8 *(Released Thu, 04 Jul 2013)*
10 Incompatible/important changes
11 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
13 - Instance policies for disk size were documented to be on a per-disk
14 basis, but hail applied them to the sum of all disks. This has been
16 - ``hbal`` will now exit with status 0 if, during job execution over
17 LUXI, early exit has been requested and all jobs are successful;
18 before, exit status 1 was used, which cannot be differentiated from
20 - Compatibility with newer versions of rbd has been fixed
21 - ``gnt-instance batch-create`` has been changed to use the bulk create
22 opcode from Ganeti. This lead to incompatible changes in the format of
23 the JSON file. It's now not a custom dict anymore but a dict
24 compatible with the ``OpInstanceCreate`` opcode.
25 - Parent directories for file storage need to be listed in
26 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
27 write the file automatically based on old configuration values, but it
28 can not distribute it across all nodes and the file contents should be
29 verified. Use ``gnt-cluster copyfile
30 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
31 upgraded. The reason for requiring this list of paths now is that
32 before it would have been possible to inject new paths via RPC,
33 allowing files to be created in arbitrary locations. The RPC protocol
34 is protected using SSL/X.509 certificates, but as a design principle
35 Ganeti does not permit arbitrary paths to be passed.
36 - The parsing of the variants file for OSes (see
37 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
38 lines and comment lines (starting with ``#``) are ignored for better
40 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
41 longer available. ``gnt-node add`` now invokes a new tool on the
42 destination node, named ``prepare-node-join``, to configure the SSH
43 daemon. Paramiko is no longer necessary to configure nodes' SSH
44 daemons via ``gnt-node add``.
45 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
46 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
47 the flag is not reset after a master failover.
48 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
49 is recommended. 2.6 will be mandatory from the 2.8 series.
55 - New network management functionality to support automatic allocation
56 of IP addresses and managing of network parameters. See
57 :manpage:`gnt-network(8)` for more details.
58 - New external storage backend, to allow managing arbitrary storage
59 systems external to the cluster. See
60 :manpage:`ganeti-extstorage-interface(7)`.
61 - New ``exclusive-storage`` node parameter added, restricted to
62 nodegroup level. When it's set to true, physical disks are assigned in
63 an exclusive fashion to instances, as documented in :doc:`Partitioned
64 Ganeti <design-partitioned>`. Currently, only instances using the
65 ``plain`` disk template are supported.
66 - The KVM hypervisor has been updated with many new hypervisor
67 parameters, including a generic one for passing arbitrary command line
68 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
69 compatible up to qemu 1.4.
70 - A new tool, called ``mon-collector``, is the stand-alone executor of
71 the data collectors for a monitoring system. As of this version, it
72 just includes the DRBD data collector, that can be executed by calling
73 ``mon-collector`` using the ``drbd`` parameter. See
74 :manpage:`mon-collector(7)`.
75 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
76 for RAPI users. It allows granting permissions to query for
77 information to a specific user without giving
78 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
79 - A new tool named ``node-cleanup`` has been added. It cleans remains of
80 a cluster from a machine by stopping all daemons, removing
81 certificates and ssconf files. Unless the ``--no-backup`` option is
82 given, copies of the certificates are made.
83 - Instance creations now support the use of opportunistic locking,
84 potentially speeding up the (parallel) creation of multiple instances.
85 This feature is currently only available via the :doc:`RAPI
86 <rapi>` interface and when an instance allocator is used. If the
87 ``opportunistic_locking`` parameter is set the opcode will try to
88 acquire as many locks as possible, but will not wait for any locks
89 held by other opcodes. If not enough resources can be found to
90 allocate the instance, the temporary error code
91 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
92 retried thereafter, with or without opportunistic locking.
93 - New experimental linux-ha resource scripts.
94 - Restricted-commands support: ganeti can now be asked (via command line
95 or rapi) to perform commands on a node. These are passed via ganeti
96 RPC rather than ssh. This functionality is restricted to commands
97 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
98 reasons. The file is not copied automatically.
104 - Diskless instances are now externally mirrored (Issue 237). This for
105 now has only been tested in conjunction with explicit target nodes for
107 - Queries not needing locks or RPC access to the node can now be
108 performed by the confd daemon, making them independent from jobs, and
109 thus faster to execute. This is selectable at configure time.
110 - The functionality for allocating multiple instances at once has been
111 overhauled and is now also available through :doc:`RAPI <rapi>`.
113 There are no significant changes from version 2.7.0~rc3.
119 *(Released Tue, 25 Jun 2013)*
121 - Fix permissions on the confd query socket (Issue 477)
122 - Fix permissions on the job archive dir (Issue 498)
123 - Fix handling of an internal exception in replace-disks (Issue 472)
124 - Fix gnt-node info handling of shortened names (Issue 497)
125 - Fix gnt-instance grow-disk when wiping is enabled
126 - Documentation improvements, and support for newer pandoc
127 - Fix hspace honoring ipolicy for disks (Issue 484)
128 - Improve handling of the ``kvm_extra`` HV parameter
134 *(Released Fri, 24 May 2013)*
136 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
138 - Disks added through ``gnt-instance modify`` or created through
139 ``gnt-instance recreate-disks`` are wiped, if the
140 ``prealloc_wipe_disks`` flag is set.
141 - If wiping newly created disks fails, the disks are removed. Also,
142 partial failures in creating disks through ``gnt-instance modify``
143 triggers a cleanup of the partially-created disks.
144 - Removing the master IP address doesn't fail if the address has been
146 - Fix ownership of the OS log dir
147 - Workaround missing SO_PEERCRED constant (Issue 191)
153 *(Released Fri, 3 May 2013)*
155 This was the first release candidate of the 2.7 series. Since beta3:
157 - Fix kvm compatibility with qemu 1.4 (Issue 389)
158 - Documentation updates (admin guide, upgrade notes, install
159 instructions) (Issue 372)
160 - Fix gnt-group list nodes and instances count (Issue 436)
161 - Fix compilation without non-mandatory libraries (Issue 441)
162 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
163 - Make confd logging more verbose at INFO level (Issue 435)
164 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
165 - Fix failure path for instance storage type conversion (Issue 229)
166 - Update htools text backend documentation
167 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
168 - Disable inter-cluster instance move for file-based instances, because
169 it is dependant on instance export, which is not supported for
170 file-based instances. (Issue 414)
171 - Fix gnt-job crashes on non-ascii characters (Issue 427)
172 - Fix volume group checks on non-vm-capable nodes (Issue 432)
178 *(Released Mon, 22 Apr 2013)*
180 This was the third beta release of the 2.7 series. Since beta2:
182 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
183 - Fix data loss on wrong usage of ``gnt-instance move``
184 - Properly export errors in confd-based job queries
185 - Add ``users-setup`` tool
186 - Fix iallocator protocol to report 0 as a disk size for diskless
187 instances. This avoids hail breaking when a diskless instance is
189 - Fix job queue directory permission problem that made confd job queries
190 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
191 for access to archived jobs (Issue 406).
192 - Limit the sizes of networks supported by ``gnt-network`` to something
193 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
194 - Fix bugs in instance disk template conversion
195 - Fix GHC 7 compatibility
196 - Fix ``burnin`` install path (Issue 426).
197 - Allow very small disk grows (Issue 347).
198 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
199 that noded doesn't import masterd code (Issue 419).
200 - Make sure the default metavg at cluster init is the same as the vg, if
201 unspecified (Issue 358).
202 - Fix cleanup of partially created disks (part of Issue 416)
208 *(Released Tue, 2 Apr 2013)*
210 This was the second beta release of the 2.7 series. Since beta1:
212 - Networks no longer have a "type" slot, since this information was
213 unused in Ganeti: instead of it tags should be used.
214 - The rapi client now has a ``target_node`` option to MigrateInstance.
215 - Fix early exit return code for hbal (Issue 386).
216 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
217 - Fix ``rbd showmapped`` output parsing (Issue 312).
218 - Networks are now referenced indexed by UUID, rather than name. This
219 will require running cfgupgrade, from 2.7.0beta1, if networks are in
221 - The OS environment now includes network information.
222 - Deleting of a network is now disallowed if any instance nic is using
223 it, to prevent dangling references.
224 - External storage is now documented in man pages.
225 - The exclusive_storage flag can now only be set at nodegroup level.
226 - Hbal can now submit an explicit priority with its jobs.
227 - Many network related locking fixes.
228 - Bump up the required pylint version to 0.25.1.
229 - Fix the ``no_remember`` option in RAPI client.
230 - Many ipolicy related tests, qa, and fixes.
231 - Many documentation improvements and fixes.
232 - Fix building with ``--disable-file-storage``.
233 - Fix ``-q`` option in htools, which was broken if passed more than
235 - Some haskell/python interaction improvements and fixes.
236 - Fix iallocator in case of missing LVM storage.
237 - Fix confd config load in case of ``--no-lvm-storage``.
238 - The confd/query functionality is now mentioned in the security
245 *(Released Wed, 6 Feb 2013)*
247 This was the first beta release of the 2.7 series. All important changes
248 are listed in the latest 2.7 entry.
254 *(Released Fri, 21 Dec 2012)*
256 Important behaviour change: hbal won't rebalance anymore instances which
257 have the ``auto_balance`` attribute set to false. This was the intention
258 all along, but until now it only skipped those from the N+1 memory
259 reservation (DRBD-specific).
261 A significant number of bug fixes in this release:
263 - Fixed disk adoption interaction with ipolicy checks.
264 - Fixed networking issues when instances are started, stopped or
265 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
266 - Fixed the warning in cluster verify for shared storage instances not
268 - Fixed removal of storage directory on shared file storage (issue 262).
269 - Fixed validation of LVM volume group name in OpClusterSetParams
270 (``gnt-cluster modify``) (issue 285).
271 - Fixed runtime memory increases (``gnt-instance modify -m``).
272 - Fixed live migration under Xen's ``xl`` mode.
273 - Fixed ``gnt-instance console`` with ``xl``.
274 - Fixed building with newer Haskell compiler/libraries.
275 - Fixed PID file writing in Haskell daemons (confd); this prevents
276 restart issues if confd was launched manually (outside of
277 ``daemon-util``) while another copy of it was running
278 - Fixed a type error when doing live migrations with KVM (issue 297) and
279 the error messages for failing migrations have been improved.
280 - Fixed opcode validation for the out-of-band commands (``gnt-node
282 - Fixed a type error when unsetting OS hypervisor parameters (issue
283 311); now it's possible to unset all OS-specific hypervisor
285 - Fixed the ``dry-run`` mode for many operations: verification of
286 results was over-zealous but didn't take into account the ``dry-run``
287 operation, resulting in "wrong" failures.
288 - Fixed bash completion in ``gnt-job list`` when the job queue has
289 hundreds of entries; especially with older ``bash`` versions, this
290 results in significant CPU usage.
292 And lastly, a few other improvements have been made:
294 - Added option to force master-failover without voting (issue 282).
295 - Clarified error message on lock conflict (issue 287).
296 - Logging of newly submitted jobs has been improved (issue 290).
297 - Hostname checks have been made uniform between instance rename and
299 - The ``--submit`` option is now supported by ``gnt-debug delay``.
300 - Shutting down the master daemon by sending SIGTERM now stops it from
301 processing jobs waiting for locks; instead, those jobs will be started
302 once again after the master daemon is started the next time (issue
304 - Support for Xen's ``xl`` program has been improved (besides the fixes
306 - Reduced logging noise in the Haskell confd daemon (only show one log
307 entry for each config reload, instead of two).
308 - Several man page updates and typo fixes.
314 *(Released Fri, 12 Oct 2012)*
316 A small bugfix release. Among the bugs fixed:
318 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
319 made the command unusable.
320 - Commands that issue many jobs don't fail anymore just because some jobs
321 take so long that other jobs are archived.
322 - Failures during ``gnt-instance reinstall`` are reflected by the exit
324 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
326 - When ``always_failover`` is set, ``--allow-failover`` is not required
327 in migrate commands anymore.
328 - ``bash_completion`` works even if extglob is disabled.
329 - Fixed bug with locks that made failover for RDB-based instances fail.
330 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
331 a random node instead of one based on the allocator metric.
332 - Support for newer versions of pylint and pep8.
333 - Hail doesn't fail anymore when trying to add an instance of type
334 ``file``, ``sharedfile`` or ``rbd``.
335 - Added new Makefile target to rebuild the whole distribution, so that
336 all files are included.
342 *(Released Fri, 27 Jul 2012)*
345 .. attention:: The ``LUXI`` protocol has been made more consistent
346 regarding its handling of command arguments. This, however, leads to
347 incompatibility issues with previous versions. Please ensure that you
348 restart Ganeti daemons soon after the upgrade, otherwise most
349 ``LUXI`` calls (job submission, setting/resetting the drain flag,
350 pausing/resuming the watcher, cancelling and archiving jobs, querying
351 the cluster configuration) will fail.
360 The current ``admin_up`` field, which used to denote whether an instance
361 should be running or not, has been removed. Instead, ``admin_state`` is
362 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
364 The rational behind this is that an instance being “down” can have
367 - it could be down during a reboot
368 - it could be temporarily be down for a reinstall
369 - or it could be down because it is deprecated and kept just for its
372 The previous Boolean state was making it difficult to do capacity
373 calculations: should Ganeti reserve memory for a down instance? Now, the
374 tri-state field makes it clear:
376 - in ``up`` and ``down`` state, all resources are reserved for the
377 instance, and it can be at any time brought up if it is down
378 - in ``offline`` state, only disk space is reserved for it, but not
381 The field can have an extra use: since the transition between ``up`` and
382 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
383 transition between ``offline`` and ``down`` is done via ``gnt-instance
384 modify``, it is possible to given different rights to users. For
385 example, owners of an instance could be allowed to start/stop it, but
386 not transition it out of the offline state.
388 Instance policies and specs
389 +++++++++++++++++++++++++++
391 In previous Ganeti versions, an instance creation request was not
392 limited on the minimum size and on the maximum size just by the cluster
393 resources. As such, any policy could be implemented only in third-party
394 clients (RAPI clients, or shell wrappers over ``gnt-*``
395 tools). Furthermore, calculating cluster capacity via ``hspace`` again
396 required external input with regards to instance sizes.
398 In order to improve these workflows and to allow for example better
399 per-node group differentiation, we introduced instance specs, which
402 - minimum instance disk size, disk count, memory size, cpu count
403 - maximum values for the above metrics
404 - and “standard” values (used in ``hspace`` to calculate the standard
407 The minimum/maximum values can be also customised at node-group level,
408 for example allowing more powerful hardware to support bigger instance
411 Beside the instance specs, there are a few other settings belonging to
412 the instance policy framework. It is possible now to customise, per
413 cluster and node-group:
415 - the list of allowed disk templates
416 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
417 - the maximum ratio of instance to spindles (see below for more
418 information) for local storage
420 All these together should allow all tools that talk to Ganeti to know
421 what are the ranges of allowed values for instances and the
422 over-subscription that is allowed.
424 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
425 instance configuration, and the physical CPU configuration from the
426 node. For the spindle ratios however, we didn't track before these
427 values, so new parameters have been added:
429 - a new node parameter ``spindle_count``, defaults to 1, customisable at
430 node group or node level
431 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
433 Note that spindles in this context doesn't need to mean actual
434 mechanical hard-drives; it's just a relative number for both the node
435 I/O capacity and instance I/O consumption.
437 Instance migration behaviour
438 ++++++++++++++++++++++++++++
440 While live-migration is in general desirable over failover, it is
441 possible that for some workloads it is actually worse, due to the
442 variable time of the “suspend” phase during live migration.
444 To allow the tools to work consistently over such instances (without
445 having to hard-code instance names), a new backend parameter
446 ``always_failover`` has been added to control the migration/failover
447 behaviour. When set to True, all migration requests for an instance will
448 instead fall-back to failover.
450 Instance memory ballooning
451 ++++++++++++++++++++++++++
453 Initial support for memory ballooning has been added. The memory for an
454 instance is no longer fixed (backend parameter ``memory``), but instead
455 can vary between minimum and maximum values (backend parameters
456 ``minmem`` and ``maxmem``). Currently we only change an instance's
459 - live migrating or failing over and instance and the target node
460 doesn't have enough memory
461 - user requests changing the memory via ``gnt-instance modify
467 In order to control the use of specific CPUs by instance, support for
468 controlling CPU pinning has been added for the Xen, HVM and LXC
469 hypervisors. This is controlled by a new hypervisor parameter
470 ``cpu_mask``; details about possible values for this are in the
471 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
472 VCPU-to-CPU mapping) form will work well only when all nodes in your
473 cluster have the same amount of CPUs.
478 Another area in which Ganeti was not customisable were the parameters
479 used for storage configuration, e.g. how many stripes to use for LVM,
480 DRBD resync configuration, etc.
482 To improve this area, we've added disks parameters, which are
483 customisable at cluster and node group level, and which allow to
484 specify various parameters for disks (DRBD has the most parameters
485 currently), for example:
487 - DRBD resync algorithm and parameters (e.g. speed)
488 - the default VG for meta-data volumes for DRBD
489 - number of stripes for LVM (plain disk template)
492 These parameters can be modified via ``gnt-cluster modify -D …`` and
493 ``gnt-group modify -D …``, and are used at either instance creation (in
494 case of LVM stripes, for example) or at disk “activation” time
497 Rados block device support
498 ++++++++++++++++++++++++++
500 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
501 denoted by the ``rbd`` disk template type. This is considered
502 experimental, feedback is welcome. For details on configuring it, see
503 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
508 The existing master IP functionality works well only in simple setups (a
509 single network shared by all nodes); however, if nodes belong to
510 different networks, then the ``/32`` setup and lack of routing
511 information is not enough.
513 To allow the master IP to function well in more complex cases, the
514 system was reworked as follows:
516 - a master IP netmask setting has been added
517 - the master IP activation/turn-down code was moved from the node daemon
519 - whether to run the Ganeti-supplied master IP script or a user-supplied
520 on is a ``gnt-cluster init`` setting
522 Details about the location of the standard and custom setup scripts are
523 in the man page :manpage:`gnt-cluster(8)`; for information about the
524 setup script protocol, look at the Ganeti-supplied script.
529 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
532 It is now possible to use TLS-protected connections, and when renewing
533 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
534 it is now possible to specify spice or spice CA certificates. Also, it
535 is possible to configure a password for SPICE sessions via the
536 hypervisor parameter ``spice_password_file``.
538 There are also new parameters to control the compression and streaming
539 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
540 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
541 for the spice parameters.
543 Lastly, it is now possible to see the SPICE connection information via
544 ``gnt-instance console``.
549 A new tool (``tools/ovfconverter``) has been added that supports
550 conversion between Ganeti and the `Open Virtualization Format
551 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
554 This relies on the ``qemu-img`` tool to convert the disk formats, so the
555 actual compatibility with other virtualization solutions depends on it.
560 The configuration query daemon (``ganeti-confd``) is now optional, and
561 has been rewritten in Haskell; whether to use the daemon at all, use the
562 Python (default) or the Haskell version is selectable at configure time
563 via the ``--enable-confd`` parameter, which can take one of the
564 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
565 daemon will result in a smaller footprint; for larger systems, we
566 welcome feedback on the Haskell version which might become the default
569 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
570 daemon running. The Python version doesn't implement the new call.
573 User interface changes
574 ~~~~~~~~~~~~~~~~~~~~~~
576 We have replaced the ``--disks`` option of ``gnt-instance
577 replace-disks`` with a more flexible ``--disk`` option, which allows
578 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
579 disk size and mode can be changed upon recreation (via ``gnt-instance
580 recreate-disks``, which accepts the same ``--disk`` option).
582 As many people are used to a ``show`` command, we have added that as an
583 alias to ``info`` on all ``gnt-*`` commands.
585 The ``gnt-instance grow-disk`` command has a new mode in which it can
586 accept the target size of the disk, instead of the delta; this can be
587 more safe since two runs in absolute mode will be idempotent, and
588 sometimes it's also easier to specify the desired size directly.
590 Also the handling of instances with regard to offline secondaries has
591 been improved. Instance operations should not fail because one of it's
592 secondary nodes is offline, even though it's safe to proceed.
594 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
595 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
596 minors to instance name.
601 RAPI coverage has improved, with (for example) new resources for
602 recreate-disks, node power-cycle, etc.
607 There is partial support for ``xl`` in the Xen hypervisor; feedback is
610 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
611 whether to still support Python 2.4 or move to Python 2.6 as minimum
614 Support for Fedora has been slightly improved; the provided example
615 init.d script should work better on it and the INSTALL file should
616 document the needed dependencies.
621 The deprecated ``QueryLocks`` LUXI request has been removed. Use
622 ``Query(what=QR_LOCK, ...)`` instead.
624 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
625 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
626 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
627 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
628 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
630 RAPI client: ``CertificateError`` now derives from
631 ``GanetiApiError``. This should make it more easy to handle Ganeti
634 Deprecation warnings due to PyCrypto/paramiko import in
635 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
636 make sure to run an up-to-date paramiko version, if you use this tool.
638 The QA scripts now depend on Python 2.5 or above (the main code base
639 still works with Python 2.4).
641 The configuration file (``config.data``) is now written without
642 indentation for performance reasons; if you want to edit it, it can be
643 re-formatted via ``tools/fmtjson``.
645 A number of bugs has been fixed in the cluster merge tool.
647 ``x509`` certification verification (used in import-export) has been
648 changed to allow the same clock skew as permitted by the cluster
649 verification. This will remove some rare but hard to diagnose errors in
656 *(Released Thu, 19 Jul 2012)*
658 Very few changes from rc4 to the final release, only bugfixes:
660 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
661 instance, fix CDROM booting for KVM instances)
662 - fixed node group modification of node parameters
663 - fixed issue in LUClusterVerifyGroup with multi-group clusters
664 - fixed generation of bash completion to ensure a stable ordering
671 *(Released Fri, 13 Jul 2012)*
673 Third release candidate for 2.6. The following changes were done from
676 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
677 - Fixed an inconsistency in the LUXI protocol with the provided
678 arguments (NOT backwards compatible)
679 - Fixed a bug with node groups ipolicy where ``min`` was greater than
680 the cluster ``std`` value
681 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
682 easier instance debugging on nodes (requires ``hconfd`` to work)
688 *(Released Tue, 03 Jul 2012)*
690 Second release candidate for 2.6. The following changes were done from
693 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
695 - Fixed a RAPI regression on missing beparams/memory
696 - Fixed redistribution of files on offline nodes
697 - Added possibility to run activate-disks even though secondaries are
698 offline. With this change it relaxes also the strictness on some other
699 commands which use activate disks internally:
700 * ``gnt-instance start|reboot|rename|backup|export``
701 - Made it possible to remove safely an instance if its secondaries are
703 - Made it possible to reinstall even though secondaries are offline
709 *(Released Mon, 25 Jun 2012)*
711 First release candidate for 2.6. The following changes were done from
714 - Fixed bugs with disk parameters and ``rbd`` templates as well as
716 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
717 behaviour. It supports now the modify operation
718 - ``hcheck`` implemented to analyze cluster health and possibility of
719 improving health by rebalance
720 - ``hbal`` has been improved in dealing with split instances
726 *(Released Mon, 11 Jun 2012)*
728 Second beta release of 2.6. The following changes were done from beta2
731 - Fixed ``daemon-util`` with non-root user models
732 - Fixed creation of plain instances with ``--no-wait-for-sync``
733 - Fix wrong iv_names when running ``cfgupgrade``
734 - Export more information in RAPI group queries
735 - Fixed bug when changing instance network interfaces
736 - Extended burnin to do NIC changes
737 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
738 - Changed default for DRBD barriers
739 - Fixed DRBD error reporting for syncer rate
740 - Verify the options on disk parameters
742 And of course various fixes to documentation and improved unittests and
749 *(Released Wed, 23 May 2012)*
751 First beta release of 2.6. The following changes were done from beta1 to
754 - integrated patch for distributions without ``start-stop-daemon``
755 - adapted example init.d script to work on Fedora
756 - fixed log handling in Haskell daemons
757 - adapted checks in the watcher for pycurl linked against libnss
758 - add partial support for ``xl`` instead of ``xm`` for Xen
759 - fixed a type issue in cluster verification
760 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
763 Plus integrated fixes from the 2.5 branch:
765 - fixed ``kvm-ifup`` to use ``/bin/bash``
766 - fixed parallel build failures
767 - KVM live migration when using a custom keymap
773 *(Released Tue, 24 Jul 2012)*
775 A small bugfix release, with no new features:
777 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
778 different default shell (e.g. Debian, Ubuntu)
779 - fixed KVM startup and live migration with a custom keymap (fixes Issue
780 243 and Debian bug #650664)
781 - fixed compatibility with KVM versions that don't support multiple boot
782 devices (fixes Issue 230 and Debian bug #624256)
784 Additionally, a few fixes were done to the build system (fixed parallel
785 build failures) and to the unittests (fixed race condition in test for
786 FileID functions, and the default enable/disable mode for QA test is now
793 *(Released Fri, 11 May 2012)*
795 A small bugfix release.
797 The main issues solved are on the topic of compatibility with newer LVM
800 - fixed parsing of ``lv_attr`` field
801 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
802 the ``--force`` flag is needed
804 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
805 to accept kernel 3.x too (was hardcoded to 2.6.*).
807 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
808 of hook results and that also made display of other errors suboptimal
809 was fixed; the code behaves now like 2.4 and earlier.
811 Another change in 2.5, the cleanup of the OS scripts environment, is too
812 aggressive: it removed even the ``PATH`` variable, which requires the OS
813 scripts to *always* need to export it. Since this is a bit too strict,
814 we now export a minimal PATH, the same that we export for hooks.
816 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
817 integrated into this release.
819 Finally, a few other miscellaneous changes were done (no new features,
820 just small improvements):
822 - Fix ``gnt-group --help`` display
823 - Fix hardcoded Xen kernel path
824 - Fix grow-disk handling of invalid units
825 - Update synopsis for ``gnt-cluster repair-disk-sizes``
826 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
832 *(Released Thu, 12 Apr 2012)*
834 Incompatible/important changes and bugfixes
835 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
837 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
838 resource's ``ip_check`` parameter changed from ``True`` to ``False``
839 to match the underlying LUXI interface.
840 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
841 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
842 not maintain backwards-compatibility as the underlying operation
843 changed in an incompatible way. The RAPI client can talk to old
844 servers, but it needs to be told so as the return value changed.
845 - When creating file-based instances via RAPI, the ``file_driver``
846 parameter no longer defaults to ``loop`` and must be specified.
847 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
849 - Support for the undocumented and deprecated RAPI instance creation
850 request format version 0 has been dropped. Use version 1, supported
851 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
852 - Pyparsing 1.4.6 or above is required, see :doc:`installation
853 documentation <install>`.
854 - The "cluster-verify" hooks are now executed per group by the
855 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
856 if you just run ``gnt-cluster verify``, which generates one opcode per
858 - The environment as passed to the OS scripts is cleared, and thus no
859 environment variables defined in the node daemon's environment will be
860 inherited by the scripts.
861 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
863 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
864 support operations involving multiple node groups.
865 - Offline nodes are ignored when failing over an instance.
866 - Support for KVM version 1.0, which changed the version reporting format
868 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
870 - ``Makefile`` is now compatible with Automake 1.11.2
871 - Includes all bugfixes made in the 2.4 series
876 - The ganeti-htools project has been merged into the ganeti-core source
877 tree and will be built as part of Ganeti (see :doc:`install-quick`).
878 - Implemented support for :doc:`shared storage <design-shared-storage>`.
879 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
880 GPT-style partition tables (requires `parted
881 <http://www.gnu.org/s/parted/>`_).
882 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
883 - Allowed adding tags on instance creation.
884 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
885 - Allow instances to be started in a paused state, enabling the user to
886 see the complete console output on boot using the console.
887 - Added new hypervisor flag to control default reboot behaviour
888 (``reboot_behavior``).
889 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
890 - Improved out-of-band management support:
892 - Added ``gnt-node health`` command reporting the health status of
894 - Added ``gnt-node power`` command to manage power status of nodes.
895 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
897 - Instance migration can fall back to failover if instance is not
899 - Filters can be used when listing nodes, instances, groups and locks;
900 see :manpage:`ganeti(7)` manpage.
901 - Added post-execution status as variables to :doc:`hooks <hooks>`
903 - Instance tags are exported/imported together with the instance.
904 - When given an explicit job ID, ``gnt-job info`` will work for archived
906 - Jobs can define dependencies on other jobs (not yet supported via
907 RAPI or command line, but used by internal commands and usable via
910 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
913 - Instance failover is now available as a RAPI resource
914 (``/2/instances/[instance_name]/failover``).
915 - ``gnt-instance info`` defaults to static information if primary node
917 - Opcodes have a new ``comment`` attribute.
918 - Added basic SPICE support to KVM hypervisor.
919 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
921 Node group improvements
922 ~~~~~~~~~~~~~~~~~~~~~~~
924 - ``gnt-cluster verify`` has been modified to check groups separately,
925 thereby improving performance.
926 - Node group support has been added to ``gnt-cluster verify-disks``,
927 which now operates per node group.
928 - Watcher has been changed to work better with node groups.
930 - One process and state file per node group.
931 - Slow watcher in one group doesn't block other group's watcher.
933 - Added new command, ``gnt-group evacuate``, to move all instances in a
934 node group to other groups.
935 - Added ``gnt-instance change-group`` to move an instance to another
937 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
938 per-group operations.
939 - Node groups can be tagged.
940 - Some operations switch from an exclusive to a shared lock as soon as
942 - Instance's primary and secondary nodes' groups are now available as
943 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
944 and ``snodes.group.uuid``).
949 - Numerous updates to documentation and manpages.
951 - :doc:`RAPI <rapi>` documentation now has detailed parameter
953 - Some opcode/job results are now also documented, see :doc:`RAPI
956 - A lockset's internal lock is now also visible in lock monitor.
957 - Log messages from job queue workers now contain information about the
958 opcode they're processing.
959 - ``gnt-instance console`` no longer requires the instance lock.
960 - A short delay when waiting for job changes reduces the number of LUXI
961 requests significantly.
962 - DRBD metadata volumes are overwritten with zeros during disk creation.
963 - Out-of-band commands no longer acquire the cluster lock in exclusive
965 - ``devel/upload`` now uses correct permissions for directories.
971 *(Released Fri, 23 Mar 2012)*
973 This was the sixth release candidate of the 2.5 series.
979 *(Released Mon, 9 Jan 2012)*
981 This was the fifth release candidate of the 2.5 series.
987 *(Released Thu, 27 Oct 2011)*
989 This was the fourth release candidate of the 2.5 series.
995 *(Released Wed, 26 Oct 2011)*
997 This was the third release candidate of the 2.5 series.
1003 *(Released Tue, 18 Oct 2011)*
1005 This was the second release candidate of the 2.5 series.
1011 *(Released Tue, 4 Oct 2011)*
1013 This was the first release candidate of the 2.5 series.
1019 *(Released Wed, 31 Aug 2011)*
1021 This was the third beta release of the 2.5 series.
1027 *(Released Mon, 22 Aug 2011)*
1029 This was the second beta release of the 2.5 series.
1035 *(Released Fri, 12 Aug 2011)*
1037 This was the first beta release of the 2.5 series.
1043 *(Released Thu, 27 Oct 2011)*
1045 - Fixed bug when parsing command line parameter values ending in
1047 - Fixed assertion error after unclean master shutdown
1048 - Disable HTTP client pool for RPC, significantly reducing memory usage
1050 - Fixed queue archive creation with wrong permissions
1056 *(Released Tue, 23 Aug 2011)*
1060 - Fixed documentation for importing with ``--src-dir`` option
1061 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1062 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1068 *(Released Fri, 5 Aug 2011)*
1070 Many bug-fixes and a few small features:
1072 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1073 issues when you tried to add an instance with two MAC addresses in one
1075 - KVM: fixed per-instance stored UID value
1076 - KVM: configure bridged NICs at migration start
1077 - KVM: Fix a bug where instance will not start with never KVM versions
1079 - Added OS search path to ``gnt-cluster info``
1080 - Fixed an issue with ``file_storage_dir`` where you were forced to
1081 provide an absolute path, but the documentation states it is a
1082 relative path, the documentation was right
1083 - Added a new parameter to instance stop/start called ``--no-remember``
1084 that will make the state change to not be remembered
1085 - Implemented ``no_remember`` at RAPI level
1086 - Improved the documentation
1087 - Node evacuation: don't call IAllocator if node is already empty
1088 - Fixed bug in DRBD8 replace disks on current nodes
1089 - Fixed bug in recreate-disks for DRBD instances
1090 - Moved assertion checking locks in ``gnt-instance replace-disks``
1091 causing it to abort with not owning the right locks for some situation
1092 - Job queue: Fixed potential race condition when cancelling queued jobs
1093 - Fixed off-by-one bug in job serial generation
1094 - ``gnt-node volumes``: Fix instance names
1095 - Fixed aliases in bash completion
1096 - Fixed a bug in reopening log files after being sent a SIGHUP
1097 - Added a flag to burnin to allow specifying VCPU count
1098 - Bugfixes to non-root Ganeti configuration
1104 *(Released Thu, 12 May 2011)*
1106 Many bug-fixes and a few new small features:
1108 - Fixed a bug related to log opening failures
1109 - Fixed a bug in instance listing with orphan instances
1110 - Fixed a bug which prevented resetting the cluster-level node parameter
1111 ``oob_program`` to the default
1112 - Many fixes related to the ``cluster-merge`` tool
1113 - Fixed a race condition in the lock monitor, which caused failures
1114 during (at least) creation of many instances in parallel
1115 - Improved output for gnt-job info
1116 - Removed the quiet flag on some ssh calls which prevented debugging
1118 - Improved the N+1 failure messages in cluster verify by actually
1119 showing the memory values (needed and available)
1120 - Increased lock attempt timeouts so that when executing long operations
1121 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1122 too early and thus prevent the use of the 'fair' mechanism
1123 - Changed instance query data (``gnt-instance info``) to not acquire
1124 locks unless needed, thus allowing its use on locked instance if only
1125 static information is asked for
1126 - Improved behaviour with filesystems that do not support rename on an
1128 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1129 kept locks on all nodes during the wipe, which is unneeded
1130 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1131 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1132 10GiB) which caused the wipe to fail right at the end in some cases
1133 - Fixed master IP activation when doing master failover with no-voting
1134 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1135 the master node itself
1136 - Fixed potential data-loss in under disk full conditions, where Ganeti
1137 wouldn't check correctly the return code and would consider
1138 partially-written files 'correct'
1139 - Fixed bug related to multiple VGs and DRBD disk replacing
1140 - Added new disk parameter ``metavg`` that allows placement of the meta
1141 device for DRBD in a different volume group
1142 - Fixed error handling in the node daemon when the system libc doesn't
1143 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1144 - Fixed lock release during replace-disks, which kept cluster-wide locks
1145 when doing disk replaces with an iallocator script
1146 - Added check for missing bridges in cluster verify
1147 - Handle EPIPE errors while writing to the terminal better, so that
1148 piping the output to e.g. ``less`` doesn't cause a backtrace
1149 - Fixed rare case where a ^C during Luxi calls could have been
1150 interpreted as server errors, instead of simply terminating
1151 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1153 - Added a few more parameters to the KVM hypervisor, allowing a second
1154 CDROM, custom disk type for CDROMs and a floppy image
1155 - Removed redundant message in instance rename when the name is given
1157 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1158 disks on new nodes, allowing recreation when the original instance
1159 nodes are completely gone
1160 - Added option when converting disk templates to DRBD to skip waiting
1161 for the resync, in order to make the instance available sooner
1162 - Added two new variables to the OS scripts environment (containing the
1164 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1165 to allow use of ``pvgrub`` as bootloader
1166 - Changed the instance memory modifications to only check out-of-memory
1167 conditions on memory increases, and turned the secondary node warnings
1168 into errors (they can still be overridden via ``--force``)
1169 - Fixed the handling of a corner case when the Python installation gets
1170 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1171 to execute a command that doesn't exist
1172 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1173 node of the instance returned failures during instance shutdown; this
1174 adds the option ``--ignore-consistency`` to gnt-instance move
1176 And as usual, various improvements to the error messages, documentation
1183 *(Released Wed, 09 Mar 2011)*
1185 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1186 the RAPI users file if run twice (even with ``--dry-run``).
1188 The release fixes that bug (nothing else changed).
1194 *(Released Mon, 07 Mar 2011)*
1196 Final 2.4.0 release. Just a few small fixes:
1198 - Fixed RAPI node evacuate
1199 - Fixed the kvm-ifup script
1200 - Fixed internal error handling for special job cases
1201 - Updated man page to specify the escaping feature for options
1207 *(Released Mon, 28 Feb 2011)*
1209 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1210 that this feature wiped the disks of the wrong instance, leading to loss
1215 - Fixed title of query field containing instance name
1216 - Expanded the glossary in the documentation
1217 - Fixed one unittest (internal issue)
1223 *(Released Mon, 21 Feb 2011)*
1225 A number of bug fixes plus just a couple functionality changes.
1227 On the user-visible side, the ``gnt-* list`` command output has changed
1228 with respect to "special" field states. The current rc1 style of display
1229 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1230 the default output mode special fields states are displayed as follows:
1232 - Offline resource: ``*``
1233 - Unavailable/not applicable: ``-``
1234 - Data missing (RPC failure): ``?``
1235 - Unknown field: ``??``
1237 Another user-visible change is the addition of ``--force-join`` to
1242 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1243 - Fixed regression in RAPI/instance reinstall where all parameters were
1244 required (instead of optional)
1245 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1246 - Fixed iallocator usage (offline nodes were not considered offline)
1247 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1248 - Fixed hypervisor and OS parameter validation with respect to
1249 non-vm_capable nodes
1250 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1252 - Fixed ``tools/listrunner`` with respect to agent-based usage
1258 *(Released Fri, 4 Feb 2011)*
1260 Many changes and fixes since the beta1 release. While there were some
1261 internal changes, the code has been mostly stabilised for the RC
1264 Note: the dumb allocator was removed in this release, as it was not kept
1265 up-to-date with the IAllocator protocol changes. It is recommended to
1266 use the ``hail`` command from the ganeti-htools package.
1268 Note: the 2.4 and up versions of Ganeti are not compatible with the
1269 0.2.x branch of ganeti-htools. You need to upgrade to
1270 ganeti-htools-0.3.0 (or later).
1272 Regressions fixed from 2.3
1273 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1275 - Fixed the ``gnt-cluster verify-disks`` command
1276 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1278 - Fixed disk adoption breakage
1279 - Fixed wrong headers in instance listing for field aliases
1284 - Fixed corner case in KVM handling of NICs
1285 - Fixed many cases of wrong handling of non-vm_capable nodes
1286 - Fixed a bug where a missing instance symlink was not possible to
1287 recreate with any ``gnt-*`` command (now ``gnt-instance
1288 activate-disks`` does it)
1289 - Fixed the volume group name as reported by ``gnt-cluster
1291 - Increased timeouts for the import-export code, hopefully leading to
1292 fewer aborts due network or instance timeouts
1293 - Fixed bug in ``gnt-node list-storage``
1294 - Fixed bug where not all daemons were started on cluster
1295 initialisation, but only at the first watcher run
1296 - Fixed many bugs in the OOB implementation
1297 - Fixed watcher behaviour in presence of instances with offline
1299 - Fixed instance list output for instances running on the wrong node
1300 - a few fixes to the cluster-merge tool, but it still cannot merge
1301 multi-node groups (currently it is not recommended to use this tool)
1307 - Improved network configuration for the KVM hypervisor
1308 - Added e1000 as a supported NIC for Xen-HVM
1309 - Improved the lvmstrap tool to also be able to use partitions, as
1310 opposed to full disks
1311 - Improved speed of disk wiping (the cluster parameter
1312 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1313 of instance creations
1314 - Added documentation for the OS parameters
1315 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1316 hypervisor is not responding
1317 - Added display of blacklisted and hidden OS information in
1318 ``gnt-cluster info``
1319 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1320 NIC and node parameters, which might create problems with currently
1321 invalid (but undetected) configuration files, but prevents validation
1322 failures when unrelated parameters are modified
1323 - Changed cluster initialisation to wait for the master daemon to become
1325 - Expanded the RAPI interface:
1327 - Added config redistribution resource
1328 - Added activation/deactivation of instance disks
1329 - Added export of console information
1331 - Implemented log file reopening on SIGHUP, which allows using
1332 logrotate(8) for the Ganeti log files
1333 - Added a basic OOB helper script as an example
1339 *(Released Fri, 14 Jan 2011)*
1344 - Fixed timezone issues when formatting timestamps
1345 - Added support for node groups, available via ``gnt-group`` and other
1347 - Added out-of-band framework and management, see :doc:`design
1348 document <design-oob>`
1349 - Removed support for roman numbers from ``gnt-node list`` and
1350 ``gnt-instance list``.
1351 - Allowed modification of master network interface via ``gnt-cluster
1352 modify --master-netdev``
1353 - Accept offline secondaries while shutting down instance disks
1354 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1355 - Added support for multiple LVM volume groups
1356 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1358 - Added commands to list available fields:
1360 - ``gnt-node list-fields``
1361 - ``gnt-group list-fields``
1362 - ``gnt-instance list-fields``
1364 - Updated documentation and man pages
1369 - Moved ``rapi_users`` file into separate directory, now named
1370 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1372 - Added new tool for running commands on many machines,
1373 ``tools/ganeti-listrunner``
1374 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1375 improving the ``gnt-instance console`` output
1376 - Allowed customisation of disk index separator at ``configure`` time
1377 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1378 - Added support for non-partitioned md disks in ``lvmstrap``
1379 - Added script to gracefully power off KVM instances
1380 - Split ``utils`` module into smaller parts
1381 - Changed query operations to return more detailed information, e.g.
1382 whether an information is unavailable due to an offline node. To use
1383 this new functionality, the LUXI call ``Query`` must be used. Field
1384 information is now stored by the master daemon and can be retrieved
1385 using ``QueryFields``. Instances, nodes and groups can also be queried
1386 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1387 exposed via RAPI). The following commands make use of this
1388 infrastructure change:
1390 - ``gnt-group list``
1391 - ``gnt-group list-fields``
1393 - ``gnt-node list-fields``
1394 - ``gnt-instance list``
1395 - ``gnt-instance list-fields``
1396 - ``gnt-debug locks``
1401 - New RAPI resources (see :doc:`rapi`):
1405 - ``/2/groups/[group_name]``
1406 - ``/2/groups/[group_name]/assign-nodes``
1407 - ``/2/groups/[group_name]/modify``
1408 - ``/2/groups/[group_name]/rename``
1409 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1413 - Implemented ``no_install`` for instance creation
1414 - Implemented OS parameters for instance reinstallation, allowing
1415 use of special settings on reinstallation (e.g. for preserving data)
1420 - Added IPv6 support in import/export
1421 - Pause DRBD synchronization while wiping disks on instance creation
1422 - Updated unittests and QA scripts
1423 - Improved network parameters passed to KVM
1424 - Converted man pages from docbook to reStructuredText
1430 *(Released Mon, 20 Dec 2010)*
1432 Released version 2.3.1~rc1 without any changes.
1438 *(Released Wed, 1 Dec 2010)*
1440 - impexpd: Disable OpenSSL compression in socat if possible (backport
1441 from master, commit e90739d625b, see :doc:`installation guide
1442 <install-quick>` for details)
1443 - Changed unittest coverage report to exclude test scripts
1444 - Added script to check version format
1450 *(Released Wed, 1 Dec 2010)*
1452 Released version 2.3.0~rc1 without any changes.
1458 *(Released Fri, 19 Nov 2010)*
1460 A number of bugfixes and documentation updates:
1462 - Update ganeti-os-interface documentation
1463 - Fixed a bug related to duplicate MACs or similar items which should be
1465 - Fix breakage in OS state modify
1466 - Reinstall instance: disallow offline secondaries (fixes bug related to
1467 OS changing but reinstall failing)
1468 - plus all the other fixes between 2.2.1 and 2.2.2
1474 *(Released Tue, 2 Nov 2010)*
1476 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1477 - Fixed master failover race with watcher
1478 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1480 - Accept previously stopped instance for export with instance removal
1481 - Simplify and extend the environment variables for instance OS scripts
1482 - Added new node flags, ``master_capable`` and ``vm_capable``
1483 - Added optional instance disk wiping prior during allocation. This is a
1484 cluster-wide option and can be set/modified using
1485 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1486 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1487 :doc:`install-quick`
1488 - Added a new watcher option (``--ignore-pause``)
1489 - Added option to ignore offline node on instance start/stop
1490 (``--ignore-offline``)
1491 - Allow overriding OS parameters with ``gnt-instance reinstall``
1492 - Added ability to change node's secondary IP address using ``gnt-node
1494 - Implemented privilege separation for all daemons except
1495 ``ganeti-noded``, see ``configure`` options
1496 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1498 - Implemented job priorities (see ``ganeti(7)`` manpage)
1499 - Ignore failures while shutting down instances during failover from
1501 - Exit daemon's bootstrap process only once daemon is ready
1502 - Export more information via ``LUInstanceQuery``/remote API
1503 - Improved documentation, QA and unittests
1504 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1505 restart if the file was created or changed
1506 - Added LUXI protocol version sent with each request and response,
1507 allowing detection of server/client mismatches
1508 - Moved the Python scripts among gnt-* and ganeti-* into modules
1509 - Moved all code related to setting up SSH to an external script,
1511 - Infrastructure changes for node group support in future versions
1517 *(Released Fri, 19 Nov 2010)*
1519 A few small bugs fixed, and some improvements to the build system:
1521 - Fix documentation regarding conversion to drbd
1522 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1524 - Fix error handling in node modify with multiple changes
1525 - Allow remote imports without checked names
1531 *(Released Tue, 19 Oct 2010)*
1533 - Disable SSL session ID cache in RPC client
1539 *(Released Thu, 14 Oct 2010)*
1541 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1542 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1544 - Fix problems with interaction between Curl and Python's HTTP server,
1545 resulting in increased speed in many RPC calls
1546 - Improve our release script to prevent breakage with older aclocal and
1553 *(Released Thu, 7 Oct 2010)*
1555 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1556 value retrieved from master
1557 - Added support for blacklisted or hidden OS definitions
1558 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1559 - Added support for -mem-path in KVM hypervisor abstraction layer
1560 - Allow overriding instance parameters in tool for inter-cluster
1561 instance moves (``tools/move-instance``)
1562 - Improved opcode summaries (e.g. in ``gnt-job list``)
1563 - Improve consistency of OS listing by sorting it
1564 - Documentation updates
1570 *(Released Fri, 8 Oct 2010)*
1572 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1578 *(Released Mon, 4 Oct 2010)*
1580 - Fixed regression in ``gnt-instance rename``
1586 *(Released Wed, 22 Sep 2010)*
1588 - Fixed OS_VARIANT variable for OS scripts
1589 - Fixed cluster tag operations via RAPI
1590 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1591 - Disabled RAPI CA checks in watcher
1597 *(Released Mon, 23 Aug 2010)*
1599 - Support DRBD versions of the format "a.b.c.d"
1601 - Re-introduce support for usage from multiple threads in RAPI client
1602 - Instance renames and modify via RAPI
1603 - Work around race condition between processing and archival in job
1605 - Mark opcodes following failed one as failed, too
1606 - Job field ``lock_status`` was removed due to difficulties making it
1607 work with the changed job queue in Ganeti 2.2; a better way to monitor
1608 locks is expected for a later 2.2.x release
1609 - Fixed dry-run behaviour with many commands
1610 - Support ``ssh-agent`` again when adding nodes
1611 - Many additional bugfixes
1617 *(Released Fri, 30 Jul 2010)*
1619 Important change: the internal RPC mechanism between Ganeti nodes has
1620 changed from using a home-grown http library (based on the Python base
1621 libraries) to use the PycURL library. This requires that PycURL is
1622 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1623 against GnuTLS by default. cURL's support for GnuTLS had known issues
1624 before cURL 7.21.0 and we recommend using the latest cURL release or
1625 linking against OpenSSL. Most other distributions already link PycURL
1626 and cURL against OpenSSL. The command::
1628 python -c 'import pycurl; print pycurl.version'
1630 can be used to determine the libraries PycURL and cURL are linked
1633 Other significant changes:
1635 - Rewrote much of the internals of the job queue, in order to achieve
1636 better parallelism; this decouples job query operations from the job
1637 processing, and it should allow much nicer behaviour of the master
1638 daemon under load, and it also has uncovered some long-standing bugs
1639 related to the job serialisation (now fixed)
1640 - Added a default iallocator setting to the cluster parameters,
1641 eliminating the need to always pass nodes or an iallocator for
1642 operations that require selection of new node(s)
1643 - Added experimental support for the LXC virtualization method
1644 - Added support for OS parameters, which allows the installation of
1645 instances to pass parameter to OS scripts in order to customise the
1647 - Added a hypervisor parameter controlling the migration type (live or
1648 non-live), since hypervisors have various levels of reliability; this
1649 has renamed the 'live' parameter to 'mode'
1650 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1651 logical volumes, meaning that cluster verify will ignore them and not
1652 flag their presence as errors
1653 - The watcher will now reset the error count for failed instances after
1654 8 hours, thus allowing self-healing if the problem that caused the
1655 instances to be down/fail to start has cleared in the meantime
1656 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1657 check for, and warn, if the drbd module parameter ``usermode_helper``
1658 is not consistent with the cluster-wide setting; this is needed to
1659 make diagnose easier of failed drbd creations
1660 - Started adding base IPv6 support, but this is not yet
1661 enabled/available for use
1662 - Rename operations (cluster, instance) will now return the new name,
1663 which is especially useful if a short name was passed in
1664 - Added support for instance migration in RAPI
1665 - Added a tool to pre-configure nodes for the SSH setup, before joining
1666 them to the cluster; this will allow in the future a simplified model
1667 for node joining (but not yet fully enabled in 2.2); this needs the
1668 paramiko python library
1669 - Fixed handling of name-resolving errors
1670 - Fixed consistency of job results on the error path
1671 - Fixed master-failover race condition when executed multiple times in
1673 - Fixed many bugs related to the job queue (mostly introduced during the
1674 2.2 development cycle, so not all are impacting 2.1)
1675 - Fixed instance migration with missing disk symlinks
1676 - Fixed handling of unknown jobs in ``gnt-job archive``
1677 - And many other small fixes/improvements
1681 - Enhanced both the unittest and the QA coverage
1682 - Switched the opcode validation to a generic model, and extended the
1683 validation to all opcode parameters
1684 - Changed more parts of the code that write shell scripts to use the
1686 - Switched the master daemon to use the asyncore library for the Luxi
1693 *(Released Thu, 17 Jun 2010)*
1695 - Added tool (``move-instance``) and infrastructure to move instances
1696 between separate clusters (see :doc:`separate documentation
1697 <move-instance>` and :doc:`design document <design-2.2>`)
1698 - Added per-request RPC timeout
1699 - RAPI now requires a Content-Type header for requests with a body (e.g.
1700 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1701 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1702 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1704 - Implemented initial support for running Ganeti daemons as separate
1705 users, see configure-time flags ``--with-user-prefix`` and
1706 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1708 - Instances can be removed after export (``gnt-backup export
1709 --remove-instance``)
1710 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1711 key (instead of 1024 bit)
1712 - Added new cluster configuration file for cluster domain secret
1713 - Import/export now use SSL instead of SSH
1714 - Added support for showing estimated time when exporting an instance,
1715 see the ``ganeti-os-interface(7)`` manpage and look for
1722 *(Released Tue, 16 Nov 2010)*
1724 Some more bugfixes. Unless critical bugs occur, this will be the last
1727 - Fix case of MAC special-values
1728 - Fix mac checker regex
1729 - backend: Fix typo causing "out of range" error
1730 - Add missing --units in gnt-instance list man page
1736 *(Released Tue, 24 Aug 2010)*
1739 - Don't ignore secondary node silently on non-mirrored disk templates
1741 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1742 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1743 - Properly document the usb_mouse parameter
1744 - Fix path in ganeti-rapi(8) (issue 116)
1745 - Adjust error message when the ganeti user's .ssh directory is
1747 - Add same-node-check when changing the disk template to drbd
1753 *(Released Fri, 16 Jul 2010)*
1756 - Add an option to only select some reboot types during qa/burnin.
1757 (on some hypervisors consequent reboots are not supported)
1758 - Fix infrequent race condition in master failover. Sometimes the old
1759 master ip address would be still detected as up for a short time
1760 after it was removed, causing failover to fail.
1761 - Decrease mlockall warnings when the ctypes module is missing. On
1762 Python 2.4 we support running even if no ctypes module is installed,
1763 but we were too verbose about this issue.
1764 - Fix building on old distributions, on which man doesn't have a
1766 - Fix RAPI not to ignore the MAC address on instance creation
1767 - Implement the old instance creation format in the RAPI client.
1773 *(Released Thu, 01 Jul 2010)*
1775 A small bugfix release:
1776 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1777 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1778 a non-existing option
1779 - Fix instance create: the --force-variant option was ignored
1780 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1781 - Fix modify node storage with non-FQDN arguments
1782 - Fix RAPI client to authenticate under Python 2.6 when used
1783 for more than 5 requests needing authentication
1784 - Fix gnt-instance modify -t (storage) giving a wrong error message
1785 when converting a non-shutdown drbd instance to plain
1791 *(Released Fri, 18 Jun 2010)*
1793 A small bugfix release:
1795 - Fix live migration of KVM instances started with older Ganeti
1796 versions which had fewer hypervisor parameters
1797 - Fix gnt-instance grow-disk on down instances
1798 - Fix an error-reporting bug during instance migration
1799 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1800 silently ignoring broken ones
1801 - Fix an RPC error reporting bug affecting, for example, RAPI client
1803 - Fix bug triggered by different API version os-es on different nodes
1804 - Fix a bug in instance startup with custom hvparams: OS level
1805 parameters would fail to be applied.
1806 - Fix the RAPI client under Python 2.6 (but more work is needed to
1807 make it work completely well with OpenSSL)
1808 - Fix handling of errors when resolving names from DNS
1814 *(Released Thu, 3 Jun 2010)*
1816 A medium sized development cycle. Some new features, and some
1817 fixes/small improvements/cleanups.
1819 Significant features
1820 ~~~~~~~~~~~~~~~~~~~~
1822 The node deamon now tries to mlock itself into memory, unless the
1823 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1824 its logs, and falls back to console logging. This allows emergency
1825 features such as ``gnt-node powercycle`` to work even in the event of a
1826 broken node disk (tested offlining the disk hosting the node's
1827 filesystem and dropping its memory caches; don't try this at home)
1829 KVM: add vhost-net acceleration support. It can be tested with a new
1830 enough version of the kernel and of qemu-kvm.
1832 KVM: Add instance chrooting feature. If you use privilege dropping for
1833 your VMs you can also now force them to chroot to an empty directory,
1834 before starting the emulated guest.
1836 KVM: Add maximum migration bandwith and maximum downtime tweaking
1837 support (requires a new-enough version of qemu-kvm).
1839 Cluster verify will now warn if the master node doesn't have the master
1840 ip configured on it.
1842 Add a new (incompatible) instance creation request format to RAPI which
1843 supports all parameters (previously only a subset was supported, and it
1844 wasn't possible to extend the old format to accomodate all the new
1845 features. The old format is still supported, and a client can check for
1846 this feature, before using it, by checking for its presence in the
1847 ``features`` RAPI resource.
1849 Now with ancient latin support. Try it passing the ``--roman`` option to
1850 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1851 (requires the python-roman module to be installed, in order to work).
1856 As usual many internal code refactorings, documentation updates, and
1859 - Lots of improvements and cleanups to the experimental Remote API
1860 (RAPI) client library.
1861 - A new unit test suite for the core daemon libraries.
1862 - A fix to creating missing directories makes sure the umask is not
1863 applied anymore. This enforces the same directory permissions
1865 - Better handling terminating daemons with ctrl+c (used when running
1866 them in debugging mode).
1867 - Fix a race condition in live migrating a KVM instance, when stat()
1868 on the old proc status file returned EINVAL, which is an unexpected
1870 - Fixed manpage checking with newer man and utf-8 charachters. But now
1871 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1877 *(Released Fri, 7 May 2010)*
1879 Fix a bug which prevented untagged KVM instances from starting.
1885 *(Released Fri, 7 May 2010)*
1887 Another release with a long development cycle, during which many
1888 different features were added.
1890 Significant features
1891 ~~~~~~~~~~~~~~~~~~~~
1893 The KVM hypervisor now can run the individual instances as non-root, to
1894 reduce the impact of a VM being hijacked due to bugs in the
1895 hypervisor. It is possible to run all instances as a single (non-root)
1896 user, to manually specify a user for each instance, or to dynamically
1897 allocate a user out of a cluster-wide pool to each instance, with the
1898 guarantee that no two instances will run under the same user ID on any
1901 An experimental RAPI client library, that can be used standalone
1902 (without the other Ganeti libraries), is provided in the source tree as
1903 ``lib/rapi/client.py``. Note this client might change its interface in
1904 the future, as we iterate on its capabilities.
1906 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1907 replace the cluster's certificates and crypto keys. This might help in
1908 case they have been compromised, or have simply expired.
1910 A new disk option for instance creation has been added that allows one
1911 to "adopt" currently existing logical volumes, with data
1912 preservation. This should allow easier migration to Ganeti from
1913 unmanaged (or managed via other software) instances.
1915 Another disk improvement is the possibility to convert between redundant
1916 (DRBD) and plain (LVM) disk configuration for an instance. This should
1917 allow better scalability (starting with one node and growing the
1918 cluster, or shrinking a two-node cluster to one node).
1920 A new feature that could help with automated node failovers has been
1921 implemented: if a node sees itself as offline (by querying the master
1922 candidates), it will try to shutdown (hard) all instances and any active
1923 DRBD devices. This reduces the risk of duplicate instances if an
1924 external script automatically failovers the instances on such nodes. To
1925 enable this, the cluster parameter ``maintain_node_health`` should be
1926 enabled; in the future this option (per the name) will enable other
1927 automatic maintenance features.
1929 Instance export/import now will reuse the original instance
1930 specifications for all parameters; that means exporting an instance,
1931 deleting it and the importing it back should give an almost identical
1932 instance. Note that the default import behaviour has changed from
1933 before, where it created only one NIC; now it recreates the original
1936 Cluster verify has added a few new checks: SSL certificates validity,
1937 /etc/hosts consistency across the cluster, etc.
1942 As usual, many internal changes were done, documentation fixes,
1945 - Fixed cluster initialization with disabled cluster storage (regression
1946 introduced in 2.1.1)
1947 - File-based storage supports growing the disks
1948 - Fixed behaviour of node role changes
1949 - Fixed cluster verify for some corner cases, plus a general rewrite of
1950 cluster verify to allow future extension with more checks
1951 - Fixed log spamming by watcher and node daemon (regression introduced
1953 - Fixed possible validation issues when changing the list of enabled
1955 - Fixed cleanup of /etc/hosts during node removal
1956 - Fixed RAPI response for invalid methods
1957 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1958 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1959 from ide disks, etc.)
1960 - Allow OS changes without re-installation (to record a changed OS
1961 outside of Ganeti, or to allow OS renames)
1962 - Allow instance creation without OS installation (useful for example if
1963 the OS will be installed manually, or restored from a backup not in
1965 - Implemented option to make cluster ``copyfile`` use the replication
1967 - Added list of enabled hypervisors to ssconf (possibly useful for
1969 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1971 - A partial form of node re-IP is possible via node readd, which now
1972 allows changed node primary IP
1973 - Command line utilities now show an informational message if the job is
1975 - The logs of the master daemon now show the PID/UID/GID of the
1982 *(Released Fri, 12 Mar 2010)*
1984 During the 2.1.0 long release candidate cycle, a lot of improvements and
1985 changes have accumulated with were released later as 2.1.1.
1990 The node evacuate command (``gnt-node evacuate``) was significantly
1991 rewritten, and as such the IAllocator protocol was changed - a new
1992 request type has been added. This unfortunate change during a stable
1993 series is designed to improve performance of node evacuations; on
1994 clusters with more than about five nodes and which are well-balanced,
1995 evacuation should proceed in parallel for all instances of the node
1996 being evacuated. As such, any existing IAllocator scripts need to be
1997 updated, otherwise the above command will fail due to the unknown
1998 request. The provided "dumb" allocator has not been updated; but the
1999 ganeti-htools package supports the new protocol since version 0.2.4.
2001 Another important change is increased validation of node and instance
2002 names. This might create problems in special cases, if invalid host
2003 names are being used.
2005 Also, a new layer of hypervisor parameters has been added, that sits at
2006 OS level between the cluster defaults and the instance ones. This allows
2007 customisation of virtualization parameters depending on the installed
2008 OS. For example instances with OS 'X' may have a different KVM kernel
2009 (or any other parameter) than the cluster defaults. This is intended to
2010 help managing a multiple OSes on the same cluster, without manual
2011 modification of each instance's parameters.
2013 A tool for merging clusters, ``cluster-merge``, has been added in the
2014 tools sub-directory.
2019 - Improved the int/float conversions that should make the code more
2020 robust in face of errors from the node daemons
2021 - Fixed the remove node code in case of internal configuration errors
2022 - Fixed the node daemon behaviour in face of inconsistent queue
2023 directory (e.g. read-only file-system where we can't open the files
2025 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2026 now it either aborts cleanly or, if given the new "auto_promote"
2027 parameter, will automatically promote other nodes as needed
2028 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2029 completely prevent Ganeti from working
2030 - Fixed bug for instance export when not all disks were successfully
2032 - Fixed behaviour of node add when the new node is slow in starting up
2034 - Fixed handling of signals in the LUXI client, which should improve
2035 behaviour of command-line scripts
2036 - Added checks for invalid node/instance names in the configuration (now
2037 flagged during cluster verify)
2038 - Fixed watcher behaviour for disk activation errors
2039 - Fixed two potentially endless loops in http library, which led to the
2040 RAPI daemon hanging and consuming 100% CPU in some cases
2041 - Fixed bug in RAPI daemon related to hashed passwords
2042 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2044 - Enhanced compatibility with non-Debian OSes, but not using absolute
2045 path in some commands and allowing customisation of the ssh
2046 configuration directory
2047 - Fixed possible future issue with new Python versions by abiding to the
2048 proper use of ``__slots__`` attribute on classes
2049 - Added checks that should prevent directory traversal attacks
2050 - Many documentation fixes based on feedback from users
2055 - Added an "early_release" more for instance replace disks and node
2056 evacuate, where we release locks earlier and thus allow higher
2057 parallelism within the cluster
2058 - Added watcher hooks, intended to allow the watcher to restart other
2059 daemons (e.g. from the ganeti-nbma project), but they can be used of
2060 course for any other purpose
2061 - Added a compile-time disable for DRBD barriers, to increase
2062 performance if the administrator trusts the power supply or the
2063 storage system to not lose writes
2064 - Added the option of using syslog for logging instead of, or in
2065 addition to, Ganeti's own log files
2066 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2067 can indeed boot from a paravirtual NIC
2068 - Added a generic debug level for many operations; while this is not
2069 used widely yet, it allows one to pass the debug value all the way to
2071 - Enhanced the hooks environment for instance moves (failovers,
2072 migrations) where the primary/secondary nodes changed during the
2073 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2074 - Enhanced data validations for many user-supplied values; one important
2075 item is the restrictions imposed on instance and node names, which
2076 might reject some (invalid) host names
2077 - Add a configure-time option to disable file-based storage, if it's not
2078 needed; this allows greater security separation between the master
2079 node and the other nodes from the point of view of the inter-node RPC
2081 - Added user notification in interactive tools if job is waiting in the
2082 job queue or trying to acquire locks
2083 - Added log messages when a job is waiting for locks
2084 - Added filtering by node tags in instance operations which admit
2085 multiple instances (start, stop, reboot, reinstall)
2086 - Added a new tool for cluster mergers, ``cluster-merge``
2087 - Parameters from command line which are of the form ``a=b,c=d`` can now
2088 use backslash escapes to pass in values which contain commas,
2089 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2091 - For KVM, the instance name is the first parameter passed to KVM, so
2092 that it's more visible in the process list
2098 *(Released Tue, 2 Mar 2010)*
2100 Ganeti 2.1 brings many improvements with it. Major changes:
2102 - Added infrastructure to ease automated disk repairs
2103 - Added new daemon to export configuration data in a cheaper way than
2104 using the remote API
2105 - Instance NICs can now be routed instead of being associated with a
2107 - Improved job locking logic to reduce impact of jobs acquiring multiple
2108 locks waiting for other long-running jobs
2110 In-depth implementation details can be found in the Ganeti 2.1 design
2116 - Added chroot hypervisor
2117 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2119 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2120 ``bootloader_path`` and ``bootloader_args``)
2121 - Added the ``use_localtime`` option for the xen-hvm and kvm
2122 hypervisors, and the default value for this has changed to false (in
2123 2.0 xen-hvm always enabled it)
2124 - Added luxi call to submit multiple jobs in one go
2125 - Added cluster initialization option to not modify ``/etc/hosts``
2127 - Added network interface parameters
2128 - Added dry run mode to some LUs
2129 - Added RAPI resources:
2131 - ``/2/instances/[instance_name]/info``
2132 - ``/2/instances/[instance_name]/replace-disks``
2133 - ``/2/nodes/[node_name]/evacuate``
2134 - ``/2/nodes/[node_name]/migrate``
2135 - ``/2/nodes/[node_name]/role``
2136 - ``/2/nodes/[node_name]/storage``
2137 - ``/2/nodes/[node_name]/storage/modify``
2138 - ``/2/nodes/[node_name]/storage/repair``
2140 - Added OpCodes to evacuate or migrate all instances on a node
2141 - Added new command to list storage elements on nodes (``gnt-node
2142 list-storage``) and modify them (``gnt-node modify-storage``)
2143 - Added new ssconf files with master candidate IP address
2144 (``ssconf_master_candidates_ips``), node primary IP address
2145 (``ssconf_node_primary_ips``) and node secondary IP address
2146 (``ssconf_node_secondary_ips``)
2147 - Added ``ganeti-confd`` and a client library to query the Ganeti
2148 configuration via UDP
2149 - Added ability to run hooks after cluster initialization and before
2151 - Added automatic mode for disk replace (``gnt-instance replace-disks
2153 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2154 after catastrophic data-loss
2155 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2157 - Added ``gnt-instance move`` command to move instances
2158 - Added ``gnt-cluster watcher`` command to control watcher
2159 - Added ``gnt-node powercycle`` command to powercycle nodes
2160 - Added new job status field ``lock_status``
2161 - Added parseable error codes to cluster verification (``gnt-cluster
2162 verify --error-codes``) and made output less verbose (use
2163 ``--verbose`` to restore previous behaviour)
2164 - Added UUIDs to the main config entities (cluster, nodes, instances)
2165 - Added support for OS variants
2166 - Added support for hashed passwords in the Ganeti remote API users file
2168 - Added option to specify maximum timeout on instance shutdown
2169 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2170 - Added new helper script to start and stop Ganeti daemons
2171 (``daemon-util``), with the intent to reduce the work necessary to
2172 adjust Ganeti for non-Debian distributions and to start/stop daemons
2174 - Added more unittests
2175 - Fixed critical bug in ganeti-masterd startup
2176 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2177 now customisable at the cluster level for both the KVM and Xen
2178 hypervisors using the new ``migration_port`` parameter
2179 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2180 reinstalling an instance
2181 - Allowed ``@`` in tag names
2182 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2183 - Many documentation updates
2184 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2185 - ``gnt-instance reinstall`` can now reinstall multiple instances
2186 - Updated many command line parameters
2187 - Introduced new OS API version 15
2188 - No longer support a default hypervisor
2189 - Treat virtual LVs as inexistent
2190 - Improved job locking logic to reduce lock contention
2191 - Match instance and node names case insensitively
2192 - Reimplemented bash completion script to be more complete
2199 *(Released Thu, 4 Feb 2010)*
2201 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2202 - Fix a string formatting bug
2203 - Improve safety of the code in some error paths
2204 - Improve data validation in the master of values returned from nodes
2210 *(Released Thu, 17 Dec 2009)*
2212 - Fix security issue due to missing validation of iallocator names; this
2213 allows local and remote execution of arbitrary executables
2214 - Fix failure of gnt-node list during instance removal
2215 - Ship the RAPI documentation in the archive
2221 *(Released Wed, 30 Sep 2009)*
2223 - Fixed many wrong messages
2224 - Fixed a few bugs related to the locking library
2225 - Fixed MAC checking at instance creation time
2226 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2227 - Fixed a few issues related to signal handling in both daemons and
2229 - Fixed the example startup script provided
2230 - Fixed insserv dependencies in the example startup script (patch from
2232 - Fixed handling of drained nodes in the iallocator framework
2233 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2235 - Fixed error related to invalid job IDs in job polling
2236 - Fixed job/opcode persistence on unclean master shutdown
2237 - Fixed handling of partial job processing after unclean master
2239 - Fixed error reporting from LUs, previously all errors were converted
2240 into execution errors
2241 - Fixed error reporting from burnin
2242 - Decreased significantly the memory usage of the job queue
2243 - Optimised slightly multi-job submission
2244 - Optimised slightly opcode loading
2245 - Backported the multi-job submit framework from the development
2246 branch; multi-instance start and stop should be faster
2247 - Added script to clean archived jobs after 21 days; this will reduce
2248 the size of the queue directory
2249 - Added some extra checks in disk size tracking
2250 - Added an example ethers hook script
2251 - Added a cluster parameter that prevents Ganeti from modifying of
2253 - Added more node information to RAPI responses
2254 - Added a ``gnt-job watch`` command that allows following the ouput of a
2256 - Added a bind-address option to ganeti-rapi
2257 - Added more checks to the configuration verify
2258 - Enhanced the burnin script such that some operations can be retried
2260 - Converted instance reinstall to multi-instance model
2266 *(Released Fri, 7 Aug 2009)*
2268 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2269 to allow using the pre-2.0.2 behaviour in activation, if any existing
2270 instances have mismatched disk sizes in the configuration
2271 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2272 any configuration mismatches for disk sizes
2273 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2274 failover to work on two-node clusters
2275 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2277 - Fixed detection of OS script errors in ``gnt-backup export``
2278 - Fixed exit code of ``gnt-backup export``
2284 *(Released Fri, 17 Jul 2009)*
2286 - Added experimental support for stripped logical volumes; this should
2287 enhance performance but comes with a higher complexity in the block
2288 device handling; stripping is only enabled when passing
2289 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2290 affected even in the non-stripped mode
2291 - Improved resiliency against transient failures at the end of DRBD
2292 resyncs, and in general of DRBD resync checks
2293 - Fixed a couple of issues with exports and snapshot errors
2294 - Fixed a couple of issues in instance listing
2295 - Added display of the disk size in ``gnt-instance info``
2296 - Fixed checking for valid OSes in instance creation
2297 - Fixed handling of the "vcpus" parameter in instance listing and in
2298 general of invalid parameters
2299 - Fixed http server library, and thus RAPI, to handle invalid
2300 username/password combinations correctly; this means that now they
2301 report unauthorized for queries too, not only for modifications,
2302 allowing earlier detect of configuration problems
2303 - Added a new "role" node list field, equivalent to the master/master
2304 candidate/drained/offline flags combinations
2305 - Fixed cluster modify and changes of candidate pool size
2306 - Fixed cluster verify error messages for wrong files on regular nodes
2307 - Fixed a couple of issues with node demotion from master candidate role
2308 - Fixed node readd issues
2309 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2310 - Added a new ``--no-voting`` option for masterfailover to fix failover
2311 on two-nodes clusters when the former master node is unreachable
2312 - Added instance reinstall over RAPI
2318 *(Released Tue, 16 Jun 2009)*
2320 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2321 allow re-adding the start in single-user option (regression from 1.2)
2322 - the watcher writes the instance status to a file, to allow monitoring
2323 to report the instance status (from the master) based on cached
2324 results of the watcher's queries; while this can get stale if the
2325 watcher is being locked due to other work on the cluster, this is
2326 still an improvement
2327 - the watcher now also restarts the node daemon and the rapi daemon if
2329 - fixed the watcher to handle full and drained queue cases
2330 - hooks export more instance data in the environment, which helps if
2331 hook scripts need to take action based on the instance's properties
2332 (no longer need to query back into ganeti)
2333 - instance failovers when the instance is stopped do not check for free
2334 RAM, so that failing over a stopped instance is possible in low memory
2336 - rapi uses queries for tags instead of jobs (for less job traffic), and
2337 for cluster tags it won't talk to masterd at all but read them from
2339 - a couple of error handling fixes in RAPI
2340 - drbd handling: improved the error handling of inconsistent disks after
2341 resync to reduce the frequency of "there are some degraded disks for
2342 this instance" messages
2343 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2344 error handling path called a wrong function name)
2350 *(Released Wed, 27 May 2009)*
2352 - no changes from rc5
2358 *(Released Wed, 20 May 2009)*
2360 - fix a couple of bugs (validation, argument checks)
2361 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2362 - some small improvements to RAPI and IAllocator
2363 - make watcher automatically start the master daemon if down
2369 *(Released Mon, 27 Apr 2009)*
2371 - change the OS list to not require locks; this helps with big clusters
2372 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2373 volume group is broken
2374 - ``gnt-instance info``, without any arguments, doesn't run for all
2375 instances anymore; either pass ``--all`` or pass the desired
2376 instances; this helps against mistakes on big clusters where listing
2377 the information for all instances takes a long time
2378 - miscellaneous doc and man pages fixes
2384 *(Released Wed, 8 Apr 2009)*
2386 - Change the internal locking model of some ``gnt-node`` commands, in
2387 order to reduce contention (and blocking of master daemon) when
2388 batching many creation/reinstall jobs
2389 - Fixes to Xen soft reboot
2390 - No longer build documentation at build time, instead distribute it in
2391 the archive, in order to reduce the need for the whole docbook/rst
2398 *(Released Fri, 27 Mar 2009)*
2400 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2401 - Fix watcher startup sequence, improves the behaviour of busy clusters
2402 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2403 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2404 bind address and other places
2405 - Some documentation fixes and updates
2411 *(Released Mon, 2 Mar 2009)*
2413 - More documentation updates, now all docs should be more-or-less
2415 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2417 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2419 - Fix an issue related to $libdir/run/ganeti and cluster creation
2425 *(Released Thu, 19 Feb 2009)*
2427 - Xen PVM and KVM have switched the default value for the instance root
2428 disk to the first partition on the first drive, instead of the whole
2429 drive; this means that the OS installation scripts must be changed
2431 - Man pages have been updated
2432 - RAPI has been switched by default to HTTPS, and the exported functions
2433 should all work correctly
2434 - RAPI v1 has been removed
2435 - Many improvements to the KVM hypervisor
2436 - Block device errors are now better reported
2437 - Many other bugfixes and small improvements
2443 *(Released Mon, 26 Jan 2009)*
2445 - Version 2 is a general rewrite of the code and therefore the
2446 differences are too many to list, see the design document for 2.0 in
2447 the ``doc/`` subdirectory for more details
2448 - In this beta version there is not yet a migration path from 1.2 (there
2449 will be one in the final 2.0 release)
2450 - A few significant changes are:
2452 - all commands are executed by a daemon (``ganeti-masterd``) and the
2453 various ``gnt-*`` commands are just front-ends to it
2454 - all the commands are entered into, and executed from a job queue,
2455 see the ``gnt-job(8)`` manpage
2456 - the RAPI daemon supports read-write operations, secured by basic
2457 HTTP authentication on top of HTTPS
2458 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2459 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2460 to migrate to DRBD 8 first while still running Ganeti 1.2)
2461 - DRBD devices are using statically allocated minor numbers, which
2462 will be assigned to existing instances during the migration process
2463 - there is support for both Xen PVM and Xen HVM instances running on
2465 - KVM virtualization is supported too
2466 - file-based storage has been implemented, which means that it is
2467 possible to run the cluster without LVM and DRBD storage, for
2468 example using a shared filesystem exported from shared storage (and
2469 still have live migration)
2475 *(Released Tue, 13 Jan 2009)*
2477 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2478 - Reuse the old instance mac address by default on instance import, if
2479 the instance name is the same.
2480 - Handle situations in which the node info rpc returns incomplete
2482 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2483 - Improved version of batcher:
2485 - state file support
2486 - instance mac address support
2487 - support for HVM clusters/instances
2489 - Add an option to show the number of cpu sockets and nodes in
2491 - Support OSes that handle more than one version of the OS api (but do
2492 not change the current API in any other way)
2493 - Fix ``gnt-node migrate``
2494 - ``gnt-debug`` man page
2495 - Fixes various more typos and small issues
2496 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2502 *(Released Wed, 24 Sep 2008)*
2504 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2505 type of disk exported to fully virtualized instances.
2506 - provide access to the serial console of HVM instances
2507 - instance auto_balance flag, set by default. If turned off it will
2508 avoid warnings on cluster verify if there is not enough memory to fail
2509 over an instance. in the future it will prevent automatically failing
2510 it over when we will support that.
2511 - batcher tool for instance creation, see ``tools/README.batcher``
2512 - ``gnt-instance reinstall --select-os`` to interactively select a new
2513 operating system when reinstalling an instance.
2514 - when changing the memory amount on instance modify a check has been
2515 added that the instance will be able to start. also warnings are
2516 emitted if the instance will not be able to fail over, if auto_balance
2518 - documentation fixes
2519 - sync fields between ``gnt-instance list/modify/add/import``
2520 - fix a race condition in drbd when the sync speed was set after giving
2521 the device a remote peer.
2527 *(Released Tue, 22 Jul 2008)*
2529 - note: the allowed size and number of tags per object were reduced
2530 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2531 - fixed twisted 8.x compatibility
2532 - fixed ``gnt-instance replace-disks`` with iallocator
2533 - add TCP keepalives on twisted connections to detect restarted nodes
2534 - disk increase support, see ``gnt-instance grow-disk``
2535 - implement bulk node/instance query for RAPI
2536 - add tags in node/instance listing (optional)
2537 - experimental migration (and live migration) support, read the man page
2538 for ``gnt-instance migrate``
2539 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2540 has some small improvements in handling its state file
2546 *(Released Fri, 13 Jun 2008)*
2548 - Experimental readonly, REST-based remote API implementation;
2549 automatically started on master node, TCP port 5080, if enabled by
2550 ``--enable-rapi`` parameter to configure script.
2551 - Instance allocator support. Add and import instance accept a
2552 ``--iallocator`` parameter, and call that instance allocator to decide
2553 which node to use for the instance. The iallocator document describes
2554 what's expected from an allocator script.
2555 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2556 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2557 node is lost there is still enough memory to fail over the instances
2559 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2560 ``gnt-cluster verify``, to check for site-specific compliance. All the
2561 hooks will run, and their output, if any, will be displayed. Any
2562 failing hook will make the verification return an error value.
2563 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2564 primary and secondary interfaces
2565 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2566 that is still declared as part of the cluster and has failed.
2567 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2568 specifying output fields, that just adds the chosen fields to the
2570 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2571 export from the filesystem.
2572 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2573 have been added. Using them you can enable/disable acpi and pae
2574 support, and specify a path for a cd image to be exported to the
2575 instance. These parameters as the name suggest only work on HVM
2577 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2578 PAE support will be set to the previously hardcoded values, but the
2579 (previously hardcoded) path to the CDROM ISO image will be unset and
2580 if required, needs to be set manually with ``gnt-instance modify``
2582 - The address to which an instance's VNC console is bound is now
2583 selectable per-instance, rather than being cluster wide. Of course
2584 this only applies to instances controlled via VNC, so currently just
2585 applies to HVM clusters.
2591 *(Released Mon, 18 Feb 2008)*
2593 - more tweaks to the disk activation code (especially helpful for DRBD)
2594 - change the default ``gnt-instance list`` output format, now there is
2595 one combined status field (see the manpage for the exact values this
2597 - some more fixes for the mac export to hooks change
2598 - make Ganeti not break with DRBD 8.2.x (which changed the version
2599 format in ``/proc/drbd``) (issue 24)
2600 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2601 template, allowing migration from DRBD0.7+MD to DRBD8
2607 *(Released Wed, 30 Jan 2008)*
2609 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2611 - add command aliases infrastructure and a few aliases
2612 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2613 man pages and the ``--help`` option of ``gnt-node
2614 list``/``gnt-instance list``
2615 - fix ``gnt-backup list`` with down nodes (issue 21)
2616 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2617 - fix the dist archive and add a check for including svn/git files in
2619 - some developer-related changes: improve the burnin and the QA suite,
2620 add an upload script for testing during development
2626 *(Released Wed, 16 Jan 2008)*
2628 - experimental HVM support, read the install document, section
2629 "Initializing the cluster"
2630 - allow for the PVM hypervisor per-instance kernel and initrd paths
2631 - add a new command ``gnt-cluster verify-disks`` which uses a new
2632 algorithm to improve the reconnection of the DRBD pairs if the device
2633 on the secondary node has gone away
2634 - make logical volume code auto-activate LVs at disk activation time
2635 - slightly improve the speed of activating disks
2636 - allow specification of the MAC address at instance creation time, and
2637 changing it later via ``gnt-instance modify``
2638 - fix handling of external commands that generate lots of output on
2640 - update documentation with regard to minimum version of DRBD8 supported
2646 *(Released Tue, 4 Dec 2007)*
2648 - Log the ``xm create`` output to the node daemon log on failure (to
2649 help diagnosing the error)
2650 - In debug mode, log all external commands output if failed to the logs
2651 - Change parsing of lvm commands to ignore stderr
2657 *(Released Wed, 28 Nov 2007)*
2659 - Another round of updates to the DRBD 8 code to deal with more failures
2660 in the replace secondary node operation
2661 - Some more logging of failures in disk operations (lvm, drbd)
2662 - A few documentation updates
2669 *(Released Tue, 13 Nov 2007)*
2671 - Change configuration file format from Python's Pickle to JSON.
2672 Upgrading is possible using the cfgupgrade utility.
2673 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2674 faster replace disks and is more stable (DRBD 8 has many improvements
2675 compared to DRBD 0.7)
2676 - Added command line tags support (see man pages for ``gnt-instance``,
2677 ``gnt-node``, ``gnt-cluster``)
2678 - Added instance rename support
2679 - Added multi-instance startup/shutdown
2680 - Added cluster rename support
2681 - Added ``gnt-node evacuate`` to simplify some node operations
2682 - Added instance reboot operation that can speedup reboot as compared to
2684 - Soften the requirement that hostnames are in FQDN format
2685 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2687 - Removed dependency on debian's patched fping that uses the
2688 non-standard ``-S`` option
2689 - Now the OS definitions are searched for in multiple, configurable
2690 paths (easier for distros to package)
2691 - Some changes to the hooks infrastructure (especially the new
2692 post-configuration update hook)
2693 - Other small bugfixes
2695 .. vim: set textwidth=72 syntax=rst :