10 - Change the connected groups format in ``gnt-network info`` output; it
11 was previously displayed as a raw list by mistake.
17 *(Released Thu, 25 Jul 2013)*
19 - Add logrotate functionality in daemon-util
20 - Add logrotate example file
21 - Add missing fields to network queries over rapi
22 - Fix network object timestamps
23 - Add support for querying network timestamps
24 - Fix a typo in the example crontab
25 - Fix a documentation typo
31 *(Released Thu, 04 Jul 2013)*
33 Incompatible/important changes
34 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
36 - Instance policies for disk size were documented to be on a per-disk
37 basis, but hail applied them to the sum of all disks. This has been
39 - ``hbal`` will now exit with status 0 if, during job execution over
40 LUXI, early exit has been requested and all jobs are successful;
41 before, exit status 1 was used, which cannot be differentiated from
43 - Compatibility with newer versions of rbd has been fixed
44 - ``gnt-instance batch-create`` has been changed to use the bulk create
45 opcode from Ganeti. This lead to incompatible changes in the format of
46 the JSON file. It's now not a custom dict anymore but a dict
47 compatible with the ``OpInstanceCreate`` opcode.
48 - Parent directories for file storage need to be listed in
49 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
50 write the file automatically based on old configuration values, but it
51 can not distribute it across all nodes and the file contents should be
52 verified. Use ``gnt-cluster copyfile
53 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
54 upgraded. The reason for requiring this list of paths now is that
55 before it would have been possible to inject new paths via RPC,
56 allowing files to be created in arbitrary locations. The RPC protocol
57 is protected using SSL/X.509 certificates, but as a design principle
58 Ganeti does not permit arbitrary paths to be passed.
59 - The parsing of the variants file for OSes (see
60 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
61 lines and comment lines (starting with ``#``) are ignored for better
63 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
64 longer available. ``gnt-node add`` now invokes a new tool on the
65 destination node, named ``prepare-node-join``, to configure the SSH
66 daemon. Paramiko is no longer necessary to configure nodes' SSH
67 daemons via ``gnt-node add``.
68 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
69 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
70 the flag is not reset after a master failover.
71 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
72 is recommended. 2.6 will be mandatory from the 2.8 series.
78 - New network management functionality to support automatic allocation
79 of IP addresses and managing of network parameters. See
80 :manpage:`gnt-network(8)` for more details.
81 - New external storage backend, to allow managing arbitrary storage
82 systems external to the cluster. See
83 :manpage:`ganeti-extstorage-interface(7)`.
84 - New ``exclusive-storage`` node parameter added, restricted to
85 nodegroup level. When it's set to true, physical disks are assigned in
86 an exclusive fashion to instances, as documented in :doc:`Partitioned
87 Ganeti <design-partitioned>`. Currently, only instances using the
88 ``plain`` disk template are supported.
89 - The KVM hypervisor has been updated with many new hypervisor
90 parameters, including a generic one for passing arbitrary command line
91 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
92 compatible up to qemu 1.4.
93 - A new tool, called ``mon-collector``, is the stand-alone executor of
94 the data collectors for a monitoring system. As of this version, it
95 just includes the DRBD data collector, that can be executed by calling
96 ``mon-collector`` using the ``drbd`` parameter. See
97 :manpage:`mon-collector(7)`.
98 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
99 for RAPI users. It allows granting permissions to query for
100 information to a specific user without giving
101 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
102 - A new tool named ``node-cleanup`` has been added. It cleans remains of
103 a cluster from a machine by stopping all daemons, removing
104 certificates and ssconf files. Unless the ``--no-backup`` option is
105 given, copies of the certificates are made.
106 - Instance creations now support the use of opportunistic locking,
107 potentially speeding up the (parallel) creation of multiple instances.
108 This feature is currently only available via the :doc:`RAPI
109 <rapi>` interface and when an instance allocator is used. If the
110 ``opportunistic_locking`` parameter is set the opcode will try to
111 acquire as many locks as possible, but will not wait for any locks
112 held by other opcodes. If not enough resources can be found to
113 allocate the instance, the temporary error code
114 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
115 retried thereafter, with or without opportunistic locking.
116 - New experimental linux-ha resource scripts.
117 - Restricted-commands support: ganeti can now be asked (via command line
118 or rapi) to perform commands on a node. These are passed via ganeti
119 RPC rather than ssh. This functionality is restricted to commands
120 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
121 reasons. The file is not copied automatically.
127 - Diskless instances are now externally mirrored (Issue 237). This for
128 now has only been tested in conjunction with explicit target nodes for
130 - Queries not needing locks or RPC access to the node can now be
131 performed by the confd daemon, making them independent from jobs, and
132 thus faster to execute. This is selectable at configure time.
133 - The functionality for allocating multiple instances at once has been
134 overhauled and is now also available through :doc:`RAPI <rapi>`.
136 There are no significant changes from version 2.7.0~rc3.
142 *(Released Tue, 25 Jun 2013)*
144 - Fix permissions on the confd query socket (Issue 477)
145 - Fix permissions on the job archive dir (Issue 498)
146 - Fix handling of an internal exception in replace-disks (Issue 472)
147 - Fix gnt-node info handling of shortened names (Issue 497)
148 - Fix gnt-instance grow-disk when wiping is enabled
149 - Documentation improvements, and support for newer pandoc
150 - Fix hspace honoring ipolicy for disks (Issue 484)
151 - Improve handling of the ``kvm_extra`` HV parameter
157 *(Released Fri, 24 May 2013)*
159 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
161 - Disks added through ``gnt-instance modify`` or created through
162 ``gnt-instance recreate-disks`` are wiped, if the
163 ``prealloc_wipe_disks`` flag is set.
164 - If wiping newly created disks fails, the disks are removed. Also,
165 partial failures in creating disks through ``gnt-instance modify``
166 triggers a cleanup of the partially-created disks.
167 - Removing the master IP address doesn't fail if the address has been
169 - Fix ownership of the OS log dir
170 - Workaround missing SO_PEERCRED constant (Issue 191)
176 *(Released Fri, 3 May 2013)*
178 This was the first release candidate of the 2.7 series. Since beta3:
180 - Fix kvm compatibility with qemu 1.4 (Issue 389)
181 - Documentation updates (admin guide, upgrade notes, install
182 instructions) (Issue 372)
183 - Fix gnt-group list nodes and instances count (Issue 436)
184 - Fix compilation without non-mandatory libraries (Issue 441)
185 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
186 - Make confd logging more verbose at INFO level (Issue 435)
187 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
188 - Fix failure path for instance storage type conversion (Issue 229)
189 - Update htools text backend documentation
190 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
191 - Disable inter-cluster instance move for file-based instances, because
192 it is dependant on instance export, which is not supported for
193 file-based instances. (Issue 414)
194 - Fix gnt-job crashes on non-ascii characters (Issue 427)
195 - Fix volume group checks on non-vm-capable nodes (Issue 432)
201 *(Released Mon, 22 Apr 2013)*
203 This was the third beta release of the 2.7 series. Since beta2:
205 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
206 - Fix data loss on wrong usage of ``gnt-instance move``
207 - Properly export errors in confd-based job queries
208 - Add ``users-setup`` tool
209 - Fix iallocator protocol to report 0 as a disk size for diskless
210 instances. This avoids hail breaking when a diskless instance is
212 - Fix job queue directory permission problem that made confd job queries
213 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
214 for access to archived jobs (Issue 406).
215 - Limit the sizes of networks supported by ``gnt-network`` to something
216 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
217 - Fix bugs in instance disk template conversion
218 - Fix GHC 7 compatibility
219 - Fix ``burnin`` install path (Issue 426).
220 - Allow very small disk grows (Issue 347).
221 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
222 that noded doesn't import masterd code (Issue 419).
223 - Make sure the default metavg at cluster init is the same as the vg, if
224 unspecified (Issue 358).
225 - Fix cleanup of partially created disks (part of Issue 416)
231 *(Released Tue, 2 Apr 2013)*
233 This was the second beta release of the 2.7 series. Since beta1:
235 - Networks no longer have a "type" slot, since this information was
236 unused in Ganeti: instead of it tags should be used.
237 - The rapi client now has a ``target_node`` option to MigrateInstance.
238 - Fix early exit return code for hbal (Issue 386).
239 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
240 - Fix ``rbd showmapped`` output parsing (Issue 312).
241 - Networks are now referenced indexed by UUID, rather than name. This
242 will require running cfgupgrade, from 2.7.0beta1, if networks are in
244 - The OS environment now includes network information.
245 - Deleting of a network is now disallowed if any instance nic is using
246 it, to prevent dangling references.
247 - External storage is now documented in man pages.
248 - The exclusive_storage flag can now only be set at nodegroup level.
249 - Hbal can now submit an explicit priority with its jobs.
250 - Many network related locking fixes.
251 - Bump up the required pylint version to 0.25.1.
252 - Fix the ``no_remember`` option in RAPI client.
253 - Many ipolicy related tests, qa, and fixes.
254 - Many documentation improvements and fixes.
255 - Fix building with ``--disable-file-storage``.
256 - Fix ``-q`` option in htools, which was broken if passed more than
258 - Some haskell/python interaction improvements and fixes.
259 - Fix iallocator in case of missing LVM storage.
260 - Fix confd config load in case of ``--no-lvm-storage``.
261 - The confd/query functionality is now mentioned in the security
268 *(Released Wed, 6 Feb 2013)*
270 This was the first beta release of the 2.7 series. All important changes
271 are listed in the latest 2.7 entry.
277 *(Released Fri, 21 Dec 2012)*
279 Important behaviour change: hbal won't rebalance anymore instances which
280 have the ``auto_balance`` attribute set to false. This was the intention
281 all along, but until now it only skipped those from the N+1 memory
282 reservation (DRBD-specific).
284 A significant number of bug fixes in this release:
286 - Fixed disk adoption interaction with ipolicy checks.
287 - Fixed networking issues when instances are started, stopped or
288 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
289 - Fixed the warning in cluster verify for shared storage instances not
291 - Fixed removal of storage directory on shared file storage (issue 262).
292 - Fixed validation of LVM volume group name in OpClusterSetParams
293 (``gnt-cluster modify``) (issue 285).
294 - Fixed runtime memory increases (``gnt-instance modify -m``).
295 - Fixed live migration under Xen's ``xl`` mode.
296 - Fixed ``gnt-instance console`` with ``xl``.
297 - Fixed building with newer Haskell compiler/libraries.
298 - Fixed PID file writing in Haskell daemons (confd); this prevents
299 restart issues if confd was launched manually (outside of
300 ``daemon-util``) while another copy of it was running
301 - Fixed a type error when doing live migrations with KVM (issue 297) and
302 the error messages for failing migrations have been improved.
303 - Fixed opcode validation for the out-of-band commands (``gnt-node
305 - Fixed a type error when unsetting OS hypervisor parameters (issue
306 311); now it's possible to unset all OS-specific hypervisor
308 - Fixed the ``dry-run`` mode for many operations: verification of
309 results was over-zealous but didn't take into account the ``dry-run``
310 operation, resulting in "wrong" failures.
311 - Fixed bash completion in ``gnt-job list`` when the job queue has
312 hundreds of entries; especially with older ``bash`` versions, this
313 results in significant CPU usage.
315 And lastly, a few other improvements have been made:
317 - Added option to force master-failover without voting (issue 282).
318 - Clarified error message on lock conflict (issue 287).
319 - Logging of newly submitted jobs has been improved (issue 290).
320 - Hostname checks have been made uniform between instance rename and
322 - The ``--submit`` option is now supported by ``gnt-debug delay``.
323 - Shutting down the master daemon by sending SIGTERM now stops it from
324 processing jobs waiting for locks; instead, those jobs will be started
325 once again after the master daemon is started the next time (issue
327 - Support for Xen's ``xl`` program has been improved (besides the fixes
329 - Reduced logging noise in the Haskell confd daemon (only show one log
330 entry for each config reload, instead of two).
331 - Several man page updates and typo fixes.
337 *(Released Fri, 12 Oct 2012)*
339 A small bugfix release. Among the bugs fixed:
341 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
342 made the command unusable.
343 - Commands that issue many jobs don't fail anymore just because some jobs
344 take so long that other jobs are archived.
345 - Failures during ``gnt-instance reinstall`` are reflected by the exit
347 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
349 - When ``always_failover`` is set, ``--allow-failover`` is not required
350 in migrate commands anymore.
351 - ``bash_completion`` works even if extglob is disabled.
352 - Fixed bug with locks that made failover for RDB-based instances fail.
353 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
354 a random node instead of one based on the allocator metric.
355 - Support for newer versions of pylint and pep8.
356 - Hail doesn't fail anymore when trying to add an instance of type
357 ``file``, ``sharedfile`` or ``rbd``.
358 - Added new Makefile target to rebuild the whole distribution, so that
359 all files are included.
365 *(Released Fri, 27 Jul 2012)*
368 .. attention:: The ``LUXI`` protocol has been made more consistent
369 regarding its handling of command arguments. This, however, leads to
370 incompatibility issues with previous versions. Please ensure that you
371 restart Ganeti daemons soon after the upgrade, otherwise most
372 ``LUXI`` calls (job submission, setting/resetting the drain flag,
373 pausing/resuming the watcher, cancelling and archiving jobs, querying
374 the cluster configuration) will fail.
383 The current ``admin_up`` field, which used to denote whether an instance
384 should be running or not, has been removed. Instead, ``admin_state`` is
385 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
387 The rational behind this is that an instance being “down” can have
390 - it could be down during a reboot
391 - it could be temporarily be down for a reinstall
392 - or it could be down because it is deprecated and kept just for its
395 The previous Boolean state was making it difficult to do capacity
396 calculations: should Ganeti reserve memory for a down instance? Now, the
397 tri-state field makes it clear:
399 - in ``up`` and ``down`` state, all resources are reserved for the
400 instance, and it can be at any time brought up if it is down
401 - in ``offline`` state, only disk space is reserved for it, but not
404 The field can have an extra use: since the transition between ``up`` and
405 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
406 transition between ``offline`` and ``down`` is done via ``gnt-instance
407 modify``, it is possible to given different rights to users. For
408 example, owners of an instance could be allowed to start/stop it, but
409 not transition it out of the offline state.
411 Instance policies and specs
412 +++++++++++++++++++++++++++
414 In previous Ganeti versions, an instance creation request was not
415 limited on the minimum size and on the maximum size just by the cluster
416 resources. As such, any policy could be implemented only in third-party
417 clients (RAPI clients, or shell wrappers over ``gnt-*``
418 tools). Furthermore, calculating cluster capacity via ``hspace`` again
419 required external input with regards to instance sizes.
421 In order to improve these workflows and to allow for example better
422 per-node group differentiation, we introduced instance specs, which
425 - minimum instance disk size, disk count, memory size, cpu count
426 - maximum values for the above metrics
427 - and “standard” values (used in ``hspace`` to calculate the standard
430 The minimum/maximum values can be also customised at node-group level,
431 for example allowing more powerful hardware to support bigger instance
434 Beside the instance specs, there are a few other settings belonging to
435 the instance policy framework. It is possible now to customise, per
436 cluster and node-group:
438 - the list of allowed disk templates
439 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
440 - the maximum ratio of instance to spindles (see below for more
441 information) for local storage
443 All these together should allow all tools that talk to Ganeti to know
444 what are the ranges of allowed values for instances and the
445 over-subscription that is allowed.
447 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
448 instance configuration, and the physical CPU configuration from the
449 node. For the spindle ratios however, we didn't track before these
450 values, so new parameters have been added:
452 - a new node parameter ``spindle_count``, defaults to 1, customisable at
453 node group or node level
454 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
456 Note that spindles in this context doesn't need to mean actual
457 mechanical hard-drives; it's just a relative number for both the node
458 I/O capacity and instance I/O consumption.
460 Instance migration behaviour
461 ++++++++++++++++++++++++++++
463 While live-migration is in general desirable over failover, it is
464 possible that for some workloads it is actually worse, due to the
465 variable time of the “suspend” phase during live migration.
467 To allow the tools to work consistently over such instances (without
468 having to hard-code instance names), a new backend parameter
469 ``always_failover`` has been added to control the migration/failover
470 behaviour. When set to True, all migration requests for an instance will
471 instead fall-back to failover.
473 Instance memory ballooning
474 ++++++++++++++++++++++++++
476 Initial support for memory ballooning has been added. The memory for an
477 instance is no longer fixed (backend parameter ``memory``), but instead
478 can vary between minimum and maximum values (backend parameters
479 ``minmem`` and ``maxmem``). Currently we only change an instance's
482 - live migrating or failing over and instance and the target node
483 doesn't have enough memory
484 - user requests changing the memory via ``gnt-instance modify
490 In order to control the use of specific CPUs by instance, support for
491 controlling CPU pinning has been added for the Xen, HVM and LXC
492 hypervisors. This is controlled by a new hypervisor parameter
493 ``cpu_mask``; details about possible values for this are in the
494 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
495 VCPU-to-CPU mapping) form will work well only when all nodes in your
496 cluster have the same amount of CPUs.
501 Another area in which Ganeti was not customisable were the parameters
502 used for storage configuration, e.g. how many stripes to use for LVM,
503 DRBD resync configuration, etc.
505 To improve this area, we've added disks parameters, which are
506 customisable at cluster and node group level, and which allow to
507 specify various parameters for disks (DRBD has the most parameters
508 currently), for example:
510 - DRBD resync algorithm and parameters (e.g. speed)
511 - the default VG for meta-data volumes for DRBD
512 - number of stripes for LVM (plain disk template)
515 These parameters can be modified via ``gnt-cluster modify -D …`` and
516 ``gnt-group modify -D …``, and are used at either instance creation (in
517 case of LVM stripes, for example) or at disk “activation” time
520 Rados block device support
521 ++++++++++++++++++++++++++
523 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
524 denoted by the ``rbd`` disk template type. This is considered
525 experimental, feedback is welcome. For details on configuring it, see
526 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
531 The existing master IP functionality works well only in simple setups (a
532 single network shared by all nodes); however, if nodes belong to
533 different networks, then the ``/32`` setup and lack of routing
534 information is not enough.
536 To allow the master IP to function well in more complex cases, the
537 system was reworked as follows:
539 - a master IP netmask setting has been added
540 - the master IP activation/turn-down code was moved from the node daemon
542 - whether to run the Ganeti-supplied master IP script or a user-supplied
543 on is a ``gnt-cluster init`` setting
545 Details about the location of the standard and custom setup scripts are
546 in the man page :manpage:`gnt-cluster(8)`; for information about the
547 setup script protocol, look at the Ganeti-supplied script.
552 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
555 It is now possible to use TLS-protected connections, and when renewing
556 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
557 it is now possible to specify spice or spice CA certificates. Also, it
558 is possible to configure a password for SPICE sessions via the
559 hypervisor parameter ``spice_password_file``.
561 There are also new parameters to control the compression and streaming
562 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
563 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
564 for the spice parameters.
566 Lastly, it is now possible to see the SPICE connection information via
567 ``gnt-instance console``.
572 A new tool (``tools/ovfconverter``) has been added that supports
573 conversion between Ganeti and the `Open Virtualization Format
574 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
577 This relies on the ``qemu-img`` tool to convert the disk formats, so the
578 actual compatibility with other virtualization solutions depends on it.
583 The configuration query daemon (``ganeti-confd``) is now optional, and
584 has been rewritten in Haskell; whether to use the daemon at all, use the
585 Python (default) or the Haskell version is selectable at configure time
586 via the ``--enable-confd`` parameter, which can take one of the
587 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
588 daemon will result in a smaller footprint; for larger systems, we
589 welcome feedback on the Haskell version which might become the default
592 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
593 daemon running. The Python version doesn't implement the new call.
596 User interface changes
597 ~~~~~~~~~~~~~~~~~~~~~~
599 We have replaced the ``--disks`` option of ``gnt-instance
600 replace-disks`` with a more flexible ``--disk`` option, which allows
601 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
602 disk size and mode can be changed upon recreation (via ``gnt-instance
603 recreate-disks``, which accepts the same ``--disk`` option).
605 As many people are used to a ``show`` command, we have added that as an
606 alias to ``info`` on all ``gnt-*`` commands.
608 The ``gnt-instance grow-disk`` command has a new mode in which it can
609 accept the target size of the disk, instead of the delta; this can be
610 more safe since two runs in absolute mode will be idempotent, and
611 sometimes it's also easier to specify the desired size directly.
613 Also the handling of instances with regard to offline secondaries has
614 been improved. Instance operations should not fail because one of it's
615 secondary nodes is offline, even though it's safe to proceed.
617 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
618 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
619 minors to instance name.
624 RAPI coverage has improved, with (for example) new resources for
625 recreate-disks, node power-cycle, etc.
630 There is partial support for ``xl`` in the Xen hypervisor; feedback is
633 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
634 whether to still support Python 2.4 or move to Python 2.6 as minimum
637 Support for Fedora has been slightly improved; the provided example
638 init.d script should work better on it and the INSTALL file should
639 document the needed dependencies.
644 The deprecated ``QueryLocks`` LUXI request has been removed. Use
645 ``Query(what=QR_LOCK, ...)`` instead.
647 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
648 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
649 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
650 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
651 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
653 RAPI client: ``CertificateError`` now derives from
654 ``GanetiApiError``. This should make it more easy to handle Ganeti
657 Deprecation warnings due to PyCrypto/paramiko import in
658 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
659 make sure to run an up-to-date paramiko version, if you use this tool.
661 The QA scripts now depend on Python 2.5 or above (the main code base
662 still works with Python 2.4).
664 The configuration file (``config.data``) is now written without
665 indentation for performance reasons; if you want to edit it, it can be
666 re-formatted via ``tools/fmtjson``.
668 A number of bugs has been fixed in the cluster merge tool.
670 ``x509`` certification verification (used in import-export) has been
671 changed to allow the same clock skew as permitted by the cluster
672 verification. This will remove some rare but hard to diagnose errors in
679 *(Released Thu, 19 Jul 2012)*
681 Very few changes from rc4 to the final release, only bugfixes:
683 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
684 instance, fix CDROM booting for KVM instances)
685 - fixed node group modification of node parameters
686 - fixed issue in LUClusterVerifyGroup with multi-group clusters
687 - fixed generation of bash completion to ensure a stable ordering
694 *(Released Fri, 13 Jul 2012)*
696 Third release candidate for 2.6. The following changes were done from
699 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
700 - Fixed an inconsistency in the LUXI protocol with the provided
701 arguments (NOT backwards compatible)
702 - Fixed a bug with node groups ipolicy where ``min`` was greater than
703 the cluster ``std`` value
704 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
705 easier instance debugging on nodes (requires ``hconfd`` to work)
711 *(Released Tue, 03 Jul 2012)*
713 Second release candidate for 2.6. The following changes were done from
716 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
718 - Fixed a RAPI regression on missing beparams/memory
719 - Fixed redistribution of files on offline nodes
720 - Added possibility to run activate-disks even though secondaries are
721 offline. With this change it relaxes also the strictness on some other
722 commands which use activate disks internally:
723 * ``gnt-instance start|reboot|rename|backup|export``
724 - Made it possible to remove safely an instance if its secondaries are
726 - Made it possible to reinstall even though secondaries are offline
732 *(Released Mon, 25 Jun 2012)*
734 First release candidate for 2.6. The following changes were done from
737 - Fixed bugs with disk parameters and ``rbd`` templates as well as
739 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
740 behaviour. It supports now the modify operation
741 - ``hcheck`` implemented to analyze cluster health and possibility of
742 improving health by rebalance
743 - ``hbal`` has been improved in dealing with split instances
749 *(Released Mon, 11 Jun 2012)*
751 Second beta release of 2.6. The following changes were done from beta2
754 - Fixed ``daemon-util`` with non-root user models
755 - Fixed creation of plain instances with ``--no-wait-for-sync``
756 - Fix wrong iv_names when running ``cfgupgrade``
757 - Export more information in RAPI group queries
758 - Fixed bug when changing instance network interfaces
759 - Extended burnin to do NIC changes
760 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
761 - Changed default for DRBD barriers
762 - Fixed DRBD error reporting for syncer rate
763 - Verify the options on disk parameters
765 And of course various fixes to documentation and improved unittests and
772 *(Released Wed, 23 May 2012)*
774 First beta release of 2.6. The following changes were done from beta1 to
777 - integrated patch for distributions without ``start-stop-daemon``
778 - adapted example init.d script to work on Fedora
779 - fixed log handling in Haskell daemons
780 - adapted checks in the watcher for pycurl linked against libnss
781 - add partial support for ``xl`` instead of ``xm`` for Xen
782 - fixed a type issue in cluster verification
783 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
786 Plus integrated fixes from the 2.5 branch:
788 - fixed ``kvm-ifup`` to use ``/bin/bash``
789 - fixed parallel build failures
790 - KVM live migration when using a custom keymap
796 *(Released Tue, 24 Jul 2012)*
798 A small bugfix release, with no new features:
800 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
801 different default shell (e.g. Debian, Ubuntu)
802 - fixed KVM startup and live migration with a custom keymap (fixes Issue
803 243 and Debian bug #650664)
804 - fixed compatibility with KVM versions that don't support multiple boot
805 devices (fixes Issue 230 and Debian bug #624256)
807 Additionally, a few fixes were done to the build system (fixed parallel
808 build failures) and to the unittests (fixed race condition in test for
809 FileID functions, and the default enable/disable mode for QA test is now
816 *(Released Fri, 11 May 2012)*
818 A small bugfix release.
820 The main issues solved are on the topic of compatibility with newer LVM
823 - fixed parsing of ``lv_attr`` field
824 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
825 the ``--force`` flag is needed
827 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
828 to accept kernel 3.x too (was hardcoded to 2.6.*).
830 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
831 of hook results and that also made display of other errors suboptimal
832 was fixed; the code behaves now like 2.4 and earlier.
834 Another change in 2.5, the cleanup of the OS scripts environment, is too
835 aggressive: it removed even the ``PATH`` variable, which requires the OS
836 scripts to *always* need to export it. Since this is a bit too strict,
837 we now export a minimal PATH, the same that we export for hooks.
839 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
840 integrated into this release.
842 Finally, a few other miscellaneous changes were done (no new features,
843 just small improvements):
845 - Fix ``gnt-group --help`` display
846 - Fix hardcoded Xen kernel path
847 - Fix grow-disk handling of invalid units
848 - Update synopsis for ``gnt-cluster repair-disk-sizes``
849 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
855 *(Released Thu, 12 Apr 2012)*
857 Incompatible/important changes and bugfixes
858 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
860 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
861 resource's ``ip_check`` parameter changed from ``True`` to ``False``
862 to match the underlying LUXI interface.
863 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
864 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
865 not maintain backwards-compatibility as the underlying operation
866 changed in an incompatible way. The RAPI client can talk to old
867 servers, but it needs to be told so as the return value changed.
868 - When creating file-based instances via RAPI, the ``file_driver``
869 parameter no longer defaults to ``loop`` and must be specified.
870 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
872 - Support for the undocumented and deprecated RAPI instance creation
873 request format version 0 has been dropped. Use version 1, supported
874 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
875 - Pyparsing 1.4.6 or above is required, see :doc:`installation
876 documentation <install>`.
877 - The "cluster-verify" hooks are now executed per group by the
878 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
879 if you just run ``gnt-cluster verify``, which generates one opcode per
881 - The environment as passed to the OS scripts is cleared, and thus no
882 environment variables defined in the node daemon's environment will be
883 inherited by the scripts.
884 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
886 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
887 support operations involving multiple node groups.
888 - Offline nodes are ignored when failing over an instance.
889 - Support for KVM version 1.0, which changed the version reporting format
891 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
893 - ``Makefile`` is now compatible with Automake 1.11.2
894 - Includes all bugfixes made in the 2.4 series
899 - The ganeti-htools project has been merged into the ganeti-core source
900 tree and will be built as part of Ganeti (see :doc:`install-quick`).
901 - Implemented support for :doc:`shared storage <design-shared-storage>`.
902 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
903 GPT-style partition tables (requires `parted
904 <http://www.gnu.org/s/parted/>`_).
905 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
906 - Allowed adding tags on instance creation.
907 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
908 - Allow instances to be started in a paused state, enabling the user to
909 see the complete console output on boot using the console.
910 - Added new hypervisor flag to control default reboot behaviour
911 (``reboot_behavior``).
912 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
913 - Improved out-of-band management support:
915 - Added ``gnt-node health`` command reporting the health status of
917 - Added ``gnt-node power`` command to manage power status of nodes.
918 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
920 - Instance migration can fall back to failover if instance is not
922 - Filters can be used when listing nodes, instances, groups and locks;
923 see :manpage:`ganeti(7)` manpage.
924 - Added post-execution status as variables to :doc:`hooks <hooks>`
926 - Instance tags are exported/imported together with the instance.
927 - When given an explicit job ID, ``gnt-job info`` will work for archived
929 - Jobs can define dependencies on other jobs (not yet supported via
930 RAPI or command line, but used by internal commands and usable via
933 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
936 - Instance failover is now available as a RAPI resource
937 (``/2/instances/[instance_name]/failover``).
938 - ``gnt-instance info`` defaults to static information if primary node
940 - Opcodes have a new ``comment`` attribute.
941 - Added basic SPICE support to KVM hypervisor.
942 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
944 Node group improvements
945 ~~~~~~~~~~~~~~~~~~~~~~~
947 - ``gnt-cluster verify`` has been modified to check groups separately,
948 thereby improving performance.
949 - Node group support has been added to ``gnt-cluster verify-disks``,
950 which now operates per node group.
951 - Watcher has been changed to work better with node groups.
953 - One process and state file per node group.
954 - Slow watcher in one group doesn't block other group's watcher.
956 - Added new command, ``gnt-group evacuate``, to move all instances in a
957 node group to other groups.
958 - Added ``gnt-instance change-group`` to move an instance to another
960 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
961 per-group operations.
962 - Node groups can be tagged.
963 - Some operations switch from an exclusive to a shared lock as soon as
965 - Instance's primary and secondary nodes' groups are now available as
966 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
967 and ``snodes.group.uuid``).
972 - Numerous updates to documentation and manpages.
974 - :doc:`RAPI <rapi>` documentation now has detailed parameter
976 - Some opcode/job results are now also documented, see :doc:`RAPI
979 - A lockset's internal lock is now also visible in lock monitor.
980 - Log messages from job queue workers now contain information about the
981 opcode they're processing.
982 - ``gnt-instance console`` no longer requires the instance lock.
983 - A short delay when waiting for job changes reduces the number of LUXI
984 requests significantly.
985 - DRBD metadata volumes are overwritten with zeros during disk creation.
986 - Out-of-band commands no longer acquire the cluster lock in exclusive
988 - ``devel/upload`` now uses correct permissions for directories.
994 *(Released Fri, 23 Mar 2012)*
996 This was the sixth release candidate of the 2.5 series.
1002 *(Released Mon, 9 Jan 2012)*
1004 This was the fifth release candidate of the 2.5 series.
1010 *(Released Thu, 27 Oct 2011)*
1012 This was the fourth release candidate of the 2.5 series.
1018 *(Released Wed, 26 Oct 2011)*
1020 This was the third release candidate of the 2.5 series.
1026 *(Released Tue, 18 Oct 2011)*
1028 This was the second release candidate of the 2.5 series.
1034 *(Released Tue, 4 Oct 2011)*
1036 This was the first release candidate of the 2.5 series.
1042 *(Released Wed, 31 Aug 2011)*
1044 This was the third beta release of the 2.5 series.
1050 *(Released Mon, 22 Aug 2011)*
1052 This was the second beta release of the 2.5 series.
1058 *(Released Fri, 12 Aug 2011)*
1060 This was the first beta release of the 2.5 series.
1066 *(Released Thu, 27 Oct 2011)*
1068 - Fixed bug when parsing command line parameter values ending in
1070 - Fixed assertion error after unclean master shutdown
1071 - Disable HTTP client pool for RPC, significantly reducing memory usage
1073 - Fixed queue archive creation with wrong permissions
1079 *(Released Tue, 23 Aug 2011)*
1083 - Fixed documentation for importing with ``--src-dir`` option
1084 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1085 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1091 *(Released Fri, 5 Aug 2011)*
1093 Many bug-fixes and a few small features:
1095 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1096 issues when you tried to add an instance with two MAC addresses in one
1098 - KVM: fixed per-instance stored UID value
1099 - KVM: configure bridged NICs at migration start
1100 - KVM: Fix a bug where instance will not start with never KVM versions
1102 - Added OS search path to ``gnt-cluster info``
1103 - Fixed an issue with ``file_storage_dir`` where you were forced to
1104 provide an absolute path, but the documentation states it is a
1105 relative path, the documentation was right
1106 - Added a new parameter to instance stop/start called ``--no-remember``
1107 that will make the state change to not be remembered
1108 - Implemented ``no_remember`` at RAPI level
1109 - Improved the documentation
1110 - Node evacuation: don't call IAllocator if node is already empty
1111 - Fixed bug in DRBD8 replace disks on current nodes
1112 - Fixed bug in recreate-disks for DRBD instances
1113 - Moved assertion checking locks in ``gnt-instance replace-disks``
1114 causing it to abort with not owning the right locks for some situation
1115 - Job queue: Fixed potential race condition when cancelling queued jobs
1116 - Fixed off-by-one bug in job serial generation
1117 - ``gnt-node volumes``: Fix instance names
1118 - Fixed aliases in bash completion
1119 - Fixed a bug in reopening log files after being sent a SIGHUP
1120 - Added a flag to burnin to allow specifying VCPU count
1121 - Bugfixes to non-root Ganeti configuration
1127 *(Released Thu, 12 May 2011)*
1129 Many bug-fixes and a few new small features:
1131 - Fixed a bug related to log opening failures
1132 - Fixed a bug in instance listing with orphan instances
1133 - Fixed a bug which prevented resetting the cluster-level node parameter
1134 ``oob_program`` to the default
1135 - Many fixes related to the ``cluster-merge`` tool
1136 - Fixed a race condition in the lock monitor, which caused failures
1137 during (at least) creation of many instances in parallel
1138 - Improved output for gnt-job info
1139 - Removed the quiet flag on some ssh calls which prevented debugging
1141 - Improved the N+1 failure messages in cluster verify by actually
1142 showing the memory values (needed and available)
1143 - Increased lock attempt timeouts so that when executing long operations
1144 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1145 too early and thus prevent the use of the 'fair' mechanism
1146 - Changed instance query data (``gnt-instance info``) to not acquire
1147 locks unless needed, thus allowing its use on locked instance if only
1148 static information is asked for
1149 - Improved behaviour with filesystems that do not support rename on an
1151 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1152 kept locks on all nodes during the wipe, which is unneeded
1153 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1154 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1155 10GiB) which caused the wipe to fail right at the end in some cases
1156 - Fixed master IP activation when doing master failover with no-voting
1157 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1158 the master node itself
1159 - Fixed potential data-loss in under disk full conditions, where Ganeti
1160 wouldn't check correctly the return code and would consider
1161 partially-written files 'correct'
1162 - Fixed bug related to multiple VGs and DRBD disk replacing
1163 - Added new disk parameter ``metavg`` that allows placement of the meta
1164 device for DRBD in a different volume group
1165 - Fixed error handling in the node daemon when the system libc doesn't
1166 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1167 - Fixed lock release during replace-disks, which kept cluster-wide locks
1168 when doing disk replaces with an iallocator script
1169 - Added check for missing bridges in cluster verify
1170 - Handle EPIPE errors while writing to the terminal better, so that
1171 piping the output to e.g. ``less`` doesn't cause a backtrace
1172 - Fixed rare case where a ^C during Luxi calls could have been
1173 interpreted as server errors, instead of simply terminating
1174 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1176 - Added a few more parameters to the KVM hypervisor, allowing a second
1177 CDROM, custom disk type for CDROMs and a floppy image
1178 - Removed redundant message in instance rename when the name is given
1180 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1181 disks on new nodes, allowing recreation when the original instance
1182 nodes are completely gone
1183 - Added option when converting disk templates to DRBD to skip waiting
1184 for the resync, in order to make the instance available sooner
1185 - Added two new variables to the OS scripts environment (containing the
1187 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1188 to allow use of ``pvgrub`` as bootloader
1189 - Changed the instance memory modifications to only check out-of-memory
1190 conditions on memory increases, and turned the secondary node warnings
1191 into errors (they can still be overridden via ``--force``)
1192 - Fixed the handling of a corner case when the Python installation gets
1193 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1194 to execute a command that doesn't exist
1195 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1196 node of the instance returned failures during instance shutdown; this
1197 adds the option ``--ignore-consistency`` to gnt-instance move
1199 And as usual, various improvements to the error messages, documentation
1206 *(Released Wed, 09 Mar 2011)*
1208 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1209 the RAPI users file if run twice (even with ``--dry-run``).
1211 The release fixes that bug (nothing else changed).
1217 *(Released Mon, 07 Mar 2011)*
1219 Final 2.4.0 release. Just a few small fixes:
1221 - Fixed RAPI node evacuate
1222 - Fixed the kvm-ifup script
1223 - Fixed internal error handling for special job cases
1224 - Updated man page to specify the escaping feature for options
1230 *(Released Mon, 28 Feb 2011)*
1232 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1233 that this feature wiped the disks of the wrong instance, leading to loss
1238 - Fixed title of query field containing instance name
1239 - Expanded the glossary in the documentation
1240 - Fixed one unittest (internal issue)
1246 *(Released Mon, 21 Feb 2011)*
1248 A number of bug fixes plus just a couple functionality changes.
1250 On the user-visible side, the ``gnt-* list`` command output has changed
1251 with respect to "special" field states. The current rc1 style of display
1252 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1253 the default output mode special fields states are displayed as follows:
1255 - Offline resource: ``*``
1256 - Unavailable/not applicable: ``-``
1257 - Data missing (RPC failure): ``?``
1258 - Unknown field: ``??``
1260 Another user-visible change is the addition of ``--force-join`` to
1265 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1266 - Fixed regression in RAPI/instance reinstall where all parameters were
1267 required (instead of optional)
1268 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1269 - Fixed iallocator usage (offline nodes were not considered offline)
1270 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1271 - Fixed hypervisor and OS parameter validation with respect to
1272 non-vm_capable nodes
1273 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1275 - Fixed ``tools/listrunner`` with respect to agent-based usage
1281 *(Released Fri, 4 Feb 2011)*
1283 Many changes and fixes since the beta1 release. While there were some
1284 internal changes, the code has been mostly stabilised for the RC
1287 Note: the dumb allocator was removed in this release, as it was not kept
1288 up-to-date with the IAllocator protocol changes. It is recommended to
1289 use the ``hail`` command from the ganeti-htools package.
1291 Note: the 2.4 and up versions of Ganeti are not compatible with the
1292 0.2.x branch of ganeti-htools. You need to upgrade to
1293 ganeti-htools-0.3.0 (or later).
1295 Regressions fixed from 2.3
1296 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1298 - Fixed the ``gnt-cluster verify-disks`` command
1299 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1301 - Fixed disk adoption breakage
1302 - Fixed wrong headers in instance listing for field aliases
1307 - Fixed corner case in KVM handling of NICs
1308 - Fixed many cases of wrong handling of non-vm_capable nodes
1309 - Fixed a bug where a missing instance symlink was not possible to
1310 recreate with any ``gnt-*`` command (now ``gnt-instance
1311 activate-disks`` does it)
1312 - Fixed the volume group name as reported by ``gnt-cluster
1314 - Increased timeouts for the import-export code, hopefully leading to
1315 fewer aborts due network or instance timeouts
1316 - Fixed bug in ``gnt-node list-storage``
1317 - Fixed bug where not all daemons were started on cluster
1318 initialisation, but only at the first watcher run
1319 - Fixed many bugs in the OOB implementation
1320 - Fixed watcher behaviour in presence of instances with offline
1322 - Fixed instance list output for instances running on the wrong node
1323 - a few fixes to the cluster-merge tool, but it still cannot merge
1324 multi-node groups (currently it is not recommended to use this tool)
1330 - Improved network configuration for the KVM hypervisor
1331 - Added e1000 as a supported NIC for Xen-HVM
1332 - Improved the lvmstrap tool to also be able to use partitions, as
1333 opposed to full disks
1334 - Improved speed of disk wiping (the cluster parameter
1335 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1336 of instance creations
1337 - Added documentation for the OS parameters
1338 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1339 hypervisor is not responding
1340 - Added display of blacklisted and hidden OS information in
1341 ``gnt-cluster info``
1342 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1343 NIC and node parameters, which might create problems with currently
1344 invalid (but undetected) configuration files, but prevents validation
1345 failures when unrelated parameters are modified
1346 - Changed cluster initialisation to wait for the master daemon to become
1348 - Expanded the RAPI interface:
1350 - Added config redistribution resource
1351 - Added activation/deactivation of instance disks
1352 - Added export of console information
1354 - Implemented log file reopening on SIGHUP, which allows using
1355 logrotate(8) for the Ganeti log files
1356 - Added a basic OOB helper script as an example
1362 *(Released Fri, 14 Jan 2011)*
1367 - Fixed timezone issues when formatting timestamps
1368 - Added support for node groups, available via ``gnt-group`` and other
1370 - Added out-of-band framework and management, see :doc:`design
1371 document <design-oob>`
1372 - Removed support for roman numbers from ``gnt-node list`` and
1373 ``gnt-instance list``.
1374 - Allowed modification of master network interface via ``gnt-cluster
1375 modify --master-netdev``
1376 - Accept offline secondaries while shutting down instance disks
1377 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1378 - Added support for multiple LVM volume groups
1379 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1381 - Added commands to list available fields:
1383 - ``gnt-node list-fields``
1384 - ``gnt-group list-fields``
1385 - ``gnt-instance list-fields``
1387 - Updated documentation and man pages
1392 - Moved ``rapi_users`` file into separate directory, now named
1393 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1395 - Added new tool for running commands on many machines,
1396 ``tools/ganeti-listrunner``
1397 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1398 improving the ``gnt-instance console`` output
1399 - Allowed customisation of disk index separator at ``configure`` time
1400 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1401 - Added support for non-partitioned md disks in ``lvmstrap``
1402 - Added script to gracefully power off KVM instances
1403 - Split ``utils`` module into smaller parts
1404 - Changed query operations to return more detailed information, e.g.
1405 whether an information is unavailable due to an offline node. To use
1406 this new functionality, the LUXI call ``Query`` must be used. Field
1407 information is now stored by the master daemon and can be retrieved
1408 using ``QueryFields``. Instances, nodes and groups can also be queried
1409 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1410 exposed via RAPI). The following commands make use of this
1411 infrastructure change:
1413 - ``gnt-group list``
1414 - ``gnt-group list-fields``
1416 - ``gnt-node list-fields``
1417 - ``gnt-instance list``
1418 - ``gnt-instance list-fields``
1419 - ``gnt-debug locks``
1424 - New RAPI resources (see :doc:`rapi`):
1428 - ``/2/groups/[group_name]``
1429 - ``/2/groups/[group_name]/assign-nodes``
1430 - ``/2/groups/[group_name]/modify``
1431 - ``/2/groups/[group_name]/rename``
1432 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1436 - Implemented ``no_install`` for instance creation
1437 - Implemented OS parameters for instance reinstallation, allowing
1438 use of special settings on reinstallation (e.g. for preserving data)
1443 - Added IPv6 support in import/export
1444 - Pause DRBD synchronization while wiping disks on instance creation
1445 - Updated unittests and QA scripts
1446 - Improved network parameters passed to KVM
1447 - Converted man pages from docbook to reStructuredText
1453 *(Released Mon, 20 Dec 2010)*
1455 Released version 2.3.1~rc1 without any changes.
1461 *(Released Wed, 1 Dec 2010)*
1463 - impexpd: Disable OpenSSL compression in socat if possible (backport
1464 from master, commit e90739d625b, see :doc:`installation guide
1465 <install-quick>` for details)
1466 - Changed unittest coverage report to exclude test scripts
1467 - Added script to check version format
1473 *(Released Wed, 1 Dec 2010)*
1475 Released version 2.3.0~rc1 without any changes.
1481 *(Released Fri, 19 Nov 2010)*
1483 A number of bugfixes and documentation updates:
1485 - Update ganeti-os-interface documentation
1486 - Fixed a bug related to duplicate MACs or similar items which should be
1488 - Fix breakage in OS state modify
1489 - Reinstall instance: disallow offline secondaries (fixes bug related to
1490 OS changing but reinstall failing)
1491 - plus all the other fixes between 2.2.1 and 2.2.2
1497 *(Released Tue, 2 Nov 2010)*
1499 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1500 - Fixed master failover race with watcher
1501 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1503 - Accept previously stopped instance for export with instance removal
1504 - Simplify and extend the environment variables for instance OS scripts
1505 - Added new node flags, ``master_capable`` and ``vm_capable``
1506 - Added optional instance disk wiping prior during allocation. This is a
1507 cluster-wide option and can be set/modified using
1508 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1509 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1510 :doc:`install-quick`
1511 - Added a new watcher option (``--ignore-pause``)
1512 - Added option to ignore offline node on instance start/stop
1513 (``--ignore-offline``)
1514 - Allow overriding OS parameters with ``gnt-instance reinstall``
1515 - Added ability to change node's secondary IP address using ``gnt-node
1517 - Implemented privilege separation for all daemons except
1518 ``ganeti-noded``, see ``configure`` options
1519 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1521 - Implemented job priorities (see ``ganeti(7)`` manpage)
1522 - Ignore failures while shutting down instances during failover from
1524 - Exit daemon's bootstrap process only once daemon is ready
1525 - Export more information via ``LUInstanceQuery``/remote API
1526 - Improved documentation, QA and unittests
1527 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1528 restart if the file was created or changed
1529 - Added LUXI protocol version sent with each request and response,
1530 allowing detection of server/client mismatches
1531 - Moved the Python scripts among gnt-* and ganeti-* into modules
1532 - Moved all code related to setting up SSH to an external script,
1534 - Infrastructure changes for node group support in future versions
1540 *(Released Fri, 19 Nov 2010)*
1542 A few small bugs fixed, and some improvements to the build system:
1544 - Fix documentation regarding conversion to drbd
1545 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1547 - Fix error handling in node modify with multiple changes
1548 - Allow remote imports without checked names
1554 *(Released Tue, 19 Oct 2010)*
1556 - Disable SSL session ID cache in RPC client
1562 *(Released Thu, 14 Oct 2010)*
1564 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1565 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1567 - Fix problems with interaction between Curl and Python's HTTP server,
1568 resulting in increased speed in many RPC calls
1569 - Improve our release script to prevent breakage with older aclocal and
1576 *(Released Thu, 7 Oct 2010)*
1578 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1579 value retrieved from master
1580 - Added support for blacklisted or hidden OS definitions
1581 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1582 - Added support for -mem-path in KVM hypervisor abstraction layer
1583 - Allow overriding instance parameters in tool for inter-cluster
1584 instance moves (``tools/move-instance``)
1585 - Improved opcode summaries (e.g. in ``gnt-job list``)
1586 - Improve consistency of OS listing by sorting it
1587 - Documentation updates
1593 *(Released Fri, 8 Oct 2010)*
1595 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1601 *(Released Mon, 4 Oct 2010)*
1603 - Fixed regression in ``gnt-instance rename``
1609 *(Released Wed, 22 Sep 2010)*
1611 - Fixed OS_VARIANT variable for OS scripts
1612 - Fixed cluster tag operations via RAPI
1613 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1614 - Disabled RAPI CA checks in watcher
1620 *(Released Mon, 23 Aug 2010)*
1622 - Support DRBD versions of the format "a.b.c.d"
1624 - Re-introduce support for usage from multiple threads in RAPI client
1625 - Instance renames and modify via RAPI
1626 - Work around race condition between processing and archival in job
1628 - Mark opcodes following failed one as failed, too
1629 - Job field ``lock_status`` was removed due to difficulties making it
1630 work with the changed job queue in Ganeti 2.2; a better way to monitor
1631 locks is expected for a later 2.2.x release
1632 - Fixed dry-run behaviour with many commands
1633 - Support ``ssh-agent`` again when adding nodes
1634 - Many additional bugfixes
1640 *(Released Fri, 30 Jul 2010)*
1642 Important change: the internal RPC mechanism between Ganeti nodes has
1643 changed from using a home-grown http library (based on the Python base
1644 libraries) to use the PycURL library. This requires that PycURL is
1645 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1646 against GnuTLS by default. cURL's support for GnuTLS had known issues
1647 before cURL 7.21.0 and we recommend using the latest cURL release or
1648 linking against OpenSSL. Most other distributions already link PycURL
1649 and cURL against OpenSSL. The command::
1651 python -c 'import pycurl; print pycurl.version'
1653 can be used to determine the libraries PycURL and cURL are linked
1656 Other significant changes:
1658 - Rewrote much of the internals of the job queue, in order to achieve
1659 better parallelism; this decouples job query operations from the job
1660 processing, and it should allow much nicer behaviour of the master
1661 daemon under load, and it also has uncovered some long-standing bugs
1662 related to the job serialisation (now fixed)
1663 - Added a default iallocator setting to the cluster parameters,
1664 eliminating the need to always pass nodes or an iallocator for
1665 operations that require selection of new node(s)
1666 - Added experimental support for the LXC virtualization method
1667 - Added support for OS parameters, which allows the installation of
1668 instances to pass parameter to OS scripts in order to customise the
1670 - Added a hypervisor parameter controlling the migration type (live or
1671 non-live), since hypervisors have various levels of reliability; this
1672 has renamed the 'live' parameter to 'mode'
1673 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1674 logical volumes, meaning that cluster verify will ignore them and not
1675 flag their presence as errors
1676 - The watcher will now reset the error count for failed instances after
1677 8 hours, thus allowing self-healing if the problem that caused the
1678 instances to be down/fail to start has cleared in the meantime
1679 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1680 check for, and warn, if the drbd module parameter ``usermode_helper``
1681 is not consistent with the cluster-wide setting; this is needed to
1682 make diagnose easier of failed drbd creations
1683 - Started adding base IPv6 support, but this is not yet
1684 enabled/available for use
1685 - Rename operations (cluster, instance) will now return the new name,
1686 which is especially useful if a short name was passed in
1687 - Added support for instance migration in RAPI
1688 - Added a tool to pre-configure nodes for the SSH setup, before joining
1689 them to the cluster; this will allow in the future a simplified model
1690 for node joining (but not yet fully enabled in 2.2); this needs the
1691 paramiko python library
1692 - Fixed handling of name-resolving errors
1693 - Fixed consistency of job results on the error path
1694 - Fixed master-failover race condition when executed multiple times in
1696 - Fixed many bugs related to the job queue (mostly introduced during the
1697 2.2 development cycle, so not all are impacting 2.1)
1698 - Fixed instance migration with missing disk symlinks
1699 - Fixed handling of unknown jobs in ``gnt-job archive``
1700 - And many other small fixes/improvements
1704 - Enhanced both the unittest and the QA coverage
1705 - Switched the opcode validation to a generic model, and extended the
1706 validation to all opcode parameters
1707 - Changed more parts of the code that write shell scripts to use the
1709 - Switched the master daemon to use the asyncore library for the Luxi
1716 *(Released Thu, 17 Jun 2010)*
1718 - Added tool (``move-instance``) and infrastructure to move instances
1719 between separate clusters (see :doc:`separate documentation
1720 <move-instance>` and :doc:`design document <design-2.2>`)
1721 - Added per-request RPC timeout
1722 - RAPI now requires a Content-Type header for requests with a body (e.g.
1723 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1724 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1725 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1727 - Implemented initial support for running Ganeti daemons as separate
1728 users, see configure-time flags ``--with-user-prefix`` and
1729 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1731 - Instances can be removed after export (``gnt-backup export
1732 --remove-instance``)
1733 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1734 key (instead of 1024 bit)
1735 - Added new cluster configuration file for cluster domain secret
1736 - Import/export now use SSL instead of SSH
1737 - Added support for showing estimated time when exporting an instance,
1738 see the ``ganeti-os-interface(7)`` manpage and look for
1745 *(Released Tue, 16 Nov 2010)*
1747 Some more bugfixes. Unless critical bugs occur, this will be the last
1750 - Fix case of MAC special-values
1751 - Fix mac checker regex
1752 - backend: Fix typo causing "out of range" error
1753 - Add missing --units in gnt-instance list man page
1759 *(Released Tue, 24 Aug 2010)*
1762 - Don't ignore secondary node silently on non-mirrored disk templates
1764 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1765 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1766 - Properly document the usb_mouse parameter
1767 - Fix path in ganeti-rapi(8) (issue 116)
1768 - Adjust error message when the ganeti user's .ssh directory is
1770 - Add same-node-check when changing the disk template to drbd
1776 *(Released Fri, 16 Jul 2010)*
1779 - Add an option to only select some reboot types during qa/burnin.
1780 (on some hypervisors consequent reboots are not supported)
1781 - Fix infrequent race condition in master failover. Sometimes the old
1782 master ip address would be still detected as up for a short time
1783 after it was removed, causing failover to fail.
1784 - Decrease mlockall warnings when the ctypes module is missing. On
1785 Python 2.4 we support running even if no ctypes module is installed,
1786 but we were too verbose about this issue.
1787 - Fix building on old distributions, on which man doesn't have a
1789 - Fix RAPI not to ignore the MAC address on instance creation
1790 - Implement the old instance creation format in the RAPI client.
1796 *(Released Thu, 01 Jul 2010)*
1798 A small bugfix release:
1799 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1800 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1801 a non-existing option
1802 - Fix instance create: the --force-variant option was ignored
1803 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1804 - Fix modify node storage with non-FQDN arguments
1805 - Fix RAPI client to authenticate under Python 2.6 when used
1806 for more than 5 requests needing authentication
1807 - Fix gnt-instance modify -t (storage) giving a wrong error message
1808 when converting a non-shutdown drbd instance to plain
1814 *(Released Fri, 18 Jun 2010)*
1816 A small bugfix release:
1818 - Fix live migration of KVM instances started with older Ganeti
1819 versions which had fewer hypervisor parameters
1820 - Fix gnt-instance grow-disk on down instances
1821 - Fix an error-reporting bug during instance migration
1822 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1823 silently ignoring broken ones
1824 - Fix an RPC error reporting bug affecting, for example, RAPI client
1826 - Fix bug triggered by different API version os-es on different nodes
1827 - Fix a bug in instance startup with custom hvparams: OS level
1828 parameters would fail to be applied.
1829 - Fix the RAPI client under Python 2.6 (but more work is needed to
1830 make it work completely well with OpenSSL)
1831 - Fix handling of errors when resolving names from DNS
1837 *(Released Thu, 3 Jun 2010)*
1839 A medium sized development cycle. Some new features, and some
1840 fixes/small improvements/cleanups.
1842 Significant features
1843 ~~~~~~~~~~~~~~~~~~~~
1845 The node deamon now tries to mlock itself into memory, unless the
1846 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1847 its logs, and falls back to console logging. This allows emergency
1848 features such as ``gnt-node powercycle`` to work even in the event of a
1849 broken node disk (tested offlining the disk hosting the node's
1850 filesystem and dropping its memory caches; don't try this at home)
1852 KVM: add vhost-net acceleration support. It can be tested with a new
1853 enough version of the kernel and of qemu-kvm.
1855 KVM: Add instance chrooting feature. If you use privilege dropping for
1856 your VMs you can also now force them to chroot to an empty directory,
1857 before starting the emulated guest.
1859 KVM: Add maximum migration bandwith and maximum downtime tweaking
1860 support (requires a new-enough version of qemu-kvm).
1862 Cluster verify will now warn if the master node doesn't have the master
1863 ip configured on it.
1865 Add a new (incompatible) instance creation request format to RAPI which
1866 supports all parameters (previously only a subset was supported, and it
1867 wasn't possible to extend the old format to accomodate all the new
1868 features. The old format is still supported, and a client can check for
1869 this feature, before using it, by checking for its presence in the
1870 ``features`` RAPI resource.
1872 Now with ancient latin support. Try it passing the ``--roman`` option to
1873 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1874 (requires the python-roman module to be installed, in order to work).
1879 As usual many internal code refactorings, documentation updates, and
1882 - Lots of improvements and cleanups to the experimental Remote API
1883 (RAPI) client library.
1884 - A new unit test suite for the core daemon libraries.
1885 - A fix to creating missing directories makes sure the umask is not
1886 applied anymore. This enforces the same directory permissions
1888 - Better handling terminating daemons with ctrl+c (used when running
1889 them in debugging mode).
1890 - Fix a race condition in live migrating a KVM instance, when stat()
1891 on the old proc status file returned EINVAL, which is an unexpected
1893 - Fixed manpage checking with newer man and utf-8 charachters. But now
1894 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1900 *(Released Fri, 7 May 2010)*
1902 Fix a bug which prevented untagged KVM instances from starting.
1908 *(Released Fri, 7 May 2010)*
1910 Another release with a long development cycle, during which many
1911 different features were added.
1913 Significant features
1914 ~~~~~~~~~~~~~~~~~~~~
1916 The KVM hypervisor now can run the individual instances as non-root, to
1917 reduce the impact of a VM being hijacked due to bugs in the
1918 hypervisor. It is possible to run all instances as a single (non-root)
1919 user, to manually specify a user for each instance, or to dynamically
1920 allocate a user out of a cluster-wide pool to each instance, with the
1921 guarantee that no two instances will run under the same user ID on any
1924 An experimental RAPI client library, that can be used standalone
1925 (without the other Ganeti libraries), is provided in the source tree as
1926 ``lib/rapi/client.py``. Note this client might change its interface in
1927 the future, as we iterate on its capabilities.
1929 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1930 replace the cluster's certificates and crypto keys. This might help in
1931 case they have been compromised, or have simply expired.
1933 A new disk option for instance creation has been added that allows one
1934 to "adopt" currently existing logical volumes, with data
1935 preservation. This should allow easier migration to Ganeti from
1936 unmanaged (or managed via other software) instances.
1938 Another disk improvement is the possibility to convert between redundant
1939 (DRBD) and plain (LVM) disk configuration for an instance. This should
1940 allow better scalability (starting with one node and growing the
1941 cluster, or shrinking a two-node cluster to one node).
1943 A new feature that could help with automated node failovers has been
1944 implemented: if a node sees itself as offline (by querying the master
1945 candidates), it will try to shutdown (hard) all instances and any active
1946 DRBD devices. This reduces the risk of duplicate instances if an
1947 external script automatically failovers the instances on such nodes. To
1948 enable this, the cluster parameter ``maintain_node_health`` should be
1949 enabled; in the future this option (per the name) will enable other
1950 automatic maintenance features.
1952 Instance export/import now will reuse the original instance
1953 specifications for all parameters; that means exporting an instance,
1954 deleting it and the importing it back should give an almost identical
1955 instance. Note that the default import behaviour has changed from
1956 before, where it created only one NIC; now it recreates the original
1959 Cluster verify has added a few new checks: SSL certificates validity,
1960 /etc/hosts consistency across the cluster, etc.
1965 As usual, many internal changes were done, documentation fixes,
1968 - Fixed cluster initialization with disabled cluster storage (regression
1969 introduced in 2.1.1)
1970 - File-based storage supports growing the disks
1971 - Fixed behaviour of node role changes
1972 - Fixed cluster verify for some corner cases, plus a general rewrite of
1973 cluster verify to allow future extension with more checks
1974 - Fixed log spamming by watcher and node daemon (regression introduced
1976 - Fixed possible validation issues when changing the list of enabled
1978 - Fixed cleanup of /etc/hosts during node removal
1979 - Fixed RAPI response for invalid methods
1980 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1981 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1982 from ide disks, etc.)
1983 - Allow OS changes without re-installation (to record a changed OS
1984 outside of Ganeti, or to allow OS renames)
1985 - Allow instance creation without OS installation (useful for example if
1986 the OS will be installed manually, or restored from a backup not in
1988 - Implemented option to make cluster ``copyfile`` use the replication
1990 - Added list of enabled hypervisors to ssconf (possibly useful for
1992 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1994 - A partial form of node re-IP is possible via node readd, which now
1995 allows changed node primary IP
1996 - Command line utilities now show an informational message if the job is
1998 - The logs of the master daemon now show the PID/UID/GID of the
2005 *(Released Fri, 12 Mar 2010)*
2007 During the 2.1.0 long release candidate cycle, a lot of improvements and
2008 changes have accumulated with were released later as 2.1.1.
2013 The node evacuate command (``gnt-node evacuate``) was significantly
2014 rewritten, and as such the IAllocator protocol was changed - a new
2015 request type has been added. This unfortunate change during a stable
2016 series is designed to improve performance of node evacuations; on
2017 clusters with more than about five nodes and which are well-balanced,
2018 evacuation should proceed in parallel for all instances of the node
2019 being evacuated. As such, any existing IAllocator scripts need to be
2020 updated, otherwise the above command will fail due to the unknown
2021 request. The provided "dumb" allocator has not been updated; but the
2022 ganeti-htools package supports the new protocol since version 0.2.4.
2024 Another important change is increased validation of node and instance
2025 names. This might create problems in special cases, if invalid host
2026 names are being used.
2028 Also, a new layer of hypervisor parameters has been added, that sits at
2029 OS level between the cluster defaults and the instance ones. This allows
2030 customisation of virtualization parameters depending on the installed
2031 OS. For example instances with OS 'X' may have a different KVM kernel
2032 (or any other parameter) than the cluster defaults. This is intended to
2033 help managing a multiple OSes on the same cluster, without manual
2034 modification of each instance's parameters.
2036 A tool for merging clusters, ``cluster-merge``, has been added in the
2037 tools sub-directory.
2042 - Improved the int/float conversions that should make the code more
2043 robust in face of errors from the node daemons
2044 - Fixed the remove node code in case of internal configuration errors
2045 - Fixed the node daemon behaviour in face of inconsistent queue
2046 directory (e.g. read-only file-system where we can't open the files
2048 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2049 now it either aborts cleanly or, if given the new "auto_promote"
2050 parameter, will automatically promote other nodes as needed
2051 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2052 completely prevent Ganeti from working
2053 - Fixed bug for instance export when not all disks were successfully
2055 - Fixed behaviour of node add when the new node is slow in starting up
2057 - Fixed handling of signals in the LUXI client, which should improve
2058 behaviour of command-line scripts
2059 - Added checks for invalid node/instance names in the configuration (now
2060 flagged during cluster verify)
2061 - Fixed watcher behaviour for disk activation errors
2062 - Fixed two potentially endless loops in http library, which led to the
2063 RAPI daemon hanging and consuming 100% CPU in some cases
2064 - Fixed bug in RAPI daemon related to hashed passwords
2065 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2067 - Enhanced compatibility with non-Debian OSes, but not using absolute
2068 path in some commands and allowing customisation of the ssh
2069 configuration directory
2070 - Fixed possible future issue with new Python versions by abiding to the
2071 proper use of ``__slots__`` attribute on classes
2072 - Added checks that should prevent directory traversal attacks
2073 - Many documentation fixes based on feedback from users
2078 - Added an "early_release" more for instance replace disks and node
2079 evacuate, where we release locks earlier and thus allow higher
2080 parallelism within the cluster
2081 - Added watcher hooks, intended to allow the watcher to restart other
2082 daemons (e.g. from the ganeti-nbma project), but they can be used of
2083 course for any other purpose
2084 - Added a compile-time disable for DRBD barriers, to increase
2085 performance if the administrator trusts the power supply or the
2086 storage system to not lose writes
2087 - Added the option of using syslog for logging instead of, or in
2088 addition to, Ganeti's own log files
2089 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2090 can indeed boot from a paravirtual NIC
2091 - Added a generic debug level for many operations; while this is not
2092 used widely yet, it allows one to pass the debug value all the way to
2094 - Enhanced the hooks environment for instance moves (failovers,
2095 migrations) where the primary/secondary nodes changed during the
2096 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2097 - Enhanced data validations for many user-supplied values; one important
2098 item is the restrictions imposed on instance and node names, which
2099 might reject some (invalid) host names
2100 - Add a configure-time option to disable file-based storage, if it's not
2101 needed; this allows greater security separation between the master
2102 node and the other nodes from the point of view of the inter-node RPC
2104 - Added user notification in interactive tools if job is waiting in the
2105 job queue or trying to acquire locks
2106 - Added log messages when a job is waiting for locks
2107 - Added filtering by node tags in instance operations which admit
2108 multiple instances (start, stop, reboot, reinstall)
2109 - Added a new tool for cluster mergers, ``cluster-merge``
2110 - Parameters from command line which are of the form ``a=b,c=d`` can now
2111 use backslash escapes to pass in values which contain commas,
2112 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2114 - For KVM, the instance name is the first parameter passed to KVM, so
2115 that it's more visible in the process list
2121 *(Released Tue, 2 Mar 2010)*
2123 Ganeti 2.1 brings many improvements with it. Major changes:
2125 - Added infrastructure to ease automated disk repairs
2126 - Added new daemon to export configuration data in a cheaper way than
2127 using the remote API
2128 - Instance NICs can now be routed instead of being associated with a
2130 - Improved job locking logic to reduce impact of jobs acquiring multiple
2131 locks waiting for other long-running jobs
2133 In-depth implementation details can be found in the Ganeti 2.1 design
2139 - Added chroot hypervisor
2140 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2142 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2143 ``bootloader_path`` and ``bootloader_args``)
2144 - Added the ``use_localtime`` option for the xen-hvm and kvm
2145 hypervisors, and the default value for this has changed to false (in
2146 2.0 xen-hvm always enabled it)
2147 - Added luxi call to submit multiple jobs in one go
2148 - Added cluster initialization option to not modify ``/etc/hosts``
2150 - Added network interface parameters
2151 - Added dry run mode to some LUs
2152 - Added RAPI resources:
2154 - ``/2/instances/[instance_name]/info``
2155 - ``/2/instances/[instance_name]/replace-disks``
2156 - ``/2/nodes/[node_name]/evacuate``
2157 - ``/2/nodes/[node_name]/migrate``
2158 - ``/2/nodes/[node_name]/role``
2159 - ``/2/nodes/[node_name]/storage``
2160 - ``/2/nodes/[node_name]/storage/modify``
2161 - ``/2/nodes/[node_name]/storage/repair``
2163 - Added OpCodes to evacuate or migrate all instances on a node
2164 - Added new command to list storage elements on nodes (``gnt-node
2165 list-storage``) and modify them (``gnt-node modify-storage``)
2166 - Added new ssconf files with master candidate IP address
2167 (``ssconf_master_candidates_ips``), node primary IP address
2168 (``ssconf_node_primary_ips``) and node secondary IP address
2169 (``ssconf_node_secondary_ips``)
2170 - Added ``ganeti-confd`` and a client library to query the Ganeti
2171 configuration via UDP
2172 - Added ability to run hooks after cluster initialization and before
2174 - Added automatic mode for disk replace (``gnt-instance replace-disks
2176 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2177 after catastrophic data-loss
2178 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2180 - Added ``gnt-instance move`` command to move instances
2181 - Added ``gnt-cluster watcher`` command to control watcher
2182 - Added ``gnt-node powercycle`` command to powercycle nodes
2183 - Added new job status field ``lock_status``
2184 - Added parseable error codes to cluster verification (``gnt-cluster
2185 verify --error-codes``) and made output less verbose (use
2186 ``--verbose`` to restore previous behaviour)
2187 - Added UUIDs to the main config entities (cluster, nodes, instances)
2188 - Added support for OS variants
2189 - Added support for hashed passwords in the Ganeti remote API users file
2191 - Added option to specify maximum timeout on instance shutdown
2192 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2193 - Added new helper script to start and stop Ganeti daemons
2194 (``daemon-util``), with the intent to reduce the work necessary to
2195 adjust Ganeti for non-Debian distributions and to start/stop daemons
2197 - Added more unittests
2198 - Fixed critical bug in ganeti-masterd startup
2199 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2200 now customisable at the cluster level for both the KVM and Xen
2201 hypervisors using the new ``migration_port`` parameter
2202 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2203 reinstalling an instance
2204 - Allowed ``@`` in tag names
2205 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2206 - Many documentation updates
2207 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2208 - ``gnt-instance reinstall`` can now reinstall multiple instances
2209 - Updated many command line parameters
2210 - Introduced new OS API version 15
2211 - No longer support a default hypervisor
2212 - Treat virtual LVs as inexistent
2213 - Improved job locking logic to reduce lock contention
2214 - Match instance and node names case insensitively
2215 - Reimplemented bash completion script to be more complete
2222 *(Released Thu, 4 Feb 2010)*
2224 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2225 - Fix a string formatting bug
2226 - Improve safety of the code in some error paths
2227 - Improve data validation in the master of values returned from nodes
2233 *(Released Thu, 17 Dec 2009)*
2235 - Fix security issue due to missing validation of iallocator names; this
2236 allows local and remote execution of arbitrary executables
2237 - Fix failure of gnt-node list during instance removal
2238 - Ship the RAPI documentation in the archive
2244 *(Released Wed, 30 Sep 2009)*
2246 - Fixed many wrong messages
2247 - Fixed a few bugs related to the locking library
2248 - Fixed MAC checking at instance creation time
2249 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2250 - Fixed a few issues related to signal handling in both daemons and
2252 - Fixed the example startup script provided
2253 - Fixed insserv dependencies in the example startup script (patch from
2255 - Fixed handling of drained nodes in the iallocator framework
2256 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2258 - Fixed error related to invalid job IDs in job polling
2259 - Fixed job/opcode persistence on unclean master shutdown
2260 - Fixed handling of partial job processing after unclean master
2262 - Fixed error reporting from LUs, previously all errors were converted
2263 into execution errors
2264 - Fixed error reporting from burnin
2265 - Decreased significantly the memory usage of the job queue
2266 - Optimised slightly multi-job submission
2267 - Optimised slightly opcode loading
2268 - Backported the multi-job submit framework from the development
2269 branch; multi-instance start and stop should be faster
2270 - Added script to clean archived jobs after 21 days; this will reduce
2271 the size of the queue directory
2272 - Added some extra checks in disk size tracking
2273 - Added an example ethers hook script
2274 - Added a cluster parameter that prevents Ganeti from modifying of
2276 - Added more node information to RAPI responses
2277 - Added a ``gnt-job watch`` command that allows following the ouput of a
2279 - Added a bind-address option to ganeti-rapi
2280 - Added more checks to the configuration verify
2281 - Enhanced the burnin script such that some operations can be retried
2283 - Converted instance reinstall to multi-instance model
2289 *(Released Fri, 7 Aug 2009)*
2291 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2292 to allow using the pre-2.0.2 behaviour in activation, if any existing
2293 instances have mismatched disk sizes in the configuration
2294 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2295 any configuration mismatches for disk sizes
2296 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2297 failover to work on two-node clusters
2298 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2300 - Fixed detection of OS script errors in ``gnt-backup export``
2301 - Fixed exit code of ``gnt-backup export``
2307 *(Released Fri, 17 Jul 2009)*
2309 - Added experimental support for stripped logical volumes; this should
2310 enhance performance but comes with a higher complexity in the block
2311 device handling; stripping is only enabled when passing
2312 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2313 affected even in the non-stripped mode
2314 - Improved resiliency against transient failures at the end of DRBD
2315 resyncs, and in general of DRBD resync checks
2316 - Fixed a couple of issues with exports and snapshot errors
2317 - Fixed a couple of issues in instance listing
2318 - Added display of the disk size in ``gnt-instance info``
2319 - Fixed checking for valid OSes in instance creation
2320 - Fixed handling of the "vcpus" parameter in instance listing and in
2321 general of invalid parameters
2322 - Fixed http server library, and thus RAPI, to handle invalid
2323 username/password combinations correctly; this means that now they
2324 report unauthorized for queries too, not only for modifications,
2325 allowing earlier detect of configuration problems
2326 - Added a new "role" node list field, equivalent to the master/master
2327 candidate/drained/offline flags combinations
2328 - Fixed cluster modify and changes of candidate pool size
2329 - Fixed cluster verify error messages for wrong files on regular nodes
2330 - Fixed a couple of issues with node demotion from master candidate role
2331 - Fixed node readd issues
2332 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2333 - Added a new ``--no-voting`` option for masterfailover to fix failover
2334 on two-nodes clusters when the former master node is unreachable
2335 - Added instance reinstall over RAPI
2341 *(Released Tue, 16 Jun 2009)*
2343 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2344 allow re-adding the start in single-user option (regression from 1.2)
2345 - the watcher writes the instance status to a file, to allow monitoring
2346 to report the instance status (from the master) based on cached
2347 results of the watcher's queries; while this can get stale if the
2348 watcher is being locked due to other work on the cluster, this is
2349 still an improvement
2350 - the watcher now also restarts the node daemon and the rapi daemon if
2352 - fixed the watcher to handle full and drained queue cases
2353 - hooks export more instance data in the environment, which helps if
2354 hook scripts need to take action based on the instance's properties
2355 (no longer need to query back into ganeti)
2356 - instance failovers when the instance is stopped do not check for free
2357 RAM, so that failing over a stopped instance is possible in low memory
2359 - rapi uses queries for tags instead of jobs (for less job traffic), and
2360 for cluster tags it won't talk to masterd at all but read them from
2362 - a couple of error handling fixes in RAPI
2363 - drbd handling: improved the error handling of inconsistent disks after
2364 resync to reduce the frequency of "there are some degraded disks for
2365 this instance" messages
2366 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2367 error handling path called a wrong function name)
2373 *(Released Wed, 27 May 2009)*
2375 - no changes from rc5
2381 *(Released Wed, 20 May 2009)*
2383 - fix a couple of bugs (validation, argument checks)
2384 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2385 - some small improvements to RAPI and IAllocator
2386 - make watcher automatically start the master daemon if down
2392 *(Released Mon, 27 Apr 2009)*
2394 - change the OS list to not require locks; this helps with big clusters
2395 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2396 volume group is broken
2397 - ``gnt-instance info``, without any arguments, doesn't run for all
2398 instances anymore; either pass ``--all`` or pass the desired
2399 instances; this helps against mistakes on big clusters where listing
2400 the information for all instances takes a long time
2401 - miscellaneous doc and man pages fixes
2407 *(Released Wed, 8 Apr 2009)*
2409 - Change the internal locking model of some ``gnt-node`` commands, in
2410 order to reduce contention (and blocking of master daemon) when
2411 batching many creation/reinstall jobs
2412 - Fixes to Xen soft reboot
2413 - No longer build documentation at build time, instead distribute it in
2414 the archive, in order to reduce the need for the whole docbook/rst
2421 *(Released Fri, 27 Mar 2009)*
2423 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2424 - Fix watcher startup sequence, improves the behaviour of busy clusters
2425 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2426 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2427 bind address and other places
2428 - Some documentation fixes and updates
2434 *(Released Mon, 2 Mar 2009)*
2436 - More documentation updates, now all docs should be more-or-less
2438 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2440 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2442 - Fix an issue related to $libdir/run/ganeti and cluster creation
2448 *(Released Thu, 19 Feb 2009)*
2450 - Xen PVM and KVM have switched the default value for the instance root
2451 disk to the first partition on the first drive, instead of the whole
2452 drive; this means that the OS installation scripts must be changed
2454 - Man pages have been updated
2455 - RAPI has been switched by default to HTTPS, and the exported functions
2456 should all work correctly
2457 - RAPI v1 has been removed
2458 - Many improvements to the KVM hypervisor
2459 - Block device errors are now better reported
2460 - Many other bugfixes and small improvements
2466 *(Released Mon, 26 Jan 2009)*
2468 - Version 2 is a general rewrite of the code and therefore the
2469 differences are too many to list, see the design document for 2.0 in
2470 the ``doc/`` subdirectory for more details
2471 - In this beta version there is not yet a migration path from 1.2 (there
2472 will be one in the final 2.0 release)
2473 - A few significant changes are:
2475 - all commands are executed by a daemon (``ganeti-masterd``) and the
2476 various ``gnt-*`` commands are just front-ends to it
2477 - all the commands are entered into, and executed from a job queue,
2478 see the ``gnt-job(8)`` manpage
2479 - the RAPI daemon supports read-write operations, secured by basic
2480 HTTP authentication on top of HTTPS
2481 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2482 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2483 to migrate to DRBD 8 first while still running Ganeti 1.2)
2484 - DRBD devices are using statically allocated minor numbers, which
2485 will be assigned to existing instances during the migration process
2486 - there is support for both Xen PVM and Xen HVM instances running on
2488 - KVM virtualization is supported too
2489 - file-based storage has been implemented, which means that it is
2490 possible to run the cluster without LVM and DRBD storage, for
2491 example using a shared filesystem exported from shared storage (and
2492 still have live migration)
2498 *(Released Tue, 13 Jan 2009)*
2500 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2501 - Reuse the old instance mac address by default on instance import, if
2502 the instance name is the same.
2503 - Handle situations in which the node info rpc returns incomplete
2505 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2506 - Improved version of batcher:
2508 - state file support
2509 - instance mac address support
2510 - support for HVM clusters/instances
2512 - Add an option to show the number of cpu sockets and nodes in
2514 - Support OSes that handle more than one version of the OS api (but do
2515 not change the current API in any other way)
2516 - Fix ``gnt-node migrate``
2517 - ``gnt-debug`` man page
2518 - Fixes various more typos and small issues
2519 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2525 *(Released Wed, 24 Sep 2008)*
2527 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2528 type of disk exported to fully virtualized instances.
2529 - provide access to the serial console of HVM instances
2530 - instance auto_balance flag, set by default. If turned off it will
2531 avoid warnings on cluster verify if there is not enough memory to fail
2532 over an instance. in the future it will prevent automatically failing
2533 it over when we will support that.
2534 - batcher tool for instance creation, see ``tools/README.batcher``
2535 - ``gnt-instance reinstall --select-os`` to interactively select a new
2536 operating system when reinstalling an instance.
2537 - when changing the memory amount on instance modify a check has been
2538 added that the instance will be able to start. also warnings are
2539 emitted if the instance will not be able to fail over, if auto_balance
2541 - documentation fixes
2542 - sync fields between ``gnt-instance list/modify/add/import``
2543 - fix a race condition in drbd when the sync speed was set after giving
2544 the device a remote peer.
2550 *(Released Tue, 22 Jul 2008)*
2552 - note: the allowed size and number of tags per object were reduced
2553 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2554 - fixed twisted 8.x compatibility
2555 - fixed ``gnt-instance replace-disks`` with iallocator
2556 - add TCP keepalives on twisted connections to detect restarted nodes
2557 - disk increase support, see ``gnt-instance grow-disk``
2558 - implement bulk node/instance query for RAPI
2559 - add tags in node/instance listing (optional)
2560 - experimental migration (and live migration) support, read the man page
2561 for ``gnt-instance migrate``
2562 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2563 has some small improvements in handling its state file
2569 *(Released Fri, 13 Jun 2008)*
2571 - Experimental readonly, REST-based remote API implementation;
2572 automatically started on master node, TCP port 5080, if enabled by
2573 ``--enable-rapi`` parameter to configure script.
2574 - Instance allocator support. Add and import instance accept a
2575 ``--iallocator`` parameter, and call that instance allocator to decide
2576 which node to use for the instance. The iallocator document describes
2577 what's expected from an allocator script.
2578 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2579 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2580 node is lost there is still enough memory to fail over the instances
2582 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2583 ``gnt-cluster verify``, to check for site-specific compliance. All the
2584 hooks will run, and their output, if any, will be displayed. Any
2585 failing hook will make the verification return an error value.
2586 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2587 primary and secondary interfaces
2588 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2589 that is still declared as part of the cluster and has failed.
2590 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2591 specifying output fields, that just adds the chosen fields to the
2593 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2594 export from the filesystem.
2595 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2596 have been added. Using them you can enable/disable acpi and pae
2597 support, and specify a path for a cd image to be exported to the
2598 instance. These parameters as the name suggest only work on HVM
2600 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2601 PAE support will be set to the previously hardcoded values, but the
2602 (previously hardcoded) path to the CDROM ISO image will be unset and
2603 if required, needs to be set manually with ``gnt-instance modify``
2605 - The address to which an instance's VNC console is bound is now
2606 selectable per-instance, rather than being cluster wide. Of course
2607 this only applies to instances controlled via VNC, so currently just
2608 applies to HVM clusters.
2614 *(Released Mon, 18 Feb 2008)*
2616 - more tweaks to the disk activation code (especially helpful for DRBD)
2617 - change the default ``gnt-instance list`` output format, now there is
2618 one combined status field (see the manpage for the exact values this
2620 - some more fixes for the mac export to hooks change
2621 - make Ganeti not break with DRBD 8.2.x (which changed the version
2622 format in ``/proc/drbd``) (issue 24)
2623 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2624 template, allowing migration from DRBD0.7+MD to DRBD8
2630 *(Released Wed, 30 Jan 2008)*
2632 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2634 - add command aliases infrastructure and a few aliases
2635 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2636 man pages and the ``--help`` option of ``gnt-node
2637 list``/``gnt-instance list``
2638 - fix ``gnt-backup list`` with down nodes (issue 21)
2639 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2640 - fix the dist archive and add a check for including svn/git files in
2642 - some developer-related changes: improve the burnin and the QA suite,
2643 add an upload script for testing during development
2649 *(Released Wed, 16 Jan 2008)*
2651 - experimental HVM support, read the install document, section
2652 "Initializing the cluster"
2653 - allow for the PVM hypervisor per-instance kernel and initrd paths
2654 - add a new command ``gnt-cluster verify-disks`` which uses a new
2655 algorithm to improve the reconnection of the DRBD pairs if the device
2656 on the secondary node has gone away
2657 - make logical volume code auto-activate LVs at disk activation time
2658 - slightly improve the speed of activating disks
2659 - allow specification of the MAC address at instance creation time, and
2660 changing it later via ``gnt-instance modify``
2661 - fix handling of external commands that generate lots of output on
2663 - update documentation with regard to minimum version of DRBD8 supported
2669 *(Released Tue, 4 Dec 2007)*
2671 - Log the ``xm create`` output to the node daemon log on failure (to
2672 help diagnosing the error)
2673 - In debug mode, log all external commands output if failed to the logs
2674 - Change parsing of lvm commands to ignore stderr
2680 *(Released Wed, 28 Nov 2007)*
2682 - Another round of updates to the DRBD 8 code to deal with more failures
2683 in the replace secondary node operation
2684 - Some more logging of failures in disk operations (lvm, drbd)
2685 - A few documentation updates
2692 *(Released Tue, 13 Nov 2007)*
2694 - Change configuration file format from Python's Pickle to JSON.
2695 Upgrading is possible using the cfgupgrade utility.
2696 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2697 faster replace disks and is more stable (DRBD 8 has many improvements
2698 compared to DRBD 0.7)
2699 - Added command line tags support (see man pages for ``gnt-instance``,
2700 ``gnt-node``, ``gnt-cluster``)
2701 - Added instance rename support
2702 - Added multi-instance startup/shutdown
2703 - Added cluster rename support
2704 - Added ``gnt-node evacuate`` to simplify some node operations
2705 - Added instance reboot operation that can speedup reboot as compared to
2707 - Soften the requirement that hostnames are in FQDN format
2708 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2710 - Removed dependency on debian's patched fping that uses the
2711 non-standard ``-S`` option
2712 - Now the OS definitions are searched for in multiple, configurable
2713 paths (easier for distros to package)
2714 - Some changes to the hooks infrastructure (especially the new
2715 post-configuration update hook)
2716 - Other small bugfixes
2718 .. vim: set textwidth=72 syntax=rst :