10 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
11 the correct command syntax. It is possible to use different DRBD versions
12 on different nodes as long as they are compatible to each other. This
13 enables rolling upgrades of DRBD with no downtime. As permanent operation
14 of different DRBD versions within a node group is discouraged,
15 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
23 - The minimum Python version needed to run Ganeti is now 2.6.
24 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
25 to always require authentication, ``--require-authentication``. It can
26 be specified in ``$sysconfdir/default/ganeti``.
27 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
28 be used to manage the disk templates to be used by instances in the cluster.
29 Initially, it will be set to a list that includes plain, drbd, if they were
30 enabled by specifying a volume group name, and file and sharedfile, if those
31 were enabled at configure time. Additionally, it will include all disk
32 templates that are currently used by instances. The order of disk templates
33 will be based on Ganeti's history of supporting them. In the future, the
34 first entry of the list will be used as a default disk template on instance
36 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
37 configuration back to the previous stable version.
38 - The cluster option '--no-lvm-storage' was removed in favor of the new option
39 '--enabled-disk-templates'.
47 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
49 - Disks added through ``gnt-instance modify`` or created through
50 ``gnt-instance recreate-disks`` are wiped, if the
51 ``prealloc_wipe_disks`` flag is set.
52 - If wiping newly created disks fails, the disks are removed. Also,
53 partial failures in creating disks through ``gnt-instance modify``
54 triggers a cleanup of the partially-created disks.
55 - Removing the master IP address doesn't fail if the address has been
62 *(Released Fri, 3 May 2013)*
64 Incompatible/important changes
65 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
67 - Instance policies for disk size were documented to be on a per-disk
68 basis, but hail applied them to the sum of all disks. This has been
70 - ``hbal`` will now exit with status 0 if, during job execution over
71 LUXI, early exit has been requested and all jobs are successful;
72 before, exit status 1 was used, which cannot be differentiated from
74 - Compatibility with newer versions of rbd has been fixed
75 - ``gnt-instance batch-create`` has been changed to use the bulk create
76 opcode from Ganeti. This lead to incompatible changes in the format of
77 the JSON file. It's now not a custom dict anymore but a dict
78 compatible with the ``OpInstanceCreate`` opcode.
79 - Parent directories for file storage need to be listed in
80 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
81 write the file automatically based on old configuration values, but it
82 can not distribute it across all nodes and the file contents should be
83 verified. Use ``gnt-cluster copyfile
84 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
85 upgraded. The reason for requiring this list of paths now is that
86 before it would have been possible to inject new paths via RPC,
87 allowing files to be created in arbitrary locations. The RPC protocol
88 is protected using SSL/X.509 certificates, but as a design principle
89 Ganeti does not permit arbitrary paths to be passed.
90 - The parsing of the variants file for OSes (see
91 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
92 lines and comment lines (starting with ``#``) are ignored for better
94 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
95 longer available. ``gnt-node add`` now invokes a new tool on the
96 destination node, named ``prepare-node-join``, to configure the SSH
97 daemon. Paramiko is no longer necessary to configure nodes' SSH
98 daemons via ``gnt-node add``.
99 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
100 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
101 the flag is not reset after a master failover.
102 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
103 is recommended. 2.6 will be mandatory from the 2.8 series.
109 - New network management functionality to support automatic allocation
110 of IP addresses and managing of network parameters. See
111 :manpage:`gnt-network(8)` for more details.
112 - New external storage backend, to allow managing arbitrary storage
113 systems external to the cluster. See
114 :manpage:`ganeti-extstorage-interface(7)`.
115 - New ``exclusive-storage`` node parameter added, restricted to
116 nodegroup level. When it's set to true, physical disks are assigned in
117 an exclusive fashion to instances, as documented in :doc:`Partitioned
118 Ganeti <design-partitioned>`. Currently, only instances using the
119 ``plain`` disk template are supported.
120 - The KVM hypervisor has been updated with many new hypervisor
121 parameters, including a generic one for passing arbitrary command line
122 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
123 compatible up to qemu 1.4.
124 - A new tool, called ``mon-collector``, is the stand-alone executor of
125 the data collectors for a monitoring system. As of this version, it
126 just includes the DRBD data collector, that can be executed by calling
127 ``mon-collector`` using the ``drbd`` parameter. See
128 :manpage:`mon-collector(7)`.
129 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
130 for RAPI users. It allows granting permissions to query for
131 information to a specific user without giving
132 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
133 - A new tool named ``node-cleanup`` has been added. It cleans remains of
134 a cluster from a machine by stopping all daemons, removing
135 certificates and ssconf files. Unless the ``--no-backup`` option is
136 given, copies of the certificates are made.
137 - Instance creations now support the use of opportunistic locking,
138 potentially speeding up the (parallel) creation of multiple instances.
139 This feature is currently only available via the :doc:`RAPI
140 <rapi>` interface and when an instance allocator is used. If the
141 ``opportunistic_locking`` parameter is set the opcode will try to
142 acquire as many locks as possible, but will not wait for any locks
143 held by other opcodes. If not enough resources can be found to
144 allocate the instance, the temporary error code
145 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
146 retried thereafter, with or without opportunistic locking.
147 - New experimental linux-ha resource scripts.
148 - Restricted-commands support: ganeti can now be asked (via command line
149 or rapi) to perform commands on a node. These are passed via ganeti
150 RPC rather than ssh. This functionality is restricted to commands
151 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
152 reasons. The file is not copied automatically.
158 - Diskless instances are now externally mirrored (Issue 237). This for
159 now has only been tested in conjunction with explicit target nodes for
161 - Queries not needing locks or RPC access to the node can now be
162 performed by the confd daemon, making them independent from jobs, and
163 thus faster to execute. This is selectable at configure time.
164 - The functionality for allocating multiple instances at once has been
165 overhauled and is now also available through :doc:`RAPI <rapi>`.
170 - Fix kvm compatibility with qemu 1.4 (Issue 389)
171 - Documentation updates (admin guide, upgrade notes, install
172 instructions) (Issue 372)
173 - Fix gnt-group list nodes and instances count (Issue 436)
174 - Fix compilation without non-mandatory libraries (Issue 441)
175 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
176 - Make confd logging more verbose at INFO level (Issue 435)
177 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
178 - Fix failure path for instance storage type conversion (Issue 229)
179 - Update htools text backend documentation
180 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
181 - Disable inter-cluster instance move for file-based instances, because
182 it is dependant on instance export, which is not supported for
183 file-based instances. (Issue 414)
184 - Fix gnt-job crashes on non-ascii characters (Issue 427)
185 - Fix volume group checks on non-vm-capable nodes (Issue 432)
191 *(Released Mon, 22 Apr 2013)*
193 This was the third beta release of the 2.7 series. Since beta2:
195 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
196 - Fix data loss on wrong usage of ``gnt-instance move``
197 - Properly export errors in confd-based job queries
198 - Add ``users-setup`` tool
199 - Fix iallocator protocol to report 0 as a disk size for diskless
200 instances. This avoids hail breaking when a diskless instance is
202 - Fix job queue directory permission problem that made confd job queries
203 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
204 for access to archived jobs (Issue 406).
205 - Limit the sizes of networks supported by ``gnt-network`` to something
206 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
207 - Fix bugs in instance disk template conversion
208 - Fix GHC 7 compatibility
209 - Fix ``burnin`` install path (Issue 426).
210 - Allow very small disk grows (Issue 347).
211 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
212 that noded doesn't import masterd code (Issue 419).
213 - Make sure the default metavg at cluster init is the same as the vg, if
214 unspecified (Issue 358).
215 - Fix cleanup of partially created disks (part of Issue 416)
221 *(Released Tue, 2 Apr 2013)*
223 This was the second beta release of the 2.7 series. Since beta1:
225 - Networks no longer have a "type" slot, since this information was
226 unused in Ganeti: instead of it tags should be used.
227 - The rapi client now has a ``target_node`` option to MigrateInstance.
228 - Fix early exit return code for hbal (Issue 386).
229 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
230 - Fix ``rbd showmapped`` output parsing (Issue 312).
231 - Networks are now referenced indexed by UUID, rather than name. This
232 will require running cfgupgrade, from 2.7.0beta1, if networks are in
234 - The OS environment now includes network information.
235 - Deleting of a network is now disallowed if any instance nic is using
236 it, to prevent dangling references.
237 - External storage is now documented in man pages.
238 - The exclusive_storage flag can now only be set at nodegroup level.
239 - Hbal can now submit an explicit priority with its jobs.
240 - Many network related locking fixes.
241 - Bump up the required pylint version to 0.25.1.
242 - Fix the ``no_remember`` option in RAPI client.
243 - Many ipolicy related tests, qa, and fixes.
244 - Many documentation improvements and fixes.
245 - Fix building with ``--disable-file-storage``.
246 - Fix ``-q`` option in htools, which was broken if passed more than
248 - Some haskell/python interaction improvements and fixes.
249 - Fix iallocator in case of missing LVM storage.
250 - Fix confd config load in case of ``--no-lvm-storage``.
251 - The confd/query functionality is now mentioned in the security
258 *(Released Wed, 6 Feb 2013)*
260 This was the first beta release of the 2.7 series. All important changes
261 are listed in the latest 2.7 entry.
267 *(Released Fri, 21 Dec 2012)*
269 Important behaviour change: hbal won't rebalance anymore instances which
270 have the ``auto_balance`` attribute set to false. This was the intention
271 all along, but until now it only skipped those from the N+1 memory
272 reservation (DRBD-specific).
274 A significant number of bug fixes in this release:
276 - Fixed disk adoption interaction with ipolicy checks.
277 - Fixed networking issues when instances are started, stopped or
278 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
279 - Fixed the warning in cluster verify for shared storage instances not
281 - Fixed removal of storage directory on shared file storage (issue 262).
282 - Fixed validation of LVM volume group name in OpClusterSetParams
283 (``gnt-cluster modify``) (issue 285).
284 - Fixed runtime memory increases (``gnt-instance modify -m``).
285 - Fixed live migration under Xen's ``xl`` mode.
286 - Fixed ``gnt-instance console`` with ``xl``.
287 - Fixed building with newer Haskell compiler/libraries.
288 - Fixed PID file writing in Haskell daemons (confd); this prevents
289 restart issues if confd was launched manually (outside of
290 ``daemon-util``) while another copy of it was running
291 - Fixed a type error when doing live migrations with KVM (issue 297) and
292 the error messages for failing migrations have been improved.
293 - Fixed opcode validation for the out-of-band commands (``gnt-node
295 - Fixed a type error when unsetting OS hypervisor parameters (issue
296 311); now it's possible to unset all OS-specific hypervisor
298 - Fixed the ``dry-run`` mode for many operations: verification of
299 results was over-zealous but didn't take into account the ``dry-run``
300 operation, resulting in "wrong" failures.
301 - Fixed bash completion in ``gnt-job list`` when the job queue has
302 hundreds of entries; especially with older ``bash`` versions, this
303 results in significant CPU usage.
305 And lastly, a few other improvements have been made:
307 - Added option to force master-failover without voting (issue 282).
308 - Clarified error message on lock conflict (issue 287).
309 - Logging of newly submitted jobs has been improved (issue 290).
310 - Hostname checks have been made uniform between instance rename and
312 - The ``--submit`` option is now supported by ``gnt-debug delay``.
313 - Shutting down the master daemon by sending SIGTERM now stops it from
314 processing jobs waiting for locks; instead, those jobs will be started
315 once again after the master daemon is started the next time (issue
317 - Support for Xen's ``xl`` program has been improved (besides the fixes
319 - Reduced logging noise in the Haskell confd daemon (only show one log
320 entry for each config reload, instead of two).
321 - Several man page updates and typo fixes.
327 *(Released Fri, 12 Oct 2012)*
329 A small bugfix release. Among the bugs fixed:
331 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
332 made the command unusable.
333 - Commands that issue many jobs don't fail anymore just because some jobs
334 take so long that other jobs are archived.
335 - Failures during ``gnt-instance reinstall`` are reflected by the exit
337 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
339 - When ``always_failover`` is set, ``--allow-failover`` is not required
340 in migrate commands anymore.
341 - ``bash_completion`` works even if extglob is disabled.
342 - Fixed bug with locks that made failover for RDB-based instances fail.
343 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
344 a random node instead of one based on the allocator metric.
345 - Support for newer versions of pylint and pep8.
346 - Hail doesn't fail anymore when trying to add an instance of type
347 ``file``, ``sharedfile`` or ``rbd``.
348 - Added new Makefile target to rebuild the whole distribution, so that
349 all files are included.
355 *(Released Fri, 27 Jul 2012)*
358 .. attention:: The ``LUXI`` protocol has been made more consistent
359 regarding its handling of command arguments. This, however, leads to
360 incompatibility issues with previous versions. Please ensure that you
361 restart Ganeti daemons soon after the upgrade, otherwise most
362 ``LUXI`` calls (job submission, setting/resetting the drain flag,
363 pausing/resuming the watcher, cancelling and archiving jobs, querying
364 the cluster configuration) will fail.
373 The current ``admin_up`` field, which used to denote whether an instance
374 should be running or not, has been removed. Instead, ``admin_state`` is
375 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
377 The rational behind this is that an instance being “down” can have
380 - it could be down during a reboot
381 - it could be temporarily be down for a reinstall
382 - or it could be down because it is deprecated and kept just for its
385 The previous Boolean state was making it difficult to do capacity
386 calculations: should Ganeti reserve memory for a down instance? Now, the
387 tri-state field makes it clear:
389 - in ``up`` and ``down`` state, all resources are reserved for the
390 instance, and it can be at any time brought up if it is down
391 - in ``offline`` state, only disk space is reserved for it, but not
394 The field can have an extra use: since the transition between ``up`` and
395 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
396 transition between ``offline`` and ``down`` is done via ``gnt-instance
397 modify``, it is possible to given different rights to users. For
398 example, owners of an instance could be allowed to start/stop it, but
399 not transition it out of the offline state.
401 Instance policies and specs
402 +++++++++++++++++++++++++++
404 In previous Ganeti versions, an instance creation request was not
405 limited on the minimum size and on the maximum size just by the cluster
406 resources. As such, any policy could be implemented only in third-party
407 clients (RAPI clients, or shell wrappers over ``gnt-*``
408 tools). Furthermore, calculating cluster capacity via ``hspace`` again
409 required external input with regards to instance sizes.
411 In order to improve these workflows and to allow for example better
412 per-node group differentiation, we introduced instance specs, which
415 - minimum instance disk size, disk count, memory size, cpu count
416 - maximum values for the above metrics
417 - and “standard” values (used in ``hspace`` to calculate the standard
420 The minimum/maximum values can be also customised at node-group level,
421 for example allowing more powerful hardware to support bigger instance
424 Beside the instance specs, there are a few other settings belonging to
425 the instance policy framework. It is possible now to customise, per
426 cluster and node-group:
428 - the list of allowed disk templates
429 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
430 - the maximum ratio of instance to spindles (see below for more
431 information) for local storage
433 All these together should allow all tools that talk to Ganeti to know
434 what are the ranges of allowed values for instances and the
435 over-subscription that is allowed.
437 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
438 instance configuration, and the physical CPU configuration from the
439 node. For the spindle ratios however, we didn't track before these
440 values, so new parameters have been added:
442 - a new node parameter ``spindle_count``, defaults to 1, customisable at
443 node group or node level
444 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
446 Note that spindles in this context doesn't need to mean actual
447 mechanical hard-drives; it's just a relative number for both the node
448 I/O capacity and instance I/O consumption.
450 Instance migration behaviour
451 ++++++++++++++++++++++++++++
453 While live-migration is in general desirable over failover, it is
454 possible that for some workloads it is actually worse, due to the
455 variable time of the “suspend” phase during live migration.
457 To allow the tools to work consistently over such instances (without
458 having to hard-code instance names), a new backend parameter
459 ``always_failover`` has been added to control the migration/failover
460 behaviour. When set to True, all migration requests for an instance will
461 instead fall-back to failover.
463 Instance memory ballooning
464 ++++++++++++++++++++++++++
466 Initial support for memory ballooning has been added. The memory for an
467 instance is no longer fixed (backend parameter ``memory``), but instead
468 can vary between minimum and maximum values (backend parameters
469 ``minmem`` and ``maxmem``). Currently we only change an instance's
472 - live migrating or failing over and instance and the target node
473 doesn't have enough memory
474 - user requests changing the memory via ``gnt-instance modify
480 In order to control the use of specific CPUs by instance, support for
481 controlling CPU pinning has been added for the Xen, HVM and LXC
482 hypervisors. This is controlled by a new hypervisor parameter
483 ``cpu_mask``; details about possible values for this are in the
484 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
485 VCPU-to-CPU mapping) form will work well only when all nodes in your
486 cluster have the same amount of CPUs.
491 Another area in which Ganeti was not customisable were the parameters
492 used for storage configuration, e.g. how many stripes to use for LVM,
493 DRBD resync configuration, etc.
495 To improve this area, we've added disks parameters, which are
496 customisable at cluster and node group level, and which allow to
497 specify various parameters for disks (DRBD has the most parameters
498 currently), for example:
500 - DRBD resync algorithm and parameters (e.g. speed)
501 - the default VG for meta-data volumes for DRBD
502 - number of stripes for LVM (plain disk template)
505 These parameters can be modified via ``gnt-cluster modify -D …`` and
506 ``gnt-group modify -D …``, and are used at either instance creation (in
507 case of LVM stripes, for example) or at disk “activation” time
510 Rados block device support
511 ++++++++++++++++++++++++++
513 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
514 denoted by the ``rbd`` disk template type. This is considered
515 experimental, feedback is welcome. For details on configuring it, see
516 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
521 The existing master IP functionality works well only in simple setups (a
522 single network shared by all nodes); however, if nodes belong to
523 different networks, then the ``/32`` setup and lack of routing
524 information is not enough.
526 To allow the master IP to function well in more complex cases, the
527 system was reworked as follows:
529 - a master IP netmask setting has been added
530 - the master IP activation/turn-down code was moved from the node daemon
532 - whether to run the Ganeti-supplied master IP script or a user-supplied
533 on is a ``gnt-cluster init`` setting
535 Details about the location of the standard and custom setup scripts are
536 in the man page :manpage:`gnt-cluster(8)`; for information about the
537 setup script protocol, look at the Ganeti-supplied script.
542 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
545 It is now possible to use TLS-protected connections, and when renewing
546 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
547 it is now possible to specify spice or spice CA certificates. Also, it
548 is possible to configure a password for SPICE sessions via the
549 hypervisor parameter ``spice_password_file``.
551 There are also new parameters to control the compression and streaming
552 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
553 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
554 for the spice parameters.
556 Lastly, it is now possible to see the SPICE connection information via
557 ``gnt-instance console``.
562 A new tool (``tools/ovfconverter``) has been added that supports
563 conversion between Ganeti and the `Open Virtualization Format
564 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
567 This relies on the ``qemu-img`` tool to convert the disk formats, so the
568 actual compatibility with other virtualization solutions depends on it.
573 The configuration query daemon (``ganeti-confd``) is now optional, and
574 has been rewritten in Haskell; whether to use the daemon at all, use the
575 Python (default) or the Haskell version is selectable at configure time
576 via the ``--enable-confd`` parameter, which can take one of the
577 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
578 daemon will result in a smaller footprint; for larger systems, we
579 welcome feedback on the Haskell version which might become the default
582 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
583 daemon running. The Python version doesn't implement the new call.
586 User interface changes
587 ~~~~~~~~~~~~~~~~~~~~~~
589 We have replaced the ``--disks`` option of ``gnt-instance
590 replace-disks`` with a more flexible ``--disk`` option, which allows
591 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
592 disk size and mode can be changed upon recreation (via ``gnt-instance
593 recreate-disks``, which accepts the same ``--disk`` option).
595 As many people are used to a ``show`` command, we have added that as an
596 alias to ``info`` on all ``gnt-*`` commands.
598 The ``gnt-instance grow-disk`` command has a new mode in which it can
599 accept the target size of the disk, instead of the delta; this can be
600 more safe since two runs in absolute mode will be idempotent, and
601 sometimes it's also easier to specify the desired size directly.
603 Also the handling of instances with regard to offline secondaries has
604 been improved. Instance operations should not fail because one of it's
605 secondary nodes is offline, even though it's safe to proceed.
607 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
608 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
609 minors to instance name.
614 RAPI coverage has improved, with (for example) new resources for
615 recreate-disks, node power-cycle, etc.
620 There is partial support for ``xl`` in the Xen hypervisor; feedback is
623 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
624 whether to still support Python 2.4 or move to Python 2.6 as minimum
627 Support for Fedora has been slightly improved; the provided example
628 init.d script should work better on it and the INSTALL file should
629 document the needed dependencies.
634 The deprecated ``QueryLocks`` LUXI request has been removed. Use
635 ``Query(what=QR_LOCK, ...)`` instead.
637 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
638 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
639 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
640 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
641 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
643 RAPI client: ``CertificateError`` now derives from
644 ``GanetiApiError``. This should make it more easy to handle Ganeti
647 Deprecation warnings due to PyCrypto/paramiko import in
648 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
649 make sure to run an up-to-date paramiko version, if you use this tool.
651 The QA scripts now depend on Python 2.5 or above (the main code base
652 still works with Python 2.4).
654 The configuration file (``config.data``) is now written without
655 indentation for performance reasons; if you want to edit it, it can be
656 re-formatted via ``tools/fmtjson``.
658 A number of bugs has been fixed in the cluster merge tool.
660 ``x509`` certification verification (used in import-export) has been
661 changed to allow the same clock skew as permitted by the cluster
662 verification. This will remove some rare but hard to diagnose errors in
669 *(Released Thu, 19 Jul 2012)*
671 Very few changes from rc4 to the final release, only bugfixes:
673 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
674 instance, fix CDROM booting for KVM instances)
675 - fixed node group modification of node parameters
676 - fixed issue in LUClusterVerifyGroup with multi-group clusters
677 - fixed generation of bash completion to ensure a stable ordering
684 *(Released Fri, 13 Jul 2012)*
686 Third release candidate for 2.6. The following changes were done from
689 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
690 - Fixed an inconsistency in the LUXI protocol with the provided
691 arguments (NOT backwards compatible)
692 - Fixed a bug with node groups ipolicy where ``min`` was greater than
693 the cluster ``std`` value
694 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
695 easier instance debugging on nodes (requires ``hconfd`` to work)
701 *(Released Tue, 03 Jul 2012)*
703 Second release candidate for 2.6. The following changes were done from
706 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
708 - Fixed a RAPI regression on missing beparams/memory
709 - Fixed redistribution of files on offline nodes
710 - Added possibility to run activate-disks even though secondaries are
711 offline. With this change it relaxes also the strictness on some other
712 commands which use activate disks internally:
713 * ``gnt-instance start|reboot|rename|backup|export``
714 - Made it possible to remove safely an instance if its secondaries are
716 - Made it possible to reinstall even though secondaries are offline
722 *(Released Mon, 25 Jun 2012)*
724 First release candidate for 2.6. The following changes were done from
727 - Fixed bugs with disk parameters and ``rbd`` templates as well as
729 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
730 behaviour. It supports now the modify operation
731 - ``hcheck`` implemented to analyze cluster health and possibility of
732 improving health by rebalance
733 - ``hbal`` has been improved in dealing with split instances
739 *(Released Mon, 11 Jun 2012)*
741 Second beta release of 2.6. The following changes were done from beta2
744 - Fixed ``daemon-util`` with non-root user models
745 - Fixed creation of plain instances with ``--no-wait-for-sync``
746 - Fix wrong iv_names when running ``cfgupgrade``
747 - Export more information in RAPI group queries
748 - Fixed bug when changing instance network interfaces
749 - Extended burnin to do NIC changes
750 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
751 - Changed default for DRBD barriers
752 - Fixed DRBD error reporting for syncer rate
753 - Verify the options on disk parameters
755 And of course various fixes to documentation and improved unittests and
762 *(Released Wed, 23 May 2012)*
764 First beta release of 2.6. The following changes were done from beta1 to
767 - integrated patch for distributions without ``start-stop-daemon``
768 - adapted example init.d script to work on Fedora
769 - fixed log handling in Haskell daemons
770 - adapted checks in the watcher for pycurl linked against libnss
771 - add partial support for ``xl`` instead of ``xm`` for Xen
772 - fixed a type issue in cluster verification
773 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
776 Plus integrated fixes from the 2.5 branch:
778 - fixed ``kvm-ifup`` to use ``/bin/bash``
779 - fixed parallel build failures
780 - KVM live migration when using a custom keymap
786 *(Released Tue, 24 Jul 2012)*
788 A small bugfix release, with no new features:
790 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
791 different default shell (e.g. Debian, Ubuntu)
792 - fixed KVM startup and live migration with a custom keymap (fixes Issue
793 243 and Debian bug #650664)
794 - fixed compatibility with KVM versions that don't support multiple boot
795 devices (fixes Issue 230 and Debian bug #624256)
797 Additionally, a few fixes were done to the build system (fixed parallel
798 build failures) and to the unittests (fixed race condition in test for
799 FileID functions, and the default enable/disable mode for QA test is now
806 *(Released Fri, 11 May 2012)*
808 A small bugfix release.
810 The main issues solved are on the topic of compatibility with newer LVM
813 - fixed parsing of ``lv_attr`` field
814 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
815 the ``--force`` flag is needed
817 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
818 to accept kernel 3.x too (was hardcoded to 2.6.*).
820 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
821 of hook results and that also made display of other errors suboptimal
822 was fixed; the code behaves now like 2.4 and earlier.
824 Another change in 2.5, the cleanup of the OS scripts environment, is too
825 aggressive: it removed even the ``PATH`` variable, which requires the OS
826 scripts to *always* need to export it. Since this is a bit too strict,
827 we now export a minimal PATH, the same that we export for hooks.
829 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
830 integrated into this release.
832 Finally, a few other miscellaneous changes were done (no new features,
833 just small improvements):
835 - Fix ``gnt-group --help`` display
836 - Fix hardcoded Xen kernel path
837 - Fix grow-disk handling of invalid units
838 - Update synopsis for ``gnt-cluster repair-disk-sizes``
839 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
845 *(Released Thu, 12 Apr 2012)*
847 Incompatible/important changes and bugfixes
848 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
850 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
851 resource's ``ip_check`` parameter changed from ``True`` to ``False``
852 to match the underlying LUXI interface.
853 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
854 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
855 not maintain backwards-compatibility as the underlying operation
856 changed in an incompatible way. The RAPI client can talk to old
857 servers, but it needs to be told so as the return value changed.
858 - When creating file-based instances via RAPI, the ``file_driver``
859 parameter no longer defaults to ``loop`` and must be specified.
860 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
862 - Support for the undocumented and deprecated RAPI instance creation
863 request format version 0 has been dropped. Use version 1, supported
864 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
865 - Pyparsing 1.4.6 or above is required, see :doc:`installation
866 documentation <install>`.
867 - The "cluster-verify" hooks are now executed per group by the
868 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
869 if you just run ``gnt-cluster verify``, which generates one opcode per
871 - The environment as passed to the OS scripts is cleared, and thus no
872 environment variables defined in the node daemon's environment will be
873 inherited by the scripts.
874 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
876 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
877 support operations involving multiple node groups.
878 - Offline nodes are ignored when failing over an instance.
879 - Support for KVM version 1.0, which changed the version reporting format
881 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
883 - ``Makefile`` is now compatible with Automake 1.11.2
884 - Includes all bugfixes made in the 2.4 series
889 - The ganeti-htools project has been merged into the ganeti-core source
890 tree and will be built as part of Ganeti (see :doc:`install-quick`).
891 - Implemented support for :doc:`shared storage <design-shared-storage>`.
892 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
893 GPT-style partition tables (requires `parted
894 <http://www.gnu.org/s/parted/>`_).
895 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
896 - Allowed adding tags on instance creation.
897 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
898 - Allow instances to be started in a paused state, enabling the user to
899 see the complete console output on boot using the console.
900 - Added new hypervisor flag to control default reboot behaviour
901 (``reboot_behavior``).
902 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
903 - Improved out-of-band management support:
905 - Added ``gnt-node health`` command reporting the health status of
907 - Added ``gnt-node power`` command to manage power status of nodes.
908 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
910 - Instance migration can fall back to failover if instance is not
912 - Filters can be used when listing nodes, instances, groups and locks;
913 see :manpage:`ganeti(7)` manpage.
914 - Added post-execution status as variables to :doc:`hooks <hooks>`
916 - Instance tags are exported/imported together with the instance.
917 - When given an explicit job ID, ``gnt-job info`` will work for archived
919 - Jobs can define dependencies on other jobs (not yet supported via
920 RAPI or command line, but used by internal commands and usable via
923 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
926 - Instance failover is now available as a RAPI resource
927 (``/2/instances/[instance_name]/failover``).
928 - ``gnt-instance info`` defaults to static information if primary node
930 - Opcodes have a new ``comment`` attribute.
931 - Added basic SPICE support to KVM hypervisor.
932 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
934 Node group improvements
935 ~~~~~~~~~~~~~~~~~~~~~~~
937 - ``gnt-cluster verify`` has been modified to check groups separately,
938 thereby improving performance.
939 - Node group support has been added to ``gnt-cluster verify-disks``,
940 which now operates per node group.
941 - Watcher has been changed to work better with node groups.
943 - One process and state file per node group.
944 - Slow watcher in one group doesn't block other group's watcher.
946 - Added new command, ``gnt-group evacuate``, to move all instances in a
947 node group to other groups.
948 - Added ``gnt-instance change-group`` to move an instance to another
950 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
951 per-group operations.
952 - Node groups can be tagged.
953 - Some operations switch from an exclusive to a shared lock as soon as
955 - Instance's primary and secondary nodes' groups are now available as
956 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
957 and ``snodes.group.uuid``).
962 - Numerous updates to documentation and manpages.
964 - :doc:`RAPI <rapi>` documentation now has detailed parameter
966 - Some opcode/job results are now also documented, see :doc:`RAPI
969 - A lockset's internal lock is now also visible in lock monitor.
970 - Log messages from job queue workers now contain information about the
971 opcode they're processing.
972 - ``gnt-instance console`` no longer requires the instance lock.
973 - A short delay when waiting for job changes reduces the number of LUXI
974 requests significantly.
975 - DRBD metadata volumes are overwritten with zeros during disk creation.
976 - Out-of-band commands no longer acquire the cluster lock in exclusive
978 - ``devel/upload`` now uses correct permissions for directories.
984 *(Released Fri, 23 Mar 2012)*
986 This was the sixth release candidate of the 2.5 series.
992 *(Released Mon, 9 Jan 2012)*
994 This was the fifth release candidate of the 2.5 series.
1000 *(Released Thu, 27 Oct 2011)*
1002 This was the fourth release candidate of the 2.5 series.
1008 *(Released Wed, 26 Oct 2011)*
1010 This was the third release candidate of the 2.5 series.
1016 *(Released Tue, 18 Oct 2011)*
1018 This was the second release candidate of the 2.5 series.
1024 *(Released Tue, 4 Oct 2011)*
1026 This was the first release candidate of the 2.5 series.
1032 *(Released Wed, 31 Aug 2011)*
1034 This was the third beta release of the 2.5 series.
1040 *(Released Mon, 22 Aug 2011)*
1042 This was the second beta release of the 2.5 series.
1048 *(Released Fri, 12 Aug 2011)*
1050 This was the first beta release of the 2.5 series.
1056 *(Released Thu, 27 Oct 2011)*
1058 - Fixed bug when parsing command line parameter values ending in
1060 - Fixed assertion error after unclean master shutdown
1061 - Disable HTTP client pool for RPC, significantly reducing memory usage
1063 - Fixed queue archive creation with wrong permissions
1069 *(Released Tue, 23 Aug 2011)*
1073 - Fixed documentation for importing with ``--src-dir`` option
1074 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1075 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1081 *(Released Fri, 5 Aug 2011)*
1083 Many bug-fixes and a few small features:
1085 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1086 issues when you tried to add an instance with two MAC addresses in one
1088 - KVM: fixed per-instance stored UID value
1089 - KVM: configure bridged NICs at migration start
1090 - KVM: Fix a bug where instance will not start with never KVM versions
1092 - Added OS search path to ``gnt-cluster info``
1093 - Fixed an issue with ``file_storage_dir`` where you were forced to
1094 provide an absolute path, but the documentation states it is a
1095 relative path, the documentation was right
1096 - Added a new parameter to instance stop/start called ``--no-remember``
1097 that will make the state change to not be remembered
1098 - Implemented ``no_remember`` at RAPI level
1099 - Improved the documentation
1100 - Node evacuation: don't call IAllocator if node is already empty
1101 - Fixed bug in DRBD8 replace disks on current nodes
1102 - Fixed bug in recreate-disks for DRBD instances
1103 - Moved assertion checking locks in ``gnt-instance replace-disks``
1104 causing it to abort with not owning the right locks for some situation
1105 - Job queue: Fixed potential race condition when cancelling queued jobs
1106 - Fixed off-by-one bug in job serial generation
1107 - ``gnt-node volumes``: Fix instance names
1108 - Fixed aliases in bash completion
1109 - Fixed a bug in reopening log files after being sent a SIGHUP
1110 - Added a flag to burnin to allow specifying VCPU count
1111 - Bugfixes to non-root Ganeti configuration
1117 *(Released Thu, 12 May 2011)*
1119 Many bug-fixes and a few new small features:
1121 - Fixed a bug related to log opening failures
1122 - Fixed a bug in instance listing with orphan instances
1123 - Fixed a bug which prevented resetting the cluster-level node parameter
1124 ``oob_program`` to the default
1125 - Many fixes related to the ``cluster-merge`` tool
1126 - Fixed a race condition in the lock monitor, which caused failures
1127 during (at least) creation of many instances in parallel
1128 - Improved output for gnt-job info
1129 - Removed the quiet flag on some ssh calls which prevented debugging
1131 - Improved the N+1 failure messages in cluster verify by actually
1132 showing the memory values (needed and available)
1133 - Increased lock attempt timeouts so that when executing long operations
1134 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1135 too early and thus prevent the use of the 'fair' mechanism
1136 - Changed instance query data (``gnt-instance info``) to not acquire
1137 locks unless needed, thus allowing its use on locked instance if only
1138 static information is asked for
1139 - Improved behaviour with filesystems that do not support rename on an
1141 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1142 kept locks on all nodes during the wipe, which is unneeded
1143 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1144 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1145 10GiB) which caused the wipe to fail right at the end in some cases
1146 - Fixed master IP activation when doing master failover with no-voting
1147 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1148 the master node itself
1149 - Fixed potential data-loss in under disk full conditions, where Ganeti
1150 wouldn't check correctly the return code and would consider
1151 partially-written files 'correct'
1152 - Fixed bug related to multiple VGs and DRBD disk replacing
1153 - Added new disk parameter ``metavg`` that allows placement of the meta
1154 device for DRBD in a different volume group
1155 - Fixed error handling in the node daemon when the system libc doesn't
1156 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1157 - Fixed lock release during replace-disks, which kept cluster-wide locks
1158 when doing disk replaces with an iallocator script
1159 - Added check for missing bridges in cluster verify
1160 - Handle EPIPE errors while writing to the terminal better, so that
1161 piping the output to e.g. ``less`` doesn't cause a backtrace
1162 - Fixed rare case where a ^C during Luxi calls could have been
1163 interpreted as server errors, instead of simply terminating
1164 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1166 - Added a few more parameters to the KVM hypervisor, allowing a second
1167 CDROM, custom disk type for CDROMs and a floppy image
1168 - Removed redundant message in instance rename when the name is given
1170 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1171 disks on new nodes, allowing recreation when the original instance
1172 nodes are completely gone
1173 - Added option when converting disk templates to DRBD to skip waiting
1174 for the resync, in order to make the instance available sooner
1175 - Added two new variables to the OS scripts environment (containing the
1177 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1178 to allow use of ``pvgrub`` as bootloader
1179 - Changed the instance memory modifications to only check out-of-memory
1180 conditions on memory increases, and turned the secondary node warnings
1181 into errors (they can still be overridden via ``--force``)
1182 - Fixed the handling of a corner case when the Python installation gets
1183 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1184 to execute a command that doesn't exist
1185 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1186 node of the instance returned failures during instance shutdown; this
1187 adds the option ``--ignore-consistency`` to gnt-instance move
1189 And as usual, various improvements to the error messages, documentation
1196 *(Released Wed, 09 Mar 2011)*
1198 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1199 the RAPI users file if run twice (even with ``--dry-run``).
1201 The release fixes that bug (nothing else changed).
1207 *(Released Mon, 07 Mar 2011)*
1209 Final 2.4.0 release. Just a few small fixes:
1211 - Fixed RAPI node evacuate
1212 - Fixed the kvm-ifup script
1213 - Fixed internal error handling for special job cases
1214 - Updated man page to specify the escaping feature for options
1220 *(Released Mon, 28 Feb 2011)*
1222 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1223 that this feature wiped the disks of the wrong instance, leading to loss
1228 - Fixed title of query field containing instance name
1229 - Expanded the glossary in the documentation
1230 - Fixed one unittest (internal issue)
1236 *(Released Mon, 21 Feb 2011)*
1238 A number of bug fixes plus just a couple functionality changes.
1240 On the user-visible side, the ``gnt-* list`` command output has changed
1241 with respect to "special" field states. The current rc1 style of display
1242 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1243 the default output mode special fields states are displayed as follows:
1245 - Offline resource: ``*``
1246 - Unavailable/not applicable: ``-``
1247 - Data missing (RPC failure): ``?``
1248 - Unknown field: ``??``
1250 Another user-visible change is the addition of ``--force-join`` to
1255 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1256 - Fixed regression in RAPI/instance reinstall where all parameters were
1257 required (instead of optional)
1258 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1259 - Fixed iallocator usage (offline nodes were not considered offline)
1260 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1261 - Fixed hypervisor and OS parameter validation with respect to
1262 non-vm_capable nodes
1263 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1265 - Fixed ``tools/listrunner`` with respect to agent-based usage
1271 *(Released Fri, 4 Feb 2011)*
1273 Many changes and fixes since the beta1 release. While there were some
1274 internal changes, the code has been mostly stabilised for the RC
1277 Note: the dumb allocator was removed in this release, as it was not kept
1278 up-to-date with the IAllocator protocol changes. It is recommended to
1279 use the ``hail`` command from the ganeti-htools package.
1281 Note: the 2.4 and up versions of Ganeti are not compatible with the
1282 0.2.x branch of ganeti-htools. You need to upgrade to
1283 ganeti-htools-0.3.0 (or later).
1285 Regressions fixed from 2.3
1286 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1288 - Fixed the ``gnt-cluster verify-disks`` command
1289 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1291 - Fixed disk adoption breakage
1292 - Fixed wrong headers in instance listing for field aliases
1297 - Fixed corner case in KVM handling of NICs
1298 - Fixed many cases of wrong handling of non-vm_capable nodes
1299 - Fixed a bug where a missing instance symlink was not possible to
1300 recreate with any ``gnt-*`` command (now ``gnt-instance
1301 activate-disks`` does it)
1302 - Fixed the volume group name as reported by ``gnt-cluster
1304 - Increased timeouts for the import-export code, hopefully leading to
1305 fewer aborts due network or instance timeouts
1306 - Fixed bug in ``gnt-node list-storage``
1307 - Fixed bug where not all daemons were started on cluster
1308 initialisation, but only at the first watcher run
1309 - Fixed many bugs in the OOB implementation
1310 - Fixed watcher behaviour in presence of instances with offline
1312 - Fixed instance list output for instances running on the wrong node
1313 - a few fixes to the cluster-merge tool, but it still cannot merge
1314 multi-node groups (currently it is not recommended to use this tool)
1320 - Improved network configuration for the KVM hypervisor
1321 - Added e1000 as a supported NIC for Xen-HVM
1322 - Improved the lvmstrap tool to also be able to use partitions, as
1323 opposed to full disks
1324 - Improved speed of disk wiping (the cluster parameter
1325 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1326 of instance creations
1327 - Added documentation for the OS parameters
1328 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1329 hypervisor is not responding
1330 - Added display of blacklisted and hidden OS information in
1331 ``gnt-cluster info``
1332 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1333 NIC and node parameters, which might create problems with currently
1334 invalid (but undetected) configuration files, but prevents validation
1335 failures when unrelated parameters are modified
1336 - Changed cluster initialisation to wait for the master daemon to become
1338 - Expanded the RAPI interface:
1340 - Added config redistribution resource
1341 - Added activation/deactivation of instance disks
1342 - Added export of console information
1344 - Implemented log file reopening on SIGHUP, which allows using
1345 logrotate(8) for the Ganeti log files
1346 - Added a basic OOB helper script as an example
1352 *(Released Fri, 14 Jan 2011)*
1357 - Fixed timezone issues when formatting timestamps
1358 - Added support for node groups, available via ``gnt-group`` and other
1360 - Added out-of-band framework and management, see :doc:`design
1361 document <design-oob>`
1362 - Removed support for roman numbers from ``gnt-node list`` and
1363 ``gnt-instance list``.
1364 - Allowed modification of master network interface via ``gnt-cluster
1365 modify --master-netdev``
1366 - Accept offline secondaries while shutting down instance disks
1367 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1368 - Added support for multiple LVM volume groups
1369 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1371 - Added commands to list available fields:
1373 - ``gnt-node list-fields``
1374 - ``gnt-group list-fields``
1375 - ``gnt-instance list-fields``
1377 - Updated documentation and man pages
1382 - Moved ``rapi_users`` file into separate directory, now named
1383 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1385 - Added new tool for running commands on many machines,
1386 ``tools/ganeti-listrunner``
1387 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1388 improving the ``gnt-instance console`` output
1389 - Allowed customisation of disk index separator at ``configure`` time
1390 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1391 - Added support for non-partitioned md disks in ``lvmstrap``
1392 - Added script to gracefully power off KVM instances
1393 - Split ``utils`` module into smaller parts
1394 - Changed query operations to return more detailed information, e.g.
1395 whether an information is unavailable due to an offline node. To use
1396 this new functionality, the LUXI call ``Query`` must be used. Field
1397 information is now stored by the master daemon and can be retrieved
1398 using ``QueryFields``. Instances, nodes and groups can also be queried
1399 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1400 exposed via RAPI). The following commands make use of this
1401 infrastructure change:
1403 - ``gnt-group list``
1404 - ``gnt-group list-fields``
1406 - ``gnt-node list-fields``
1407 - ``gnt-instance list``
1408 - ``gnt-instance list-fields``
1409 - ``gnt-debug locks``
1414 - New RAPI resources (see :doc:`rapi`):
1418 - ``/2/groups/[group_name]``
1419 - ``/2/groups/[group_name]/assign-nodes``
1420 - ``/2/groups/[group_name]/modify``
1421 - ``/2/groups/[group_name]/rename``
1422 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1426 - Implemented ``no_install`` for instance creation
1427 - Implemented OS parameters for instance reinstallation, allowing
1428 use of special settings on reinstallation (e.g. for preserving data)
1433 - Added IPv6 support in import/export
1434 - Pause DRBD synchronization while wiping disks on instance creation
1435 - Updated unittests and QA scripts
1436 - Improved network parameters passed to KVM
1437 - Converted man pages from docbook to reStructuredText
1443 *(Released Mon, 20 Dec 2010)*
1445 Released version 2.3.1~rc1 without any changes.
1451 *(Released Wed, 1 Dec 2010)*
1453 - impexpd: Disable OpenSSL compression in socat if possible (backport
1454 from master, commit e90739d625b, see :doc:`installation guide
1455 <install-quick>` for details)
1456 - Changed unittest coverage report to exclude test scripts
1457 - Added script to check version format
1463 *(Released Wed, 1 Dec 2010)*
1465 Released version 2.3.0~rc1 without any changes.
1471 *(Released Fri, 19 Nov 2010)*
1473 A number of bugfixes and documentation updates:
1475 - Update ganeti-os-interface documentation
1476 - Fixed a bug related to duplicate MACs or similar items which should be
1478 - Fix breakage in OS state modify
1479 - Reinstall instance: disallow offline secondaries (fixes bug related to
1480 OS changing but reinstall failing)
1481 - plus all the other fixes between 2.2.1 and 2.2.2
1487 *(Released Tue, 2 Nov 2010)*
1489 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1490 - Fixed master failover race with watcher
1491 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1493 - Accept previously stopped instance for export with instance removal
1494 - Simplify and extend the environment variables for instance OS scripts
1495 - Added new node flags, ``master_capable`` and ``vm_capable``
1496 - Added optional instance disk wiping prior during allocation. This is a
1497 cluster-wide option and can be set/modified using
1498 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1499 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1500 :doc:`install-quick`
1501 - Added a new watcher option (``--ignore-pause``)
1502 - Added option to ignore offline node on instance start/stop
1503 (``--ignore-offline``)
1504 - Allow overriding OS parameters with ``gnt-instance reinstall``
1505 - Added ability to change node's secondary IP address using ``gnt-node
1507 - Implemented privilege separation for all daemons except
1508 ``ganeti-noded``, see ``configure`` options
1509 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1511 - Implemented job priorities (see ``ganeti(7)`` manpage)
1512 - Ignore failures while shutting down instances during failover from
1514 - Exit daemon's bootstrap process only once daemon is ready
1515 - Export more information via ``LUInstanceQuery``/remote API
1516 - Improved documentation, QA and unittests
1517 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1518 restart if the file was created or changed
1519 - Added LUXI protocol version sent with each request and response,
1520 allowing detection of server/client mismatches
1521 - Moved the Python scripts among gnt-* and ganeti-* into modules
1522 - Moved all code related to setting up SSH to an external script,
1524 - Infrastructure changes for node group support in future versions
1530 *(Released Fri, 19 Nov 2010)*
1532 A few small bugs fixed, and some improvements to the build system:
1534 - Fix documentation regarding conversion to drbd
1535 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1537 - Fix error handling in node modify with multiple changes
1538 - Allow remote imports without checked names
1544 *(Released Tue, 19 Oct 2010)*
1546 - Disable SSL session ID cache in RPC client
1552 *(Released Thu, 14 Oct 2010)*
1554 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1555 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1557 - Fix problems with interaction between Curl and Python's HTTP server,
1558 resulting in increased speed in many RPC calls
1559 - Improve our release script to prevent breakage with older aclocal and
1566 *(Released Thu, 7 Oct 2010)*
1568 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1569 value retrieved from master
1570 - Added support for blacklisted or hidden OS definitions
1571 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1572 - Added support for -mem-path in KVM hypervisor abstraction layer
1573 - Allow overriding instance parameters in tool for inter-cluster
1574 instance moves (``tools/move-instance``)
1575 - Improved opcode summaries (e.g. in ``gnt-job list``)
1576 - Improve consistency of OS listing by sorting it
1577 - Documentation updates
1583 *(Released Fri, 8 Oct 2010)*
1585 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1591 *(Released Mon, 4 Oct 2010)*
1593 - Fixed regression in ``gnt-instance rename``
1599 *(Released Wed, 22 Sep 2010)*
1601 - Fixed OS_VARIANT variable for OS scripts
1602 - Fixed cluster tag operations via RAPI
1603 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1604 - Disabled RAPI CA checks in watcher
1610 *(Released Mon, 23 Aug 2010)*
1612 - Support DRBD versions of the format "a.b.c.d"
1614 - Re-introduce support for usage from multiple threads in RAPI client
1615 - Instance renames and modify via RAPI
1616 - Work around race condition between processing and archival in job
1618 - Mark opcodes following failed one as failed, too
1619 - Job field ``lock_status`` was removed due to difficulties making it
1620 work with the changed job queue in Ganeti 2.2; a better way to monitor
1621 locks is expected for a later 2.2.x release
1622 - Fixed dry-run behaviour with many commands
1623 - Support ``ssh-agent`` again when adding nodes
1624 - Many additional bugfixes
1630 *(Released Fri, 30 Jul 2010)*
1632 Important change: the internal RPC mechanism between Ganeti nodes has
1633 changed from using a home-grown http library (based on the Python base
1634 libraries) to use the PycURL library. This requires that PycURL is
1635 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1636 against GnuTLS by default. cURL's support for GnuTLS had known issues
1637 before cURL 7.21.0 and we recommend using the latest cURL release or
1638 linking against OpenSSL. Most other distributions already link PycURL
1639 and cURL against OpenSSL. The command::
1641 python -c 'import pycurl; print pycurl.version'
1643 can be used to determine the libraries PycURL and cURL are linked
1646 Other significant changes:
1648 - Rewrote much of the internals of the job queue, in order to achieve
1649 better parallelism; this decouples job query operations from the job
1650 processing, and it should allow much nicer behaviour of the master
1651 daemon under load, and it also has uncovered some long-standing bugs
1652 related to the job serialisation (now fixed)
1653 - Added a default iallocator setting to the cluster parameters,
1654 eliminating the need to always pass nodes or an iallocator for
1655 operations that require selection of new node(s)
1656 - Added experimental support for the LXC virtualization method
1657 - Added support for OS parameters, which allows the installation of
1658 instances to pass parameter to OS scripts in order to customise the
1660 - Added a hypervisor parameter controlling the migration type (live or
1661 non-live), since hypervisors have various levels of reliability; this
1662 has renamed the 'live' parameter to 'mode'
1663 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1664 logical volumes, meaning that cluster verify will ignore them and not
1665 flag their presence as errors
1666 - The watcher will now reset the error count for failed instances after
1667 8 hours, thus allowing self-healing if the problem that caused the
1668 instances to be down/fail to start has cleared in the meantime
1669 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1670 check for, and warn, if the drbd module parameter ``usermode_helper``
1671 is not consistent with the cluster-wide setting; this is needed to
1672 make diagnose easier of failed drbd creations
1673 - Started adding base IPv6 support, but this is not yet
1674 enabled/available for use
1675 - Rename operations (cluster, instance) will now return the new name,
1676 which is especially useful if a short name was passed in
1677 - Added support for instance migration in RAPI
1678 - Added a tool to pre-configure nodes for the SSH setup, before joining
1679 them to the cluster; this will allow in the future a simplified model
1680 for node joining (but not yet fully enabled in 2.2); this needs the
1681 paramiko python library
1682 - Fixed handling of name-resolving errors
1683 - Fixed consistency of job results on the error path
1684 - Fixed master-failover race condition when executed multiple times in
1686 - Fixed many bugs related to the job queue (mostly introduced during the
1687 2.2 development cycle, so not all are impacting 2.1)
1688 - Fixed instance migration with missing disk symlinks
1689 - Fixed handling of unknown jobs in ``gnt-job archive``
1690 - And many other small fixes/improvements
1694 - Enhanced both the unittest and the QA coverage
1695 - Switched the opcode validation to a generic model, and extended the
1696 validation to all opcode parameters
1697 - Changed more parts of the code that write shell scripts to use the
1699 - Switched the master daemon to use the asyncore library for the Luxi
1706 *(Released Thu, 17 Jun 2010)*
1708 - Added tool (``move-instance``) and infrastructure to move instances
1709 between separate clusters (see :doc:`separate documentation
1710 <move-instance>` and :doc:`design document <design-2.2>`)
1711 - Added per-request RPC timeout
1712 - RAPI now requires a Content-Type header for requests with a body (e.g.
1713 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1714 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1715 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1717 - Implemented initial support for running Ganeti daemons as separate
1718 users, see configure-time flags ``--with-user-prefix`` and
1719 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1721 - Instances can be removed after export (``gnt-backup export
1722 --remove-instance``)
1723 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1724 key (instead of 1024 bit)
1725 - Added new cluster configuration file for cluster domain secret
1726 - Import/export now use SSL instead of SSH
1727 - Added support for showing estimated time when exporting an instance,
1728 see the ``ganeti-os-interface(7)`` manpage and look for
1735 *(Released Tue, 16 Nov 2010)*
1737 Some more bugfixes. Unless critical bugs occur, this will be the last
1740 - Fix case of MAC special-values
1741 - Fix mac checker regex
1742 - backend: Fix typo causing "out of range" error
1743 - Add missing --units in gnt-instance list man page
1749 *(Released Tue, 24 Aug 2010)*
1752 - Don't ignore secondary node silently on non-mirrored disk templates
1754 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1755 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1756 - Properly document the usb_mouse parameter
1757 - Fix path in ganeti-rapi(8) (issue 116)
1758 - Adjust error message when the ganeti user's .ssh directory is
1760 - Add same-node-check when changing the disk template to drbd
1766 *(Released Fri, 16 Jul 2010)*
1769 - Add an option to only select some reboot types during qa/burnin.
1770 (on some hypervisors consequent reboots are not supported)
1771 - Fix infrequent race condition in master failover. Sometimes the old
1772 master ip address would be still detected as up for a short time
1773 after it was removed, causing failover to fail.
1774 - Decrease mlockall warnings when the ctypes module is missing. On
1775 Python 2.4 we support running even if no ctypes module is installed,
1776 but we were too verbose about this issue.
1777 - Fix building on old distributions, on which man doesn't have a
1779 - Fix RAPI not to ignore the MAC address on instance creation
1780 - Implement the old instance creation format in the RAPI client.
1786 *(Released Thu, 01 Jul 2010)*
1788 A small bugfix release:
1789 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1790 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1791 a non-existing option
1792 - Fix instance create: the --force-variant option was ignored
1793 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1794 - Fix modify node storage with non-FQDN arguments
1795 - Fix RAPI client to authenticate under Python 2.6 when used
1796 for more than 5 requests needing authentication
1797 - Fix gnt-instance modify -t (storage) giving a wrong error message
1798 when converting a non-shutdown drbd instance to plain
1804 *(Released Fri, 18 Jun 2010)*
1806 A small bugfix release:
1808 - Fix live migration of KVM instances started with older Ganeti
1809 versions which had fewer hypervisor parameters
1810 - Fix gnt-instance grow-disk on down instances
1811 - Fix an error-reporting bug during instance migration
1812 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1813 silently ignoring broken ones
1814 - Fix an RPC error reporting bug affecting, for example, RAPI client
1816 - Fix bug triggered by different API version os-es on different nodes
1817 - Fix a bug in instance startup with custom hvparams: OS level
1818 parameters would fail to be applied.
1819 - Fix the RAPI client under Python 2.6 (but more work is needed to
1820 make it work completely well with OpenSSL)
1821 - Fix handling of errors when resolving names from DNS
1827 *(Released Thu, 3 Jun 2010)*
1829 A medium sized development cycle. Some new features, and some
1830 fixes/small improvements/cleanups.
1832 Significant features
1833 ~~~~~~~~~~~~~~~~~~~~
1835 The node deamon now tries to mlock itself into memory, unless the
1836 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1837 its logs, and falls back to console logging. This allows emergency
1838 features such as ``gnt-node powercycle`` to work even in the event of a
1839 broken node disk (tested offlining the disk hosting the node's
1840 filesystem and dropping its memory caches; don't try this at home)
1842 KVM: add vhost-net acceleration support. It can be tested with a new
1843 enough version of the kernel and of qemu-kvm.
1845 KVM: Add instance chrooting feature. If you use privilege dropping for
1846 your VMs you can also now force them to chroot to an empty directory,
1847 before starting the emulated guest.
1849 KVM: Add maximum migration bandwith and maximum downtime tweaking
1850 support (requires a new-enough version of qemu-kvm).
1852 Cluster verify will now warn if the master node doesn't have the master
1853 ip configured on it.
1855 Add a new (incompatible) instance creation request format to RAPI which
1856 supports all parameters (previously only a subset was supported, and it
1857 wasn't possible to extend the old format to accomodate all the new
1858 features. The old format is still supported, and a client can check for
1859 this feature, before using it, by checking for its presence in the
1860 ``features`` RAPI resource.
1862 Now with ancient latin support. Try it passing the ``--roman`` option to
1863 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1864 (requires the python-roman module to be installed, in order to work).
1869 As usual many internal code refactorings, documentation updates, and
1872 - Lots of improvements and cleanups to the experimental Remote API
1873 (RAPI) client library.
1874 - A new unit test suite for the core daemon libraries.
1875 - A fix to creating missing directories makes sure the umask is not
1876 applied anymore. This enforces the same directory permissions
1878 - Better handling terminating daemons with ctrl+c (used when running
1879 them in debugging mode).
1880 - Fix a race condition in live migrating a KVM instance, when stat()
1881 on the old proc status file returned EINVAL, which is an unexpected
1883 - Fixed manpage checking with newer man and utf-8 charachters. But now
1884 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1890 *(Released Fri, 7 May 2010)*
1892 Fix a bug which prevented untagged KVM instances from starting.
1898 *(Released Fri, 7 May 2010)*
1900 Another release with a long development cycle, during which many
1901 different features were added.
1903 Significant features
1904 ~~~~~~~~~~~~~~~~~~~~
1906 The KVM hypervisor now can run the individual instances as non-root, to
1907 reduce the impact of a VM being hijacked due to bugs in the
1908 hypervisor. It is possible to run all instances as a single (non-root)
1909 user, to manually specify a user for each instance, or to dynamically
1910 allocate a user out of a cluster-wide pool to each instance, with the
1911 guarantee that no two instances will run under the same user ID on any
1914 An experimental RAPI client library, that can be used standalone
1915 (without the other Ganeti libraries), is provided in the source tree as
1916 ``lib/rapi/client.py``. Note this client might change its interface in
1917 the future, as we iterate on its capabilities.
1919 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1920 replace the cluster's certificates and crypto keys. This might help in
1921 case they have been compromised, or have simply expired.
1923 A new disk option for instance creation has been added that allows one
1924 to "adopt" currently existing logical volumes, with data
1925 preservation. This should allow easier migration to Ganeti from
1926 unmanaged (or managed via other software) instances.
1928 Another disk improvement is the possibility to convert between redundant
1929 (DRBD) and plain (LVM) disk configuration for an instance. This should
1930 allow better scalability (starting with one node and growing the
1931 cluster, or shrinking a two-node cluster to one node).
1933 A new feature that could help with automated node failovers has been
1934 implemented: if a node sees itself as offline (by querying the master
1935 candidates), it will try to shutdown (hard) all instances and any active
1936 DRBD devices. This reduces the risk of duplicate instances if an
1937 external script automatically failovers the instances on such nodes. To
1938 enable this, the cluster parameter ``maintain_node_health`` should be
1939 enabled; in the future this option (per the name) will enable other
1940 automatic maintenance features.
1942 Instance export/import now will reuse the original instance
1943 specifications for all parameters; that means exporting an instance,
1944 deleting it and the importing it back should give an almost identical
1945 instance. Note that the default import behaviour has changed from
1946 before, where it created only one NIC; now it recreates the original
1949 Cluster verify has added a few new checks: SSL certificates validity,
1950 /etc/hosts consistency across the cluster, etc.
1955 As usual, many internal changes were done, documentation fixes,
1958 - Fixed cluster initialization with disabled cluster storage (regression
1959 introduced in 2.1.1)
1960 - File-based storage supports growing the disks
1961 - Fixed behaviour of node role changes
1962 - Fixed cluster verify for some corner cases, plus a general rewrite of
1963 cluster verify to allow future extension with more checks
1964 - Fixed log spamming by watcher and node daemon (regression introduced
1966 - Fixed possible validation issues when changing the list of enabled
1968 - Fixed cleanup of /etc/hosts during node removal
1969 - Fixed RAPI response for invalid methods
1970 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1971 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1972 from ide disks, etc.)
1973 - Allow OS changes without re-installation (to record a changed OS
1974 outside of Ganeti, or to allow OS renames)
1975 - Allow instance creation without OS installation (useful for example if
1976 the OS will be installed manually, or restored from a backup not in
1978 - Implemented option to make cluster ``copyfile`` use the replication
1980 - Added list of enabled hypervisors to ssconf (possibly useful for
1982 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1984 - A partial form of node re-IP is possible via node readd, which now
1985 allows changed node primary IP
1986 - Command line utilities now show an informational message if the job is
1988 - The logs of the master daemon now show the PID/UID/GID of the
1995 *(Released Fri, 12 Mar 2010)*
1997 During the 2.1.0 long release candidate cycle, a lot of improvements and
1998 changes have accumulated with were released later as 2.1.1.
2003 The node evacuate command (``gnt-node evacuate``) was significantly
2004 rewritten, and as such the IAllocator protocol was changed - a new
2005 request type has been added. This unfortunate change during a stable
2006 series is designed to improve performance of node evacuations; on
2007 clusters with more than about five nodes and which are well-balanced,
2008 evacuation should proceed in parallel for all instances of the node
2009 being evacuated. As such, any existing IAllocator scripts need to be
2010 updated, otherwise the above command will fail due to the unknown
2011 request. The provided "dumb" allocator has not been updated; but the
2012 ganeti-htools package supports the new protocol since version 0.2.4.
2014 Another important change is increased validation of node and instance
2015 names. This might create problems in special cases, if invalid host
2016 names are being used.
2018 Also, a new layer of hypervisor parameters has been added, that sits at
2019 OS level between the cluster defaults and the instance ones. This allows
2020 customisation of virtualization parameters depending on the installed
2021 OS. For example instances with OS 'X' may have a different KVM kernel
2022 (or any other parameter) than the cluster defaults. This is intended to
2023 help managing a multiple OSes on the same cluster, without manual
2024 modification of each instance's parameters.
2026 A tool for merging clusters, ``cluster-merge``, has been added in the
2027 tools sub-directory.
2032 - Improved the int/float conversions that should make the code more
2033 robust in face of errors from the node daemons
2034 - Fixed the remove node code in case of internal configuration errors
2035 - Fixed the node daemon behaviour in face of inconsistent queue
2036 directory (e.g. read-only file-system where we can't open the files
2038 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2039 now it either aborts cleanly or, if given the new "auto_promote"
2040 parameter, will automatically promote other nodes as needed
2041 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2042 completely prevent Ganeti from working
2043 - Fixed bug for instance export when not all disks were successfully
2045 - Fixed behaviour of node add when the new node is slow in starting up
2047 - Fixed handling of signals in the LUXI client, which should improve
2048 behaviour of command-line scripts
2049 - Added checks for invalid node/instance names in the configuration (now
2050 flagged during cluster verify)
2051 - Fixed watcher behaviour for disk activation errors
2052 - Fixed two potentially endless loops in http library, which led to the
2053 RAPI daemon hanging and consuming 100% CPU in some cases
2054 - Fixed bug in RAPI daemon related to hashed passwords
2055 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2057 - Enhanced compatibility with non-Debian OSes, but not using absolute
2058 path in some commands and allowing customisation of the ssh
2059 configuration directory
2060 - Fixed possible future issue with new Python versions by abiding to the
2061 proper use of ``__slots__`` attribute on classes
2062 - Added checks that should prevent directory traversal attacks
2063 - Many documentation fixes based on feedback from users
2068 - Added an "early_release" more for instance replace disks and node
2069 evacuate, where we release locks earlier and thus allow higher
2070 parallelism within the cluster
2071 - Added watcher hooks, intended to allow the watcher to restart other
2072 daemons (e.g. from the ganeti-nbma project), but they can be used of
2073 course for any other purpose
2074 - Added a compile-time disable for DRBD barriers, to increase
2075 performance if the administrator trusts the power supply or the
2076 storage system to not lose writes
2077 - Added the option of using syslog for logging instead of, or in
2078 addition to, Ganeti's own log files
2079 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2080 can indeed boot from a paravirtual NIC
2081 - Added a generic debug level for many operations; while this is not
2082 used widely yet, it allows one to pass the debug value all the way to
2084 - Enhanced the hooks environment for instance moves (failovers,
2085 migrations) where the primary/secondary nodes changed during the
2086 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2087 - Enhanced data validations for many user-supplied values; one important
2088 item is the restrictions imposed on instance and node names, which
2089 might reject some (invalid) host names
2090 - Add a configure-time option to disable file-based storage, if it's not
2091 needed; this allows greater security separation between the master
2092 node and the other nodes from the point of view of the inter-node RPC
2094 - Added user notification in interactive tools if job is waiting in the
2095 job queue or trying to acquire locks
2096 - Added log messages when a job is waiting for locks
2097 - Added filtering by node tags in instance operations which admit
2098 multiple instances (start, stop, reboot, reinstall)
2099 - Added a new tool for cluster mergers, ``cluster-merge``
2100 - Parameters from command line which are of the form ``a=b,c=d`` can now
2101 use backslash escapes to pass in values which contain commas,
2102 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2104 - For KVM, the instance name is the first parameter passed to KVM, so
2105 that it's more visible in the process list
2111 *(Released Tue, 2 Mar 2010)*
2113 Ganeti 2.1 brings many improvements with it. Major changes:
2115 - Added infrastructure to ease automated disk repairs
2116 - Added new daemon to export configuration data in a cheaper way than
2117 using the remote API
2118 - Instance NICs can now be routed instead of being associated with a
2120 - Improved job locking logic to reduce impact of jobs acquiring multiple
2121 locks waiting for other long-running jobs
2123 In-depth implementation details can be found in the Ganeti 2.1 design
2129 - Added chroot hypervisor
2130 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2132 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2133 ``bootloader_path`` and ``bootloader_args``)
2134 - Added the ``use_localtime`` option for the xen-hvm and kvm
2135 hypervisors, and the default value for this has changed to false (in
2136 2.0 xen-hvm always enabled it)
2137 - Added luxi call to submit multiple jobs in one go
2138 - Added cluster initialization option to not modify ``/etc/hosts``
2140 - Added network interface parameters
2141 - Added dry run mode to some LUs
2142 - Added RAPI resources:
2144 - ``/2/instances/[instance_name]/info``
2145 - ``/2/instances/[instance_name]/replace-disks``
2146 - ``/2/nodes/[node_name]/evacuate``
2147 - ``/2/nodes/[node_name]/migrate``
2148 - ``/2/nodes/[node_name]/role``
2149 - ``/2/nodes/[node_name]/storage``
2150 - ``/2/nodes/[node_name]/storage/modify``
2151 - ``/2/nodes/[node_name]/storage/repair``
2153 - Added OpCodes to evacuate or migrate all instances on a node
2154 - Added new command to list storage elements on nodes (``gnt-node
2155 list-storage``) and modify them (``gnt-node modify-storage``)
2156 - Added new ssconf files with master candidate IP address
2157 (``ssconf_master_candidates_ips``), node primary IP address
2158 (``ssconf_node_primary_ips``) and node secondary IP address
2159 (``ssconf_node_secondary_ips``)
2160 - Added ``ganeti-confd`` and a client library to query the Ganeti
2161 configuration via UDP
2162 - Added ability to run hooks after cluster initialization and before
2164 - Added automatic mode for disk replace (``gnt-instance replace-disks
2166 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2167 after catastrophic data-loss
2168 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2170 - Added ``gnt-instance move`` command to move instances
2171 - Added ``gnt-cluster watcher`` command to control watcher
2172 - Added ``gnt-node powercycle`` command to powercycle nodes
2173 - Added new job status field ``lock_status``
2174 - Added parseable error codes to cluster verification (``gnt-cluster
2175 verify --error-codes``) and made output less verbose (use
2176 ``--verbose`` to restore previous behaviour)
2177 - Added UUIDs to the main config entities (cluster, nodes, instances)
2178 - Added support for OS variants
2179 - Added support for hashed passwords in the Ganeti remote API users file
2181 - Added option to specify maximum timeout on instance shutdown
2182 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2183 - Added new helper script to start and stop Ganeti daemons
2184 (``daemon-util``), with the intent to reduce the work necessary to
2185 adjust Ganeti for non-Debian distributions and to start/stop daemons
2187 - Added more unittests
2188 - Fixed critical bug in ganeti-masterd startup
2189 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2190 now customisable at the cluster level for both the KVM and Xen
2191 hypervisors using the new ``migration_port`` parameter
2192 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2193 reinstalling an instance
2194 - Allowed ``@`` in tag names
2195 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2196 - Many documentation updates
2197 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2198 - ``gnt-instance reinstall`` can now reinstall multiple instances
2199 - Updated many command line parameters
2200 - Introduced new OS API version 15
2201 - No longer support a default hypervisor
2202 - Treat virtual LVs as inexistent
2203 - Improved job locking logic to reduce lock contention
2204 - Match instance and node names case insensitively
2205 - Reimplemented bash completion script to be more complete
2212 *(Released Thu, 4 Feb 2010)*
2214 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2215 - Fix a string formatting bug
2216 - Improve safety of the code in some error paths
2217 - Improve data validation in the master of values returned from nodes
2223 *(Released Thu, 17 Dec 2009)*
2225 - Fix security issue due to missing validation of iallocator names; this
2226 allows local and remote execution of arbitrary executables
2227 - Fix failure of gnt-node list during instance removal
2228 - Ship the RAPI documentation in the archive
2234 *(Released Wed, 30 Sep 2009)*
2236 - Fixed many wrong messages
2237 - Fixed a few bugs related to the locking library
2238 - Fixed MAC checking at instance creation time
2239 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2240 - Fixed a few issues related to signal handling in both daemons and
2242 - Fixed the example startup script provided
2243 - Fixed insserv dependencies in the example startup script (patch from
2245 - Fixed handling of drained nodes in the iallocator framework
2246 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2248 - Fixed error related to invalid job IDs in job polling
2249 - Fixed job/opcode persistence on unclean master shutdown
2250 - Fixed handling of partial job processing after unclean master
2252 - Fixed error reporting from LUs, previously all errors were converted
2253 into execution errors
2254 - Fixed error reporting from burnin
2255 - Decreased significantly the memory usage of the job queue
2256 - Optimised slightly multi-job submission
2257 - Optimised slightly opcode loading
2258 - Backported the multi-job submit framework from the development
2259 branch; multi-instance start and stop should be faster
2260 - Added script to clean archived jobs after 21 days; this will reduce
2261 the size of the queue directory
2262 - Added some extra checks in disk size tracking
2263 - Added an example ethers hook script
2264 - Added a cluster parameter that prevents Ganeti from modifying of
2266 - Added more node information to RAPI responses
2267 - Added a ``gnt-job watch`` command that allows following the ouput of a
2269 - Added a bind-address option to ganeti-rapi
2270 - Added more checks to the configuration verify
2271 - Enhanced the burnin script such that some operations can be retried
2273 - Converted instance reinstall to multi-instance model
2279 *(Released Fri, 7 Aug 2009)*
2281 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2282 to allow using the pre-2.0.2 behaviour in activation, if any existing
2283 instances have mismatched disk sizes in the configuration
2284 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2285 any configuration mismatches for disk sizes
2286 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2287 failover to work on two-node clusters
2288 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2290 - Fixed detection of OS script errors in ``gnt-backup export``
2291 - Fixed exit code of ``gnt-backup export``
2297 *(Released Fri, 17 Jul 2009)*
2299 - Added experimental support for stripped logical volumes; this should
2300 enhance performance but comes with a higher complexity in the block
2301 device handling; stripping is only enabled when passing
2302 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2303 affected even in the non-stripped mode
2304 - Improved resiliency against transient failures at the end of DRBD
2305 resyncs, and in general of DRBD resync checks
2306 - Fixed a couple of issues with exports and snapshot errors
2307 - Fixed a couple of issues in instance listing
2308 - Added display of the disk size in ``gnt-instance info``
2309 - Fixed checking for valid OSes in instance creation
2310 - Fixed handling of the "vcpus" parameter in instance listing and in
2311 general of invalid parameters
2312 - Fixed http server library, and thus RAPI, to handle invalid
2313 username/password combinations correctly; this means that now they
2314 report unauthorized for queries too, not only for modifications,
2315 allowing earlier detect of configuration problems
2316 - Added a new "role" node list field, equivalent to the master/master
2317 candidate/drained/offline flags combinations
2318 - Fixed cluster modify and changes of candidate pool size
2319 - Fixed cluster verify error messages for wrong files on regular nodes
2320 - Fixed a couple of issues with node demotion from master candidate role
2321 - Fixed node readd issues
2322 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2323 - Added a new ``--no-voting`` option for masterfailover to fix failover
2324 on two-nodes clusters when the former master node is unreachable
2325 - Added instance reinstall over RAPI
2331 *(Released Tue, 16 Jun 2009)*
2333 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2334 allow re-adding the start in single-user option (regression from 1.2)
2335 - the watcher writes the instance status to a file, to allow monitoring
2336 to report the instance status (from the master) based on cached
2337 results of the watcher's queries; while this can get stale if the
2338 watcher is being locked due to other work on the cluster, this is
2339 still an improvement
2340 - the watcher now also restarts the node daemon and the rapi daemon if
2342 - fixed the watcher to handle full and drained queue cases
2343 - hooks export more instance data in the environment, which helps if
2344 hook scripts need to take action based on the instance's properties
2345 (no longer need to query back into ganeti)
2346 - instance failovers when the instance is stopped do not check for free
2347 RAM, so that failing over a stopped instance is possible in low memory
2349 - rapi uses queries for tags instead of jobs (for less job traffic), and
2350 for cluster tags it won't talk to masterd at all but read them from
2352 - a couple of error handling fixes in RAPI
2353 - drbd handling: improved the error handling of inconsistent disks after
2354 resync to reduce the frequency of "there are some degraded disks for
2355 this instance" messages
2356 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2357 error handling path called a wrong function name)
2363 *(Released Wed, 27 May 2009)*
2365 - no changes from rc5
2371 *(Released Wed, 20 May 2009)*
2373 - fix a couple of bugs (validation, argument checks)
2374 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2375 - some small improvements to RAPI and IAllocator
2376 - make watcher automatically start the master daemon if down
2382 *(Released Mon, 27 Apr 2009)*
2384 - change the OS list to not require locks; this helps with big clusters
2385 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2386 volume group is broken
2387 - ``gnt-instance info``, without any arguments, doesn't run for all
2388 instances anymore; either pass ``--all`` or pass the desired
2389 instances; this helps against mistakes on big clusters where listing
2390 the information for all instances takes a long time
2391 - miscellaneous doc and man pages fixes
2397 *(Released Wed, 8 Apr 2009)*
2399 - Change the internal locking model of some ``gnt-node`` commands, in
2400 order to reduce contention (and blocking of master daemon) when
2401 batching many creation/reinstall jobs
2402 - Fixes to Xen soft reboot
2403 - No longer build documentation at build time, instead distribute it in
2404 the archive, in order to reduce the need for the whole docbook/rst
2411 *(Released Fri, 27 Mar 2009)*
2413 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2414 - Fix watcher startup sequence, improves the behaviour of busy clusters
2415 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2416 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2417 bind address and other places
2418 - Some documentation fixes and updates
2424 *(Released Mon, 2 Mar 2009)*
2426 - More documentation updates, now all docs should be more-or-less
2428 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2430 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2432 - Fix an issue related to $libdir/run/ganeti and cluster creation
2438 *(Released Thu, 19 Feb 2009)*
2440 - Xen PVM and KVM have switched the default value for the instance root
2441 disk to the first partition on the first drive, instead of the whole
2442 drive; this means that the OS installation scripts must be changed
2444 - Man pages have been updated
2445 - RAPI has been switched by default to HTTPS, and the exported functions
2446 should all work correctly
2447 - RAPI v1 has been removed
2448 - Many improvements to the KVM hypervisor
2449 - Block device errors are now better reported
2450 - Many other bugfixes and small improvements
2456 *(Released Mon, 26 Jan 2009)*
2458 - Version 2 is a general rewrite of the code and therefore the
2459 differences are too many to list, see the design document for 2.0 in
2460 the ``doc/`` subdirectory for more details
2461 - In this beta version there is not yet a migration path from 1.2 (there
2462 will be one in the final 2.0 release)
2463 - A few significant changes are:
2465 - all commands are executed by a daemon (``ganeti-masterd``) and the
2466 various ``gnt-*`` commands are just front-ends to it
2467 - all the commands are entered into, and executed from a job queue,
2468 see the ``gnt-job(8)`` manpage
2469 - the RAPI daemon supports read-write operations, secured by basic
2470 HTTP authentication on top of HTTPS
2471 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2472 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2473 to migrate to DRBD 8 first while still running Ganeti 1.2)
2474 - DRBD devices are using statically allocated minor numbers, which
2475 will be assigned to existing instances during the migration process
2476 - there is support for both Xen PVM and Xen HVM instances running on
2478 - KVM virtualization is supported too
2479 - file-based storage has been implemented, which means that it is
2480 possible to run the cluster without LVM and DRBD storage, for
2481 example using a shared filesystem exported from shared storage (and
2482 still have live migration)
2488 *(Released Tue, 13 Jan 2009)*
2490 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2491 - Reuse the old instance mac address by default on instance import, if
2492 the instance name is the same.
2493 - Handle situations in which the node info rpc returns incomplete
2495 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2496 - Improved version of batcher:
2498 - state file support
2499 - instance mac address support
2500 - support for HVM clusters/instances
2502 - Add an option to show the number of cpu sockets and nodes in
2504 - Support OSes that handle more than one version of the OS api (but do
2505 not change the current API in any other way)
2506 - Fix ``gnt-node migrate``
2507 - ``gnt-debug`` man page
2508 - Fixes various more typos and small issues
2509 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2515 *(Released Wed, 24 Sep 2008)*
2517 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2518 type of disk exported to fully virtualized instances.
2519 - provide access to the serial console of HVM instances
2520 - instance auto_balance flag, set by default. If turned off it will
2521 avoid warnings on cluster verify if there is not enough memory to fail
2522 over an instance. in the future it will prevent automatically failing
2523 it over when we will support that.
2524 - batcher tool for instance creation, see ``tools/README.batcher``
2525 - ``gnt-instance reinstall --select-os`` to interactively select a new
2526 operating system when reinstalling an instance.
2527 - when changing the memory amount on instance modify a check has been
2528 added that the instance will be able to start. also warnings are
2529 emitted if the instance will not be able to fail over, if auto_balance
2531 - documentation fixes
2532 - sync fields between ``gnt-instance list/modify/add/import``
2533 - fix a race condition in drbd when the sync speed was set after giving
2534 the device a remote peer.
2540 *(Released Tue, 22 Jul 2008)*
2542 - note: the allowed size and number of tags per object were reduced
2543 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2544 - fixed twisted 8.x compatibility
2545 - fixed ``gnt-instance replace-disks`` with iallocator
2546 - add TCP keepalives on twisted connections to detect restarted nodes
2547 - disk increase support, see ``gnt-instance grow-disk``
2548 - implement bulk node/instance query for RAPI
2549 - add tags in node/instance listing (optional)
2550 - experimental migration (and live migration) support, read the man page
2551 for ``gnt-instance migrate``
2552 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2553 has some small improvements in handling its state file
2559 *(Released Fri, 13 Jun 2008)*
2561 - Experimental readonly, REST-based remote API implementation;
2562 automatically started on master node, TCP port 5080, if enabled by
2563 ``--enable-rapi`` parameter to configure script.
2564 - Instance allocator support. Add and import instance accept a
2565 ``--iallocator`` parameter, and call that instance allocator to decide
2566 which node to use for the instance. The iallocator document describes
2567 what's expected from an allocator script.
2568 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2569 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2570 node is lost there is still enough memory to fail over the instances
2572 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2573 ``gnt-cluster verify``, to check for site-specific compliance. All the
2574 hooks will run, and their output, if any, will be displayed. Any
2575 failing hook will make the verification return an error value.
2576 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2577 primary and secondary interfaces
2578 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2579 that is still declared as part of the cluster and has failed.
2580 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2581 specifying output fields, that just adds the chosen fields to the
2583 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2584 export from the filesystem.
2585 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2586 have been added. Using them you can enable/disable acpi and pae
2587 support, and specify a path for a cd image to be exported to the
2588 instance. These parameters as the name suggest only work on HVM
2590 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2591 PAE support will be set to the previously hardcoded values, but the
2592 (previously hardcoded) path to the CDROM ISO image will be unset and
2593 if required, needs to be set manually with ``gnt-instance modify``
2595 - The address to which an instance's VNC console is bound is now
2596 selectable per-instance, rather than being cluster wide. Of course
2597 this only applies to instances controlled via VNC, so currently just
2598 applies to HVM clusters.
2604 *(Released Mon, 18 Feb 2008)*
2606 - more tweaks to the disk activation code (especially helpful for DRBD)
2607 - change the default ``gnt-instance list`` output format, now there is
2608 one combined status field (see the manpage for the exact values this
2610 - some more fixes for the mac export to hooks change
2611 - make Ganeti not break with DRBD 8.2.x (which changed the version
2612 format in ``/proc/drbd``) (issue 24)
2613 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2614 template, allowing migration from DRBD0.7+MD to DRBD8
2620 *(Released Wed, 30 Jan 2008)*
2622 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2624 - add command aliases infrastructure and a few aliases
2625 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2626 man pages and the ``--help`` option of ``gnt-node
2627 list``/``gnt-instance list``
2628 - fix ``gnt-backup list`` with down nodes (issue 21)
2629 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2630 - fix the dist archive and add a check for including svn/git files in
2632 - some developer-related changes: improve the burnin and the QA suite,
2633 add an upload script for testing during development
2639 *(Released Wed, 16 Jan 2008)*
2641 - experimental HVM support, read the install document, section
2642 "Initializing the cluster"
2643 - allow for the PVM hypervisor per-instance kernel and initrd paths
2644 - add a new command ``gnt-cluster verify-disks`` which uses a new
2645 algorithm to improve the reconnection of the DRBD pairs if the device
2646 on the secondary node has gone away
2647 - make logical volume code auto-activate LVs at disk activation time
2648 - slightly improve the speed of activating disks
2649 - allow specification of the MAC address at instance creation time, and
2650 changing it later via ``gnt-instance modify``
2651 - fix handling of external commands that generate lots of output on
2653 - update documentation with regard to minimum version of DRBD8 supported
2659 *(Released Tue, 4 Dec 2007)*
2661 - Log the ``xm create`` output to the node daemon log on failure (to
2662 help diagnosing the error)
2663 - In debug mode, log all external commands output if failed to the logs
2664 - Change parsing of lvm commands to ignore stderr
2670 *(Released Wed, 28 Nov 2007)*
2672 - Another round of updates to the DRBD 8 code to deal with more failures
2673 in the replace secondary node operation
2674 - Some more logging of failures in disk operations (lvm, drbd)
2675 - A few documentation updates
2682 *(Released Tue, 13 Nov 2007)*
2684 - Change configuration file format from Python's Pickle to JSON.
2685 Upgrading is possible using the cfgupgrade utility.
2686 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2687 faster replace disks and is more stable (DRBD 8 has many improvements
2688 compared to DRBD 0.7)
2689 - Added command line tags support (see man pages for ``gnt-instance``,
2690 ``gnt-node``, ``gnt-cluster``)
2691 - Added instance rename support
2692 - Added multi-instance startup/shutdown
2693 - Added cluster rename support
2694 - Added ``gnt-node evacuate`` to simplify some node operations
2695 - Added instance reboot operation that can speedup reboot as compared to
2697 - Soften the requirement that hostnames are in FQDN format
2698 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2700 - Removed dependency on debian's patched fping that uses the
2701 non-standard ``-S`` option
2702 - Now the OS definitions are searched for in multiple, configurable
2703 paths (easier for distros to package)
2704 - Some changes to the hooks infrastructure (especially the new
2705 post-configuration update hook)
2706 - Other small bugfixes
2708 .. vim: set textwidth=72 syntax=rst :