smaller footprint; for larger systems, we
580 welcome feedback on the Haskell version which might become the default
583 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
584 daemon running. The Python version doesn't implement the new call.
587 User interface changes
588 ~~~~~~~~~~~~~~~~~~~~~~
590 We have replaced the ``--disks`` option of ``gnt-instance
591 replace-disks`` with a more flexible ``--disk`` option, which allows
592 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
593 disk size and mode can be changed upon recreation (via ``gnt-instance
594 recreate-disks``, which accepts the same ``--disk`` option).
596 As many people are used to a ``show`` command, we have added that as an
597 alias to ``info`` on all ``gnt-*`` commands.
599 The ``gnt-instance grow-disk`` command has a new mode in which it can
600 accept the target size of the disk, instead of the delta; this can be
601 more safe since two runs in absolute mode will be idempotent, and
602 sometimes it's also easier to specify the desired size directly.
604 Also the handling of instances with regard to offline secondaries has
605 been improved. Instance operations should not fail because one of it's
606 secondary nodes is offline, even though it's safe to proceed.
608 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
609 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
610 minors to instance name.
615 RAPI coverage has improved, with (for example) new resources for
616 recreate-disks, node power-cycle, etc.
621 There is partial support for ``xl`` in the Xen hypervisor; feedback is
624 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
625 whether to still support Python 2.4 or move to Python 2.6 as minimum
628 Support for Fedora has been slightly improved; the provided example
629 init.d script should work better on it and the INSTALL file should
630 document the needed dependencies.
635 The deprecated ``QueryLocks`` LUXI request has been removed. Use
636 ``Query(what=QR_LOCK, ...)`` instead.
638 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
639 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
640 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
641 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
642 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
644 RAPI client: ``CertificateError`` now derives from
645 ``GanetiApiError``. This should make it more easy to handle Ganeti
648 Deprecation warnings due to PyCrypto/paramiko import in
649 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
650 make sure to run an up-to-date paramiko version, if you use this tool.
652 The QA scripts now depend on Python 2.5 or above (the main code base
653 still works with Python 2.4).
655 The configuration file (``config.data``) is now written without
656 indentation for performance reasons; if you want to edit it, it can be
657 re-formatted via ``tools/fmtjson``.
659 A number of bugs has been fixed in the cluster merge tool.
661 ``x509`` certification verification (used in import-export) has been
662 changed to allow the same clock skew as permitted by the cluster
663 verification. This will remove some rare but hard to diagnose errors in
670 *(Released Thu, 19 Jul 2012)*
672 Very few changes from rc4 to the final release, only bugfixes:
674 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
675 instance, fix CDROM booting for KVM instances)
676 - fixed node group modification of node parameters
677 - fixed issue in LUClusterVerifyGroup with multi-group clusters
678 - fixed generation of bash completion to ensure a stable ordering
685 *(Released Fri, 13 Jul 2012)*
687 Third release candidate for 2.6. The following changes were done from
690 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
691 - Fixed an inconsistency in the LUXI protocol with the provided
692 arguments (NOT backwards compatible)
693 - Fixed a bug with node groups ipolicy where ``min`` was greater than
694 the cluster ``std`` value
695 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
696 easier instance debugging on nodes (requires ``hconfd`` to work)
702 *(Released Tue, 03 Jul 2012)*
704 Second release candidate for 2.6. The following changes were done from
707 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
709 - Fixed a RAPI regression on missing beparams/memory
710 - Fixed redistribution of files on offline nodes
711 - Added possibility to run activate-disks even though secondaries are
712 offline. With this change it relaxes also the strictness on some other
713 commands which use activate disks internally:
714 * ``gnt-instance start|reboot|rename|backup|export``
715 - Made it possible to remove safely an instance if its secondaries are
717 - Made it possible to reinstall even though secondaries are offline
723 *(Released Mon, 25 Jun 2012)*
725 First release candidate for 2.6. The following changes were done from
728 - Fixed bugs with disk parameters and ``rbd`` templates as well as
730 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
731 behaviour. It supports now the modify operation
732 - ``hcheck`` implemented to analyze cluster health and possibility of
733 improving health by rebalance
734 - ``hbal`` has been improved in dealing with split instances
740 *(Released Mon, 11 Jun 2012)*
742 Second beta release of 2.6. The following changes were done from beta2
745 - Fixed ``daemon-util`` with non-root user models
746 - Fixed creation of plain instances with ``--no-wait-for-sync``
747 - Fix wrong iv_names when running ``cfgupgrade``
748 - Export more information in RAPI group queries
749 - Fixed bug when changing instance network interfaces
750 - Extended burnin to do NIC changes
751 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
752 - Changed default for DRBD barriers
753 - Fixed DRBD error reporting for syncer rate
754 - Verify the options on disk parameters
756 And of course various fixes to documentation and improved unittests and
763 *(Released Wed, 23 May 2012)*
765 First beta release of 2.6. The following changes were done from beta1 to
768 - integrated patch for distributions without ``start-stop-daemon``
769 - adapted example init.d script to work on Fedora
770 - fixed log handling in Haskell daemons
771 - adapted checks in the watcher for pycurl linked against libnss
772 - add partial support for ``xl`` instead of ``xm`` for Xen
773 - fixed a type issue in cluster verification
774 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
777 Plus integrated fixes from the 2.5 branch:
779 - fixed ``kvm-ifup`` to use ``/bin/bash``
780 - fixed parallel build failures
781 - KVM live migration when using a custom keymap
787 *(Released Tue, 24 Jul 2012)*
789 A small bugfix release, with no new features:
791 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
792 different default shell (e.g. Debian, Ubuntu)
793 - fixed KVM startup and live migration with a custom keymap (fixes Issue
794 243 and Debian bug #650664)
795 - fixed compatibility with KVM versions that don't support multiple boot
796 devices (fixes Issue 230 and Debian bug #624256)
798 Additionally, a few fixes were done to the build system (fixed parallel
799 build failures) and to the unittests (fixed race condition in test for
800 FileID functions, and the default enable/disable mode for QA test is now
807 *(Released Fri, 11 May 2012)*
809 A small bugfix release.
811 The main issues solved are on the topic of compatibility with newer LVM
814 - fixed parsing of ``lv_attr`` field
815 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
816 the ``--force`` flag is needed
818 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
819 to accept kernel 3.x too (was hardcoded to 2.6.*).
821 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
822 of hook results and that also made display of other errors suboptimal
823 was fixed; the code behaves now like 2.4 and earlier.
825 Another change in 2.5, the cleanup of the OS scripts environment, is too
826 aggressive: it removed even the ``PATH`` variable, which requires the OS
827 scripts to *always* need to export it. Since this is a bit too strict,
828 we now export a minimal PATH, the same that we export for hooks.
830 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
831 integrated into this release.
833 Finally, a few other miscellaneous changes were done (no new features,
834 just small improvements):
836 - Fix ``gnt-group --help`` display
837 - Fix hardcoded Xen kernel path
838 - Fix grow-disk handling of invalid units
839 - Update synopsis for ``gnt-cluster repair-disk-sizes``
840 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
846 *(Released Thu, 12 Apr 2012)*
848 Incompatible/important changes and bugfixes
849 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
851 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
852 resource's ``ip_check`` parameter changed from ``True`` to ``False``
853 to match the underlying LUXI interface.
854 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
855 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
856 not maintain backwards-compatibility as the underlying operation
857 changed in an incompatible way. The RAPI client can talk to old
858 servers, but it needs to be told so as the return value changed.
859 - When creating file-based instances via RAPI, the ``file_driver``
860 parameter no longer defaults to ``loop`` and must be specified.
861 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
863 - Support for the undocumented and deprecated RAPI instance creation
864 request format version 0 has been dropped. Use version 1, supported
865 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
866 - Pyparsing 1.4.6 or above is required, see :doc:`installation
867 documentation <install>`.
868 - The "cluster-verify" hooks are now executed per group by the
869 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
870 if you just run ``gnt-cluster verify``, which generates one opcode per
872 - The environment as passed to the OS scripts is cleared, and thus no
873 environment variables defined in the node daemon's environment will be
874 inherited by the scripts.
875 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
877 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
878 support operations involving multiple node groups.
879 - Offline nodes are ignored when failing over an instance.
880 - Support for KVM version 1.0, which changed the version reporting format
882 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
884 - ``Makefile`` is now compatible with Automake 1.11.2
885 - Includes all bugfixes made in the 2.4 series
890 - The ganeti-htools project has been merged into the ganeti-core source
891 tree and will be built as part of Ganeti (see :doc:`install-quick`).
892 - Implemented support for :doc:`shared storage <design-shared-storage>`.
893 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
894 GPT-style partition tables (requires `parted
895 <http://www.gnu.org/s/parted/>`_).
896 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
897 - Allowed adding tags on instance creation.
898 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
899 - Allow instances to be started in a paused state, enabling the user to
900 see the complete console output on boot using the console.
901 - Added new hypervisor flag to control default reboot behaviour
902 (``reboot_behavior``).
903 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
904 - Improved out-of-band management support:
906 - Added ``gnt-node health`` command reporting the health status of
908 - Added ``gnt-node power`` command to manage power status of nodes.
909 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
911 - Instance migration can fall back to failover if instance is not
913 - Filters can be used when listing nodes, instances, groups and locks;
914 see :manpage:`ganeti(7)` manpage.
915 - Added post-execution status as variables to :doc:`hooks <hooks>`
917 - Instance tags are exported/imported together with the instance.
918 - When given an explicit job ID, ``gnt-job info`` will work for archived
920 - Jobs can define dependencies on other jobs (not yet supported via
921 RAPI or command line, but used by internal commands and usable via
924 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
927 - Instance failover is now available as a RAPI resource
928 (``/2/instances/[instance_name]/failover``).
929 - ``gnt-instance info`` defaults to static information if primary node
931 - Opcodes have a new ``comment`` attribute.
932 - Added basic SPICE support to KVM hypervisor.
933 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
935 Node group improvements
936 ~~~~~~~~~~~~~~~~~~~~~~~
938 - ``gnt-cluster verify`` has been modified to check groups separately,
939 thereby improving performance.
940 - Node group support has been added to ``gnt-cluster verify-disks``,
941 which now operates per node group.
942 - Watcher has been changed to work better with node groups.
944 - One process and state file per node group.
945 - Slow watcher in one group doesn't block other group's watcher.
947 - Added new command, ``gnt-group evacuate``, to move all instances in a
948 node group to other groups.
949 - Added ``gnt-instance change-group`` to move an instance to another
951 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
952 per-group operations.
953 - Node groups can be tagged.
954 - Some operations switch from an exclusive to a shared lock as soon as
956 - Instance's primary and secondary nodes' groups are now available as
957 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
958 and ``snodes.group.uuid``).
963 - Numerous updates to documentation and manpages.
965 - :doc:`RAPI <rapi>` documentation now has detailed parameter
967 - Some opcode/job results are now also documented, see :doc:`RAPI
970 - A lockset's internal lock is now also visible in lock monitor.
971 - Log messages from job queue workers now contain information about the
972 opcode they're processing.
973 - ``gnt-instance console`` no longer requires the instance lock.
974 - A short delay when waiting for job changes reduces the number of LUXI
975 requests significantly.
976 - DRBD metadata volumes are overwritten with zeros during disk creation.
977 - Out-of-band commands no longer acquire the cluster lock in exclusive
979 - ``devel/upload`` now uses correct permissions for directories.
985 *(Released Fri, 23 Mar 2012)*
987 This was the sixth release candidate of the 2.5 series.
993 *(Released Mon, 9 Jan 2012)*
995 This was the fifth release candidate of the 2.5 series.
1001 *(Released Thu, 27 Oct 2011)*
1003 This was the fourth release candidate of the 2.5 series.
1009 *(Released Wed, 26 Oct 2011)*
1011 This was the third release candidate of the 2.5 series.
1017 *(Released Tue, 18 Oct 2011)*
1019 This was the second release candidate of the 2.5 series.
1025 *(Released Tue, 4 Oct 2011)*
1027 This was the first release candidate of the 2.5 series.
1033 *(Released Wed, 31 Aug 2011)*
1035 This was the third beta release of the 2.5 series.
1041 *(Released Mon, 22 Aug 2011)*
1043 This was the second beta release of the 2.5 series.
1049 *(Released Fri, 12 Aug 2011)*
1051 This was the first beta release of the 2.5 series.
1057 *(Released Thu, 27 Oct 2011)*
1059 - Fixed bug when parsing command line parameter values ending in
1061 - Fixed assertion error after unclean master shutdown
1062 - Disable HTTP client pool for RPC, significantly reducing memory usage
1064 - Fixed queue archive creation with wrong permissions
1070 *(Released Tue, 23 Aug 2011)*
1074 - Fixed documentation for importing with ``--src-dir`` option
1075 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1076 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1082 *(Released Fri, 5 Aug 2011)*
1084 Many bug-fixes and a few small features:
1086 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1087 issues when you tried to add an instance with two MAC addresses in one
1089 - KVM: fixed per-instance stored UID value
1090 - KVM: configure bridged NICs at migration start
1091 - KVM: Fix a bug where instance will not start with never KVM versions
1093 - Added OS search path to ``gnt-cluster info``
1094 - Fixed an issue with ``file_storage_dir`` where you were forced to
1095 provide an absolute path, but the documentation states it is a
1096 relative path, the documentation was right
1097 - Added a new parameter to instance stop/start called ``--no-remember``
1098 that will make the state change to not be remembered
1099 - Implemented ``no_remember`` at RAPI level
1100 - Improved the documentation
1101 - Node evacuation: don't call IAllocator if node is already empty
1102 - Fixed bug in DRBD8 replace disks on current nodes
1103 - Fixed bug in recreate-disks for DRBD instances
1104 - Moved assertion checking locks in ``gnt-instance replace-disks``
1105 causing it to abort with not owning the right locks for some situation
1106 - Job queue: Fixed potential race condition when cancelling queued jobs
1107 - Fixed off-by-one bug in job serial generation
1108 - ``gnt-node volumes``: Fix instance names
1109 - Fixed aliases in bash completion
1110 - Fixed a bug in reopening log files after being sent a SIGHUP
1111 - Added a flag to burnin to allow specifying VCPU count
1112 - Bugfixes to non-root Ganeti configuration
1118 *(Released Thu, 12 May 2011)*
1120 Many bug-fixes and a few new small features:
1122 - Fixed a bug related to log opening failures
1123 - Fixed a bug in instance listing with orphan instances
1124 - Fixed a bug which prevented resetting the cluster-level node parameter
1125 ``oob_program`` to the default
1126 - Many fixes related to the ``cluster-merge`` tool
1127 - Fixed a race condition in the lock monitor, which caused failures
1128 during (at least) creation of many instances in parallel
1129 - Improved output for gnt-job info
1130 - Removed the quiet flag on some ssh calls which prevented debugging
1132 - Improved the N+1 failure messages in cluster verify by actually
1133 showing the memory values (needed and available)
1134 - Increased lock attempt timeouts so that when executing long operations
1135 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1136 too early and thus prevent the use of the 'fair' mechanism
1137 - Changed instance query data (``gnt-instance info``) to not acquire
1138 locks unless needed, thus allowing its use on locked instance if only
1139 static information is asked for
1140 - Improved behaviour with filesystems that do not support rename on an
1142 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1143 kept locks on all nodes during the wipe, which is unneeded
1144 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1145 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1146 10GiB) which caused the wipe to fail right at the end in some cases
1147 - Fixed master IP activation when doing master failover with no-voting
1148 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1149 the master node itself
1150 - Fixed potential data-loss in under disk full conditions, where Ganeti
1151 wouldn't check correctly the return code and would consider
1152 partially-written files 'correct'
1153 - Fixed bug related to multiple VGs and DRBD disk replacing
1154 - Added new disk parameter ``metavg`` that allows placement of the meta
1155 device for DRBD in a different volume group
1156 - Fixed error handling in the node daemon when the system libc doesn't
1157 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1158 - Fixed lock release during replace-disks, which kept cluster-wide locks
1159 when doing disk replaces with an iallocator script
1160 - Added check for missing bridges in cluster verify
1161 - Handle EPIPE errors while writing to the terminal better, so that
1162 piping the output to e.g. ``less`` doesn't cause a backtrace
1163 - Fixed rare case where a ^C during Luxi calls could have been
1164 interpreted as server errors, instead of simply terminating
1165 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1167 - Added a few more parameters to the KVM hypervisor, allowing a second
1168 CDROM, custom disk type for CDROMs and a floppy image
1169 - Removed redundant message in instance rename when the name is given
1171 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1172 disks on new nodes, allowing recreation when the original instance
1173 nodes are completely gone
1174 - Added option when converting disk templates to DRBD to skip waiting
1175 for the resync, in order to make the instance available sooner
1176 - Added two new variables to the OS scripts environment (containing the
1178 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1179 to allow use of ``pvgrub`` as bootloader
1180 - Changed the instance memory modifications to only check out-of-memory
1181 conditions on memory increases, and turned the secondary node warnings
1182 into errors (they can still be overridden via ``--force``)
1183 - Fixed the handling of a corner case when the Python installation gets
1184 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1185 to execute a command that doesn't exist
1186 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1187 node of the instance returned failures during instance shutdown; this
1188 adds the option ``--ignore-consistency`` to gnt-instance move
1190 And as usual, various improvements to the error messages, documentation
1197 *(Released Wed, 09 Mar 2011)*
1199 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1200 the RAPI users file if run twice (even with ``--dry-run``).
1202 The release fixes that bug (nothing else changed).
1208 *(Released Mon, 07 Mar 2011)*
1210 Final 2.4.0 release. Just a few small fixes:
1212 - Fixed RAPI node evacuate
1213 - Fixed the kvm-ifup script
1214 - Fixed internal error handling for special job cases
1215 - Updated man page to specify the escaping feature for options
1221 *(Released Mon, 28 Feb 2011)*
1223 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1224 that this feature wiped the disks of the wrong instance, leading to loss
1229 - Fixed title of query field containing instance name
1230 - Expanded the glossary in the documentation
1231 - Fixed one unittest (internal issue)
1237 *(Released Mon, 21 Feb 2011)*
1239 A number of bug fixes plus just a couple functionality changes.
1241 On the user-visible side, the ``gnt-* list`` command output has changed
1242 with respect to "special" field states. The current rc1 style of display
1243 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1244 the default output mode special fields states are displayed as follows:
1246 - Offline resource: ``*``
1247 - Unavailable/not applicable: ``-``
1248 - Data missing (RPC failure): ``?``
1249 - Unknown field: ``??``
1251 Another user-visible change is the addition of ``--force-join`` to
1256 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1257 - Fixed regression in RAPI/instance reinstall where all parameters were
1258 required (instead of optional)
1259 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1260 - Fixed iallocator usage (offline nodes were not considered offline)
1261 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1262 - Fixed hypervisor and OS parameter validation with respect to
1263 non-vm_capable nodes
1264 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1266 - Fixed ``tools/listrunner`` with respect to agent-based usage
1272 *(Released Fri, 4 Feb 2011)*
1274 Many changes and fixes since the beta1 release. While there were some
1275 internal changes, the code has been mostly stabilised for the RC
1278 Note: the dumb allocator was removed in this release, as it was not kept
1279 up-to-date with the IAllocator protocol changes. It is recommended to
1280 use the ``hail`` command from the ganeti-htools package.
1282 Note: the 2.4 and up versions of Ganeti are not compatible with the
1283 0.2.x branch of ganeti-htools. You need to upgrade to
1284 ganeti-htools-0.3.0 (or later).
1286 Regressions fixed from 2.3
1287 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1289 - Fixed the ``gnt-cluster verify-disks`` command
1290 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1292 - Fixed disk adoption breakage
1293 - Fixed wrong headers in instance listing for field aliases
1298 - Fixed corner case in KVM handling of NICs
1299 - Fixed many cases of wrong handling of non-vm_capable nodes
1300 - Fixed a bug where a missing instance symlink was not possible to
1301 recreate with any ``gnt-*`` command (now ``gnt-instance
1302 activate-disks`` does it)
1303 - Fixed the volume group name as reported by ``gnt-cluster
1305 - Increased timeouts for the import-export code, hopefully leading to
1306 fewer aborts due network or instance timeouts
1307 - Fixed bug in ``gnt-node list-storage``
1308 - Fixed bug where not all daemons were started on cluster
1309 initialisation, but only at the first watcher run
1310 - Fixed many bugs in the OOB implementation
1311 - Fixed watcher behaviour in presence of instances with offline
1313 - Fixed instance list output for instances running on the wrong node
1314 - a few fixes to the cluster-merge tool, but it still cannot merge
1315 multi-node groups (currently it is not recommended to use this tool)
1321 - Improved network configuration for the KVM hypervisor
1322 - Added e1000 as a supported NIC for Xen-HVM
1323 - Improved the lvmstrap tool to also be able to use partitions, as
1324 opposed to full disks
1325 - Improved speed of disk wiping (the cluster parameter
1326 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1327 of instance creations
1328 - Added documentation for the OS parameters
1329 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1330 hypervisor is not responding
1331 - Added display of blacklisted and hidden OS information in
1332 ``gnt-cluster info``
1333 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1334 NIC and node parameters, which might create problems with currently
1335 invalid (but undetected) configuration files, but prevents validation
1336 failures when unrelated parameters are modified
1337 - Changed cluster initialisation to wait for the master daemon to become
1339 - Expanded the RAPI interface:
1341 - Added config redistribution resource
1342 - Added activation/deactivation of instance disks
1343 - Added export of console information
1345 - Implemented log file reopening on SIGHUP, which allows using
1346 logrotate(8) for the Ganeti log files
1347 - Added a basic OOB helper script as an example
1353 *(Released Fri, 14 Jan 2011)*
1358 - Fixed timezone issues when formatting timestamps
1359 - Added support for node groups, available via ``gnt-group`` and other
1361 - Added out-of-band framework and management, see :doc:`design
1362 document <design-oob>`
1363 - Removed support for roman numbers from ``gnt-node list`` and
1364 ``gnt-instance list``.
1365 - Allowed modification of master network interface via ``gnt-cluster
1366 modify --master-netdev``
1367 - Accept offline secondaries while shutting down instance disks
1368 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1369 - Added support for multiple LVM volume groups
1370 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1372 - Added commands to list available fields:
1374 - ``gnt-node list-fields``
1375 - ``gnt-group list-fields``
1376 - ``gnt-instance list-fields``
1378 - Updated documentation and man pages
1383 - Moved ``rapi_users`` file into separate directory, now named
1384 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1386 - Added new tool for running commands on many machines,
1387 ``tools/ganeti-listrunner``
1388 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1389 improving the ``gnt-instance console`` output
1390 - Allowed customisation of disk index separator at ``configure`` time
1391 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1392 - Added support for non-partitioned md disks in ``lvmstrap``
1393 - Added script to gracefully power off KVM instances
1394 - Split ``utils`` module into smaller parts
1395 - Changed query operations to return more detailed information, e.g.
1396 whether an information is unavailable due to an offline node. To use
1397 this new functionality, the LUXI call ``Query`` must be used. Field
1398 information is now stored by the master daemon and can be retrieved
1399 using ``QueryFields``. Instances, nodes and groups can also be queried
1400 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1401 exposed via RAPI). The following commands make use of this
1402 infrastructure change:
1404 - ``gnt-group list``
1405 - ``gnt-group list-fields``
1407 - ``gnt-node list-fields``
1408 - ``gnt-instance list``
1409 - ``gnt-instance list-fields``
1410 - ``gnt-debug locks``
1415 - New RAPI resources (see :doc:`rapi`):
1419 - ``/2/groups/[group_name]``
1420 - ``/2/groups/[group_name]/assign-nodes``
1421 - ``/2/groups/[group_name]/modify``
1422 - ``/2/groups/[group_name]/rename``
1423 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1427 - Implemented ``no_install`` for instance creation
1428 - Implemented OS parameters for instance reinstallation, allowing
1429 use of special settings on reinstallation (e.g. for preserving data)
1434 - Added IPv6 support in import/export
1435 - Pause DRBD synchronization while wiping disks on instance creation
1436 - Updated unittests and QA scripts
1437 - Improved network parameters passed to KVM
1438 - Converted man pages from docbook to reStructuredText
1444 *(Released Mon, 20 Dec 2010)*
1446 Released version 2.3.1~rc1 without any changes.
1452 *(Released Wed, 1 Dec 2010)*
1454 - impexpd: Disable OpenSSL compression in socat if possible (backport
1455 from master, commit e90739d625b, see :doc:`installation guide
1456 <install-quick>` for details)
1457 - Changed unittest coverage report to exclude test scripts
1458 - Added script to check version format
1464 *(Released Wed, 1 Dec 2010)*
1466 Released version 2.3.0~rc1 without any changes.
1472 *(Released Fri, 19 Nov 2010)*
1474 A number of bugfixes and documentation updates:
1476 - Update ganeti-os-interface documentation
1477 - Fixed a bug related to duplicate MACs or similar items which should be
1479 - Fix breakage in OS state modify
1480 - Reinstall instance: disallow offline secondaries (fixes bug related to
1481 OS changing but reinstall failing)
1482 - plus all the other fixes between 2.2.1 and 2.2.2
1488 *(Released Tue, 2 Nov 2010)*
1490 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1491 - Fixed master failover race with watcher
1492 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1494 - Accept previously stopped instance for export with instance removal
1495 - Simplify and extend the environment variables for instance OS scripts
1496 - Added new node flags, ``master_capable`` and ``vm_capable``
1497 - Added optional instance disk wiping prior during allocation. This is a
1498 cluster-wide option and can be set/modified using
1499 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1500 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1501 :doc:`install-quick`
1502 - Added a new watcher option (``--ignore-pause``)
1503 - Added option to ignore offline node on instance start/stop
1504 (``--ignore-offline``)
1505 - Allow overriding OS parameters with ``gnt-instance reinstall``
1506 - Added ability to change node's secondary IP address using ``gnt-node
1508 - Implemented privilege separation for all daemons except
1509 ``ganeti-noded``, see ``configure`` options
1510 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1512 - Implemented job priorities (see ``ganeti(7)`` manpage)
1513 - Ignore failures while shutting down instances during failover from
1515 - Exit daemon's bootstrap process only once daemon is ready
1516 - Export more information via ``LUInstanceQuery``/remote API
1517 - Improved documentation, QA and unittests
1518 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1519 restart if the file was created or changed
1520 - Added LUXI protocol version sent with each request and response,
1521 allowing detection of server/client mismatches
1522 - Moved the Python scripts among gnt-* and ganeti-* into modules
1523 - Moved all code related to setting up SSH to an external script,
1525 - Infrastructure changes for node group support in future versions
1531 *(Released Fri, 19 Nov 2010)*
1533 A few small bugs fixed, and some improvements to the build system:
1535 - Fix documentation regarding conversion to drbd
1536 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1538 - Fix error handling in node modify with multiple changes
1539 - Allow remote imports without checked names
1545 *(Released Tue, 19 Oct 2010)*
1547 - Disable SSL session ID cache in RPC client
1553 *(Released Thu, 14 Oct 2010)*
1555 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1556 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1558 - Fix problems with interaction between Curl and Python's HTTP server,
1559 resulting in increased speed in many RPC calls
1560 - Improve our release script to prevent breakage with older aclocal and
1567 *(Released Thu, 7 Oct 2010)*
1569 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1570 value retrieved from master
1571 - Added support for blacklisted or hidden OS definitions
1572 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1573 - Added support for -mem-path in KVM hypervisor abstraction layer
1574 - Allow overriding instance parameters in tool for inter-cluster
1575 instance moves (``tools/move-instance``)
1576 - Improved opcode summaries (e.g. in ``gnt-job list``)
1577 - Improve consistency of OS listing by sorting it
1578 - Documentation updates
1584 *(Released Fri, 8 Oct 2010)*
1586 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1592 *(Released Mon, 4 Oct 2010)*
1594 - Fixed regression in ``gnt-instance rename``
1600 *(Released Wed, 22 Sep 2010)*
1602 - Fixed OS_VARIANT variable for OS scripts
1603 - Fixed cluster tag operations via RAPI
1604 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1605 - Disabled RAPI CA checks in watcher
1611 *(Released Mon, 23 Aug 2010)*
1613 - Support DRBD versions of the format "a.b.c.d"
1615 - Re-introduce support for usage from multiple threads in RAPI client
1616 - Instance renames and modify via RAPI
1617 - Work around race condition between processing and archival in job
1619 - Mark opcodes following failed one as failed, too
1620 - Job field ``lock_status`` was removed due to difficulties making it
1621 work with the changed job queue in Ganeti 2.2; a better way to monitor
1622 locks is expected for a later 2.2.x release
1623 - Fixed dry-run behaviour with many commands
1624 - Support ``ssh-agent`` again when adding nodes
1625 - Many additional bugfixes
1631 *(Released Fri, 30 Jul 2010)*
1633 Important change: the internal RPC mechanism between Ganeti nodes has
1634 changed from using a home-grown http library (based on the Python base
1635 libraries) to use the PycURL library. This requires that PycURL is
1636 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1637 against GnuTLS by default. cURL's support for GnuTLS had known issues
1638 before cURL 7.21.0 and we recommend using the latest cURL release or
1639 linking against OpenSSL. Most other distributions already link PycURL
1640 and cURL against OpenSSL. The command::
1642 python -c 'import pycurl; print pycurl.version'
1644 can be used to determine the libraries PycURL and cURL are linked
1647 Other significant changes:
1649 - Rewrote much of the internals of the job queue, in order to achieve
1650 better parallelism; this decouples job query operations from the job
1651 processing, and it should allow much nicer behaviour of the master
1652 daemon under load, and it also has uncovered some long-standing bugs
1653 related to the job serialisation (now fixed)
1654 - Added a default iallocator setting to the cluster parameters,
1655 eliminating the need to always pass nodes or an iallocator for
1656 operations that require selection of new node(s)
1657 - Added experimental support for the LXC virtualization method
1658 - Added support for OS parameters, which allows the installation of
1659 instances to pass parameter to OS scripts in order to customise the
1661 - Added a hypervisor parameter controlling the migration type (live or
1662 non-live), since hypervisors have various levels of reliability; this
1663 has renamed the 'live' parameter to 'mode'
1664 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1665 logical volumes, meaning that cluster verify will ignore them and not
1666 flag their presence as errors
1667 - The watcher will now reset the error count for failed instances after
1668 8 hours, thus allowing self-healing if the problem that caused the
1669 instances to be down/fail to start has cleared in the meantime
1670 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1671 check for, and warn, if the drbd module parameter ``usermode_helper``
1672 is not consistent with the cluster-wide setting; this is needed to
1673 make diagnose easier of failed drbd creations
1674 - Started adding base IPv6 support, but this is not yet
1675 enabled/available for use
1676 - Rename operations (cluster, instance) will now return the new name,
1677 which is especially useful if a short name was passed in
1678 - Added support for instance migration in RAPI
1679 - Added a tool to pre-configure nodes for the SSH setup, before joining
1680 them to the cluster; this will allow in the future a simplified model
1681 for node joining (but not yet fully enabled in 2.2); this needs the
1682 paramiko python library
1683 - Fixed handling of name-resolving errors
1684 - Fixed consistency of job results on the error path
1685 - Fixed master-failover race condition when executed multiple times in
1687 - Fixed many bugs related to the job queue (mostly introduced during the
1688 2.2 development cycle, so not all are impacting 2.1)
1689 - Fixed instance migration with missing disk symlinks
1690 - Fixed handling of unknown jobs in ``gnt-job archive``
1691 - And many other small fixes/improvements
1695 - Enhanced both the unittest and the QA coverage
1696 - Switched the opcode validation to a generic model, and extended the
1697 validation to all opcode parameters
1698 - Changed more parts of the code that write shell scripts to use the
1700 - Switched the master daemon to use the asyncore library for the Luxi
1707 *(Released Thu, 17 Jun 2010)*
1709 - Added tool (``move-instance``) and infrastructure to move instances
1710 between separate clusters (see :doc:`separate documentation
1711 <move-instance>` and :doc:`design document <design-2.2>`)
1712 - Added per-request RPC timeout
1713 - RAPI now requires a Content-Type header for requests with a body (e.g.
1714 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1715 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1716 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1718 - Implemented initial support for running Ganeti daemons as separate
1719 users, see configure-time flags ``--with-user-prefix`` and
1720 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1722 - Instances can be removed after export (``gnt-backup export
1723 --remove-instance``)
1724 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1725 key (instead of 1024 bit)
1726 - Added new cluster configuration file for cluster domain secret
1727 - Import/export now use SSL instead of SSH
1728 - Added support for showing estimated time when exporting an instance,
1729 see the ``ganeti-os-interface(7)`` manpage and look for
1736 *(Released Tue, 16 Nov 2010)*
1738 Some more bugfixes. Unless critical bugs occur, this will be the last
1741 - Fix case of MAC special-values
1742 - Fix mac checker regex
1743 - backend: Fix typo causing "out of range" error
1744 - Add missing --units in gnt-instance list man page
1750 *(Released Tue, 24 Aug 2010)*
1753 - Don't ignore secondary node silently on non-mirrored disk templates
1755 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1756 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1757 - Properly document the usb_mouse parameter
1758 - Fix path in ganeti-rapi(8) (issue 116)
1759 - Adjust error message when the ganeti user's .ssh directory is
1761 - Add same-node-check when changing the disk template to drbd
1767 *(Released Fri, 16 Jul 2010)*
1770 - Add an option to only select some reboot types during qa/burnin.
1771 (on some hypervisors consequent reboots are not supported)
1772 - Fix infrequent race condition in master failover. Sometimes the old
1773 master ip address would be still detected as up for a short time
1774 after it was removed, causing failover to fail.
1775 - Decrease mlockall warnings when the ctypes module is missing. On
1776 Python 2.4 we support running even if no ctypes module is installed,
1777 but we were too verbose about this issue.
1778 - Fix building on old distributions, on which man doesn't have a
1780 - Fix RAPI not to ignore the MAC address on instance creation
1781 - Implement the old instance creation format in the RAPI client.
1787 *(Released Thu, 01 Jul 2010)*
1789 A small bugfix release:
1790 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1791 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1792 a non-existing option
1793 - Fix instance create: the --force-variant option was ignored
1794 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1795 - Fix modify node storage with non-FQDN arguments
1796 - Fix RAPI client to authenticate under Python 2.6 when used
1797 for more than 5 requests needing authentication
1798 - Fix gnt-instance modify -t (storage) giving a wrong error message
1799 when converting a non-shutdown drbd instance to plain
1805 *(Released Fri, 18 Jun 2010)*
1807 A small bugfix release:
1809 - Fix live migration of KVM instances started with older Ganeti
1810 versions which had fewer hypervisor parameters
1811 - Fix gnt-instance grow-disk on down instances
1812 - Fix an error-reporting bug during instance migration
1813 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1814 silently ignoring broken ones
1815 - Fix an RPC error reporting bug affecting, for example, RAPI client
1817 - Fix bug triggered by different API version os-es on different nodes
1818 - Fix a bug in instance startup with custom hvparams: OS level
1819 parameters would fail to be applied.
1820 - Fix the RAPI client under Python 2.6 (but more work is needed to
1821 make it work completely well with OpenSSL)
1822 - Fix handling of errors when resolving names from DNS
1828 *(Released Thu, 3 Jun 2010)*
1830 A medium sized development cycle. Some new features, and some
1831 fixes/small improvements/cleanups.
1833 Significant features
1834 ~~~~~~~~~~~~~~~~~~~~
1836 The node deamon now tries to mlock itself into memory, unless the
1837 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1838 its logs, and falls back to console logging. This allows emergency
1839 features such as ``gnt-node powercycle`` to work even in the event of a
1840 broken node disk (tested offlining the disk hosting the node's
1841 filesystem and dropping its memory caches; don't try this at home)
1843 KVM: add vhost-net acceleration support. It can be tested with a new
1844 enough version of the kernel and of qemu-kvm.
1846 KVM: Add instance chrooting feature. If you use privilege dropping for
1847 your VMs you can also now force them to chroot to an empty directory,
1848 before starting the emulated guest.
1850 KVM: Add maximum migration bandwith and maximum downtime tweaking
1851 support (requires a new-enough version of qemu-kvm).
1853 Cluster verify will now warn if the master node doesn't have the master
1854 ip configured on it.
1856 Add a new (incompatible) instance creation request format to RAPI which
1857 supports all parameters (previously only a subset was supported, and it
1858 wasn't possible to extend the old format to accomodate all the new
1859 features. The old format is still supported, and a client can check for
1860 this feature, before using it, by checking for its presence in the
1861 ``features`` RAPI resource.
1863 Now with ancient latin support. Try it passing the ``--roman`` option to
1864 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1865 (requires the python-roman module to be installed, in order to work).
1870 As usual many internal code refactorings, documentation updates, and
1873 - Lots of improvements and cleanups to the experimental Remote API
1874 (RAPI) client library.
1875 - A new unit test suite for the core daemon libraries.
1876 - A fix to creating missing directories makes sure the umask is not
1877 applied anymore. This enforces the same directory permissions
1879 - Better handling terminating daemons with ctrl+c (used when running
1880 them in debugging mode).
1881 - Fix a race condition in live migrating a KVM instance, when stat()
1882 on the old proc status file returned EINVAL, which is an unexpected
1884 - Fixed manpage checking with newer man and utf-8 charachters. But now
1885 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1891 *(Released Fri, 7 May 2010)*
1893 Fix a bug which prevented untagged KVM instances from starting.
1899 *(Released Fri, 7 May 2010)*
1901 Another release with a long development cycle, during which many
1902 different features were added.
1904 Significant features
1905 ~~~~~~~~~~~~~~~~~~~~
1907 The KVM hypervisor now can run the individual instances as non-root, to
1908 reduce the impact of a VM being hijacked due to bugs in the
1909 hypervisor. It is possible to run all instances as a single (non-root)
1910 user, to manually specify a user for each instance, or to dynamically
1911 allocate a user out of a cluster-wide pool to each instance, with the
1912 guarantee that no two instances will run under the same user ID on any
1915 An experimental RAPI client library, that can be used standalone
1916 (without the other Ganeti libraries), is provided in the source tree as
1917 ``lib/rapi/client.py``. Note this client might change its interface in
1918 the future, as we iterate on its capabilities.
1920 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1921 replace the cluster's certificates and crypto keys. This might help in
1922 case they have been compromised, or have simply expired.
1924 A new disk option for instance creation has been added that allows one
1925 to "adopt" currently existing logical volumes, with data
1926 preservation. This should allow easier migration to Ganeti from
1927 unmanaged (or managed via other software) instances.
1929 Another disk improvement is the possibility to convert between redundant
1930 (DRBD) and plain (LVM) disk configuration for an instance. This should
1931 allow better scalability (starting with one node and growing the
1932 cluster, or shrinking a two-node cluster to one node).
1934 A new feature that could help with automated node failovers has been
1935 implemented: if a node sees itself as offline (by querying the master
1936 candidates), it will try to shutdown (hard) all instances and any active
1937 DRBD devices. This reduces the risk of duplicate instances if an
1938 external script automatically failovers the instances on such nodes. To
1939 enable this, the cluster parameter ``maintain_node_health`` should be
1940 enabled; in the future this option (per the name) will enable other
1941 automatic maintenance features.
1943 Instance export/import now will reuse the original instance
1944 specifications for all parameters; that means exporting an instance,
1945 deleting it and the importing it back should give an almost identical
1946 instance. Note that the default import behaviour has changed from
1947 before, where it created only one NIC; now it recreates the original
1950 Cluster verify has added a few new checks: SSL certificates validity,
1951 /etc/hosts consistency across the cluster, etc.
1956 As usual, many internal changes were done, documentation fixes,
1959 - Fixed cluster initialization with disabled cluster storage (regression
1960 introduced in 2.1.1)
1961 - File-based storage supports growing the disks
1962 - Fixed behaviour of node role changes
1963 - Fixed cluster verify for some corner cases, plus a general rewrite of
1964 cluster verify to allow future extension with more checks
1965 - Fixed log spamming by watcher and node daemon (regression introduced
1967 - Fixed possible validation issues when changing the list of enabled
1969 - Fixed cleanup of /etc/hosts during node removal
1970 - Fixed RAPI response for invalid methods
1971 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1972 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1973 from ide disks, etc.)
1974 - Allow OS changes without re-installation (to record a changed OS
1975 outside of Ganeti, or to allow OS renames)
1976 - Allow instance creation without OS installation (useful for example if
1977 the OS will be installed manually, or restored from a backup not in
1979 - Implemented option to make cluster ``copyfile`` use the replication
1981 - Added list of enabled hypervisors to ssconf (possibly useful for
1983 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1985 - A partial form of node re-IP is possible via node readd, which now
1986 allows changed node primary IP
1987 - Command line utilities now show an informational message if the job is
1989 - The logs of the master daemon now show the PID/UID/GID of the
1996 *(Released Fri, 12 Mar 2010)*
1998 During the 2.1.0 long release candidate cycle, a lot of improvements and
1999 changes have accumulated with were released later as 2.1.1.
2004 The node evacuate command (``gnt-node evacuate``) was significantly
2005 rewritten, and as such the IAllocator protocol was changed - a new
2006 request type has been added. This unfortunate change during a stable
2007 series is designed to improve performance of node evacuations; on
2008 clusters with more than about five nodes and which are well-balanced,
2009 evacuation should proceed in parallel for all instances of the node
2010 being evacuated. As such, any existing IAllocator scripts need to be
2011 updated, otherwise the above command will fail due to the unknown
2012 request. The provided "dumb" allocator has not been updated; but the
2013 ganeti-htools package supports the new protocol since version 0.2.4.
2015 Another important change is increased validation of node and instance
2016 names. This might create problems in special cases, if invalid host
2017 names are being used.
2019 Also, a new layer of hypervisor parameters has been added, that sits at
2020 OS level between the cluster defaults and the instance ones. This allows
2021 customisation of virtualization parameters depending on the installed
2022 OS. For example instances with OS 'X' may have a different KVM kernel
2023 (or any other parameter) than the cluster defaults. This is intended to
2024 help managing a multiple OSes on the same cluster, without manual
2025 modification of each instance's parameters.
2027 A tool for merging clusters, ``cluster-merge``, has been added in the
2028 tools sub-directory.
2033 - Improved the int/float conversions that should make the code more
2034 robust in face of errors from the node daemons
2035 - Fixed the remove node code in case of internal configuration errors
2036 - Fixed the node daemon behaviour in face of inconsistent queue
2037 directory (e.g. read-only file-system where we can't open the files
2039 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2040 now it either aborts cleanly or, if given the new "auto_promote"
2041 parameter, will automatically promote other nodes as needed
2042 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2043 completely prevent Ganeti from working
2044 - Fixed bug for instance export when not all disks were successfully
2046 - Fixed behaviour of node add when the new node is slow in starting up
2048 - Fixed handling of signals in the LUXI client, which should improve
2049 behaviour of command-line scripts
2050 - Added checks for invalid node/instance names in the configuration (now
2051 flagged during cluster verify)
2052 - Fixed watcher behaviour for disk activation errors
2053 - Fixed two potentially endless loops in http library, which led to the
2054 RAPI daemon hanging and consuming 100% CPU in some cases
2055 - Fixed bug in RAPI daemon related to hashed passwords
2056 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2058 - Enhanced compatibility with non-Debian OSes, but not using absolute
2059 path in some commands and allowing customisation of the ssh
2060 configuration directory
2061 - Fixed possible future issue with new Python versions by abiding to the
2062 proper use of ``__slots__`` attribute on classes
2063 - Added checks that should prevent directory traversal attacks
2064 - Many documentation fixes based on feedback from users
2069 - Added an "early_release" more for instance replace disks and node
2070 evacuate, where we release locks earlier and thus allow higher
2071 parallelism within the cluster
2072 - Added watcher hooks, intended to allow the watcher to restart other
2073 daemons (e.g. from the ganeti-nbma project), but they can be used of
2074 course for any other purpose
2075 - Added a compile-time disable for DRBD barriers, to increase
2076 performance if the administrator trusts the power supply or the
2077 storage system to not lose writes
2078 - Added the option of using syslog for logging instead of, or in
2079 addition to, Ganeti's own log files
2080 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2081 can indeed boot from a paravirtual NIC
2082 - Added a generic debug level for many operations; while this is not
2083 used widely yet, it allows one to pass the debug value all the way to
2085 - Enhanced the hooks environment for instance moves (failovers,
2086 migrations) where the primary/secondary nodes changed during the
2087 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2088 - Enhanced data validations for many user-supplied values; one important
2089 item is the restrictions imposed on instance and node names, which
2090 might reject some (invalid) host names
2091 - Add a configure-time option to disable file-based storage, if it's not
2092 needed; this allows greater security separation between the master
2093 node and the other nodes from the point of view of the inter-node RPC
2095 - Added user notification in interactive tools if job is waiting in the
2096 job queue or trying to acquire locks
2097 - Added log messages when a job is waiting for locks
2098 - Added filtering by node tags in instance operations which admit
2099 multiple instances (start, stop, reboot, reinstall)
2100 - Added a new tool for cluster mergers, ``cluster-merge``
2101 - Parameters from command line which are of the form ``a=b,c=d`` can now
2102 use backslash escapes to pass in values which contain commas,
2103 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2105 - For KVM, the instance name is the first parameter passed to KVM, so
2106 that it's more visible in the process list
2112 *(Released Tue, 2 Mar 2010)*
2114 Ganeti 2.1 brings many improvements with it. Major changes:
2116 - Added infrastructure to ease automated disk repairs
2117 - Added new daemon to export configuration data in a cheaper way than
2118 using the remote API
2119 - Instance NICs can now be routed instead of being associated with a
2121 - Improved job locking logic to reduce impact of jobs acquiring multiple
2122 locks waiting for other long-running jobs
2124 In-depth implementation details can be found in the Ganeti 2.1 design
2130 - Added chroot hypervisor
2131 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2133 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2134 ``bootloader_path`` and ``bootloader_args``)
2135 - Added the ``use_localtime`` option for the xen-hvm and kvm
2136 hypervisors, and the default value for this has changed to false (in
2137 2.0 xen-hvm always enabled it)
2138 - Added luxi call to submit multiple jobs in one go
2139 - Added cluster initialization option to not modify ``/etc/hosts``
2141 - Added network interface parameters
2142 - Added dry run mode to some LUs
2143 - Added RAPI resources:
2145 - ``/2/instances/[instance_name]/info``
2146 - ``/2/instances/[instance_name]/replace-disks``
2147 - ``/2/nodes/[node_name]/evacuate``
2148 - ``/2/nodes/[node_name]/migrate``
2149 - ``/2/nodes/[node_name]/role``
2150 - ``/2/nodes/[node_name]/storage``
2151 - ``/2/nodes/[node_name]/storage/modify``
2152 - ``/2/nodes/[node_name]/storage/repair``
2154 - Added OpCodes to evacuate or migrate all instances on a node
2155 - Added new command to list storage elements on nodes (``gnt-node
2156 list-storage``) and modify them (``gnt-node modify-storage``)
2157 - Added new ssconf files with master candidate IP address
2158 (``ssconf_master_candidates_ips``), node primary IP address
2159 (``ssconf_node_primary_ips``) and node secondary IP address
2160 (``ssconf_node_secondary_ips``)
2161 - Added ``ganeti-confd`` and a client library to query the Ganeti
2162 configuration via UDP
2163 - Added ability to run hooks after cluster initialization and before
2165 - Added automatic mode for disk replace (``gnt-instance replace-disks
2167 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2168 after catastrophic data-loss
2169 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2171 - Added ``gnt-instance move`` command to move instances
2172 - Added ``gnt-cluster watcher`` command to control watcher
2173 - Added ``gnt-node powercycle`` command to powercycle nodes
2174 - Added new job status field ``lock_status``
2175 - Added parseable error codes to cluster verification (``gnt-cluster
2176 verify --error-codes``) and made output less verbose (use
2177 ``--verbose`` to restore previous behaviour)
2178 - Added UUIDs to the main config entities (cluster, nodes, instances)
2179 - Added support for OS variants
2180 - Added support for hashed passwords in the Ganeti remote API users file
2182 - Added option to specify maximum timeout on instance shutdown
2183 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2184 - Added new helper script to start and stop Ganeti daemons
2185 (``daemon-util``), with the intent to reduce the work necessary to
2186 adjust Ganeti for non-Debian distributions and to start/stop daemons
2188 - Added more unittests
2189 - Fixed critical bug in ganeti-masterd startup
2190 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2191 now customisable at the cluster level for both the KVM and Xen
2192 hypervisors using the new ``migration_port`` parameter
2193 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2194 reinstalling an instance
2195 - Allowed ``@`` in tag names
2196 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2197 - Many documentation updates
2198 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2199 - ``gnt-instance reinstall`` can now reinstall multiple instances
2200 - Updated many command line parameters
2201 - Introduced new OS API version 15
2202 - No longer support a default hypervisor
2203 - Treat virtual LVs as inexistent
2204 - Improved job locking logic to reduce lock contention
2205 - Match instance and node names case insensitively
2206 - Reimplemented bash completion script to be more complete
2213 *(Released Thu, 4 Feb 2010)*
2215 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2216 - Fix a string formatting bug
2217 - Improve safety of the code in some error paths
2218 - Improve data validation in the master of values returned from nodes
2224 *(Released Thu, 17 Dec 2009)*
2226 - Fix security issue due to missing validation of iallocator names; this
2227 allows local and remote execution of arbitrary executables
2228 - Fix failure of gnt-node list during instance removal
2229 - Ship the RAPI documentation in the archive
2235 *(Released Wed, 30 Sep 2009)*
2237 - Fixed many wrong messages
2238 - Fixed a few bugs related to the locking library
2239 - Fixed MAC checking at instance creation time
2240 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2241 - Fixed a few issues related to signal handling in both daemons and
2243 - Fixed the example startup script provided
2244 - Fixed insserv dependencies in the example startup script (patch from
2246 - Fixed handling of drained nodes in the iallocator framework
2247 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2249 - Fixed error related to invalid job IDs in job polling
2250 - Fixed job/opcode persistence on unclean master shutdown
2251 - Fixed handling of partial job processing after unclean master
2253 - Fixed error reporting from LUs, previously all errors were converted
2254 into execution errors
2255 - Fixed error reporting from burnin
2256 - Decreased significantly the memory usage of the job queue
2257 - Optimised slightly multi-job submission
2258 - Optimised slightly opcode loading
2259 - Backported the multi-job submit framework from the development
2260 branch; multi-instance start and stop should be faster
2261 - Added script to clean archived jobs after 21 days; this will reduce
2262 the size of the queue directory
2263 - Added some extra checks in disk size tracking
2264 - Added an example ethers hook script
2265 - Added a cluster parameter that prevents Ganeti from modifying of
2267 - Added more node information to RAPI responses
2268 - Added a ``gnt-job watch`` command that allows following the ouput of a
2270 - Added a bind-address option to ganeti-rapi
2271 - Added more checks to the configuration verify
2272 - Enhanced the burnin script such that some operations can be retried
2274 - Converted instance reinstall to multi-instance model
2280 *(Released Fri, 7 Aug 2009)*
2282 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2283 to allow using the pre-2.0.2 behaviour in activation, if any existing
2284 instances have mismatched disk sizes in the configuration
2285 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2286 any configuration mismatches for disk sizes
2287 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2288 failover to work on two-node clusters
2289 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2291 - Fixed detection of OS script errors in ``gnt-backup export``
2292 - Fixed exit code of ``gnt-backup export``
2298 *(Released Fri, 17 Jul 2009)*
2300 - Added experimental support for stripped logical volumes; this should
2301 enhance performance but comes with a higher complexity in the block
2302 device handling; stripping is only enabled when passing
2303 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2304 affected even in the non-stripped mode
2305 - Improved resiliency against transient failures at the end of DRBD
2306 resyncs, and in general of DRBD resync checks
2307 - Fixed a couple of issues with exports and snapshot errors
2308 - Fixed a couple of issues in instance listing
2309 - Added display of the disk size in ``gnt-instance info``
2310 - Fixed checking for valid OSes in instance creation
2311 - Fixed handling of the "vcpus" parameter in instance listing and in
2312 general of invalid parameters
2313 - Fixed http server library, and thus RAPI, to handle invalid
2314 username/password combinations correctly; this means that now they
2315 report unauthorized for queries too, not only for modifications,
2316 allowing earlier detect of configuration problems
2317 - Added a new "role" node list field, equivalent to the master/master
2318 candidate/drained/offline flags combinations
2319 - Fixed cluster modify and changes of candidate pool size
2320 - Fixed cluster verify error messages for wrong files on regular nodes
2321 - Fixed a couple of issues with node demotion from master candidate role
2322 - Fixed node readd issues
2323 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2324 - Added a new ``--no-voting`` option for masterfailover to fix failover
2325 on two-nodes clusters when the former master node is unreachable
2326 - Added instance reinstall over RAPI
2332 *(Released Tue, 16 Jun 2009)*
2334 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2335 allow re-adding the start in single-user option (regression from 1.2)
2336 - the watcher writes the instance status to a file, to allow monitoring
2337 to report the instance status (from the master) based on cached
2338 results of the watcher's queries; while this can get stale if the
2339 watcher is being locked due to other work on the cluster, this is
2340 still an improvement
2341 - the watcher now also restarts the node daemon and the rapi daemon if
2343 - fixed the watcher to handle full and drained queue cases
2344 - hooks export more instance data in the environment, which helps if
2345 hook scripts need to take action based on the instance's properties
2346 (no longer need to query back into ganeti)
2347 - instance failovers when the instance is stopped do not check for free
2348 RAM, so that failing over a stopped instance is possible in low memory
2350 - rapi uses queries for tags instead of jobs (for less job traffic), and
2351 for cluster tags it won't talk to masterd at all but read them from
2353 - a couple of error handling fixes in RAPI
2354 - drbd handling: improved the error handling of inconsistent disks after
2355 resync to reduce the frequency of "there are some degraded disks for
2356 this instance" messages
2357 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2358 error handling path called a wrong function name)
2364 *(Released Wed, 27 May 2009)*
2366 - no changes from rc5
2372 *(Released Wed, 20 May 2009)*
2374 - fix a couple of bugs (validation, argument checks)
2375 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2376 - some small improvements to RAPI and IAllocator
2377 - make watcher automatically start the master daemon if down
2383 *(Released Mon, 27 Apr 2009)*
2385 - change the OS list to not require locks; this helps with big clusters
2386 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2387 volume group is broken
2388 - ``gnt-instance info``, without any arguments, doesn't run for all
2389 instances anymore; either pass ``--all`` or pass the desired
2390 instances; this helps against mistakes on big clusters where listing
2391 the information for all instances takes a long time
2392 - miscellaneous doc and man pages fixes
2398 *(Released Wed, 8 Apr 2009)*
2400 - Change the internal locking model of some ``gnt-node`` commands, in
2401 order to reduce contention (and blocking of master daemon) when
2402 batching many creation/reinstall jobs
2403 - Fixes to Xen soft reboot
2404 - No longer build documentation at build time, instead distribute it in
2405 the archive, in order to reduce the need for the whole docbook/rst
2412 *(Released Fri, 27 Mar 2009)*
2414 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2415 - Fix watcher startup sequence, improves the behaviour of busy clusters
2416 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2417 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2418 bind address and other places
2419 - Some documentation fixes and updates
2425 *(Released Mon, 2 Mar 2009)*
2427 - More documentation updates, now all docs should be more-or-less
2429 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2431 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2433 - Fix an issue related to $libdir/run/ganeti and cluster creation
2439 *(Released Thu, 19 Feb 2009)*
2441 - Xen PVM and KVM have switched the default value for the instance root
2442 disk to the first partition on the first drive, instead of the whole
2443 drive; this means that the OS installation scripts must be changed
2445 - Man pages have been updated
2446 - RAPI has been switched by default to HTTPS, and the exported functions
2447 should all work correctly
2448 - RAPI v1 has been removed
2449 - Many improvements to the KVM hypervisor
2450 - Block device errors are now better reported
2451 - Many other bugfixes and small improvements
2457 *(Released Mon, 26 Jan 2009)*
2459 - Version 2 is a general rewrite of the code and therefore the
2460 differences are too many to list, see the design document for 2.0 in
2461 the ``doc/`` subdirectory for more details
2462 - In this beta version there is not yet a migration path from 1.2 (there
2463 will be one in the final 2.0 release)
2464 - A few significant changes are:
2466 - all commands are executed by a daemon (``ganeti-masterd``) and the
2467 various ``gnt-*`` commands are just front-ends to it
2468 - all the commands are entered into, and executed from a job queue,
2469 see the ``gnt-job(8)`` manpage
2470 - the RAPI daemon supports read-write operations, secured by basic
2471 HTTP authentication on top of HTTPS
2472 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2473 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2474 to migrate to DRBD 8 first while still running Ganeti 1.2)
2475 - DRBD devices are using statically allocated minor numbers, which
2476 will be assigned to existing instances during the migration process
2477 - there is support for both Xen PVM and Xen HVM instances running on
2479 - KVM virtualization is supported too
2480 - file-based storage has been implemented, which means that it is
2481 possible to run the cluster without LVM and DRBD storage, for
2482 example using a shared filesystem exported from shared storage (and
2483 still have live migration)
2489 *(Released Tue, 13 Jan 2009)*
2491 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2492 - Reuse the old instance mac address by default on instance import, if
2493 the instance name is the same.
2494 - Handle situations in which the node info rpc returns incomplete
2496 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2497 - Improved version of batcher:
2499 - state file support
2500 - instance mac address support
2501 - support for HVM clusters/instances
2503 - Add an option to show the number of cpu sockets and nodes in
2505 - Support OSes that handle more than one version of the OS api (but do
2506 not change the current API in any other way)
2507 - Fix ``gnt-node migrate``
2508 - ``gnt-debug`` man page
2509 - Fixes various more typos and small issues
2510 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2516 *(Released Wed, 24 Sep 2008)*
2518 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2519 type of disk exported to fully virtualized instances.
2520 - provide access to the serial console of HVM instances
2521 - instance auto_balance flag, set by default. If turned off it will
2522 avoid warnings on cluster verify if there is not enough memory to fail
2523 over an instance. in the future it will prevent automatically failing
2524 it over when we will support that.
2525 - batcher tool for instance creation, see ``tools/README.batcher``
2526 - ``gnt-instance reinstall --select-os`` to interactively select a new
2527 operating system when reinstalling an instance.
2528 - when changing the memory amount on instance modify a check has been
2529 added that the instance will be able to start. also warnings are
2530 emitted if the instance will not be able to fail over, if auto_balance
2532 - documentation fixes
2533 - sync fields between ``gnt-instance list/modify/add/import``
2534 - fix a race condition in drbd when the sync speed was set after giving
2535 the device a remote peer.
2541 *(Released Tue, 22 Jul 2008)*
2543 - note: the allowed size and number of tags per object were reduced
2544 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2545 - fixed twisted 8.x compatibility
2546 - fixed ``gnt-instance replace-disks`` with iallocator
2547 - add TCP keepalives on twisted connections to detect restarted nodes
2548 - disk increase support, see ``gnt-instance grow-disk``
2549 - implement bulk node/instance query for RAPI
2550 - add tags in node/instance listing (optional)
2551 - experimental migration (and live migration) support, read the man page
2552 for ``gnt-instance migrate``
2553 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2554 has some small improvements in handling its state file
2560 *(Released Fri, 13 Jun 2008)*
2562 - Experimental readonly, REST-based remote API implementation;
2563 automatically started on master node, TCP port 5080, if enabled by
2564 ``--enable-rapi`` parameter to configure script.
2565 - Instance allocator support. Add and import instance accept a
2566 ``--iallocator`` parameter, and call that instance allocator to decide
2567 which node to use for the instance. The iallocator document describes
2568 what's expected from an allocator script.
2569 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2570 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2571 node is lost there is still enough memory to fail over the instances
2573 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2574 ``gnt-cluster verify``, to check for site-specific compliance. All the
2575 hooks will run, and their output, if any, will be displayed. Any
2576 failing hook will make the verification return an error value.
2577 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2578 primary and secondary interfaces
2579 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2580 that is still declared as part of the cluster and has failed.
2581 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2582 specifying output fields, that just adds the chosen fields to the
2584 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2585 export from the filesystem.
2586 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2587 have been added. Using them you can enable/disable acpi and pae
2588 support, and specify a path for a cd image to be exported to the
2589 instance. These parameters as the name suggest only work on HVM
2591 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2592 PAE support will be set to the previously hardcoded values, but the
2593 (previously hardcoded) path to the CDROM ISO image will be unset and
2594 if required, needs to be set manually with ``gnt-instance modify``
2596 - The address to which an instance's VNC console is bound is now
2597 selectable per-instance, rather than being cluster wide. Of course
2598 this only applies to instances controlled via VNC, so currently just
2599 applies to HVM clusters.
2605 *(Released Mon, 18 Feb 2008)*
2607 - more tweaks to the disk activation code (especially helpful for DRBD)
2608 - change the default ``gnt-instance list`` output format, now there is
2609 one combined status field (see the manpage for the exact values this
2611 - some more fixes for the mac export to hooks change
2612 - make Ganeti not break with DRBD 8.2.x (which changed the version
2613 format in ``/proc/drbd``) (issue 24)
2614 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2615 template, allowing migration from DRBD0.7+MD to DRBD8
2621 *(Released Wed, 30 Jan 2008)*
2623 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2625 - add command aliases infrastructure and a few aliases
2626 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2627 man pages and the ``--help`` option of ``gnt-node
2628 list``/``gnt-instance list``
2629 - fix ``gnt-backup list`` with down nodes (issue 21)
2630 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2631 - fix the dist archive and add a check for including svn/git files in
2633 - some developer-related changes: improve the burnin and the QA suite,
2634 add an upload script for testing during development
2640 *(Released Wed, 16 Jan 2008)*
2642 - experimental HVM support, read the install document, section
2643 "Initializing the cluster"
2644 - allow for the PVM hypervisor per-instance kernel and initrd paths
2645 - add a new command ``gnt-cluster verify-disks`` which uses a new
2646 algorithm to improve the reconnection of the DRBD pairs if the device
2647 on the secondary node has gone away
2648 - make logical volume code auto-activate LVs at disk activation time
2649 - slightly improve the speed of activating disks
2650 - allow specification of the MAC address at instance creation time, and
2651 changing it later via ``gnt-instance modify``
2652 - fix handling of external commands that generate lots of output on
2654 - update documentation with regard to minimum version of DRBD8 supported
2660 *(Released Tue, 4 Dec 2007)*
2662 - Log the ``xm create`` output to the node daemon log on failure (to
2663 help diagnosing the error)
2664 - In debug mode, log all external commands output if failed to the logs
2665 - Change parsing of lvm commands to ignore stderr
2671 *(Released Wed, 28 Nov 2007)*
2673 - Another round of updates to the DRBD 8 code to deal with more failures
2674 in the replace secondary node operation
2675 - Some more logging of failures in disk operations (lvm, drbd)
2676 - A few documentation updates
2683 *(Released Tue, 13 Nov 2007)*
2685 - Change configuration file format from Python's Pickle to JSON.
2686 Upgrading is possible using the cfgupgrade utility.
2687 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2688 faster replace disks and is more stable (DRBD 8 has many improvements
2689 compared to DRBD 0.7)
2690 - Added command line tags support (see man pages for ``gnt-instance``,
2691 ``gnt-node``, ``gnt-cluster``)
2692 - Added instance rename support
2693 - Added multi-instance startup/shutdown
2694 - Added cluster rename support
2695 - Added ``gnt-node evacuate`` to simplify some node operations
2696 - Added instance reboot operation that can speedup reboot as compared to
2698 - Soften the requirement that hostnames are in FQDN format
2699 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2701 - Removed dependency on debian's patched fping that uses the
2702 non-standard ``-S`` option
2703 - Now the OS definitions are searched for in multiple, configurable
2704 paths (easier for distros to package)
2705 - Some changes to the hooks infrastructure (especially the new
2706 post-configuration update hook)
2707 - Other small bugfixes
2709 .. vim: set textwidth=72 syntax=rst :