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>`.
116 - Fix kvm compatibility with qemu 1.4 (Issue 389)
117 - Documentation updates (admin guide, upgrade notes, install
118 instructions) (Issue 372)
119 - Fix gnt-group list nodes and instances count (Issue 436)
120 - Fix compilation without non-mandatory libraries (Issue 441)
121 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
122 - Make confd logging more verbose at INFO level (Issue 435)
123 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
124 - Fix failure path for instance storage type conversion (Issue 229)
125 - Update htools text backend documentation
126 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
127 - Disable inter-cluster instance move for file-based instances, because
128 it is dependant on instance export, which is not supported for
129 file-based instances. (Issue 414)
135 *(Released Mon, 22 Apr 2013)*
137 This was the third beta release of the 2.7 series. Since beta2:
139 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
140 - Fix data loss on wrong usage of ``gnt-instance move``
141 - Properly export errors in confd-based job queries
142 - Add ``users-setup`` tool
143 - Fix iallocator protocol to report 0 as a disk size for diskless
144 instances. This avoids hail breaking when a diskless instance is
146 - Fix job queue directory permission problem that made confd job queries
147 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
148 for access to archived jobs (Issue 406).
149 - Limit the sizes of networks supported by ``gnt-network`` to something
150 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
151 - Fix bugs in instance disk template conversion
152 - Fix GHC 7 compatibility
153 - Fix ``burnin`` install path (Issue 426).
154 - Allow very small disk grows (Issue 347).
155 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
156 that noded doesn't import masterd code (Issue 419).
157 - Make sure the default metavg at cluster init is the same as the vg, if
158 unspecified (Issue 358).
159 - Fix cleanup of partially created disks (part of Issue 416)
165 *(Released Tue, 2 Apr 2013)*
167 This was the second beta release of the 2.7 series. Since beta1:
169 - Networks no longer have a "type" slot, since this information was
170 unused in Ganeti: instead of it tags should be used.
171 - The rapi client now has a ``target_node`` option to MigrateInstance.
172 - Fix early exit return code for hbal (Issue 386).
173 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
174 - Fix ``rbd showmapped`` output parsing (Issue 312).
175 - Networks are now referenced indexed by UUID, rather than name. This
176 will require running cfgupgrade, from 2.7.0beta1, if networks are in
178 - The OS environment now includes network information.
179 - Deleting of a network is now disallowed if any instance nic is using
180 it, to prevent dangling references.
181 - External storage is now documented in man pages.
182 - The exclusive_storage flag can now only be set at nodegroup level.
183 - Hbal can now submit an explicit priority with its jobs.
184 - Many network related locking fixes.
185 - Bump up the required pylint version to 0.25.1.
186 - Fix the ``no_remember`` option in RAPI client.
187 - Many ipolicy related tests, qa, and fixes.
188 - Many documentation improvements and fixes.
189 - Fix building with ``--disable-file-storage``.
190 - Fix ``-q`` option in htools, which was broken if passed more than
192 - Some haskell/python interaction improvements and fixes.
193 - Fix iallocator in case of missing LVM storage.
194 - Fix confd config load in case of ``--no-lvm-storage``.
195 - The confd/query functionality is now mentioned in the security
202 *(Released Wed, 6 Feb 2013)*
204 This was the first beta release of the 2.7 series. All important changes
205 are listed in the latest 2.7 entry.
211 *(Released Fri, 21 Dec 2012)*
213 Important behaviour change: hbal won't rebalance anymore instances which
214 have the ``auto_balance`` attribute set to false. This was the intention
215 all along, but until now it only skipped those from the N+1 memory
216 reservation (DRBD-specific).
218 A significant number of bug fixes in this release:
220 - Fixed disk adoption interaction with ipolicy checks.
221 - Fixed networking issues when instances are started, stopped or
222 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
223 - Fixed the warning in cluster verify for shared storage instances not
225 - Fixed removal of storage directory on shared file storage (issue 262).
226 - Fixed validation of LVM volume group name in OpClusterSetParams
227 (``gnt-cluster modify``) (issue 285).
228 - Fixed runtime memory increases (``gnt-instance modify -m``).
229 - Fixed live migration under Xen's ``xl`` mode.
230 - Fixed ``gnt-instance console`` with ``xl``.
231 - Fixed building with newer Haskell compiler/libraries.
232 - Fixed PID file writing in Haskell daemons (confd); this prevents
233 restart issues if confd was launched manually (outside of
234 ``daemon-util``) while another copy of it was running
235 - Fixed a type error when doing live migrations with KVM (issue 297) and
236 the error messages for failing migrations have been improved.
237 - Fixed opcode validation for the out-of-band commands (``gnt-node
239 - Fixed a type error when unsetting OS hypervisor parameters (issue
240 311); now it's possible to unset all OS-specific hypervisor
242 - Fixed the ``dry-run`` mode for many operations: verification of
243 results was over-zealous but didn't take into account the ``dry-run``
244 operation, resulting in "wrong" failures.
245 - Fixed bash completion in ``gnt-job list`` when the job queue has
246 hundreds of entries; especially with older ``bash`` versions, this
247 results in significant CPU usage.
249 And lastly, a few other improvements have been made:
251 - Added option to force master-failover without voting (issue 282).
252 - Clarified error message on lock conflict (issue 287).
253 - Logging of newly submitted jobs has been improved (issue 290).
254 - Hostname checks have been made uniform between instance rename and
256 - The ``--submit`` option is now supported by ``gnt-debug delay``.
257 - Shutting down the master daemon by sending SIGTERM now stops it from
258 processing jobs waiting for locks; instead, those jobs will be started
259 once again after the master daemon is started the next time (issue
261 - Support for Xen's ``xl`` program has been improved (besides the fixes
263 - Reduced logging noise in the Haskell confd daemon (only show one log
264 entry for each config reload, instead of two).
265 - Several man page updates and typo fixes.
271 *(Released Fri, 12 Oct 2012)*
273 A small bugfix release. Among the bugs fixed:
275 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
276 made the command unusable.
277 - Commands that issue many jobs don't fail anymore just because some jobs
278 take so long that other jobs are archived.
279 - Failures during ``gnt-instance reinstall`` are reflected by the exit
281 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
283 - When ``always_failover`` is set, ``--allow-failover`` is not required
284 in migrate commands anymore.
285 - ``bash_completion`` works even if extglob is disabled.
286 - Fixed bug with locks that made failover for RDB-based instances fail.
287 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
288 a random node instead of one based on the allocator metric.
289 - Support for newer versions of pylint and pep8.
290 - Hail doesn't fail anymore when trying to add an instance of type
291 ``file``, ``sharedfile`` or ``rbd``.
292 - Added new Makefile target to rebuild the whole distribution, so that
293 all files are included.
299 *(Released Fri, 27 Jul 2012)*
302 .. attention:: The ``LUXI`` protocol has been made more consistent
303 regarding its handling of command arguments. This, however, leads to
304 incompatibility issues with previous versions. Please ensure that you
305 restart Ganeti daemons soon after the upgrade, otherwise most
306 ``LUXI`` calls (job submission, setting/resetting the drain flag,
307 pausing/resuming the watcher, cancelling and archiving jobs, querying
308 the cluster configuration) will fail.
317 The current ``admin_up`` field, which used to denote whether an instance
318 should be running or not, has been removed. Instead, ``admin_state`` is
319 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
321 The rational behind this is that an instance being “down” can have
324 - it could be down during a reboot
325 - it could be temporarily be down for a reinstall
326 - or it could be down because it is deprecated and kept just for its
329 The previous Boolean state was making it difficult to do capacity
330 calculations: should Ganeti reserve memory for a down instance? Now, the
331 tri-state field makes it clear:
333 - in ``up`` and ``down`` state, all resources are reserved for the
334 instance, and it can be at any time brought up if it is down
335 - in ``offline`` state, only disk space is reserved for it, but not
338 The field can have an extra use: since the transition between ``up`` and
339 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
340 transition between ``offline`` and ``down`` is done via ``gnt-instance
341 modify``, it is possible to given different rights to users. For
342 example, owners of an instance could be allowed to start/stop it, but
343 not transition it out of the offline state.
345 Instance policies and specs
346 +++++++++++++++++++++++++++
348 In previous Ganeti versions, an instance creation request was not
349 limited on the minimum size and on the maximum size just by the cluster
350 resources. As such, any policy could be implemented only in third-party
351 clients (RAPI clients, or shell wrappers over ``gnt-*``
352 tools). Furthermore, calculating cluster capacity via ``hspace`` again
353 required external input with regards to instance sizes.
355 In order to improve these workflows and to allow for example better
356 per-node group differentiation, we introduced instance specs, which
359 - minimum instance disk size, disk count, memory size, cpu count
360 - maximum values for the above metrics
361 - and “standard” values (used in ``hspace`` to calculate the standard
364 The minimum/maximum values can be also customised at node-group level,
365 for example allowing more powerful hardware to support bigger instance
368 Beside the instance specs, there are a few other settings belonging to
369 the instance policy framework. It is possible now to customise, per
370 cluster and node-group:
372 - the list of allowed disk templates
373 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
374 - the maximum ratio of instance to spindles (see below for more
375 information) for local storage
377 All these together should allow all tools that talk to Ganeti to know
378 what are the ranges of allowed values for instances and the
379 over-subscription that is allowed.
381 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
382 instance configuration, and the physical CPU configuration from the
383 node. For the spindle ratios however, we didn't track before these
384 values, so new parameters have been added:
386 - a new node parameter ``spindle_count``, defaults to 1, customisable at
387 node group or node level
388 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
390 Note that spindles in this context doesn't need to mean actual
391 mechanical hard-drives; it's just a relative number for both the node
392 I/O capacity and instance I/O consumption.
394 Instance migration behaviour
395 ++++++++++++++++++++++++++++
397 While live-migration is in general desirable over failover, it is
398 possible that for some workloads it is actually worse, due to the
399 variable time of the “suspend” phase during live migration.
401 To allow the tools to work consistently over such instances (without
402 having to hard-code instance names), a new backend parameter
403 ``always_failover`` has been added to control the migration/failover
404 behaviour. When set to True, all migration requests for an instance will
405 instead fall-back to failover.
407 Instance memory ballooning
408 ++++++++++++++++++++++++++
410 Initial support for memory ballooning has been added. The memory for an
411 instance is no longer fixed (backend parameter ``memory``), but instead
412 can vary between minimum and maximum values (backend parameters
413 ``minmem`` and ``maxmem``). Currently we only change an instance's
416 - live migrating or failing over and instance and the target node
417 doesn't have enough memory
418 - user requests changing the memory via ``gnt-instance modify
424 In order to control the use of specific CPUs by instance, support for
425 controlling CPU pinning has been added for the Xen, HVM and LXC
426 hypervisors. This is controlled by a new hypervisor parameter
427 ``cpu_mask``; details about possible values for this are in the
428 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
429 VCPU-to-CPU mapping) form will work well only when all nodes in your
430 cluster have the same amount of CPUs.
435 Another area in which Ganeti was not customisable were the parameters
436 used for storage configuration, e.g. how many stripes to use for LVM,
437 DRBD resync configuration, etc.
439 To improve this area, we've added disks parameters, which are
440 customisable at cluster and node group level, and which allow to
441 specify various parameters for disks (DRBD has the most parameters
442 currently), for example:
444 - DRBD resync algorithm and parameters (e.g. speed)
445 - the default VG for meta-data volumes for DRBD
446 - number of stripes for LVM (plain disk template)
449 These parameters can be modified via ``gnt-cluster modify -D …`` and
450 ``gnt-group modify -D …``, and are used at either instance creation (in
451 case of LVM stripes, for example) or at disk “activation” time
454 Rados block device support
455 ++++++++++++++++++++++++++
457 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
458 denoted by the ``rbd`` disk template type. This is considered
459 experimental, feedback is welcome. For details on configuring it, see
460 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
465 The existing master IP functionality works well only in simple setups (a
466 single network shared by all nodes); however, if nodes belong to
467 different networks, then the ``/32`` setup and lack of routing
468 information is not enough.
470 To allow the master IP to function well in more complex cases, the
471 system was reworked as follows:
473 - a master IP netmask setting has been added
474 - the master IP activation/turn-down code was moved from the node daemon
476 - whether to run the Ganeti-supplied master IP script or a user-supplied
477 on is a ``gnt-cluster init`` setting
479 Details about the location of the standard and custom setup scripts are
480 in the man page :manpage:`gnt-cluster(8)`; for information about the
481 setup script protocol, look at the Ganeti-supplied script.
486 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
489 It is now possible to use TLS-protected connections, and when renewing
490 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
491 it is now possible to specify spice or spice CA certificates. Also, it
492 is possible to configure a password for SPICE sessions via the
493 hypervisor parameter ``spice_password_file``.
495 There are also new parameters to control the compression and streaming
496 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
497 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
498 for the spice parameters.
500 Lastly, it is now possible to see the SPICE connection information via
501 ``gnt-instance console``.
506 A new tool (``tools/ovfconverter``) has been added that supports
507 conversion between Ganeti and the `Open Virtualization Format
508 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
511 This relies on the ``qemu-img`` tool to convert the disk formats, so the
512 actual compatibility with other virtualization solutions depends on it.
517 The configuration query daemon (``ganeti-confd``) is now optional, and
518 has been rewritten in Haskell; whether to use the daemon at all, use the
519 Python (default) or the Haskell version is selectable at configure time
520 via the ``--enable-confd`` parameter, which can take one of the
521 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
522 daemon will result in a smaller footprint; for larger systems, we
523 welcome feedback on the Haskell version which might become the default
526 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
527 daemon running. The Python version doesn't implement the new call.
530 User interface changes
531 ~~~~~~~~~~~~~~~~~~~~~~
533 We have replaced the ``--disks`` option of ``gnt-instance
534 replace-disks`` with a more flexible ``--disk`` option, which allows
535 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
536 disk size and mode can be changed upon recreation (via ``gnt-instance
537 recreate-disks``, which accepts the same ``--disk`` option).
539 As many people are used to a ``show`` command, we have added that as an
540 alias to ``info`` on all ``gnt-*`` commands.
542 The ``gnt-instance grow-disk`` command has a new mode in which it can
543 accept the target size of the disk, instead of the delta; this can be
544 more safe since two runs in absolute mode will be idempotent, and
545 sometimes it's also easier to specify the desired size directly.
547 Also the handling of instances with regard to offline secondaries has
548 been improved. Instance operations should not fail because one of it's
549 secondary nodes is offline, even though it's safe to proceed.
551 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
552 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
553 minors to instance name.
558 RAPI coverage has improved, with (for example) new resources for
559 recreate-disks, node power-cycle, etc.
564 There is partial support for ``xl`` in the Xen hypervisor; feedback is
567 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
568 whether to still support Python 2.4 or move to Python 2.6 as minimum
571 Support for Fedora has been slightly improved; the provided example
572 init.d script should work better on it and the INSTALL file should
573 document the needed dependencies.
578 The deprecated ``QueryLocks`` LUXI request has been removed. Use
579 ``Query(what=QR_LOCK, ...)`` instead.
581 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
582 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
583 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
584 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
585 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
587 RAPI client: ``CertificateError`` now derives from
588 ``GanetiApiError``. This should make it more easy to handle Ganeti
591 Deprecation warnings due to PyCrypto/paramiko import in
592 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
593 make sure to run an up-to-date paramiko version, if you use this tool.
595 The QA scripts now depend on Python 2.5 or above (the main code base
596 still works with Python 2.4).
598 The configuration file (``config.data``) is now written without
599 indentation for performance reasons; if you want to edit it, it can be
600 re-formatted via ``tools/fmtjson``.
602 A number of bugs has been fixed in the cluster merge tool.
604 ``x509`` certification verification (used in import-export) has been
605 changed to allow the same clock skew as permitted by the cluster
606 verification. This will remove some rare but hard to diagnose errors in
613 *(Released Thu, 19 Jul 2012)*
615 Very few changes from rc4 to the final release, only bugfixes:
617 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
618 instance, fix CDROM booting for KVM instances)
619 - fixed node group modification of node parameters
620 - fixed issue in LUClusterVerifyGroup with multi-group clusters
621 - fixed generation of bash completion to ensure a stable ordering
628 *(Released Fri, 13 Jul 2012)*
630 Third release candidate for 2.6. The following changes were done from
633 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
634 - Fixed an inconsistency in the LUXI protocol with the provided
635 arguments (NOT backwards compatible)
636 - Fixed a bug with node groups ipolicy where ``min`` was greater than
637 the cluster ``std`` value
638 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
639 easier instance debugging on nodes (requires ``hconfd`` to work)
645 *(Released Tue, 03 Jul 2012)*
647 Second release candidate for 2.6. The following changes were done from
650 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
652 - Fixed a RAPI regression on missing beparams/memory
653 - Fixed redistribution of files on offline nodes
654 - Added possibility to run activate-disks even though secondaries are
655 offline. With this change it relaxes also the strictness on some other
656 commands which use activate disks internally:
657 * ``gnt-instance start|reboot|rename|backup|export``
658 - Made it possible to remove safely an instance if its secondaries are
660 - Made it possible to reinstall even though secondaries are offline
666 *(Released Mon, 25 Jun 2012)*
668 First release candidate for 2.6. The following changes were done from
671 - Fixed bugs with disk parameters and ``rbd`` templates as well as
673 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
674 behaviour. It supports now the modify operation
675 - ``hcheck`` implemented to analyze cluster health and possibility of
676 improving health by rebalance
677 - ``hbal`` has been improved in dealing with split instances
683 *(Released Mon, 11 Jun 2012)*
685 Second beta release of 2.6. The following changes were done from beta2
688 - Fixed ``daemon-util`` with non-root user models
689 - Fixed creation of plain instances with ``--no-wait-for-sync``
690 - Fix wrong iv_names when running ``cfgupgrade``
691 - Export more information in RAPI group queries
692 - Fixed bug when changing instance network interfaces
693 - Extended burnin to do NIC changes
694 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
695 - Changed default for DRBD barriers
696 - Fixed DRBD error reporting for syncer rate
697 - Verify the options on disk parameters
699 And of course various fixes to documentation and improved unittests and
706 *(Released Wed, 23 May 2012)*
708 First beta release of 2.6. The following changes were done from beta1 to
711 - integrated patch for distributions without ``start-stop-daemon``
712 - adapted example init.d script to work on Fedora
713 - fixed log handling in Haskell daemons
714 - adapted checks in the watcher for pycurl linked against libnss
715 - add partial support for ``xl`` instead of ``xm`` for Xen
716 - fixed a type issue in cluster verification
717 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
720 Plus integrated fixes from the 2.5 branch:
722 - fixed ``kvm-ifup`` to use ``/bin/bash``
723 - fixed parallel build failures
724 - KVM live migration when using a custom keymap
730 *(Released Tue, 24 Jul 2012)*
732 A small bugfix release, with no new features:
734 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
735 different default shell (e.g. Debian, Ubuntu)
736 - fixed KVM startup and live migration with a custom keymap (fixes Issue
737 243 and Debian bug #650664)
738 - fixed compatibility with KVM versions that don't support multiple boot
739 devices (fixes Issue 230 and Debian bug #624256)
741 Additionally, a few fixes were done to the build system (fixed parallel
742 build failures) and to the unittests (fixed race condition in test for
743 FileID functions, and the default enable/disable mode for QA test is now
750 *(Released Fri, 11 May 2012)*
752 A small bugfix release.
754 The main issues solved are on the topic of compatibility with newer LVM
757 - fixed parsing of ``lv_attr`` field
758 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
759 the ``--force`` flag is needed
761 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
762 to accept kernel 3.x too (was hardcoded to 2.6.*).
764 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
765 of hook results and that also made display of other errors suboptimal
766 was fixed; the code behaves now like 2.4 and earlier.
768 Another change in 2.5, the cleanup of the OS scripts environment, is too
769 aggressive: it removed even the ``PATH`` variable, which requires the OS
770 scripts to *always* need to export it. Since this is a bit too strict,
771 we now export a minimal PATH, the same that we export for hooks.
773 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
774 integrated into this release.
776 Finally, a few other miscellaneous changes were done (no new features,
777 just small improvements):
779 - Fix ``gnt-group --help`` display
780 - Fix hardcoded Xen kernel path
781 - Fix grow-disk handling of invalid units
782 - Update synopsis for ``gnt-cluster repair-disk-sizes``
783 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
789 *(Released Thu, 12 Apr 2012)*
791 Incompatible/important changes and bugfixes
792 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
794 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
795 resource's ``ip_check`` parameter changed from ``True`` to ``False``
796 to match the underlying LUXI interface.
797 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
798 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
799 not maintain backwards-compatibility as the underlying operation
800 changed in an incompatible way. The RAPI client can talk to old
801 servers, but it needs to be told so as the return value changed.
802 - When creating file-based instances via RAPI, the ``file_driver``
803 parameter no longer defaults to ``loop`` and must be specified.
804 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
806 - Support for the undocumented and deprecated RAPI instance creation
807 request format version 0 has been dropped. Use version 1, supported
808 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
809 - Pyparsing 1.4.6 or above is required, see :doc:`installation
810 documentation <install>`.
811 - The "cluster-verify" hooks are now executed per group by the
812 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
813 if you just run ``gnt-cluster verify``, which generates one opcode per
815 - The environment as passed to the OS scripts is cleared, and thus no
816 environment variables defined in the node daemon's environment will be
817 inherited by the scripts.
818 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
820 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
821 support operations involving multiple node groups.
822 - Offline nodes are ignored when failing over an instance.
823 - Support for KVM version 1.0, which changed the version reporting format
825 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
827 - ``Makefile`` is now compatible with Automake 1.11.2
828 - Includes all bugfixes made in the 2.4 series
833 - The ganeti-htools project has been merged into the ganeti-core source
834 tree and will be built as part of Ganeti (see :doc:`install-quick`).
835 - Implemented support for :doc:`shared storage <design-shared-storage>`.
836 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
837 GPT-style partition tables (requires `parted
838 <http://www.gnu.org/s/parted/>`_).
839 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
840 - Allowed adding tags on instance creation.
841 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
842 - Allow instances to be started in a paused state, enabling the user to
843 see the complete console output on boot using the console.
844 - Added new hypervisor flag to control default reboot behaviour
845 (``reboot_behavior``).
846 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
847 - Improved out-of-band management support:
849 - Added ``gnt-node health`` command reporting the health status of
851 - Added ``gnt-node power`` command to manage power status of nodes.
852 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
854 - Instance migration can fall back to failover if instance is not
856 - Filters can be used when listing nodes, instances, groups and locks;
857 see :manpage:`ganeti(7)` manpage.
858 - Added post-execution status as variables to :doc:`hooks <hooks>`
860 - Instance tags are exported/imported together with the instance.
861 - When given an explicit job ID, ``gnt-job info`` will work for archived
863 - Jobs can define dependencies on other jobs (not yet supported via
864 RAPI or command line, but used by internal commands and usable via
867 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
870 - Instance failover is now available as a RAPI resource
871 (``/2/instances/[instance_name]/failover``).
872 - ``gnt-instance info`` defaults to static information if primary node
874 - Opcodes have a new ``comment`` attribute.
875 - Added basic SPICE support to KVM hypervisor.
876 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
878 Node group improvements
879 ~~~~~~~~~~~~~~~~~~~~~~~
881 - ``gnt-cluster verify`` has been modified to check groups separately,
882 thereby improving performance.
883 - Node group support has been added to ``gnt-cluster verify-disks``,
884 which now operates per node group.
885 - Watcher has been changed to work better with node groups.
887 - One process and state file per node group.
888 - Slow watcher in one group doesn't block other group's watcher.
890 - Added new command, ``gnt-group evacuate``, to move all instances in a
891 node group to other groups.
892 - Added ``gnt-instance change-group`` to move an instance to another
894 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
895 per-group operations.
896 - Node groups can be tagged.
897 - Some operations switch from an exclusive to a shared lock as soon as
899 - Instance's primary and secondary nodes' groups are now available as
900 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
901 and ``snodes.group.uuid``).
906 - Numerous updates to documentation and manpages.
908 - :doc:`RAPI <rapi>` documentation now has detailed parameter
910 - Some opcode/job results are now also documented, see :doc:`RAPI
913 - A lockset's internal lock is now also visible in lock monitor.
914 - Log messages from job queue workers now contain information about the
915 opcode they're processing.
916 - ``gnt-instance console`` no longer requires the instance lock.
917 - A short delay when waiting for job changes reduces the number of LUXI
918 requests significantly.
919 - DRBD metadata volumes are overwritten with zeros during disk creation.
920 - Out-of-band commands no longer acquire the cluster lock in exclusive
922 - ``devel/upload`` now uses correct permissions for directories.
928 *(Released Fri, 23 Mar 2012)*
930 This was the sixth release candidate of the 2.5 series.
936 *(Released Mon, 9 Jan 2012)*
938 This was the fifth release candidate of the 2.5 series.
944 *(Released Thu, 27 Oct 2011)*
946 This was the fourth release candidate of the 2.5 series.
952 *(Released Wed, 26 Oct 2011)*
954 This was the third release candidate of the 2.5 series.
960 *(Released Tue, 18 Oct 2011)*
962 This was the second release candidate of the 2.5 series.
968 *(Released Tue, 4 Oct 2011)*
970 This was the first release candidate of the 2.5 series.
976 *(Released Wed, 31 Aug 2011)*
978 This was the third beta release of the 2.5 series.
984 *(Released Mon, 22 Aug 2011)*
986 This was the second beta release of the 2.5 series.
992 *(Released Fri, 12 Aug 2011)*
994 This was the first beta release of the 2.5 series.
1000 *(Released Thu, 27 Oct 2011)*
1002 - Fixed bug when parsing command line parameter values ending in
1004 - Fixed assertion error after unclean master shutdown
1005 - Disable HTTP client pool for RPC, significantly reducing memory usage
1007 - Fixed queue archive creation with wrong permissions
1013 *(Released Tue, 23 Aug 2011)*
1017 - Fixed documentation for importing with ``--src-dir`` option
1018 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1019 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1025 *(Released Fri, 5 Aug 2011)*
1027 Many bug-fixes and a few small features:
1029 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1030 issues when you tried to add an instance with two MAC addresses in one
1032 - KVM: fixed per-instance stored UID value
1033 - KVM: configure bridged NICs at migration start
1034 - KVM: Fix a bug where instance will not start with never KVM versions
1036 - Added OS search path to ``gnt-cluster info``
1037 - Fixed an issue with ``file_storage_dir`` where you were forced to
1038 provide an absolute path, but the documentation states it is a
1039 relative path, the documentation was right
1040 - Added a new parameter to instance stop/start called ``--no-remember``
1041 that will make the state change to not be remembered
1042 - Implemented ``no_remember`` at RAPI level
1043 - Improved the documentation
1044 - Node evacuation: don't call IAllocator if node is already empty
1045 - Fixed bug in DRBD8 replace disks on current nodes
1046 - Fixed bug in recreate-disks for DRBD instances
1047 - Moved assertion checking locks in ``gnt-instance replace-disks``
1048 causing it to abort with not owning the right locks for some situation
1049 - Job queue: Fixed potential race condition when cancelling queued jobs
1050 - Fixed off-by-one bug in job serial generation
1051 - ``gnt-node volumes``: Fix instance names
1052 - Fixed aliases in bash completion
1053 - Fixed a bug in reopening log files after being sent a SIGHUP
1054 - Added a flag to burnin to allow specifying VCPU count
1055 - Bugfixes to non-root Ganeti configuration
1061 *(Released Thu, 12 May 2011)*
1063 Many bug-fixes and a few new small features:
1065 - Fixed a bug related to log opening failures
1066 - Fixed a bug in instance listing with orphan instances
1067 - Fixed a bug which prevented resetting the cluster-level node parameter
1068 ``oob_program`` to the default
1069 - Many fixes related to the ``cluster-merge`` tool
1070 - Fixed a race condition in the lock monitor, which caused failures
1071 during (at least) creation of many instances in parallel
1072 - Improved output for gnt-job info
1073 - Removed the quiet flag on some ssh calls which prevented debugging
1075 - Improved the N+1 failure messages in cluster verify by actually
1076 showing the memory values (needed and available)
1077 - Increased lock attempt timeouts so that when executing long operations
1078 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1079 too early and thus prevent the use of the 'fair' mechanism
1080 - Changed instance query data (``gnt-instance info``) to not acquire
1081 locks unless needed, thus allowing its use on locked instance if only
1082 static information is asked for
1083 - Improved behaviour with filesystems that do not support rename on an
1085 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1086 kept locks on all nodes during the wipe, which is unneeded
1087 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1088 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1089 10GiB) which caused the wipe to fail right at the end in some cases
1090 - Fixed master IP activation when doing master failover with no-voting
1091 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1092 the master node itself
1093 - Fixed potential data-loss in under disk full conditions, where Ganeti
1094 wouldn't check correctly the return code and would consider
1095 partially-written files 'correct'
1096 - Fixed bug related to multiple VGs and DRBD disk replacing
1097 - Added new disk parameter ``metavg`` that allows placement of the meta
1098 device for DRBD in a different volume group
1099 - Fixed error handling in the node daemon when the system libc doesn't
1100 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1101 - Fixed lock release during replace-disks, which kept cluster-wide locks
1102 when doing disk replaces with an iallocator script
1103 - Added check for missing bridges in cluster verify
1104 - Handle EPIPE errors while writing to the terminal better, so that
1105 piping the output to e.g. ``less`` doesn't cause a backtrace
1106 - Fixed rare case where a ^C during Luxi calls could have been
1107 interpreted as server errors, instead of simply terminating
1108 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1110 - Added a few more parameters to the KVM hypervisor, allowing a second
1111 CDROM, custom disk type for CDROMs and a floppy image
1112 - Removed redundant message in instance rename when the name is given
1114 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1115 disks on new nodes, allowing recreation when the original instance
1116 nodes are completely gone
1117 - Added option when converting disk templates to DRBD to skip waiting
1118 for the resync, in order to make the instance available sooner
1119 - Added two new variables to the OS scripts environment (containing the
1121 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1122 to allow use of ``pvgrub`` as bootloader
1123 - Changed the instance memory modifications to only check out-of-memory
1124 conditions on memory increases, and turned the secondary node warnings
1125 into errors (they can still be overridden via ``--force``)
1126 - Fixed the handling of a corner case when the Python installation gets
1127 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1128 to execute a command that doesn't exist
1129 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1130 node of the instance returned failures during instance shutdown; this
1131 adds the option ``--ignore-consistency`` to gnt-instance move
1133 And as usual, various improvements to the error messages, documentation
1140 *(Released Wed, 09 Mar 2011)*
1142 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1143 the RAPI users file if run twice (even with ``--dry-run``).
1145 The release fixes that bug (nothing else changed).
1151 *(Released Mon, 07 Mar 2011)*
1153 Final 2.4.0 release. Just a few small fixes:
1155 - Fixed RAPI node evacuate
1156 - Fixed the kvm-ifup script
1157 - Fixed internal error handling for special job cases
1158 - Updated man page to specify the escaping feature for options
1164 *(Released Mon, 28 Feb 2011)*
1166 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1167 that this feature wiped the disks of the wrong instance, leading to loss
1172 - Fixed title of query field containing instance name
1173 - Expanded the glossary in the documentation
1174 - Fixed one unittest (internal issue)
1180 *(Released Mon, 21 Feb 2011)*
1182 A number of bug fixes plus just a couple functionality changes.
1184 On the user-visible side, the ``gnt-* list`` command output has changed
1185 with respect to "special" field states. The current rc1 style of display
1186 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1187 the default output mode special fields states are displayed as follows:
1189 - Offline resource: ``*``
1190 - Unavailable/not applicable: ``-``
1191 - Data missing (RPC failure): ``?``
1192 - Unknown field: ``??``
1194 Another user-visible change is the addition of ``--force-join`` to
1199 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1200 - Fixed regression in RAPI/instance reinstall where all parameters were
1201 required (instead of optional)
1202 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1203 - Fixed iallocator usage (offline nodes were not considered offline)
1204 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1205 - Fixed hypervisor and OS parameter validation with respect to
1206 non-vm_capable nodes
1207 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1209 - Fixed ``tools/listrunner`` with respect to agent-based usage
1215 *(Released Fri, 4 Feb 2011)*
1217 Many changes and fixes since the beta1 release. While there were some
1218 internal changes, the code has been mostly stabilised for the RC
1221 Note: the dumb allocator was removed in this release, as it was not kept
1222 up-to-date with the IAllocator protocol changes. It is recommended to
1223 use the ``hail`` command from the ganeti-htools package.
1225 Note: the 2.4 and up versions of Ganeti are not compatible with the
1226 0.2.x branch of ganeti-htools. You need to upgrade to
1227 ganeti-htools-0.3.0 (or later).
1229 Regressions fixed from 2.3
1230 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1232 - Fixed the ``gnt-cluster verify-disks`` command
1233 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1235 - Fixed disk adoption breakage
1236 - Fixed wrong headers in instance listing for field aliases
1241 - Fixed corner case in KVM handling of NICs
1242 - Fixed many cases of wrong handling of non-vm_capable nodes
1243 - Fixed a bug where a missing instance symlink was not possible to
1244 recreate with any ``gnt-*`` command (now ``gnt-instance
1245 activate-disks`` does it)
1246 - Fixed the volume group name as reported by ``gnt-cluster
1248 - Increased timeouts for the import-export code, hopefully leading to
1249 fewer aborts due network or instance timeouts
1250 - Fixed bug in ``gnt-node list-storage``
1251 - Fixed bug where not all daemons were started on cluster
1252 initialisation, but only at the first watcher run
1253 - Fixed many bugs in the OOB implementation
1254 - Fixed watcher behaviour in presence of instances with offline
1256 - Fixed instance list output for instances running on the wrong node
1257 - a few fixes to the cluster-merge tool, but it still cannot merge
1258 multi-node groups (currently it is not recommended to use this tool)
1264 - Improved network configuration for the KVM hypervisor
1265 - Added e1000 as a supported NIC for Xen-HVM
1266 - Improved the lvmstrap tool to also be able to use partitions, as
1267 opposed to full disks
1268 - Improved speed of disk wiping (the cluster parameter
1269 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1270 of instance creations
1271 - Added documentation for the OS parameters
1272 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1273 hypervisor is not responding
1274 - Added display of blacklisted and hidden OS information in
1275 ``gnt-cluster info``
1276 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1277 NIC and node parameters, which might create problems with currently
1278 invalid (but undetected) configuration files, but prevents validation
1279 failures when unrelated parameters are modified
1280 - Changed cluster initialisation to wait for the master daemon to become
1282 - Expanded the RAPI interface:
1284 - Added config redistribution resource
1285 - Added activation/deactivation of instance disks
1286 - Added export of console information
1288 - Implemented log file reopening on SIGHUP, which allows using
1289 logrotate(8) for the Ganeti log files
1290 - Added a basic OOB helper script as an example
1296 *(Released Fri, 14 Jan 2011)*
1301 - Fixed timezone issues when formatting timestamps
1302 - Added support for node groups, available via ``gnt-group`` and other
1304 - Added out-of-band framework and management, see :doc:`design
1305 document <design-oob>`
1306 - Removed support for roman numbers from ``gnt-node list`` and
1307 ``gnt-instance list``.
1308 - Allowed modification of master network interface via ``gnt-cluster
1309 modify --master-netdev``
1310 - Accept offline secondaries while shutting down instance disks
1311 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1312 - Added support for multiple LVM volume groups
1313 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1315 - Added commands to list available fields:
1317 - ``gnt-node list-fields``
1318 - ``gnt-group list-fields``
1319 - ``gnt-instance list-fields``
1321 - Updated documentation and man pages
1326 - Moved ``rapi_users`` file into separate directory, now named
1327 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1329 - Added new tool for running commands on many machines,
1330 ``tools/ganeti-listrunner``
1331 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1332 improving the ``gnt-instance console`` output
1333 - Allowed customisation of disk index separator at ``configure`` time
1334 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1335 - Added support for non-partitioned md disks in ``lvmstrap``
1336 - Added script to gracefully power off KVM instances
1337 - Split ``utils`` module into smaller parts
1338 - Changed query operations to return more detailed information, e.g.
1339 whether an information is unavailable due to an offline node. To use
1340 this new functionality, the LUXI call ``Query`` must be used. Field
1341 information is now stored by the master daemon and can be retrieved
1342 using ``QueryFields``. Instances, nodes and groups can also be queried
1343 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1344 exposed via RAPI). The following commands make use of this
1345 infrastructure change:
1347 - ``gnt-group list``
1348 - ``gnt-group list-fields``
1350 - ``gnt-node list-fields``
1351 - ``gnt-instance list``
1352 - ``gnt-instance list-fields``
1353 - ``gnt-debug locks``
1358 - New RAPI resources (see :doc:`rapi`):
1362 - ``/2/groups/[group_name]``
1363 - ``/2/groups/[group_name]/assign-nodes``
1364 - ``/2/groups/[group_name]/modify``
1365 - ``/2/groups/[group_name]/rename``
1366 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1370 - Implemented ``no_install`` for instance creation
1371 - Implemented OS parameters for instance reinstallation, allowing
1372 use of special settings on reinstallation (e.g. for preserving data)
1377 - Added IPv6 support in import/export
1378 - Pause DRBD synchronization while wiping disks on instance creation
1379 - Updated unittests and QA scripts
1380 - Improved network parameters passed to KVM
1381 - Converted man pages from docbook to reStructuredText
1387 *(Released Mon, 20 Dec 2010)*
1389 Released version 2.3.1~rc1 without any changes.
1395 *(Released Wed, 1 Dec 2010)*
1397 - impexpd: Disable OpenSSL compression in socat if possible (backport
1398 from master, commit e90739d625b, see :doc:`installation guide
1399 <install-quick>` for details)
1400 - Changed unittest coverage report to exclude test scripts
1401 - Added script to check version format
1407 *(Released Wed, 1 Dec 2010)*
1409 Released version 2.3.0~rc1 without any changes.
1415 *(Released Fri, 19 Nov 2010)*
1417 A number of bugfixes and documentation updates:
1419 - Update ganeti-os-interface documentation
1420 - Fixed a bug related to duplicate MACs or similar items which should be
1422 - Fix breakage in OS state modify
1423 - Reinstall instance: disallow offline secondaries (fixes bug related to
1424 OS changing but reinstall failing)
1425 - plus all the other fixes between 2.2.1 and 2.2.2
1431 *(Released Tue, 2 Nov 2010)*
1433 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1434 - Fixed master failover race with watcher
1435 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1437 - Accept previously stopped instance for export with instance removal
1438 - Simplify and extend the environment variables for instance OS scripts
1439 - Added new node flags, ``master_capable`` and ``vm_capable``
1440 - Added optional instance disk wiping prior during allocation. This is a
1441 cluster-wide option and can be set/modified using
1442 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1443 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1444 :doc:`install-quick`
1445 - Added a new watcher option (``--ignore-pause``)
1446 - Added option to ignore offline node on instance start/stop
1447 (``--ignore-offline``)
1448 - Allow overriding OS parameters with ``gnt-instance reinstall``
1449 - Added ability to change node's secondary IP address using ``gnt-node
1451 - Implemented privilege separation for all daemons except
1452 ``ganeti-noded``, see ``configure`` options
1453 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1455 - Implemented job priorities (see ``ganeti(7)`` manpage)
1456 - Ignore failures while shutting down instances during failover from
1458 - Exit daemon's bootstrap process only once daemon is ready
1459 - Export more information via ``LUInstanceQuery``/remote API
1460 - Improved documentation, QA and unittests
1461 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1462 restart if the file was created or changed
1463 - Added LUXI protocol version sent with each request and response,
1464 allowing detection of server/client mismatches
1465 - Moved the Python scripts among gnt-* and ganeti-* into modules
1466 - Moved all code related to setting up SSH to an external script,
1468 - Infrastructure changes for node group support in future versions
1474 *(Released Fri, 19 Nov 2010)*
1476 A few small bugs fixed, and some improvements to the build system:
1478 - Fix documentation regarding conversion to drbd
1479 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1481 - Fix error handling in node modify with multiple changes
1482 - Allow remote imports without checked names
1488 *(Released Tue, 19 Oct 2010)*
1490 - Disable SSL session ID cache in RPC client
1496 *(Released Thu, 14 Oct 2010)*
1498 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1499 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1501 - Fix problems with interaction between Curl and Python's HTTP server,
1502 resulting in increased speed in many RPC calls
1503 - Improve our release script to prevent breakage with older aclocal and
1510 *(Released Thu, 7 Oct 2010)*
1512 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1513 value retrieved from master
1514 - Added support for blacklisted or hidden OS definitions
1515 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1516 - Added support for -mem-path in KVM hypervisor abstraction layer
1517 - Allow overriding instance parameters in tool for inter-cluster
1518 instance moves (``tools/move-instance``)
1519 - Improved opcode summaries (e.g. in ``gnt-job list``)
1520 - Improve consistency of OS listing by sorting it
1521 - Documentation updates
1527 *(Released Fri, 8 Oct 2010)*
1529 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1535 *(Released Mon, 4 Oct 2010)*
1537 - Fixed regression in ``gnt-instance rename``
1543 *(Released Wed, 22 Sep 2010)*
1545 - Fixed OS_VARIANT variable for OS scripts
1546 - Fixed cluster tag operations via RAPI
1547 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1548 - Disabled RAPI CA checks in watcher
1554 *(Released Mon, 23 Aug 2010)*
1556 - Support DRBD versions of the format "a.b.c.d"
1558 - Re-introduce support for usage from multiple threads in RAPI client
1559 - Instance renames and modify via RAPI
1560 - Work around race condition between processing and archival in job
1562 - Mark opcodes following failed one as failed, too
1563 - Job field ``lock_status`` was removed due to difficulties making it
1564 work with the changed job queue in Ganeti 2.2; a better way to monitor
1565 locks is expected for a later 2.2.x release
1566 - Fixed dry-run behaviour with many commands
1567 - Support ``ssh-agent`` again when adding nodes
1568 - Many additional bugfixes
1574 *(Released Fri, 30 Jul 2010)*
1576 Important change: the internal RPC mechanism between Ganeti nodes has
1577 changed from using a home-grown http library (based on the Python base
1578 libraries) to use the PycURL library. This requires that PycURL is
1579 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1580 against GnuTLS by default. cURL's support for GnuTLS had known issues
1581 before cURL 7.21.0 and we recommend using the latest cURL release or
1582 linking against OpenSSL. Most other distributions already link PycURL
1583 and cURL against OpenSSL. The command::
1585 python -c 'import pycurl; print pycurl.version'
1587 can be used to determine the libraries PycURL and cURL are linked
1590 Other significant changes:
1592 - Rewrote much of the internals of the job queue, in order to achieve
1593 better parallelism; this decouples job query operations from the job
1594 processing, and it should allow much nicer behaviour of the master
1595 daemon under load, and it also has uncovered some long-standing bugs
1596 related to the job serialisation (now fixed)
1597 - Added a default iallocator setting to the cluster parameters,
1598 eliminating the need to always pass nodes or an iallocator for
1599 operations that require selection of new node(s)
1600 - Added experimental support for the LXC virtualization method
1601 - Added support for OS parameters, which allows the installation of
1602 instances to pass parameter to OS scripts in order to customise the
1604 - Added a hypervisor parameter controlling the migration type (live or
1605 non-live), since hypervisors have various levels of reliability; this
1606 has renamed the 'live' parameter to 'mode'
1607 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1608 logical volumes, meaning that cluster verify will ignore them and not
1609 flag their presence as errors
1610 - The watcher will now reset the error count for failed instances after
1611 8 hours, thus allowing self-healing if the problem that caused the
1612 instances to be down/fail to start has cleared in the meantime
1613 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1614 check for, and warn, if the drbd module parameter ``usermode_helper``
1615 is not consistent with the cluster-wide setting; this is needed to
1616 make diagnose easier of failed drbd creations
1617 - Started adding base IPv6 support, but this is not yet
1618 enabled/available for use
1619 - Rename operations (cluster, instance) will now return the new name,
1620 which is especially useful if a short name was passed in
1621 - Added support for instance migration in RAPI
1622 - Added a tool to pre-configure nodes for the SSH setup, before joining
1623 them to the cluster; this will allow in the future a simplified model
1624 for node joining (but not yet fully enabled in 2.2); this needs the
1625 paramiko python library
1626 - Fixed handling of name-resolving errors
1627 - Fixed consistency of job results on the error path
1628 - Fixed master-failover race condition when executed multiple times in
1630 - Fixed many bugs related to the job queue (mostly introduced during the
1631 2.2 development cycle, so not all are impacting 2.1)
1632 - Fixed instance migration with missing disk symlinks
1633 - Fixed handling of unknown jobs in ``gnt-job archive``
1634 - And many other small fixes/improvements
1638 - Enhanced both the unittest and the QA coverage
1639 - Switched the opcode validation to a generic model, and extended the
1640 validation to all opcode parameters
1641 - Changed more parts of the code that write shell scripts to use the
1643 - Switched the master daemon to use the asyncore library for the Luxi
1650 *(Released Thu, 17 Jun 2010)*
1652 - Added tool (``move-instance``) and infrastructure to move instances
1653 between separate clusters (see :doc:`separate documentation
1654 <move-instance>` and :doc:`design document <design-2.2>`)
1655 - Added per-request RPC timeout
1656 - RAPI now requires a Content-Type header for requests with a body (e.g.
1657 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1658 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1659 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1661 - Implemented initial support for running Ganeti daemons as separate
1662 users, see configure-time flags ``--with-user-prefix`` and
1663 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1665 - Instances can be removed after export (``gnt-backup export
1666 --remove-instance``)
1667 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1668 key (instead of 1024 bit)
1669 - Added new cluster configuration file for cluster domain secret
1670 - Import/export now use SSL instead of SSH
1671 - Added support for showing estimated time when exporting an instance,
1672 see the ``ganeti-os-interface(7)`` manpage and look for
1679 *(Released Tue, 16 Nov 2010)*
1681 Some more bugfixes. Unless critical bugs occur, this will be the last
1684 - Fix case of MAC special-values
1685 - Fix mac checker regex
1686 - backend: Fix typo causing "out of range" error
1687 - Add missing --units in gnt-instance list man page
1693 *(Released Tue, 24 Aug 2010)*
1696 - Don't ignore secondary node silently on non-mirrored disk templates
1698 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1699 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1700 - Properly document the usb_mouse parameter
1701 - Fix path in ganeti-rapi(8) (issue 116)
1702 - Adjust error message when the ganeti user's .ssh directory is
1704 - Add same-node-check when changing the disk template to drbd
1710 *(Released Fri, 16 Jul 2010)*
1713 - Add an option to only select some reboot types during qa/burnin.
1714 (on some hypervisors consequent reboots are not supported)
1715 - Fix infrequent race condition in master failover. Sometimes the old
1716 master ip address would be still detected as up for a short time
1717 after it was removed, causing failover to fail.
1718 - Decrease mlockall warnings when the ctypes module is missing. On
1719 Python 2.4 we support running even if no ctypes module is installed,
1720 but we were too verbose about this issue.
1721 - Fix building on old distributions, on which man doesn't have a
1723 - Fix RAPI not to ignore the MAC address on instance creation
1724 - Implement the old instance creation format in the RAPI client.
1730 *(Released Thu, 01 Jul 2010)*
1732 A small bugfix release:
1733 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1734 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1735 a non-existing option
1736 - Fix instance create: the --force-variant option was ignored
1737 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1738 - Fix modify node storage with non-FQDN arguments
1739 - Fix RAPI client to authenticate under Python 2.6 when used
1740 for more than 5 requests needing authentication
1741 - Fix gnt-instance modify -t (storage) giving a wrong error message
1742 when converting a non-shutdown drbd instance to plain
1748 *(Released Fri, 18 Jun 2010)*
1750 A small bugfix release:
1752 - Fix live migration of KVM instances started with older Ganeti
1753 versions which had fewer hypervisor parameters
1754 - Fix gnt-instance grow-disk on down instances
1755 - Fix an error-reporting bug during instance migration
1756 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1757 silently ignoring broken ones
1758 - Fix an RPC error reporting bug affecting, for example, RAPI client
1760 - Fix bug triggered by different API version os-es on different nodes
1761 - Fix a bug in instance startup with custom hvparams: OS level
1762 parameters would fail to be applied.
1763 - Fix the RAPI client under Python 2.6 (but more work is needed to
1764 make it work completely well with OpenSSL)
1765 - Fix handling of errors when resolving names from DNS
1771 *(Released Thu, 3 Jun 2010)*
1773 A medium sized development cycle. Some new features, and some
1774 fixes/small improvements/cleanups.
1776 Significant features
1777 ~~~~~~~~~~~~~~~~~~~~
1779 The node deamon now tries to mlock itself into memory, unless the
1780 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1781 its logs, and falls back to console logging. This allows emergency
1782 features such as ``gnt-node powercycle`` to work even in the event of a
1783 broken node disk (tested offlining the disk hosting the node's
1784 filesystem and dropping its memory caches; don't try this at home)
1786 KVM: add vhost-net acceleration support. It can be tested with a new
1787 enough version of the kernel and of qemu-kvm.
1789 KVM: Add instance chrooting feature. If you use privilege dropping for
1790 your VMs you can also now force them to chroot to an empty directory,
1791 before starting the emulated guest.
1793 KVM: Add maximum migration bandwith and maximum downtime tweaking
1794 support (requires a new-enough version of qemu-kvm).
1796 Cluster verify will now warn if the master node doesn't have the master
1797 ip configured on it.
1799 Add a new (incompatible) instance creation request format to RAPI which
1800 supports all parameters (previously only a subset was supported, and it
1801 wasn't possible to extend the old format to accomodate all the new
1802 features. The old format is still supported, and a client can check for
1803 this feature, before using it, by checking for its presence in the
1804 ``features`` RAPI resource.
1806 Now with ancient latin support. Try it passing the ``--roman`` option to
1807 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1808 (requires the python-roman module to be installed, in order to work).
1813 As usual many internal code refactorings, documentation updates, and
1816 - Lots of improvements and cleanups to the experimental Remote API
1817 (RAPI) client library.
1818 - A new unit test suite for the core daemon libraries.
1819 - A fix to creating missing directories makes sure the umask is not
1820 applied anymore. This enforces the same directory permissions
1822 - Better handling terminating daemons with ctrl+c (used when running
1823 them in debugging mode).
1824 - Fix a race condition in live migrating a KVM instance, when stat()
1825 on the old proc status file returned EINVAL, which is an unexpected
1827 - Fixed manpage checking with newer man and utf-8 charachters. But now
1828 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1834 *(Released Fri, 7 May 2010)*
1836 Fix a bug which prevented untagged KVM instances from starting.
1842 *(Released Fri, 7 May 2010)*
1844 Another release with a long development cycle, during which many
1845 different features were added.
1847 Significant features
1848 ~~~~~~~~~~~~~~~~~~~~
1850 The KVM hypervisor now can run the individual instances as non-root, to
1851 reduce the impact of a VM being hijacked due to bugs in the
1852 hypervisor. It is possible to run all instances as a single (non-root)
1853 user, to manually specify a user for each instance, or to dynamically
1854 allocate a user out of a cluster-wide pool to each instance, with the
1855 guarantee that no two instances will run under the same user ID on any
1858 An experimental RAPI client library, that can be used standalone
1859 (without the other Ganeti libraries), is provided in the source tree as
1860 ``lib/rapi/client.py``. Note this client might change its interface in
1861 the future, as we iterate on its capabilities.
1863 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1864 replace the cluster's certificates and crypto keys. This might help in
1865 case they have been compromised, or have simply expired.
1867 A new disk option for instance creation has been added that allows one
1868 to "adopt" currently existing logical volumes, with data
1869 preservation. This should allow easier migration to Ganeti from
1870 unmanaged (or managed via other software) instances.
1872 Another disk improvement is the possibility to convert between redundant
1873 (DRBD) and plain (LVM) disk configuration for an instance. This should
1874 allow better scalability (starting with one node and growing the
1875 cluster, or shrinking a two-node cluster to one node).
1877 A new feature that could help with automated node failovers has been
1878 implemented: if a node sees itself as offline (by querying the master
1879 candidates), it will try to shutdown (hard) all instances and any active
1880 DRBD devices. This reduces the risk of duplicate instances if an
1881 external script automatically failovers the instances on such nodes. To
1882 enable this, the cluster parameter ``maintain_node_health`` should be
1883 enabled; in the future this option (per the name) will enable other
1884 automatic maintenance features.
1886 Instance export/import now will reuse the original instance
1887 specifications for all parameters; that means exporting an instance,
1888 deleting it and the importing it back should give an almost identical
1889 instance. Note that the default import behaviour has changed from
1890 before, where it created only one NIC; now it recreates the original
1893 Cluster verify has added a few new checks: SSL certificates validity,
1894 /etc/hosts consistency across the cluster, etc.
1899 As usual, many internal changes were done, documentation fixes,
1902 - Fixed cluster initialization with disabled cluster storage (regression
1903 introduced in 2.1.1)
1904 - File-based storage supports growing the disks
1905 - Fixed behaviour of node role changes
1906 - Fixed cluster verify for some corner cases, plus a general rewrite of
1907 cluster verify to allow future extension with more checks
1908 - Fixed log spamming by watcher and node daemon (regression introduced
1910 - Fixed possible validation issues when changing the list of enabled
1912 - Fixed cleanup of /etc/hosts during node removal
1913 - Fixed RAPI response for invalid methods
1914 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1915 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1916 from ide disks, etc.)
1917 - Allow OS changes without re-installation (to record a changed OS
1918 outside of Ganeti, or to allow OS renames)
1919 - Allow instance creation without OS installation (useful for example if
1920 the OS will be installed manually, or restored from a backup not in
1922 - Implemented option to make cluster ``copyfile`` use the replication
1924 - Added list of enabled hypervisors to ssconf (possibly useful for
1926 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1928 - A partial form of node re-IP is possible via node readd, which now
1929 allows changed node primary IP
1930 - Command line utilities now show an informational message if the job is
1932 - The logs of the master daemon now show the PID/UID/GID of the
1939 *(Released Fri, 12 Mar 2010)*
1941 During the 2.1.0 long release candidate cycle, a lot of improvements and
1942 changes have accumulated with were released later as 2.1.1.
1947 The node evacuate command (``gnt-node evacuate``) was significantly
1948 rewritten, and as such the IAllocator protocol was changed - a new
1949 request type has been added. This unfortunate change during a stable
1950 series is designed to improve performance of node evacuations; on
1951 clusters with more than about five nodes and which are well-balanced,
1952 evacuation should proceed in parallel for all instances of the node
1953 being evacuated. As such, any existing IAllocator scripts need to be
1954 updated, otherwise the above command will fail due to the unknown
1955 request. The provided "dumb" allocator has not been updated; but the
1956 ganeti-htools package supports the new protocol since version 0.2.4.
1958 Another important change is increased validation of node and instance
1959 names. This might create problems in special cases, if invalid host
1960 names are being used.
1962 Also, a new layer of hypervisor parameters has been added, that sits at
1963 OS level between the cluster defaults and the instance ones. This allows
1964 customisation of virtualization parameters depending on the installed
1965 OS. For example instances with OS 'X' may have a different KVM kernel
1966 (or any other parameter) than the cluster defaults. This is intended to
1967 help managing a multiple OSes on the same cluster, without manual
1968 modification of each instance's parameters.
1970 A tool for merging clusters, ``cluster-merge``, has been added in the
1971 tools sub-directory.
1976 - Improved the int/float conversions that should make the code more
1977 robust in face of errors from the node daemons
1978 - Fixed the remove node code in case of internal configuration errors
1979 - Fixed the node daemon behaviour in face of inconsistent queue
1980 directory (e.g. read-only file-system where we can't open the files
1982 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1983 now it either aborts cleanly or, if given the new "auto_promote"
1984 parameter, will automatically promote other nodes as needed
1985 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1986 completely prevent Ganeti from working
1987 - Fixed bug for instance export when not all disks were successfully
1989 - Fixed behaviour of node add when the new node is slow in starting up
1991 - Fixed handling of signals in the LUXI client, which should improve
1992 behaviour of command-line scripts
1993 - Added checks for invalid node/instance names in the configuration (now
1994 flagged during cluster verify)
1995 - Fixed watcher behaviour for disk activation errors
1996 - Fixed two potentially endless loops in http library, which led to the
1997 RAPI daemon hanging and consuming 100% CPU in some cases
1998 - Fixed bug in RAPI daemon related to hashed passwords
1999 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2001 - Enhanced compatibility with non-Debian OSes, but not using absolute
2002 path in some commands and allowing customisation of the ssh
2003 configuration directory
2004 - Fixed possible future issue with new Python versions by abiding to the
2005 proper use of ``__slots__`` attribute on classes
2006 - Added checks that should prevent directory traversal attacks
2007 - Many documentation fixes based on feedback from users
2012 - Added an "early_release" more for instance replace disks and node
2013 evacuate, where we release locks earlier and thus allow higher
2014 parallelism within the cluster
2015 - Added watcher hooks, intended to allow the watcher to restart other
2016 daemons (e.g. from the ganeti-nbma project), but they can be used of
2017 course for any other purpose
2018 - Added a compile-time disable for DRBD barriers, to increase
2019 performance if the administrator trusts the power supply or the
2020 storage system to not lose writes
2021 - Added the option of using syslog for logging instead of, or in
2022 addition to, Ganeti's own log files
2023 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2024 can indeed boot from a paravirtual NIC
2025 - Added a generic debug level for many operations; while this is not
2026 used widely yet, it allows one to pass the debug value all the way to
2028 - Enhanced the hooks environment for instance moves (failovers,
2029 migrations) where the primary/secondary nodes changed during the
2030 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2031 - Enhanced data validations for many user-supplied values; one important
2032 item is the restrictions imposed on instance and node names, which
2033 might reject some (invalid) host names
2034 - Add a configure-time option to disable file-based storage, if it's not
2035 needed; this allows greater security separation between the master
2036 node and the other nodes from the point of view of the inter-node RPC
2038 - Added user notification in interactive tools if job is waiting in the
2039 job queue or trying to acquire locks
2040 - Added log messages when a job is waiting for locks
2041 - Added filtering by node tags in instance operations which admit
2042 multiple instances (start, stop, reboot, reinstall)
2043 - Added a new tool for cluster mergers, ``cluster-merge``
2044 - Parameters from command line which are of the form ``a=b,c=d`` can now
2045 use backslash escapes to pass in values which contain commas,
2046 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2048 - For KVM, the instance name is the first parameter passed to KVM, so
2049 that it's more visible in the process list
2055 *(Released Tue, 2 Mar 2010)*
2057 Ganeti 2.1 brings many improvements with it. Major changes:
2059 - Added infrastructure to ease automated disk repairs
2060 - Added new daemon to export configuration data in a cheaper way than
2061 using the remote API
2062 - Instance NICs can now be routed instead of being associated with a
2064 - Improved job locking logic to reduce impact of jobs acquiring multiple
2065 locks waiting for other long-running jobs
2067 In-depth implementation details can be found in the Ganeti 2.1 design
2073 - Added chroot hypervisor
2074 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2076 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2077 ``bootloader_path`` and ``bootloader_args``)
2078 - Added the ``use_localtime`` option for the xen-hvm and kvm
2079 hypervisors, and the default value for this has changed to false (in
2080 2.0 xen-hvm always enabled it)
2081 - Added luxi call to submit multiple jobs in one go
2082 - Added cluster initialization option to not modify ``/etc/hosts``
2084 - Added network interface parameters
2085 - Added dry run mode to some LUs
2086 - Added RAPI resources:
2088 - ``/2/instances/[instance_name]/info``
2089 - ``/2/instances/[instance_name]/replace-disks``
2090 - ``/2/nodes/[node_name]/evacuate``
2091 - ``/2/nodes/[node_name]/migrate``
2092 - ``/2/nodes/[node_name]/role``
2093 - ``/2/nodes/[node_name]/storage``
2094 - ``/2/nodes/[node_name]/storage/modify``
2095 - ``/2/nodes/[node_name]/storage/repair``
2097 - Added OpCodes to evacuate or migrate all instances on a node
2098 - Added new command to list storage elements on nodes (``gnt-node
2099 list-storage``) and modify them (``gnt-node modify-storage``)
2100 - Added new ssconf files with master candidate IP address
2101 (``ssconf_master_candidates_ips``), node primary IP address
2102 (``ssconf_node_primary_ips``) and node secondary IP address
2103 (``ssconf_node_secondary_ips``)
2104 - Added ``ganeti-confd`` and a client library to query the Ganeti
2105 configuration via UDP
2106 - Added ability to run hooks after cluster initialization and before
2108 - Added automatic mode for disk replace (``gnt-instance replace-disks
2110 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2111 after catastrophic data-loss
2112 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2114 - Added ``gnt-instance move`` command to move instances
2115 - Added ``gnt-cluster watcher`` command to control watcher
2116 - Added ``gnt-node powercycle`` command to powercycle nodes
2117 - Added new job status field ``lock_status``
2118 - Added parseable error codes to cluster verification (``gnt-cluster
2119 verify --error-codes``) and made output less verbose (use
2120 ``--verbose`` to restore previous behaviour)
2121 - Added UUIDs to the main config entities (cluster, nodes, instances)
2122 - Added support for OS variants
2123 - Added support for hashed passwords in the Ganeti remote API users file
2125 - Added option to specify maximum timeout on instance shutdown
2126 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2127 - Added new helper script to start and stop Ganeti daemons
2128 (``daemon-util``), with the intent to reduce the work necessary to
2129 adjust Ganeti for non-Debian distributions and to start/stop daemons
2131 - Added more unittests
2132 - Fixed critical bug in ganeti-masterd startup
2133 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2134 now customisable at the cluster level for both the KVM and Xen
2135 hypervisors using the new ``migration_port`` parameter
2136 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2137 reinstalling an instance
2138 - Allowed ``@`` in tag names
2139 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2140 - Many documentation updates
2141 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2142 - ``gnt-instance reinstall`` can now reinstall multiple instances
2143 - Updated many command line parameters
2144 - Introduced new OS API version 15
2145 - No longer support a default hypervisor
2146 - Treat virtual LVs as inexistent
2147 - Improved job locking logic to reduce lock contention
2148 - Match instance and node names case insensitively
2149 - Reimplemented bash completion script to be more complete
2156 *(Released Thu, 4 Feb 2010)*
2158 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2159 - Fix a string formatting bug
2160 - Improve safety of the code in some error paths
2161 - Improve data validation in the master of values returned from nodes
2167 *(Released Thu, 17 Dec 2009)*
2169 - Fix security issue due to missing validation of iallocator names; this
2170 allows local and remote execution of arbitrary executables
2171 - Fix failure of gnt-node list during instance removal
2172 - Ship the RAPI documentation in the archive
2178 *(Released Wed, 30 Sep 2009)*
2180 - Fixed many wrong messages
2181 - Fixed a few bugs related to the locking library
2182 - Fixed MAC checking at instance creation time
2183 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2184 - Fixed a few issues related to signal handling in both daemons and
2186 - Fixed the example startup script provided
2187 - Fixed insserv dependencies in the example startup script (patch from
2189 - Fixed handling of drained nodes in the iallocator framework
2190 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2192 - Fixed error related to invalid job IDs in job polling
2193 - Fixed job/opcode persistence on unclean master shutdown
2194 - Fixed handling of partial job processing after unclean master
2196 - Fixed error reporting from LUs, previously all errors were converted
2197 into execution errors
2198 - Fixed error reporting from burnin
2199 - Decreased significantly the memory usage of the job queue
2200 - Optimised slightly multi-job submission
2201 - Optimised slightly opcode loading
2202 - Backported the multi-job submit framework from the development
2203 branch; multi-instance start and stop should be faster
2204 - Added script to clean archived jobs after 21 days; this will reduce
2205 the size of the queue directory
2206 - Added some extra checks in disk size tracking
2207 - Added an example ethers hook script
2208 - Added a cluster parameter that prevents Ganeti from modifying of
2210 - Added more node information to RAPI responses
2211 - Added a ``gnt-job watch`` command that allows following the ouput of a
2213 - Added a bind-address option to ganeti-rapi
2214 - Added more checks to the configuration verify
2215 - Enhanced the burnin script such that some operations can be retried
2217 - Converted instance reinstall to multi-instance model
2223 *(Released Fri, 7 Aug 2009)*
2225 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2226 to allow using the pre-2.0.2 behaviour in activation, if any existing
2227 instances have mismatched disk sizes in the configuration
2228 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2229 any configuration mismatches for disk sizes
2230 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2231 failover to work on two-node clusters
2232 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2234 - Fixed detection of OS script errors in ``gnt-backup export``
2235 - Fixed exit code of ``gnt-backup export``
2241 *(Released Fri, 17 Jul 2009)*
2243 - Added experimental support for stripped logical volumes; this should
2244 enhance performance but comes with a higher complexity in the block
2245 device handling; stripping is only enabled when passing
2246 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2247 affected even in the non-stripped mode
2248 - Improved resiliency against transient failures at the end of DRBD
2249 resyncs, and in general of DRBD resync checks
2250 - Fixed a couple of issues with exports and snapshot errors
2251 - Fixed a couple of issues in instance listing
2252 - Added display of the disk size in ``gnt-instance info``
2253 - Fixed checking for valid OSes in instance creation
2254 - Fixed handling of the "vcpus" parameter in instance listing and in
2255 general of invalid parameters
2256 - Fixed http server library, and thus RAPI, to handle invalid
2257 username/password combinations correctly; this means that now they
2258 report unauthorized for queries too, not only for modifications,
2259 allowing earlier detect of configuration problems
2260 - Added a new "role" node list field, equivalent to the master/master
2261 candidate/drained/offline flags combinations
2262 - Fixed cluster modify and changes of candidate pool size
2263 - Fixed cluster verify error messages for wrong files on regular nodes
2264 - Fixed a couple of issues with node demotion from master candidate role
2265 - Fixed node readd issues
2266 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2267 - Added a new ``--no-voting`` option for masterfailover to fix failover
2268 on two-nodes clusters when the former master node is unreachable
2269 - Added instance reinstall over RAPI
2275 *(Released Tue, 16 Jun 2009)*
2277 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2278 allow re-adding the start in single-user option (regression from 1.2)
2279 - the watcher writes the instance status to a file, to allow monitoring
2280 to report the instance status (from the master) based on cached
2281 results of the watcher's queries; while this can get stale if the
2282 watcher is being locked due to other work on the cluster, this is
2283 still an improvement
2284 - the watcher now also restarts the node daemon and the rapi daemon if
2286 - fixed the watcher to handle full and drained queue cases
2287 - hooks export more instance data in the environment, which helps if
2288 hook scripts need to take action based on the instance's properties
2289 (no longer need to query back into ganeti)
2290 - instance failovers when the instance is stopped do not check for free
2291 RAM, so that failing over a stopped instance is possible in low memory
2293 - rapi uses queries for tags instead of jobs (for less job traffic), and
2294 for cluster tags it won't talk to masterd at all but read them from
2296 - a couple of error handling fixes in RAPI
2297 - drbd handling: improved the error handling of inconsistent disks after
2298 resync to reduce the frequency of "there are some degraded disks for
2299 this instance" messages
2300 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2301 error handling path called a wrong function name)
2307 *(Released Wed, 27 May 2009)*
2309 - no changes from rc5
2315 *(Released Wed, 20 May 2009)*
2317 - fix a couple of bugs (validation, argument checks)
2318 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2319 - some small improvements to RAPI and IAllocator
2320 - make watcher automatically start the master daemon if down
2326 *(Released Mon, 27 Apr 2009)*
2328 - change the OS list to not require locks; this helps with big clusters
2329 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2330 volume group is broken
2331 - ``gnt-instance info``, without any arguments, doesn't run for all
2332 instances anymore; either pass ``--all`` or pass the desired
2333 instances; this helps against mistakes on big clusters where listing
2334 the information for all instances takes a long time
2335 - miscellaneous doc and man pages fixes
2341 *(Released Wed, 8 Apr 2009)*
2343 - Change the internal locking model of some ``gnt-node`` commands, in
2344 order to reduce contention (and blocking of master daemon) when
2345 batching many creation/reinstall jobs
2346 - Fixes to Xen soft reboot
2347 - No longer build documentation at build time, instead distribute it in
2348 the archive, in order to reduce the need for the whole docbook/rst
2355 *(Released Fri, 27 Mar 2009)*
2357 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2358 - Fix watcher startup sequence, improves the behaviour of busy clusters
2359 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2360 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2361 bind address and other places
2362 - Some documentation fixes and updates
2368 *(Released Mon, 2 Mar 2009)*
2370 - More documentation updates, now all docs should be more-or-less
2372 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2374 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2376 - Fix an issue related to $libdir/run/ganeti and cluster creation
2382 *(Released Thu, 19 Feb 2009)*
2384 - Xen PVM and KVM have switched the default value for the instance root
2385 disk to the first partition on the first drive, instead of the whole
2386 drive; this means that the OS installation scripts must be changed
2388 - Man pages have been updated
2389 - RAPI has been switched by default to HTTPS, and the exported functions
2390 should all work correctly
2391 - RAPI v1 has been removed
2392 - Many improvements to the KVM hypervisor
2393 - Block device errors are now better reported
2394 - Many other bugfixes and small improvements
2400 *(Released Mon, 26 Jan 2009)*
2402 - Version 2 is a general rewrite of the code and therefore the
2403 differences are too many to list, see the design document for 2.0 in
2404 the ``doc/`` subdirectory for more details
2405 - In this beta version there is not yet a migration path from 1.2 (there
2406 will be one in the final 2.0 release)
2407 - A few significant changes are:
2409 - all commands are executed by a daemon (``ganeti-masterd``) and the
2410 various ``gnt-*`` commands are just front-ends to it
2411 - all the commands are entered into, and executed from a job queue,
2412 see the ``gnt-job(8)`` manpage
2413 - the RAPI daemon supports read-write operations, secured by basic
2414 HTTP authentication on top of HTTPS
2415 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2416 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2417 to migrate to DRBD 8 first while still running Ganeti 1.2)
2418 - DRBD devices are using statically allocated minor numbers, which
2419 will be assigned to existing instances during the migration process
2420 - there is support for both Xen PVM and Xen HVM instances running on
2422 - KVM virtualization is supported too
2423 - file-based storage has been implemented, which means that it is
2424 possible to run the cluster without LVM and DRBD storage, for
2425 example using a shared filesystem exported from shared storage (and
2426 still have live migration)
2432 *(Released Tue, 13 Jan 2009)*
2434 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2435 - Reuse the old instance mac address by default on instance import, if
2436 the instance name is the same.
2437 - Handle situations in which the node info rpc returns incomplete
2439 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2440 - Improved version of batcher:
2442 - state file support
2443 - instance mac address support
2444 - support for HVM clusters/instances
2446 - Add an option to show the number of cpu sockets and nodes in
2448 - Support OSes that handle more than one version of the OS api (but do
2449 not change the current API in any other way)
2450 - Fix ``gnt-node migrate``
2451 - ``gnt-debug`` man page
2452 - Fixes various more typos and small issues
2453 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2459 *(Released Wed, 24 Sep 2008)*
2461 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2462 type of disk exported to fully virtualized instances.
2463 - provide access to the serial console of HVM instances
2464 - instance auto_balance flag, set by default. If turned off it will
2465 avoid warnings on cluster verify if there is not enough memory to fail
2466 over an instance. in the future it will prevent automatically failing
2467 it over when we will support that.
2468 - batcher tool for instance creation, see ``tools/README.batcher``
2469 - ``gnt-instance reinstall --select-os`` to interactively select a new
2470 operating system when reinstalling an instance.
2471 - when changing the memory amount on instance modify a check has been
2472 added that the instance will be able to start. also warnings are
2473 emitted if the instance will not be able to fail over, if auto_balance
2475 - documentation fixes
2476 - sync fields between ``gnt-instance list/modify/add/import``
2477 - fix a race condition in drbd when the sync speed was set after giving
2478 the device a remote peer.
2484 *(Released Tue, 22 Jul 2008)*
2486 - note: the allowed size and number of tags per object were reduced
2487 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2488 - fixed twisted 8.x compatibility
2489 - fixed ``gnt-instance replace-disks`` with iallocator
2490 - add TCP keepalives on twisted connections to detect restarted nodes
2491 - disk increase support, see ``gnt-instance grow-disk``
2492 - implement bulk node/instance query for RAPI
2493 - add tags in node/instance listing (optional)
2494 - experimental migration (and live migration) support, read the man page
2495 for ``gnt-instance migrate``
2496 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2497 has some small improvements in handling its state file
2503 *(Released Fri, 13 Jun 2008)*
2505 - Experimental readonly, REST-based remote API implementation;
2506 automatically started on master node, TCP port 5080, if enabled by
2507 ``--enable-rapi`` parameter to configure script.
2508 - Instance allocator support. Add and import instance accept a
2509 ``--iallocator`` parameter, and call that instance allocator to decide
2510 which node to use for the instance. The iallocator document describes
2511 what's expected from an allocator script.
2512 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2513 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2514 node is lost there is still enough memory to fail over the instances
2516 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2517 ``gnt-cluster verify``, to check for site-specific compliance. All the
2518 hooks will run, and their output, if any, will be displayed. Any
2519 failing hook will make the verification return an error value.
2520 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2521 primary and secondary interfaces
2522 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2523 that is still declared as part of the cluster and has failed.
2524 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2525 specifying output fields, that just adds the chosen fields to the
2527 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2528 export from the filesystem.
2529 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2530 have been added. Using them you can enable/disable acpi and pae
2531 support, and specify a path for a cd image to be exported to the
2532 instance. These parameters as the name suggest only work on HVM
2534 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2535 PAE support will be set to the previously hardcoded values, but the
2536 (previously hardcoded) path to the CDROM ISO image will be unset and
2537 if required, needs to be set manually with ``gnt-instance modify``
2539 - The address to which an instance's VNC console is bound is now
2540 selectable per-instance, rather than being cluster wide. Of course
2541 this only applies to instances controlled via VNC, so currently just
2542 applies to HVM clusters.
2548 *(Released Mon, 18 Feb 2008)*
2550 - more tweaks to the disk activation code (especially helpful for DRBD)
2551 - change the default ``gnt-instance list`` output format, now there is
2552 one combined status field (see the manpage for the exact values this
2554 - some more fixes for the mac export to hooks change
2555 - make Ganeti not break with DRBD 8.2.x (which changed the version
2556 format in ``/proc/drbd``) (issue 24)
2557 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2558 template, allowing migration from DRBD0.7+MD to DRBD8
2564 *(Released Wed, 30 Jan 2008)*
2566 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2568 - add command aliases infrastructure and a few aliases
2569 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2570 man pages and the ``--help`` option of ``gnt-node
2571 list``/``gnt-instance list``
2572 - fix ``gnt-backup list`` with down nodes (issue 21)
2573 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2574 - fix the dist archive and add a check for including svn/git files in
2576 - some developer-related changes: improve the burnin and the QA suite,
2577 add an upload script for testing during development
2583 *(Released Wed, 16 Jan 2008)*
2585 - experimental HVM support, read the install document, section
2586 "Initializing the cluster"
2587 - allow for the PVM hypervisor per-instance kernel and initrd paths
2588 - add a new command ``gnt-cluster verify-disks`` which uses a new
2589 algorithm to improve the reconnection of the DRBD pairs if the device
2590 on the secondary node has gone away
2591 - make logical volume code auto-activate LVs at disk activation time
2592 - slightly improve the speed of activating disks
2593 - allow specification of the MAC address at instance creation time, and
2594 changing it later via ``gnt-instance modify``
2595 - fix handling of external commands that generate lots of output on
2597 - update documentation with regard to minimum version of DRBD8 supported
2603 *(Released Tue, 4 Dec 2007)*
2605 - Log the ``xm create`` output to the node daemon log on failure (to
2606 help diagnosing the error)
2607 - In debug mode, log all external commands output if failed to the logs
2608 - Change parsing of lvm commands to ignore stderr
2614 *(Released Wed, 28 Nov 2007)*
2616 - Another round of updates to the DRBD 8 code to deal with more failures
2617 in the replace secondary node operation
2618 - Some more logging of failures in disk operations (lvm, drbd)
2619 - A few documentation updates
2626 *(Released Tue, 13 Nov 2007)*
2628 - Change configuration file format from Python's Pickle to JSON.
2629 Upgrading is possible using the cfgupgrade utility.
2630 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2631 faster replace disks and is more stable (DRBD 8 has many improvements
2632 compared to DRBD 0.7)
2633 - Added command line tags support (see man pages for ``gnt-instance``,
2634 ``gnt-node``, ``gnt-cluster``)
2635 - Added instance rename support
2636 - Added multi-instance startup/shutdown
2637 - Added cluster rename support
2638 - Added ``gnt-node evacuate`` to simplify some node operations
2639 - Added instance reboot operation that can speedup reboot as compared to
2641 - Soften the requirement that hostnames are in FQDN format
2642 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2644 - Removed dependency on debian's patched fping that uses the
2645 non-standard ``-S`` option
2646 - Now the OS definitions are searched for in multiple, configurable
2647 paths (easier for distros to package)
2648 - Some changes to the hooks infrastructure (especially the new
2649 post-configuration update hook)
2650 - Other small bugfixes
2652 .. vim: set textwidth=72 syntax=rst :