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.
16 - hroller now also plans for capacity to move non-redundant instances off
17 any node to be rebooted; the old behavior of completely ignoring any
18 non-redundant instances can be restored by adding the --ignore-non-redundant
27 - The minimum Python version needed to run Ganeti is now 2.6.
28 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
29 to always require authentication, ``--require-authentication``. It can
30 be specified in ``$sysconfdir/default/ganeti``.
31 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
32 be used to manage the disk templates to be used by instances in the cluster.
33 Initially, it will be set to a list that includes plain, drbd, if they were
34 enabled by specifying a volume group name, and file and sharedfile, if those
35 were enabled at configure time. Additionally, it will include all disk
36 templates that are currently used by instances. The order of disk templates
37 will be based on Ganeti's history of supporting them. In the future, the
38 first entry of the list will be used as a default disk template on instance
40 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
41 configuration back to the previous stable version.
42 - The cluster option '--no-lvm-storage' was removed in favor of the new option
43 '--enabled-disk-templates'.
49 *(Released Fri, 24 May 2013)*
51 Incompatible/important changes
52 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
54 - Instance policies for disk size were documented to be on a per-disk
55 basis, but hail applied them to the sum of all disks. This has been
57 - ``hbal`` will now exit with status 0 if, during job execution over
58 LUXI, early exit has been requested and all jobs are successful;
59 before, exit status 1 was used, which cannot be differentiated from
61 - Compatibility with newer versions of rbd has been fixed
62 - ``gnt-instance batch-create`` has been changed to use the bulk create
63 opcode from Ganeti. This lead to incompatible changes in the format of
64 the JSON file. It's now not a custom dict anymore but a dict
65 compatible with the ``OpInstanceCreate`` opcode.
66 - Parent directories for file storage need to be listed in
67 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
68 write the file automatically based on old configuration values, but it
69 can not distribute it across all nodes and the file contents should be
70 verified. Use ``gnt-cluster copyfile
71 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
72 upgraded. The reason for requiring this list of paths now is that
73 before it would have been possible to inject new paths via RPC,
74 allowing files to be created in arbitrary locations. The RPC protocol
75 is protected using SSL/X.509 certificates, but as a design principle
76 Ganeti does not permit arbitrary paths to be passed.
77 - The parsing of the variants file for OSes (see
78 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
79 lines and comment lines (starting with ``#``) are ignored for better
81 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
82 longer available. ``gnt-node add`` now invokes a new tool on the
83 destination node, named ``prepare-node-join``, to configure the SSH
84 daemon. Paramiko is no longer necessary to configure nodes' SSH
85 daemons via ``gnt-node add``.
86 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
87 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
88 the flag is not reset after a master failover.
89 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
90 is recommended. 2.6 will be mandatory from the 2.8 series.
96 - New network management functionality to support automatic allocation
97 of IP addresses and managing of network parameters. See
98 :manpage:`gnt-network(8)` for more details.
99 - New external storage backend, to allow managing arbitrary storage
100 systems external to the cluster. See
101 :manpage:`ganeti-extstorage-interface(7)`.
102 - New ``exclusive-storage`` node parameter added, restricted to
103 nodegroup level. When it's set to true, physical disks are assigned in
104 an exclusive fashion to instances, as documented in :doc:`Partitioned
105 Ganeti <design-partitioned>`. Currently, only instances using the
106 ``plain`` disk template are supported.
107 - The KVM hypervisor has been updated with many new hypervisor
108 parameters, including a generic one for passing arbitrary command line
109 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
110 compatible up to qemu 1.4.
111 - A new tool, called ``mon-collector``, is the stand-alone executor of
112 the data collectors for a monitoring system. As of this version, it
113 just includes the DRBD data collector, that can be executed by calling
114 ``mon-collector`` using the ``drbd`` parameter. See
115 :manpage:`mon-collector(7)`.
116 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
117 for RAPI users. It allows granting permissions to query for
118 information to a specific user without giving
119 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
120 - A new tool named ``node-cleanup`` has been added. It cleans remains of
121 a cluster from a machine by stopping all daemons, removing
122 certificates and ssconf files. Unless the ``--no-backup`` option is
123 given, copies of the certificates are made.
124 - Instance creations now support the use of opportunistic locking,
125 potentially speeding up the (parallel) creation of multiple instances.
126 This feature is currently only available via the :doc:`RAPI
127 <rapi>` interface and when an instance allocator is used. If the
128 ``opportunistic_locking`` parameter is set the opcode will try to
129 acquire as many locks as possible, but will not wait for any locks
130 held by other opcodes. If not enough resources can be found to
131 allocate the instance, the temporary error code
132 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
133 retried thereafter, with or without opportunistic locking.
134 - New experimental linux-ha resource scripts.
135 - Restricted-commands support: ganeti can now be asked (via command line
136 or rapi) to perform commands on a node. These are passed via ganeti
137 RPC rather than ssh. This functionality is restricted to commands
138 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
139 reasons. The file is not copied automatically.
145 - Diskless instances are now externally mirrored (Issue 237). This for
146 now has only been tested in conjunction with explicit target nodes for
148 - Queries not needing locks or RPC access to the node can now be
149 performed by the confd daemon, making them independent from jobs, and
150 thus faster to execute. This is selectable at configure time.
151 - The functionality for allocating multiple instances at once has been
152 overhauled and is now also available through :doc:`RAPI <rapi>`.
156 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
158 - Disks added through ``gnt-instance modify`` or created through
159 ``gnt-instance recreate-disks`` are wiped, if the
160 ``prealloc_wipe_disks`` flag is set.
161 - If wiping newly created disks fails, the disks are removed. Also,
162 partial failures in creating disks through ``gnt-instance modify``
163 triggers a cleanup of the partially-created disks.
164 - Removing the master IP address doesn't fail if the address has been
166 - Fix ownership of the OS log dir
167 - Workaround missing SO_PEERCRED constant (Issue 191)
173 *(Released Fri, 3 May 2013)*
175 This was the first release candidate of the 2.7 series. Since beta3:
177 - Fix kvm compatibility with qemu 1.4 (Issue 389)
178 - Documentation updates (admin guide, upgrade notes, install
179 instructions) (Issue 372)
180 - Fix gnt-group list nodes and instances count (Issue 436)
181 - Fix compilation without non-mandatory libraries (Issue 441)
182 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
183 - Make confd logging more verbose at INFO level (Issue 435)
184 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
185 - Fix failure path for instance storage type conversion (Issue 229)
186 - Update htools text backend documentation
187 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
188 - Disable inter-cluster instance move for file-based instances, because
189 it is dependant on instance export, which is not supported for
190 file-based instances. (Issue 414)
191 - Fix gnt-job crashes on non-ascii characters (Issue 427)
192 - Fix volume group checks on non-vm-capable nodes (Issue 432)
198 *(Released Mon, 22 Apr 2013)*
200 This was the third beta release of the 2.7 series. Since beta2:
202 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
203 - Fix data loss on wrong usage of ``gnt-instance move``
204 - Properly export errors in confd-based job queries
205 - Add ``users-setup`` tool
206 - Fix iallocator protocol to report 0 as a disk size for diskless
207 instances. This avoids hail breaking when a diskless instance is
209 - Fix job queue directory permission problem that made confd job queries
210 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
211 for access to archived jobs (Issue 406).
212 - Limit the sizes of networks supported by ``gnt-network`` to something
213 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
214 - Fix bugs in instance disk template conversion
215 - Fix GHC 7 compatibility
216 - Fix ``burnin`` install path (Issue 426).
217 - Allow very small disk grows (Issue 347).
218 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
219 that noded doesn't import masterd code (Issue 419).
220 - Make sure the default metavg at cluster init is the same as the vg, if
221 unspecified (Issue 358).
222 - Fix cleanup of partially created disks (part of Issue 416)
228 *(Released Tue, 2 Apr 2013)*
230 This was the second beta release of the 2.7 series. Since beta1:
232 - Networks no longer have a "type" slot, since this information was
233 unused in Ganeti: instead of it tags should be used.
234 - The rapi client now has a ``target_node`` option to MigrateInstance.
235 - Fix early exit return code for hbal (Issue 386).
236 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
237 - Fix ``rbd showmapped`` output parsing (Issue 312).
238 - Networks are now referenced indexed by UUID, rather than name. This
239 will require running cfgupgrade, from 2.7.0beta1, if networks are in
241 - The OS environment now includes network information.
242 - Deleting of a network is now disallowed if any instance nic is using
243 it, to prevent dangling references.
244 - External storage is now documented in man pages.
245 - The exclusive_storage flag can now only be set at nodegroup level.
246 - Hbal can now submit an explicit priority with its jobs.
247 - Many network related locking fixes.
248 - Bump up the required pylint version to 0.25.1.
249 - Fix the ``no_remember`` option in RAPI client.
250 - Many ipolicy related tests, qa, and fixes.
251 - Many documentation improvements and fixes.
252 - Fix building with ``--disable-file-storage``.
253 - Fix ``-q`` option in htools, which was broken if passed more than
255 - Some haskell/python interaction improvements and fixes.
256 - Fix iallocator in case of missing LVM storage.
257 - Fix confd config load in case of ``--no-lvm-storage``.
258 - The confd/query functionality is now mentioned in the security
265 *(Released Wed, 6 Feb 2013)*
267 This was the first beta release of the 2.7 series. All important changes
268 are listed in the latest 2.7 entry.
274 *(Released Fri, 21 Dec 2012)*
276 Important behaviour change: hbal won't rebalance anymore instances which
277 have the ``auto_balance`` attribute set to false. This was the intention
278 all along, but until now it only skipped those from the N+1 memory
279 reservation (DRBD-specific).
281 A significant number of bug fixes in this release:
283 - Fixed disk adoption interaction with ipolicy checks.
284 - Fixed networking issues when instances are started, stopped or
285 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
286 - Fixed the warning in cluster verify for shared storage instances not
288 - Fixed removal of storage directory on shared file storage (issue 262).
289 - Fixed validation of LVM volume group name in OpClusterSetParams
290 (``gnt-cluster modify``) (issue 285).
291 - Fixed runtime memory increases (``gnt-instance modify -m``).
292 - Fixed live migration under Xen's ``xl`` mode.
293 - Fixed ``gnt-instance console`` with ``xl``.
294 - Fixed building with newer Haskell compiler/libraries.
295 - Fixed PID file writing in Haskell daemons (confd); this prevents
296 restart issues if confd was launched manually (outside of
297 ``daemon-util``) while another copy of it was running
298 - Fixed a type error when doing live migrations with KVM (issue 297) and
299 the error messages for failing migrations have been improved.
300 - Fixed opcode validation for the out-of-band commands (``gnt-node
302 - Fixed a type error when unsetting OS hypervisor parameters (issue
303 311); now it's possible to unset all OS-specific hypervisor
305 - Fixed the ``dry-run`` mode for many operations: verification of
306 results was over-zealous but didn't take into account the ``dry-run``
307 operation, resulting in "wrong" failures.
308 - Fixed bash completion in ``gnt-job list`` when the job queue has
309 hundreds of entries; especially with older ``bash`` versions, this
310 results in significant CPU usage.
312 And lastly, a few other improvements have been made:
314 - Added option to force master-failover without voting (issue 282).
315 - Clarified error message on lock conflict (issue 287).
316 - Logging of newly submitted jobs has been improved (issue 290).
317 - Hostname checks have been made uniform between instance rename and
319 - The ``--submit`` option is now supported by ``gnt-debug delay``.
320 - Shutting down the master daemon by sending SIGTERM now stops it from
321 processing jobs waiting for locks; instead, those jobs will be started
322 once again after the master daemon is started the next time (issue
324 - Support for Xen's ``xl`` program has been improved (besides the fixes
326 - Reduced logging noise in the Haskell confd daemon (only show one log
327 entry for each config reload, instead of two).
328 - Several man page updates and typo fixes.
334 *(Released Fri, 12 Oct 2012)*
336 A small bugfix release. Among the bugs fixed:
338 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
339 made the command unusable.
340 - Commands that issue many jobs don't fail anymore just because some jobs
341 take so long that other jobs are archived.
342 - Failures during ``gnt-instance reinstall`` are reflected by the exit
344 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
346 - When ``always_failover`` is set, ``--allow-failover`` is not required
347 in migrate commands anymore.
348 - ``bash_completion`` works even if extglob is disabled.
349 - Fixed bug with locks that made failover for RDB-based instances fail.
350 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
351 a random node instead of one based on the allocator metric.
352 - Support for newer versions of pylint and pep8.
353 - Hail doesn't fail anymore when trying to add an instance of type
354 ``file``, ``sharedfile`` or ``rbd``.
355 - Added new Makefile target to rebuild the whole distribution, so that
356 all files are included.
362 *(Released Fri, 27 Jul 2012)*
365 .. attention:: The ``LUXI`` protocol has been made more consistent
366 regarding its handling of command arguments. This, however, leads to
367 incompatibility issues with previous versions. Please ensure that you
368 restart Ganeti daemons soon after the upgrade, otherwise most
369 ``LUXI`` calls (job submission, setting/resetting the drain flag,
370 pausing/resuming the watcher, cancelling and archiving jobs, querying
371 the cluster configuration) will fail.
380 The current ``admin_up`` field, which used to denote whether an instance
381 should be running or not, has been removed. Instead, ``admin_state`` is
382 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
384 The rational behind this is that an instance being “down” can have
387 - it could be down during a reboot
388 - it could be temporarily be down for a reinstall
389 - or it could be down because it is deprecated and kept just for its
392 The previous Boolean state was making it difficult to do capacity
393 calculations: should Ganeti reserve memory for a down instance? Now, the
394 tri-state field makes it clear:
396 - in ``up`` and ``down`` state, all resources are reserved for the
397 instance, and it can be at any time brought up if it is down
398 - in ``offline`` state, only disk space is reserved for it, but not
401 The field can have an extra use: since the transition between ``up`` and
402 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
403 transition between ``offline`` and ``down`` is done via ``gnt-instance
404 modify``, it is possible to given different rights to users. For
405 example, owners of an instance could be allowed to start/stop it, but
406 not transition it out of the offline state.
408 Instance policies and specs
409 +++++++++++++++++++++++++++
411 In previous Ganeti versions, an instance creation request was not
412 limited on the minimum size and on the maximum size just by the cluster
413 resources. As such, any policy could be implemented only in third-party
414 clients (RAPI clients, or shell wrappers over ``gnt-*``
415 tools). Furthermore, calculating cluster capacity via ``hspace`` again
416 required external input with regards to instance sizes.
418 In order to improve these workflows and to allow for example better
419 per-node group differentiation, we introduced instance specs, which
422 - minimum instance disk size, disk count, memory size, cpu count
423 - maximum values for the above metrics
424 - and “standard” values (used in ``hspace`` to calculate the standard
427 The minimum/maximum values can be also customised at node-group level,
428 for example allowing more powerful hardware to support bigger instance
431 Beside the instance specs, there are a few other settings belonging to
432 the instance policy framework. It is possible now to customise, per
433 cluster and node-group:
435 - the list of allowed disk templates
436 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
437 - the maximum ratio of instance to spindles (see below for more
438 information) for local storage
440 All these together should allow all tools that talk to Ganeti to know
441 what are the ranges of allowed values for instances and the
442 over-subscription that is allowed.
444 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
445 instance configuration, and the physical CPU configuration from the
446 node. For the spindle ratios however, we didn't track before these
447 values, so new parameters have been added:
449 - a new node parameter ``spindle_count``, defaults to 1, customisable at
450 node group or node level
451 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
453 Note that spindles in this context doesn't need to mean actual
454 mechanical hard-drives; it's just a relative number for both the node
455 I/O capacity and instance I/O consumption.
457 Instance migration behaviour
458 ++++++++++++++++++++++++++++
460 While live-migration is in general desirable over failover, it is
461 possible that for some workloads it is actually worse, due to the
462 variable time of the “suspend” phase during live migration.
464 To allow the tools to work consistently over such instances (without
465 having to hard-code instance names), a new backend parameter
466 ``always_failover`` has been added to control the migration/failover
467 behaviour. When set to True, all migration requests for an instance will
468 instead fall-back to failover.
470 Instance memory ballooning
471 ++++++++++++++++++++++++++
473 Initial support for memory ballooning has been added. The memory for an
474 instance is no longer fixed (backend parameter ``memory``), but instead
475 can vary between minimum and maximum values (backend parameters
476 ``minmem`` and ``maxmem``). Currently we only change an instance's
479 - live migrating or failing over and instance and the target node
480 doesn't have enough memory
481 - user requests changing the memory via ``gnt-instance modify
487 In order to control the use of specific CPUs by instance, support for
488 controlling CPU pinning has been added for the Xen, HVM and LXC
489 hypervisors. This is controlled by a new hypervisor parameter
490 ``cpu_mask``; details about possible values for this are in the
491 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
492 VCPU-to-CPU mapping) form will work well only when all nodes in your
493 cluster have the same amount of CPUs.
498 Another area in which Ganeti was not customisable were the parameters
499 used for storage configuration, e.g. how many stripes to use for LVM,
500 DRBD resync configuration, etc.
502 To improve this area, we've added disks parameters, which are
503 customisable at cluster and node group level, and which allow to
504 specify various parameters for disks (DRBD has the most parameters
505 currently), for example:
507 - DRBD resync algorithm and parameters (e.g. speed)
508 - the default VG for meta-data volumes for DRBD
509 - number of stripes for LVM (plain disk template)
512 These parameters can be modified via ``gnt-cluster modify -D …`` and
513 ``gnt-group modify -D …``, and are used at either instance creation (in
514 case of LVM stripes, for example) or at disk “activation” time
517 Rados block device support
518 ++++++++++++++++++++++++++
520 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
521 denoted by the ``rbd`` disk template type. This is considered
522 experimental, feedback is welcome. For details on configuring it, see
523 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
528 The existing master IP functionality works well only in simple setups (a
529 single network shared by all nodes); however, if nodes belong to
530 different networks, then the ``/32`` setup and lack of routing
531 information is not enough.
533 To allow the master IP to function well in more complex cases, the
534 system was reworked as follows:
536 - a master IP netmask setting has been added
537 - the master IP activation/turn-down code was moved from the node daemon
539 - whether to run the Ganeti-supplied master IP script or a user-supplied
540 on is a ``gnt-cluster init`` setting
542 Details about the location of the standard and custom setup scripts are
543 in the man page :manpage:`gnt-cluster(8)`; for information about the
544 setup script protocol, look at the Ganeti-supplied script.
549 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
552 It is now possible to use TLS-protected connections, and when renewing
553 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
554 it is now possible to specify spice or spice CA certificates. Also, it
555 is possible to configure a password for SPICE sessions via the
556 hypervisor parameter ``spice_password_file``.
558 There are also new parameters to control the compression and streaming
559 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
560 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
561 for the spice parameters.
563 Lastly, it is now possible to see the SPICE connection information via
564 ``gnt-instance console``.
569 A new tool (``tools/ovfconverter``) has been added that supports
570 conversion between Ganeti and the `Open Virtualization Format
571 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
574 This relies on the ``qemu-img`` tool to convert the disk formats, so the
575 actual compatibility with other virtualization solutions depends on it.
580 The configuration query daemon (``ganeti-confd``) is now optional, and
581 has been rewritten in Haskell; whether to use the daemon at all, use the
582 Python (default) or the Haskell version is selectable at configure time
583 via the ``--enable-confd`` parameter, which can take one of the
584 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
585 daemon will result in a smaller footprint; for larger systems, we
586 welcome feedback on the Haskell version which might become the default
589 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
590 daemon running. The Python version doesn't implement the new call.
593 User interface changes
594 ~~~~~~~~~~~~~~~~~~~~~~
596 We have replaced the ``--disks`` option of ``gnt-instance
597 replace-disks`` with a more flexible ``--disk`` option, which allows
598 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
599 disk size and mode can be changed upon recreation (via ``gnt-instance
600 recreate-disks``, which accepts the same ``--disk`` option).
602 As many people are used to a ``show`` command, we have added that as an
603 alias to ``info`` on all ``gnt-*`` commands.
605 The ``gnt-instance grow-disk`` command has a new mode in which it can
606 accept the target size of the disk, instead of the delta; this can be
607 more safe since two runs in absolute mode will be idempotent, and
608 sometimes it's also easier to specify the desired size directly.
610 Also the handling of instances with regard to offline secondaries has
611 been improved. Instance operations should not fail because one of it's
612 secondary nodes is offline, even though it's safe to proceed.
614 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
615 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
616 minors to instance name.
621 RAPI coverage has improved, with (for example) new resources for
622 recreate-disks, node power-cycle, etc.
627 There is partial support for ``xl`` in the Xen hypervisor; feedback is
630 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
631 whether to still support Python 2.4 or move to Python 2.6 as minimum
634 Support for Fedora has been slightly improved; the provided example
635 init.d script should work better on it and the INSTALL file should
636 document the needed dependencies.
641 The deprecated ``QueryLocks`` LUXI request has been removed. Use
642 ``Query(what=QR_LOCK, ...)`` instead.
644 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
645 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
646 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
647 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
648 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
650 RAPI client: ``CertificateError`` now derives from
651 ``GanetiApiError``. This should make it more easy to handle Ganeti
654 Deprecation warnings due to PyCrypto/paramiko import in
655 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
656 make sure to run an up-to-date paramiko version, if you use this tool.
658 The QA scripts now depend on Python 2.5 or above (the main code base
659 still works with Python 2.4).
661 The configuration file (``config.data``) is now written without
662 indentation for performance reasons; if you want to edit it, it can be
663 re-formatted via ``tools/fmtjson``.
665 A number of bugs has been fixed in the cluster merge tool.
667 ``x509`` certification verification (used in import-export) has been
668 changed to allow the same clock skew as permitted by the cluster
669 verification. This will remove some rare but hard to diagnose errors in
676 *(Released Thu, 19 Jul 2012)*
678 Very few changes from rc4 to the final release, only bugfixes:
680 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
681 instance, fix CDROM booting for KVM instances)
682 - fixed node group modification of node parameters
683 - fixed issue in LUClusterVerifyGroup with multi-group clusters
684 - fixed generation of bash completion to ensure a stable ordering
691 *(Released Fri, 13 Jul 2012)*
693 Third release candidate for 2.6. The following changes were done from
696 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
697 - Fixed an inconsistency in the LUXI protocol with the provided
698 arguments (NOT backwards compatible)
699 - Fixed a bug with node groups ipolicy where ``min`` was greater than
700 the cluster ``std`` value
701 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
702 easier instance debugging on nodes (requires ``hconfd`` to work)
708 *(Released Tue, 03 Jul 2012)*
710 Second release candidate for 2.6. The following changes were done from
713 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
715 - Fixed a RAPI regression on missing beparams/memory
716 - Fixed redistribution of files on offline nodes
717 - Added possibility to run activate-disks even though secondaries are
718 offline. With this change it relaxes also the strictness on some other
719 commands which use activate disks internally:
720 * ``gnt-instance start|reboot|rename|backup|export``
721 - Made it possible to remove safely an instance if its secondaries are
723 - Made it possible to reinstall even though secondaries are offline
729 *(Released Mon, 25 Jun 2012)*
731 First release candidate for 2.6. The following changes were done from
734 - Fixed bugs with disk parameters and ``rbd`` templates as well as
736 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
737 behaviour. It supports now the modify operation
738 - ``hcheck`` implemented to analyze cluster health and possibility of
739 improving health by rebalance
740 - ``hbal`` has been improved in dealing with split instances
746 *(Released Mon, 11 Jun 2012)*
748 Second beta release of 2.6. The following changes were done from beta2
751 - Fixed ``daemon-util`` with non-root user models
752 - Fixed creation of plain instances with ``--no-wait-for-sync``
753 - Fix wrong iv_names when running ``cfgupgrade``
754 - Export more information in RAPI group queries
755 - Fixed bug when changing instance network interfaces
756 - Extended burnin to do NIC changes
757 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
758 - Changed default for DRBD barriers
759 - Fixed DRBD error reporting for syncer rate
760 - Verify the options on disk parameters
762 And of course various fixes to documentation and improved unittests and
769 *(Released Wed, 23 May 2012)*
771 First beta release of 2.6. The following changes were done from beta1 to
774 - integrated patch for distributions without ``start-stop-daemon``
775 - adapted example init.d script to work on Fedora
776 - fixed log handling in Haskell daemons
777 - adapted checks in the watcher for pycurl linked against libnss
778 - add partial support for ``xl`` instead of ``xm`` for Xen
779 - fixed a type issue in cluster verification
780 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
783 Plus integrated fixes from the 2.5 branch:
785 - fixed ``kvm-ifup`` to use ``/bin/bash``
786 - fixed parallel build failures
787 - KVM live migration when using a custom keymap
793 *(Released Tue, 24 Jul 2012)*
795 A small bugfix release, with no new features:
797 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
798 different default shell (e.g. Debian, Ubuntu)
799 - fixed KVM startup and live migration with a custom keymap (fixes Issue
800 243 and Debian bug #650664)
801 - fixed compatibility with KVM versions that don't support multiple boot
802 devices (fixes Issue 230 and Debian bug #624256)
804 Additionally, a few fixes were done to the build system (fixed parallel
805 build failures) and to the unittests (fixed race condition in test for
806 FileID functions, and the default enable/disable mode for QA test is now
813 *(Released Fri, 11 May 2012)*
815 A small bugfix release.
817 The main issues solved are on the topic of compatibility with newer LVM
820 - fixed parsing of ``lv_attr`` field
821 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
822 the ``--force`` flag is needed
824 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
825 to accept kernel 3.x too (was hardcoded to 2.6.*).
827 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
828 of hook results and that also made display of other errors suboptimal
829 was fixed; the code behaves now like 2.4 and earlier.
831 Another change in 2.5, the cleanup of the OS scripts environment, is too
832 aggressive: it removed even the ``PATH`` variable, which requires the OS
833 scripts to *always* need to export it. Since this is a bit too strict,
834 we now export a minimal PATH, the same that we export for hooks.
836 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
837 integrated into this release.
839 Finally, a few other miscellaneous changes were done (no new features,
840 just small improvements):
842 - Fix ``gnt-group --help`` display
843 - Fix hardcoded Xen kernel path
844 - Fix grow-disk handling of invalid units
845 - Update synopsis for ``gnt-cluster repair-disk-sizes``
846 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
852 *(Released Thu, 12 Apr 2012)*
854 Incompatible/important changes and bugfixes
855 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
857 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
858 resource's ``ip_check`` parameter changed from ``True`` to ``False``
859 to match the underlying LUXI interface.
860 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
861 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
862 not maintain backwards-compatibility as the underlying operation
863 changed in an incompatible way. The RAPI client can talk to old
864 servers, but it needs to be told so as the return value changed.
865 - When creating file-based instances via RAPI, the ``file_driver``
866 parameter no longer defaults to ``loop`` and must be specified.
867 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
869 - Support for the undocumented and deprecated RAPI instance creation
870 request format version 0 has been dropped. Use version 1, supported
871 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
872 - Pyparsing 1.4.6 or above is required, see :doc:`installation
873 documentation <install>`.
874 - The "cluster-verify" hooks are now executed per group by the
875 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
876 if you just run ``gnt-cluster verify``, which generates one opcode per
878 - The environment as passed to the OS scripts is cleared, and thus no
879 environment variables defined in the node daemon's environment will be
880 inherited by the scripts.
881 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
883 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
884 support operations involving multiple node groups.
885 - Offline nodes are ignored when failing over an instance.
886 - Support for KVM version 1.0, which changed the version reporting format
888 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
890 - ``Makefile`` is now compatible with Automake 1.11.2
891 - Includes all bugfixes made in the 2.4 series
896 - The ganeti-htools project has been merged into the ganeti-core source
897 tree and will be built as part of Ganeti (see :doc:`install-quick`).
898 - Implemented support for :doc:`shared storage <design-shared-storage>`.
899 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
900 GPT-style partition tables (requires `parted
901 <http://www.gnu.org/s/parted/>`_).
902 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
903 - Allowed adding tags on instance creation.
904 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
905 - Allow instances to be started in a paused state, enabling the user to
906 see the complete console output on boot using the console.
907 - Added new hypervisor flag to control default reboot behaviour
908 (``reboot_behavior``).
909 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
910 - Improved out-of-band management support:
912 - Added ``gnt-node health`` command reporting the health status of
914 - Added ``gnt-node power`` command to manage power status of nodes.
915 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
917 - Instance migration can fall back to failover if instance is not
919 - Filters can be used when listing nodes, instances, groups and locks;
920 see :manpage:`ganeti(7)` manpage.
921 - Added post-execution status as variables to :doc:`hooks <hooks>`
923 - Instance tags are exported/imported together with the instance.
924 - When given an explicit job ID, ``gnt-job info`` will work for archived
926 - Jobs can define dependencies on other jobs (not yet supported via
927 RAPI or command line, but used by internal commands and usable via
930 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
933 - Instance failover is now available as a RAPI resource
934 (``/2/instances/[instance_name]/failover``).
935 - ``gnt-instance info`` defaults to static information if primary node
937 - Opcodes have a new ``comment`` attribute.
938 - Added basic SPICE support to KVM hypervisor.
939 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
941 Node group improvements
942 ~~~~~~~~~~~~~~~~~~~~~~~
944 - ``gnt-cluster verify`` has been modified to check groups separately,
945 thereby improving performance.
946 - Node group support has been added to ``gnt-cluster verify-disks``,
947 which now operates per node group.
948 - Watcher has been changed to work better with node groups.
950 - One process and state file per node group.
951 - Slow watcher in one group doesn't block other group's watcher.
953 - Added new command, ``gnt-group evacuate``, to move all instances in a
954 node group to other groups.
955 - Added ``gnt-instance change-group`` to move an instance to another
957 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
958 per-group operations.
959 - Node groups can be tagged.
960 - Some operations switch from an exclusive to a shared lock as soon as
962 - Instance's primary and secondary nodes' groups are now available as
963 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
964 and ``snodes.group.uuid``).
969 - Numerous updates to documentation and manpages.
971 - :doc:`RAPI <rapi>` documentation now has detailed parameter
973 - Some opcode/job results are now also documented, see :doc:`RAPI
976 - A lockset's internal lock is now also visible in lock monitor.
977 - Log messages from job queue workers now contain information about the
978 opcode they're processing.
979 - ``gnt-instance console`` no longer requires the instance lock.
980 - A short delay when waiting for job changes reduces the number of LUXI
981 requests significantly.
982 - DRBD metadata volumes are overwritten with zeros during disk creation.
983 - Out-of-band commands no longer acquire the cluster lock in exclusive
985 - ``devel/upload`` now uses correct permissions for directories.
991 *(Released Fri, 23 Mar 2012)*
993 This was the sixth release candidate of the 2.5 series.
999 *(Released Mon, 9 Jan 2012)*
1001 This was the fifth release candidate of the 2.5 series.
1007 *(Released Thu, 27 Oct 2011)*
1009 This was the fourth release candidate of the 2.5 series.
1015 *(Released Wed, 26 Oct 2011)*
1017 This was the third release candidate of the 2.5 series.
1023 *(Released Tue, 18 Oct 2011)*
1025 This was the second release candidate of the 2.5 series.
1031 *(Released Tue, 4 Oct 2011)*
1033 This was the first release candidate of the 2.5 series.
1039 *(Released Wed, 31 Aug 2011)*
1041 This was the third beta release of the 2.5 series.
1047 *(Released Mon, 22 Aug 2011)*
1049 This was the second beta release of the 2.5 series.
1055 *(Released Fri, 12 Aug 2011)*
1057 This was the first beta release of the 2.5 series.
1063 *(Released Thu, 27 Oct 2011)*
1065 - Fixed bug when parsing command line parameter values ending in
1067 - Fixed assertion error after unclean master shutdown
1068 - Disable HTTP client pool for RPC, significantly reducing memory usage
1070 - Fixed queue archive creation with wrong permissions
1076 *(Released Tue, 23 Aug 2011)*
1080 - Fixed documentation for importing with ``--src-dir`` option
1081 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1082 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1088 *(Released Fri, 5 Aug 2011)*
1090 Many bug-fixes and a few small features:
1092 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1093 issues when you tried to add an instance with two MAC addresses in one
1095 - KVM: fixed per-instance stored UID value
1096 - KVM: configure bridged NICs at migration start
1097 - KVM: Fix a bug where instance will not start with never KVM versions
1099 - Added OS search path to ``gnt-cluster info``
1100 - Fixed an issue with ``file_storage_dir`` where you were forced to
1101 provide an absolute path, but the documentation states it is a
1102 relative path, the documentation was right
1103 - Added a new parameter to instance stop/start called ``--no-remember``
1104 that will make the state change to not be remembered
1105 - Implemented ``no_remember`` at RAPI level
1106 - Improved the documentation
1107 - Node evacuation: don't call IAllocator if node is already empty
1108 - Fixed bug in DRBD8 replace disks on current nodes
1109 - Fixed bug in recreate-disks for DRBD instances
1110 - Moved assertion checking locks in ``gnt-instance replace-disks``
1111 causing it to abort with not owning the right locks for some situation
1112 - Job queue: Fixed potential race condition when cancelling queued jobs
1113 - Fixed off-by-one bug in job serial generation
1114 - ``gnt-node volumes``: Fix instance names
1115 - Fixed aliases in bash completion
1116 - Fixed a bug in reopening log files after being sent a SIGHUP
1117 - Added a flag to burnin to allow specifying VCPU count
1118 - Bugfixes to non-root Ganeti configuration
1124 *(Released Thu, 12 May 2011)*
1126 Many bug-fixes and a few new small features:
1128 - Fixed a bug related to log opening failures
1129 - Fixed a bug in instance listing with orphan instances
1130 - Fixed a bug which prevented resetting the cluster-level node parameter
1131 ``oob_program`` to the default
1132 - Many fixes related to the ``cluster-merge`` tool
1133 - Fixed a race condition in the lock monitor, which caused failures
1134 during (at least) creation of many instances in parallel
1135 - Improved output for gnt-job info
1136 - Removed the quiet flag on some ssh calls which prevented debugging
1138 - Improved the N+1 failure messages in cluster verify by actually
1139 showing the memory values (needed and available)
1140 - Increased lock attempt timeouts so that when executing long operations
1141 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1142 too early and thus prevent the use of the 'fair' mechanism
1143 - Changed instance query data (``gnt-instance info``) to not acquire
1144 locks unless needed, thus allowing its use on locked instance if only
1145 static information is asked for
1146 - Improved behaviour with filesystems that do not support rename on an
1148 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1149 kept locks on all nodes during the wipe, which is unneeded
1150 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1151 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1152 10GiB) which caused the wipe to fail right at the end in some cases
1153 - Fixed master IP activation when doing master failover with no-voting
1154 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1155 the master node itself
1156 - Fixed potential data-loss in under disk full conditions, where Ganeti
1157 wouldn't check correctly the return code and would consider
1158 partially-written files 'correct'
1159 - Fixed bug related to multiple VGs and DRBD disk replacing
1160 - Added new disk parameter ``metavg`` that allows placement of the meta
1161 device for DRBD in a different volume group
1162 - Fixed error handling in the node daemon when the system libc doesn't
1163 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1164 - Fixed lock release during replace-disks, which kept cluster-wide locks
1165 when doing disk replaces with an iallocator script
1166 - Added check for missing bridges in cluster verify
1167 - Handle EPIPE errors while writing to the terminal better, so that
1168 piping the output to e.g. ``less`` doesn't cause a backtrace
1169 - Fixed rare case where a ^C during Luxi calls could have been
1170 interpreted as server errors, instead of simply terminating
1171 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1173 - Added a few more parameters to the KVM hypervisor, allowing a second
1174 CDROM, custom disk type for CDROMs and a floppy image
1175 - Removed redundant message in instance rename when the name is given
1177 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1178 disks on new nodes, allowing recreation when the original instance
1179 nodes are completely gone
1180 - Added option when converting disk templates to DRBD to skip waiting
1181 for the resync, in order to make the instance available sooner
1182 - Added two new variables to the OS scripts environment (containing the
1184 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1185 to allow use of ``pvgrub`` as bootloader
1186 - Changed the instance memory modifications to only check out-of-memory
1187 conditions on memory increases, and turned the secondary node warnings
1188 into errors (they can still be overridden via ``--force``)
1189 - Fixed the handling of a corner case when the Python installation gets
1190 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1191 to execute a command that doesn't exist
1192 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1193 node of the instance returned failures during instance shutdown; this
1194 adds the option ``--ignore-consistency`` to gnt-instance move
1196 And as usual, various improvements to the error messages, documentation
1203 *(Released Wed, 09 Mar 2011)*
1205 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1206 the RAPI users file if run twice (even with ``--dry-run``).
1208 The release fixes that bug (nothing else changed).
1214 *(Released Mon, 07 Mar 2011)*
1216 Final 2.4.0 release. Just a few small fixes:
1218 - Fixed RAPI node evacuate
1219 - Fixed the kvm-ifup script
1220 - Fixed internal error handling for special job cases
1221 - Updated man page to specify the escaping feature for options
1227 *(Released Mon, 28 Feb 2011)*
1229 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1230 that this feature wiped the disks of the wrong instance, leading to loss
1235 - Fixed title of query field containing instance name
1236 - Expanded the glossary in the documentation
1237 - Fixed one unittest (internal issue)
1243 *(Released Mon, 21 Feb 2011)*
1245 A number of bug fixes plus just a couple functionality changes.
1247 On the user-visible side, the ``gnt-* list`` command output has changed
1248 with respect to "special" field states. The current rc1 style of display
1249 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1250 the default output mode special fields states are displayed as follows:
1252 - Offline resource: ``*``
1253 - Unavailable/not applicable: ``-``
1254 - Data missing (RPC failure): ``?``
1255 - Unknown field: ``??``
1257 Another user-visible change is the addition of ``--force-join`` to
1262 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1263 - Fixed regression in RAPI/instance reinstall where all parameters were
1264 required (instead of optional)
1265 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1266 - Fixed iallocator usage (offline nodes were not considered offline)
1267 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1268 - Fixed hypervisor and OS parameter validation with respect to
1269 non-vm_capable nodes
1270 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1272 - Fixed ``tools/listrunner`` with respect to agent-based usage
1278 *(Released Fri, 4 Feb 2011)*
1280 Many changes and fixes since the beta1 release. While there were some
1281 internal changes, the code has been mostly stabilised for the RC
1284 Note: the dumb allocator was removed in this release, as it was not kept
1285 up-to-date with the IAllocator protocol changes. It is recommended to
1286 use the ``hail`` command from the ganeti-htools package.
1288 Note: the 2.4 and up versions of Ganeti are not compatible with the
1289 0.2.x branch of ganeti-htools. You need to upgrade to
1290 ganeti-htools-0.3.0 (or later).
1292 Regressions fixed from 2.3
1293 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1295 - Fixed the ``gnt-cluster verify-disks`` command
1296 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1298 - Fixed disk adoption breakage
1299 - Fixed wrong headers in instance listing for field aliases
1304 - Fixed corner case in KVM handling of NICs
1305 - Fixed many cases of wrong handling of non-vm_capable nodes
1306 - Fixed a bug where a missing instance symlink was not possible to
1307 recreate with any ``gnt-*`` command (now ``gnt-instance
1308 activate-disks`` does it)
1309 - Fixed the volume group name as reported by ``gnt-cluster
1311 - Increased timeouts for the import-export code, hopefully leading to
1312 fewer aborts due network or instance timeouts
1313 - Fixed bug in ``gnt-node list-storage``
1314 - Fixed bug where not all daemons were started on cluster
1315 initialisation, but only at the first watcher run
1316 - Fixed many bugs in the OOB implementation
1317 - Fixed watcher behaviour in presence of instances with offline
1319 - Fixed instance list output for instances running on the wrong node
1320 - a few fixes to the cluster-merge tool, but it still cannot merge
1321 multi-node groups (currently it is not recommended to use this tool)
1327 - Improved network configuration for the KVM hypervisor
1328 - Added e1000 as a supported NIC for Xen-HVM
1329 - Improved the lvmstrap tool to also be able to use partitions, as
1330 opposed to full disks
1331 - Improved speed of disk wiping (the cluster parameter
1332 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1333 of instance creations
1334 - Added documentation for the OS parameters
1335 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1336 hypervisor is not responding
1337 - Added display of blacklisted and hidden OS information in
1338 ``gnt-cluster info``
1339 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1340 NIC and node parameters, which might create problems with currently
1341 invalid (but undetected) configuration files, but prevents validation
1342 failures when unrelated parameters are modified
1343 - Changed cluster initialisation to wait for the master daemon to become
1345 - Expanded the RAPI interface:
1347 - Added config redistribution resource
1348 - Added activation/deactivation of instance disks
1349 - Added export of console information
1351 - Implemented log file reopening on SIGHUP, which allows using
1352 logrotate(8) for the Ganeti log files
1353 - Added a basic OOB helper script as an example
1359 *(Released Fri, 14 Jan 2011)*
1364 - Fixed timezone issues when formatting timestamps
1365 - Added support for node groups, available via ``gnt-group`` and other
1367 - Added out-of-band framework and management, see :doc:`design
1368 document <design-oob>`
1369 - Removed support for roman numbers from ``gnt-node list`` and
1370 ``gnt-instance list``.
1371 - Allowed modification of master network interface via ``gnt-cluster
1372 modify --master-netdev``
1373 - Accept offline secondaries while shutting down instance disks
1374 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1375 - Added support for multiple LVM volume groups
1376 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1378 - Added commands to list available fields:
1380 - ``gnt-node list-fields``
1381 - ``gnt-group list-fields``
1382 - ``gnt-instance list-fields``
1384 - Updated documentation and man pages
1389 - Moved ``rapi_users`` file into separate directory, now named
1390 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1392 - Added new tool for running commands on many machines,
1393 ``tools/ganeti-listrunner``
1394 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1395 improving the ``gnt-instance console`` output
1396 - Allowed customisation of disk index separator at ``configure`` time
1397 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1398 - Added support for non-partitioned md disks in ``lvmstrap``
1399 - Added script to gracefully power off KVM instances
1400 - Split ``utils`` module into smaller parts
1401 - Changed query operations to return more detailed information, e.g.
1402 whether an information is unavailable due to an offline node. To use
1403 this new functionality, the LUXI call ``Query`` must be used. Field
1404 information is now stored by the master daemon and can be retrieved
1405 using ``QueryFields``. Instances, nodes and groups can also be queried
1406 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1407 exposed via RAPI). The following commands make use of this
1408 infrastructure change:
1410 - ``gnt-group list``
1411 - ``gnt-group list-fields``
1413 - ``gnt-node list-fields``
1414 - ``gnt-instance list``
1415 - ``gnt-instance list-fields``
1416 - ``gnt-debug locks``
1421 - New RAPI resources (see :doc:`rapi`):
1425 - ``/2/groups/[group_name]``
1426 - ``/2/groups/[group_name]/assign-nodes``
1427 - ``/2/groups/[group_name]/modify``
1428 - ``/2/groups/[group_name]/rename``
1429 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1433 - Implemented ``no_install`` for instance creation
1434 - Implemented OS parameters for instance reinstallation, allowing
1435 use of special settings on reinstallation (e.g. for preserving data)
1440 - Added IPv6 support in import/export
1441 - Pause DRBD synchronization while wiping disks on instance creation
1442 - Updated unittests and QA scripts
1443 - Improved network parameters passed to KVM
1444 - Converted man pages from docbook to reStructuredText
1450 *(Released Mon, 20 Dec 2010)*
1452 Released version 2.3.1~rc1 without any changes.
1458 *(Released Wed, 1 Dec 2010)*
1460 - impexpd: Disable OpenSSL compression in socat if possible (backport
1461 from master, commit e90739d625b, see :doc:`installation guide
1462 <install-quick>` for details)
1463 - Changed unittest coverage report to exclude test scripts
1464 - Added script to check version format
1470 *(Released Wed, 1 Dec 2010)*
1472 Released version 2.3.0~rc1 without any changes.
1478 *(Released Fri, 19 Nov 2010)*
1480 A number of bugfixes and documentation updates:
1482 - Update ganeti-os-interface documentation
1483 - Fixed a bug related to duplicate MACs or similar items which should be
1485 - Fix breakage in OS state modify
1486 - Reinstall instance: disallow offline secondaries (fixes bug related to
1487 OS changing but reinstall failing)
1488 - plus all the other fixes between 2.2.1 and 2.2.2
1494 *(Released Tue, 2 Nov 2010)*
1496 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1497 - Fixed master failover race with watcher
1498 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1500 - Accept previously stopped instance for export with instance removal
1501 - Simplify and extend the environment variables for instance OS scripts
1502 - Added new node flags, ``master_capable`` and ``vm_capable``
1503 - Added optional instance disk wiping prior during allocation. This is a
1504 cluster-wide option and can be set/modified using
1505 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1506 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1507 :doc:`install-quick`
1508 - Added a new watcher option (``--ignore-pause``)
1509 - Added option to ignore offline node on instance start/stop
1510 (``--ignore-offline``)
1511 - Allow overriding OS parameters with ``gnt-instance reinstall``
1512 - Added ability to change node's secondary IP address using ``gnt-node
1514 - Implemented privilege separation for all daemons except
1515 ``ganeti-noded``, see ``configure`` options
1516 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1518 - Implemented job priorities (see ``ganeti(7)`` manpage)
1519 - Ignore failures while shutting down instances during failover from
1521 - Exit daemon's bootstrap process only once daemon is ready
1522 - Export more information via ``LUInstanceQuery``/remote API
1523 - Improved documentation, QA and unittests
1524 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1525 restart if the file was created or changed
1526 - Added LUXI protocol version sent with each request and response,
1527 allowing detection of server/client mismatches
1528 - Moved the Python scripts among gnt-* and ganeti-* into modules
1529 - Moved all code related to setting up SSH to an external script,
1531 - Infrastructure changes for node group support in future versions
1537 *(Released Fri, 19 Nov 2010)*
1539 A few small bugs fixed, and some improvements to the build system:
1541 - Fix documentation regarding conversion to drbd
1542 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1544 - Fix error handling in node modify with multiple changes
1545 - Allow remote imports without checked names
1551 *(Released Tue, 19 Oct 2010)*
1553 - Disable SSL session ID cache in RPC client
1559 *(Released Thu, 14 Oct 2010)*
1561 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1562 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1564 - Fix problems with interaction between Curl and Python's HTTP server,
1565 resulting in increased speed in many RPC calls
1566 - Improve our release script to prevent breakage with older aclocal and
1573 *(Released Thu, 7 Oct 2010)*
1575 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1576 value retrieved from master
1577 - Added support for blacklisted or hidden OS definitions
1578 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1579 - Added support for -mem-path in KVM hypervisor abstraction layer
1580 - Allow overriding instance parameters in tool for inter-cluster
1581 instance moves (``tools/move-instance``)
1582 - Improved opcode summaries (e.g. in ``gnt-job list``)
1583 - Improve consistency of OS listing by sorting it
1584 - Documentation updates
1590 *(Released Fri, 8 Oct 2010)*
1592 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1598 *(Released Mon, 4 Oct 2010)*
1600 - Fixed regression in ``gnt-instance rename``
1606 *(Released Wed, 22 Sep 2010)*
1608 - Fixed OS_VARIANT variable for OS scripts
1609 - Fixed cluster tag operations via RAPI
1610 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1611 - Disabled RAPI CA checks in watcher
1617 *(Released Mon, 23 Aug 2010)*
1619 - Support DRBD versions of the format "a.b.c.d"
1621 - Re-introduce support for usage from multiple threads in RAPI client
1622 - Instance renames and modify via RAPI
1623 - Work around race condition between processing and archival in job
1625 - Mark opcodes following failed one as failed, too
1626 - Job field ``lock_status`` was removed due to difficulties making it
1627 work with the changed job queue in Ganeti 2.2; a better way to monitor
1628 locks is expected for a later 2.2.x release
1629 - Fixed dry-run behaviour with many commands
1630 - Support ``ssh-agent`` again when adding nodes
1631 - Many additional bugfixes
1637 *(Released Fri, 30 Jul 2010)*
1639 Important change: the internal RPC mechanism between Ganeti nodes has
1640 changed from using a home-grown http library (based on the Python base
1641 libraries) to use the PycURL library. This requires that PycURL is
1642 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1643 against GnuTLS by default. cURL's support for GnuTLS had known issues
1644 before cURL 7.21.0 and we recommend using the latest cURL release or
1645 linking against OpenSSL. Most other distributions already link PycURL
1646 and cURL against OpenSSL. The command::
1648 python -c 'import pycurl; print pycurl.version'
1650 can be used to determine the libraries PycURL and cURL are linked
1653 Other significant changes:
1655 - Rewrote much of the internals of the job queue, in order to achieve
1656 better parallelism; this decouples job query operations from the job
1657 processing, and it should allow much nicer behaviour of the master
1658 daemon under load, and it also has uncovered some long-standing bugs
1659 related to the job serialisation (now fixed)
1660 - Added a default iallocator setting to the cluster parameters,
1661 eliminating the need to always pass nodes or an iallocator for
1662 operations that require selection of new node(s)
1663 - Added experimental support for the LXC virtualization method
1664 - Added support for OS parameters, which allows the installation of
1665 instances to pass parameter to OS scripts in order to customise the
1667 - Added a hypervisor parameter controlling the migration type (live or
1668 non-live), since hypervisors have various levels of reliability; this
1669 has renamed the 'live' parameter to 'mode'
1670 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1671 logical volumes, meaning that cluster verify will ignore them and not
1672 flag their presence as errors
1673 - The watcher will now reset the error count for failed instances after
1674 8 hours, thus allowing self-healing if the problem that caused the
1675 instances to be down/fail to start has cleared in the meantime
1676 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1677 check for, and warn, if the drbd module parameter ``usermode_helper``
1678 is not consistent with the cluster-wide setting; this is needed to
1679 make diagnose easier of failed drbd creations
1680 - Started adding base IPv6 support, but this is not yet
1681 enabled/available for use
1682 - Rename operations (cluster, instance) will now return the new name,
1683 which is especially useful if a short name was passed in
1684 - Added support for instance migration in RAPI
1685 - Added a tool to pre-configure nodes for the SSH setup, before joining
1686 them to the cluster; this will allow in the future a simplified model
1687 for node joining (but not yet fully enabled in 2.2); this needs the
1688 paramiko python library
1689 - Fixed handling of name-resolving errors
1690 - Fixed consistency of job results on the error path
1691 - Fixed master-failover race condition when executed multiple times in
1693 - Fixed many bugs related to the job queue (mostly introduced during the
1694 2.2 development cycle, so not all are impacting 2.1)
1695 - Fixed instance migration with missing disk symlinks
1696 - Fixed handling of unknown jobs in ``gnt-job archive``
1697 - And many other small fixes/improvements
1701 - Enhanced both the unittest and the QA coverage
1702 - Switched the opcode validation to a generic model, and extended the
1703 validation to all opcode parameters
1704 - Changed more parts of the code that write shell scripts to use the
1706 - Switched the master daemon to use the asyncore library for the Luxi
1713 *(Released Thu, 17 Jun 2010)*
1715 - Added tool (``move-instance``) and infrastructure to move instances
1716 between separate clusters (see :doc:`separate documentation
1717 <move-instance>` and :doc:`design document <design-2.2>`)
1718 - Added per-request RPC timeout
1719 - RAPI now requires a Content-Type header for requests with a body (e.g.
1720 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1721 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1722 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1724 - Implemented initial support for running Ganeti daemons as separate
1725 users, see configure-time flags ``--with-user-prefix`` and
1726 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1728 - Instances can be removed after export (``gnt-backup export
1729 --remove-instance``)
1730 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1731 key (instead of 1024 bit)
1732 - Added new cluster configuration file for cluster domain secret
1733 - Import/export now use SSL instead of SSH
1734 - Added support for showing estimated time when exporting an instance,
1735 see the ``ganeti-os-interface(7)`` manpage and look for
1742 *(Released Tue, 16 Nov 2010)*
1744 Some more bugfixes. Unless critical bugs occur, this will be the last
1747 - Fix case of MAC special-values
1748 - Fix mac checker regex
1749 - backend: Fix typo causing "out of range" error
1750 - Add missing --units in gnt-instance list man page
1756 *(Released Tue, 24 Aug 2010)*
1759 - Don't ignore secondary node silently on non-mirrored disk templates
1761 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1762 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1763 - Properly document the usb_mouse parameter
1764 - Fix path in ganeti-rapi(8) (issue 116)
1765 - Adjust error message when the ganeti user's .ssh directory is
1767 - Add same-node-check when changing the disk template to drbd
1773 *(Released Fri, 16 Jul 2010)*
1776 - Add an option to only select some reboot types during qa/burnin.
1777 (on some hypervisors consequent reboots are not supported)
1778 - Fix infrequent race condition in master failover. Sometimes the old
1779 master ip address would be still detected as up for a short time
1780 after it was removed, causing failover to fail.
1781 - Decrease mlockall warnings when the ctypes module is missing. On
1782 Python 2.4 we support running even if no ctypes module is installed,
1783 but we were too verbose about this issue.
1784 - Fix building on old distributions, on which man doesn't have a
1786 - Fix RAPI not to ignore the MAC address on instance creation
1787 - Implement the old instance creation format in the RAPI client.
1793 *(Released Thu, 01 Jul 2010)*
1795 A small bugfix release:
1796 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1797 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1798 a non-existing option
1799 - Fix instance create: the --force-variant option was ignored
1800 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1801 - Fix modify node storage with non-FQDN arguments
1802 - Fix RAPI client to authenticate under Python 2.6 when used
1803 for more than 5 requests needing authentication
1804 - Fix gnt-instance modify -t (storage) giving a wrong error message
1805 when converting a non-shutdown drbd instance to plain
1811 *(Released Fri, 18 Jun 2010)*
1813 A small bugfix release:
1815 - Fix live migration of KVM instances started with older Ganeti
1816 versions which had fewer hypervisor parameters
1817 - Fix gnt-instance grow-disk on down instances
1818 - Fix an error-reporting bug during instance migration
1819 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1820 silently ignoring broken ones
1821 - Fix an RPC error reporting bug affecting, for example, RAPI client
1823 - Fix bug triggered by different API version os-es on different nodes
1824 - Fix a bug in instance startup with custom hvparams: OS level
1825 parameters would fail to be applied.
1826 - Fix the RAPI client under Python 2.6 (but more work is needed to
1827 make it work completely well with OpenSSL)
1828 - Fix handling of errors when resolving names from DNS
1834 *(Released Thu, 3 Jun 2010)*
1836 A medium sized development cycle. Some new features, and some
1837 fixes/small improvements/cleanups.
1839 Significant features
1840 ~~~~~~~~~~~~~~~~~~~~
1842 The node deamon now tries to mlock itself into memory, unless the
1843 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1844 its logs, and falls back to console logging. This allows emergency
1845 features such as ``gnt-node powercycle`` to work even in the event of a
1846 broken node disk (tested offlining the disk hosting the node's
1847 filesystem and dropping its memory caches; don't try this at home)
1849 KVM: add vhost-net acceleration support. It can be tested with a new
1850 enough version of the kernel and of qemu-kvm.
1852 KVM: Add instance chrooting feature. If you use privilege dropping for
1853 your VMs you can also now force them to chroot to an empty directory,
1854 before starting the emulated guest.
1856 KVM: Add maximum migration bandwith and maximum downtime tweaking
1857 support (requires a new-enough version of qemu-kvm).
1859 Cluster verify will now warn if the master node doesn't have the master
1860 ip configured on it.
1862 Add a new (incompatible) instance creation request format to RAPI which
1863 supports all parameters (previously only a subset was supported, and it
1864 wasn't possible to extend the old format to accomodate all the new
1865 features. The old format is still supported, and a client can check for
1866 this feature, before using it, by checking for its presence in the
1867 ``features`` RAPI resource.
1869 Now with ancient latin support. Try it passing the ``--roman`` option to
1870 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1871 (requires the python-roman module to be installed, in order to work).
1876 As usual many internal code refactorings, documentation updates, and
1879 - Lots of improvements and cleanups to the experimental Remote API
1880 (RAPI) client library.
1881 - A new unit test suite for the core daemon libraries.
1882 - A fix to creating missing directories makes sure the umask is not
1883 applied anymore. This enforces the same directory permissions
1885 - Better handling terminating daemons with ctrl+c (used when running
1886 them in debugging mode).
1887 - Fix a race condition in live migrating a KVM instance, when stat()
1888 on the old proc status file returned EINVAL, which is an unexpected
1890 - Fixed manpage checking with newer man and utf-8 charachters. But now
1891 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1897 *(Released Fri, 7 May 2010)*
1899 Fix a bug which prevented untagged KVM instances from starting.
1905 *(Released Fri, 7 May 2010)*
1907 Another release with a long development cycle, during which many
1908 different features were added.
1910 Significant features
1911 ~~~~~~~~~~~~~~~~~~~~
1913 The KVM hypervisor now can run the individual instances as non-root, to
1914 reduce the impact of a VM being hijacked due to bugs in the
1915 hypervisor. It is possible to run all instances as a single (non-root)
1916 user, to manually specify a user for each instance, or to dynamically
1917 allocate a user out of a cluster-wide pool to each instance, with the
1918 guarantee that no two instances will run under the same user ID on any
1921 An experimental RAPI client library, that can be used standalone
1922 (without the other Ganeti libraries), is provided in the source tree as
1923 ``lib/rapi/client.py``. Note this client might change its interface in
1924 the future, as we iterate on its capabilities.
1926 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1927 replace the cluster's certificates and crypto keys. This might help in
1928 case they have been compromised, or have simply expired.
1930 A new disk option for instance creation has been added that allows one
1931 to "adopt" currently existing logical volumes, with data
1932 preservation. This should allow easier migration to Ganeti from
1933 unmanaged (or managed via other software) instances.
1935 Another disk improvement is the possibility to convert between redundant
1936 (DRBD) and plain (LVM) disk configuration for an instance. This should
1937 allow better scalability (starting with one node and growing the
1938 cluster, or shrinking a two-node cluster to one node).
1940 A new feature that could help with automated node failovers has been
1941 implemented: if a node sees itself as offline (by querying the master
1942 candidates), it will try to shutdown (hard) all instances and any active
1943 DRBD devices. This reduces the risk of duplicate instances if an
1944 external script automatically failovers the instances on such nodes. To
1945 enable this, the cluster parameter ``maintain_node_health`` should be
1946 enabled; in the future this option (per the name) will enable other
1947 automatic maintenance features.
1949 Instance export/import now will reuse the original instance
1950 specifications for all parameters; that means exporting an instance,
1951 deleting it and the importing it back should give an almost identical
1952 instance. Note that the default import behaviour has changed from
1953 before, where it created only one NIC; now it recreates the original
1956 Cluster verify has added a few new checks: SSL certificates validity,
1957 /etc/hosts consistency across the cluster, etc.
1962 As usual, many internal changes were done, documentation fixes,
1965 - Fixed cluster initialization with disabled cluster storage (regression
1966 introduced in 2.1.1)
1967 - File-based storage supports growing the disks
1968 - Fixed behaviour of node role changes
1969 - Fixed cluster verify for some corner cases, plus a general rewrite of
1970 cluster verify to allow future extension with more checks
1971 - Fixed log spamming by watcher and node daemon (regression introduced
1973 - Fixed possible validation issues when changing the list of enabled
1975 - Fixed cleanup of /etc/hosts during node removal
1976 - Fixed RAPI response for invalid methods
1977 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1978 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1979 from ide disks, etc.)
1980 - Allow OS changes without re-installation (to record a changed OS
1981 outside of Ganeti, or to allow OS renames)
1982 - Allow instance creation without OS installation (useful for example if
1983 the OS will be installed manually, or restored from a backup not in
1985 - Implemented option to make cluster ``copyfile`` use the replication
1987 - Added list of enabled hypervisors to ssconf (possibly useful for
1989 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1991 - A partial form of node re-IP is possible via node readd, which now
1992 allows changed node primary IP
1993 - Command line utilities now show an informational message if the job is
1995 - The logs of the master daemon now show the PID/UID/GID of the
2002 *(Released Fri, 12 Mar 2010)*
2004 During the 2.1.0 long release candidate cycle, a lot of improvements and
2005 changes have accumulated with were released later as 2.1.1.
2010 The node evacuate command (``gnt-node evacuate``) was significantly
2011 rewritten, and as such the IAllocator protocol was changed - a new
2012 request type has been added. This unfortunate change during a stable
2013 series is designed to improve performance of node evacuations; on
2014 clusters with more than about five nodes and which are well-balanced,
2015 evacuation should proceed in parallel for all instances of the node
2016 being evacuated. As such, any existing IAllocator scripts need to be
2017 updated, otherwise the above command will fail due to the unknown
2018 request. The provided "dumb" allocator has not been updated; but the
2019 ganeti-htools package supports the new protocol since version 0.2.4.
2021 Another important change is increased validation of node and instance
2022 names. This might create problems in special cases, if invalid host
2023 names are being used.
2025 Also, a new layer of hypervisor parameters has been added, that sits at
2026 OS level between the cluster defaults and the instance ones. This allows
2027 customisation of virtualization parameters depending on the installed
2028 OS. For example instances with OS 'X' may have a different KVM kernel
2029 (or any other parameter) than the cluster defaults. This is intended to
2030 help managing a multiple OSes on the same cluster, without manual
2031 modification of each instance's parameters.
2033 A tool for merging clusters, ``cluster-merge``, has been added in the
2034 tools sub-directory.
2039 - Improved the int/float conversions that should make the code more
2040 robust in face of errors from the node daemons
2041 - Fixed the remove node code in case of internal configuration errors
2042 - Fixed the node daemon behaviour in face of inconsistent queue
2043 directory (e.g. read-only file-system where we can't open the files
2045 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2046 now it either aborts cleanly or, if given the new "auto_promote"
2047 parameter, will automatically promote other nodes as needed
2048 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2049 completely prevent Ganeti from working
2050 - Fixed bug for instance export when not all disks were successfully
2052 - Fixed behaviour of node add when the new node is slow in starting up
2054 - Fixed handling of signals in the LUXI client, which should improve
2055 behaviour of command-line scripts
2056 - Added checks for invalid node/instance names in the configuration (now
2057 flagged during cluster verify)
2058 - Fixed watcher behaviour for disk activation errors
2059 - Fixed two potentially endless loops in http library, which led to the
2060 RAPI daemon hanging and consuming 100% CPU in some cases
2061 - Fixed bug in RAPI daemon related to hashed passwords
2062 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2064 - Enhanced compatibility with non-Debian OSes, but not using absolute
2065 path in some commands and allowing customisation of the ssh
2066 configuration directory
2067 - Fixed possible future issue with new Python versions by abiding to the
2068 proper use of ``__slots__`` attribute on classes
2069 - Added checks that should prevent directory traversal attacks
2070 - Many documentation fixes based on feedback from users
2075 - Added an "early_release" more for instance replace disks and node
2076 evacuate, where we release locks earlier and thus allow higher
2077 parallelism within the cluster
2078 - Added watcher hooks, intended to allow the watcher to restart other
2079 daemons (e.g. from the ganeti-nbma project), but they can be used of
2080 course for any other purpose
2081 - Added a compile-time disable for DRBD barriers, to increase
2082 performance if the administrator trusts the power supply or the
2083 storage system to not lose writes
2084 - Added the option of using syslog for logging instead of, or in
2085 addition to, Ganeti's own log files
2086 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2087 can indeed boot from a paravirtual NIC
2088 - Added a generic debug level for many operations; while this is not
2089 used widely yet, it allows one to pass the debug value all the way to
2091 - Enhanced the hooks environment for instance moves (failovers,
2092 migrations) where the primary/secondary nodes changed during the
2093 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2094 - Enhanced data validations for many user-supplied values; one important
2095 item is the restrictions imposed on instance and node names, which
2096 might reject some (invalid) host names
2097 - Add a configure-time option to disable file-based storage, if it's not
2098 needed; this allows greater security separation between the master
2099 node and the other nodes from the point of view of the inter-node RPC
2101 - Added user notification in interactive tools if job is waiting in the
2102 job queue or trying to acquire locks
2103 - Added log messages when a job is waiting for locks
2104 - Added filtering by node tags in instance operations which admit
2105 multiple instances (start, stop, reboot, reinstall)
2106 - Added a new tool for cluster mergers, ``cluster-merge``
2107 - Parameters from command line which are of the form ``a=b,c=d`` can now
2108 use backslash escapes to pass in values which contain commas,
2109 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2111 - For KVM, the instance name is the first parameter passed to KVM, so
2112 that it's more visible in the process list
2118 *(Released Tue, 2 Mar 2010)*
2120 Ganeti 2.1 brings many improvements with it. Major changes:
2122 - Added infrastructure to ease automated disk repairs
2123 - Added new daemon to export configuration data in a cheaper way than
2124 using the remote API
2125 - Instance NICs can now be routed instead of being associated with a
2127 - Improved job locking logic to reduce impact of jobs acquiring multiple
2128 locks waiting for other long-running jobs
2130 In-depth implementation details can be found in the Ganeti 2.1 design
2136 - Added chroot hypervisor
2137 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2139 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2140 ``bootloader_path`` and ``bootloader_args``)
2141 - Added the ``use_localtime`` option for the xen-hvm and kvm
2142 hypervisors, and the default value for this has changed to false (in
2143 2.0 xen-hvm always enabled it)
2144 - Added luxi call to submit multiple jobs in one go
2145 - Added cluster initialization option to not modify ``/etc/hosts``
2147 - Added network interface parameters
2148 - Added dry run mode to some LUs
2149 - Added RAPI resources:
2151 - ``/2/instances/[instance_name]/info``
2152 - ``/2/instances/[instance_name]/replace-disks``
2153 - ``/2/nodes/[node_name]/evacuate``
2154 - ``/2/nodes/[node_name]/migrate``
2155 - ``/2/nodes/[node_name]/role``
2156 - ``/2/nodes/[node_name]/storage``
2157 - ``/2/nodes/[node_name]/storage/modify``
2158 - ``/2/nodes/[node_name]/storage/repair``
2160 - Added OpCodes to evacuate or migrate all instances on a node
2161 - Added new command to list storage elements on nodes (``gnt-node
2162 list-storage``) and modify them (``gnt-node modify-storage``)
2163 - Added new ssconf files with master candidate IP address
2164 (``ssconf_master_candidates_ips``), node primary IP address
2165 (``ssconf_node_primary_ips``) and node secondary IP address
2166 (``ssconf_node_secondary_ips``)
2167 - Added ``ganeti-confd`` and a client library to query the Ganeti
2168 configuration via UDP
2169 - Added ability to run hooks after cluster initialization and before
2171 - Added automatic mode for disk replace (``gnt-instance replace-disks
2173 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2174 after catastrophic data-loss
2175 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2177 - Added ``gnt-instance move`` command to move instances
2178 - Added ``gnt-cluster watcher`` command to control watcher
2179 - Added ``gnt-node powercycle`` command to powercycle nodes
2180 - Added new job status field ``lock_status``
2181 - Added parseable error codes to cluster verification (``gnt-cluster
2182 verify --error-codes``) and made output less verbose (use
2183 ``--verbose`` to restore previous behaviour)
2184 - Added UUIDs to the main config entities (cluster, nodes, instances)
2185 - Added support for OS variants
2186 - Added support for hashed passwords in the Ganeti remote API users file
2188 - Added option to specify maximum timeout on instance shutdown
2189 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2190 - Added new helper script to start and stop Ganeti daemons
2191 (``daemon-util``), with the intent to reduce the work necessary to
2192 adjust Ganeti for non-Debian distributions and to start/stop daemons
2194 - Added more unittests
2195 - Fixed critical bug in ganeti-masterd startup
2196 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2197 now customisable at the cluster level for both the KVM and Xen
2198 hypervisors using the new ``migration_port`` parameter
2199 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2200 reinstalling an instance
2201 - Allowed ``@`` in tag names
2202 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2203 - Many documentation updates
2204 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2205 - ``gnt-instance reinstall`` can now reinstall multiple instances
2206 - Updated many command line parameters
2207 - Introduced new OS API version 15
2208 - No longer support a default hypervisor
2209 - Treat virtual LVs as inexistent
2210 - Improved job locking logic to reduce lock contention
2211 - Match instance and node names case insensitively
2212 - Reimplemented bash completion script to be more complete
2219 *(Released Thu, 4 Feb 2010)*
2221 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2222 - Fix a string formatting bug
2223 - Improve safety of the code in some error paths
2224 - Improve data validation in the master of values returned from nodes
2230 *(Released Thu, 17 Dec 2009)*
2232 - Fix security issue due to missing validation of iallocator names; this
2233 allows local and remote execution of arbitrary executables
2234 - Fix failure of gnt-node list during instance removal
2235 - Ship the RAPI documentation in the archive
2241 *(Released Wed, 30 Sep 2009)*
2243 - Fixed many wrong messages
2244 - Fixed a few bugs related to the locking library
2245 - Fixed MAC checking at instance creation time
2246 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2247 - Fixed a few issues related to signal handling in both daemons and
2249 - Fixed the example startup script provided
2250 - Fixed insserv dependencies in the example startup script (patch from
2252 - Fixed handling of drained nodes in the iallocator framework
2253 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2255 - Fixed error related to invalid job IDs in job polling
2256 - Fixed job/opcode persistence on unclean master shutdown
2257 - Fixed handling of partial job processing after unclean master
2259 - Fixed error reporting from LUs, previously all errors were converted
2260 into execution errors
2261 - Fixed error reporting from burnin
2262 - Decreased significantly the memory usage of the job queue
2263 - Optimised slightly multi-job submission
2264 - Optimised slightly opcode loading
2265 - Backported the multi-job submit framework from the development
2266 branch; multi-instance start and stop should be faster
2267 - Added script to clean archived jobs after 21 days; this will reduce
2268 the size of the queue directory
2269 - Added some extra checks in disk size tracking
2270 - Added an example ethers hook script
2271 - Added a cluster parameter that prevents Ganeti from modifying of
2273 - Added more node information to RAPI responses
2274 - Added a ``gnt-job watch`` command that allows following the ouput of a
2276 - Added a bind-address option to ganeti-rapi
2277 - Added more checks to the configuration verify
2278 - Enhanced the burnin script such that some operations can be retried
2280 - Converted instance reinstall to multi-instance model
2286 *(Released Fri, 7 Aug 2009)*
2288 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2289 to allow using the pre-2.0.2 behaviour in activation, if any existing
2290 instances have mismatched disk sizes in the configuration
2291 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2292 any configuration mismatches for disk sizes
2293 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2294 failover to work on two-node clusters
2295 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2297 - Fixed detection of OS script errors in ``gnt-backup export``
2298 - Fixed exit code of ``gnt-backup export``
2304 *(Released Fri, 17 Jul 2009)*
2306 - Added experimental support for stripped logical volumes; this should
2307 enhance performance but comes with a higher complexity in the block
2308 device handling; stripping is only enabled when passing
2309 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2310 affected even in the non-stripped mode
2311 - Improved resiliency against transient failures at the end of DRBD
2312 resyncs, and in general of DRBD resync checks
2313 - Fixed a couple of issues with exports and snapshot errors
2314 - Fixed a couple of issues in instance listing
2315 - Added display of the disk size in ``gnt-instance info``
2316 - Fixed checking for valid OSes in instance creation
2317 - Fixed handling of the "vcpus" parameter in instance listing and in
2318 general of invalid parameters
2319 - Fixed http server library, and thus RAPI, to handle invalid
2320 username/password combinations correctly; this means that now they
2321 report unauthorized for queries too, not only for modifications,
2322 allowing earlier detect of configuration problems
2323 - Added a new "role" node list field, equivalent to the master/master
2324 candidate/drained/offline flags combinations
2325 - Fixed cluster modify and changes of candidate pool size
2326 - Fixed cluster verify error messages for wrong files on regular nodes
2327 - Fixed a couple of issues with node demotion from master candidate role
2328 - Fixed node readd issues
2329 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2330 - Added a new ``--no-voting`` option for masterfailover to fix failover
2331 on two-nodes clusters when the former master node is unreachable
2332 - Added instance reinstall over RAPI
2338 *(Released Tue, 16 Jun 2009)*
2340 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2341 allow re-adding the start in single-user option (regression from 1.2)
2342 - the watcher writes the instance status to a file, to allow monitoring
2343 to report the instance status (from the master) based on cached
2344 results of the watcher's queries; while this can get stale if the
2345 watcher is being locked due to other work on the cluster, this is
2346 still an improvement
2347 - the watcher now also restarts the node daemon and the rapi daemon if
2349 - fixed the watcher to handle full and drained queue cases
2350 - hooks export more instance data in the environment, which helps if
2351 hook scripts need to take action based on the instance's properties
2352 (no longer need to query back into ganeti)
2353 - instance failovers when the instance is stopped do not check for free
2354 RAM, so that failing over a stopped instance is possible in low memory
2356 - rapi uses queries for tags instead of jobs (for less job traffic), and
2357 for cluster tags it won't talk to masterd at all but read them from
2359 - a couple of error handling fixes in RAPI
2360 - drbd handling: improved the error handling of inconsistent disks after
2361 resync to reduce the frequency of "there are some degraded disks for
2362 this instance" messages
2363 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2364 error handling path called a wrong function name)
2370 *(Released Wed, 27 May 2009)*
2372 - no changes from rc5
2378 *(Released Wed, 20 May 2009)*
2380 - fix a couple of bugs (validation, argument checks)
2381 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2382 - some small improvements to RAPI and IAllocator
2383 - make watcher automatically start the master daemon if down
2389 *(Released Mon, 27 Apr 2009)*
2391 - change the OS list to not require locks; this helps with big clusters
2392 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2393 volume group is broken
2394 - ``gnt-instance info``, without any arguments, doesn't run for all
2395 instances anymore; either pass ``--all`` or pass the desired
2396 instances; this helps against mistakes on big clusters where listing
2397 the information for all instances takes a long time
2398 - miscellaneous doc and man pages fixes
2404 *(Released Wed, 8 Apr 2009)*
2406 - Change the internal locking model of some ``gnt-node`` commands, in
2407 order to reduce contention (and blocking of master daemon) when
2408 batching many creation/reinstall jobs
2409 - Fixes to Xen soft reboot
2410 - No longer build documentation at build time, instead distribute it in
2411 the archive, in order to reduce the need for the whole docbook/rst
2418 *(Released Fri, 27 Mar 2009)*
2420 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2421 - Fix watcher startup sequence, improves the behaviour of busy clusters
2422 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2423 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2424 bind address and other places
2425 - Some documentation fixes and updates
2431 *(Released Mon, 2 Mar 2009)*
2433 - More documentation updates, now all docs should be more-or-less
2435 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2437 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2439 - Fix an issue related to $libdir/run/ganeti and cluster creation
2445 *(Released Thu, 19 Feb 2009)*
2447 - Xen PVM and KVM have switched the default value for the instance root
2448 disk to the first partition on the first drive, instead of the whole
2449 drive; this means that the OS installation scripts must be changed
2451 - Man pages have been updated
2452 - RAPI has been switched by default to HTTPS, and the exported functions
2453 should all work correctly
2454 - RAPI v1 has been removed
2455 - Many improvements to the KVM hypervisor
2456 - Block device errors are now better reported
2457 - Many other bugfixes and small improvements
2463 *(Released Mon, 26 Jan 2009)*
2465 - Version 2 is a general rewrite of the code and therefore the
2466 differences are too many to list, see the design document for 2.0 in
2467 the ``doc/`` subdirectory for more details
2468 - In this beta version there is not yet a migration path from 1.2 (there
2469 will be one in the final 2.0 release)
2470 - A few significant changes are:
2472 - all commands are executed by a daemon (``ganeti-masterd``) and the
2473 various ``gnt-*`` commands are just front-ends to it
2474 - all the commands are entered into, and executed from a job queue,
2475 see the ``gnt-job(8)`` manpage
2476 - the RAPI daemon supports read-write operations, secured by basic
2477 HTTP authentication on top of HTTPS
2478 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2479 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2480 to migrate to DRBD 8 first while still running Ganeti 1.2)
2481 - DRBD devices are using statically allocated minor numbers, which
2482 will be assigned to existing instances during the migration process
2483 - there is support for both Xen PVM and Xen HVM instances running on
2485 - KVM virtualization is supported too
2486 - file-based storage has been implemented, which means that it is
2487 possible to run the cluster without LVM and DRBD storage, for
2488 example using a shared filesystem exported from shared storage (and
2489 still have live migration)
2495 *(Released Tue, 13 Jan 2009)*
2497 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2498 - Reuse the old instance mac address by default on instance import, if
2499 the instance name is the same.
2500 - Handle situations in which the node info rpc returns incomplete
2502 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2503 - Improved version of batcher:
2505 - state file support
2506 - instance mac address support
2507 - support for HVM clusters/instances
2509 - Add an option to show the number of cpu sockets and nodes in
2511 - Support OSes that handle more than one version of the OS api (but do
2512 not change the current API in any other way)
2513 - Fix ``gnt-node migrate``
2514 - ``gnt-debug`` man page
2515 - Fixes various more typos and small issues
2516 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2522 *(Released Wed, 24 Sep 2008)*
2524 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2525 type of disk exported to fully virtualized instances.
2526 - provide access to the serial console of HVM instances
2527 - instance auto_balance flag, set by default. If turned off it will
2528 avoid warnings on cluster verify if there is not enough memory to fail
2529 over an instance. in the future it will prevent automatically failing
2530 it over when we will support that.
2531 - batcher tool for instance creation, see ``tools/README.batcher``
2532 - ``gnt-instance reinstall --select-os`` to interactively select a new
2533 operating system when reinstalling an instance.
2534 - when changing the memory amount on instance modify a check has been
2535 added that the instance will be able to start. also warnings are
2536 emitted if the instance will not be able to fail over, if auto_balance
2538 - documentation fixes
2539 - sync fields between ``gnt-instance list/modify/add/import``
2540 - fix a race condition in drbd when the sync speed was set after giving
2541 the device a remote peer.
2547 *(Released Tue, 22 Jul 2008)*
2549 - note: the allowed size and number of tags per object were reduced
2550 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2551 - fixed twisted 8.x compatibility
2552 - fixed ``gnt-instance replace-disks`` with iallocator
2553 - add TCP keepalives on twisted connections to detect restarted nodes
2554 - disk increase support, see ``gnt-instance grow-disk``
2555 - implement bulk node/instance query for RAPI
2556 - add tags in node/instance listing (optional)
2557 - experimental migration (and live migration) support, read the man page
2558 for ``gnt-instance migrate``
2559 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2560 has some small improvements in handling its state file
2566 *(Released Fri, 13 Jun 2008)*
2568 - Experimental readonly, REST-based remote API implementation;
2569 automatically started on master node, TCP port 5080, if enabled by
2570 ``--enable-rapi`` parameter to configure script.
2571 - Instance allocator support. Add and import instance accept a
2572 ``--iallocator`` parameter, and call that instance allocator to decide
2573 which node to use for the instance. The iallocator document describes
2574 what's expected from an allocator script.
2575 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2576 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2577 node is lost there is still enough memory to fail over the instances
2579 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2580 ``gnt-cluster verify``, to check for site-specific compliance. All the
2581 hooks will run, and their output, if any, will be displayed. Any
2582 failing hook will make the verification return an error value.
2583 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2584 primary and secondary interfaces
2585 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2586 that is still declared as part of the cluster and has failed.
2587 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2588 specifying output fields, that just adds the chosen fields to the
2590 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2591 export from the filesystem.
2592 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2593 have been added. Using them you can enable/disable acpi and pae
2594 support, and specify a path for a cd image to be exported to the
2595 instance. These parameters as the name suggest only work on HVM
2597 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2598 PAE support will be set to the previously hardcoded values, but the
2599 (previously hardcoded) path to the CDROM ISO image will be unset and
2600 if required, needs to be set manually with ``gnt-instance modify``
2602 - The address to which an instance's VNC console is bound is now
2603 selectable per-instance, rather than being cluster wide. Of course
2604 this only applies to instances controlled via VNC, so currently just
2605 applies to HVM clusters.
2611 *(Released Mon, 18 Feb 2008)*
2613 - more tweaks to the disk activation code (especially helpful for DRBD)
2614 - change the default ``gnt-instance list`` output format, now there is
2615 one combined status field (see the manpage for the exact values this
2617 - some more fixes for the mac export to hooks change
2618 - make Ganeti not break with DRBD 8.2.x (which changed the version
2619 format in ``/proc/drbd``) (issue 24)
2620 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2621 template, allowing migration from DRBD0.7+MD to DRBD8
2627 *(Released Wed, 30 Jan 2008)*
2629 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2631 - add command aliases infrastructure and a few aliases
2632 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2633 man pages and the ``--help`` option of ``gnt-node
2634 list``/``gnt-instance list``
2635 - fix ``gnt-backup list`` with down nodes (issue 21)
2636 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2637 - fix the dist archive and add a check for including svn/git files in
2639 - some developer-related changes: improve the burnin and the QA suite,
2640 add an upload script for testing during development
2646 *(Released Wed, 16 Jan 2008)*
2648 - experimental HVM support, read the install document, section
2649 "Initializing the cluster"
2650 - allow for the PVM hypervisor per-instance kernel and initrd paths
2651 - add a new command ``gnt-cluster verify-disks`` which uses a new
2652 algorithm to improve the reconnection of the DRBD pairs if the device
2653 on the secondary node has gone away
2654 - make logical volume code auto-activate LVs at disk activation time
2655 - slightly improve the speed of activating disks
2656 - allow specification of the MAC address at instance creation time, and
2657 changing it later via ``gnt-instance modify``
2658 - fix handling of external commands that generate lots of output on
2660 - update documentation with regard to minimum version of DRBD8 supported
2666 *(Released Tue, 4 Dec 2007)*
2668 - Log the ``xm create`` output to the node daemon log on failure (to
2669 help diagnosing the error)
2670 - In debug mode, log all external commands output if failed to the logs
2671 - Change parsing of lvm commands to ignore stderr
2677 *(Released Wed, 28 Nov 2007)*
2679 - Another round of updates to the DRBD 8 code to deal with more failures
2680 in the replace secondary node operation
2681 - Some more logging of failures in disk operations (lvm, drbd)
2682 - A few documentation updates
2689 *(Released Tue, 13 Nov 2007)*
2691 - Change configuration file format from Python's Pickle to JSON.
2692 Upgrading is possible using the cfgupgrade utility.
2693 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2694 faster replace disks and is more stable (DRBD 8 has many improvements
2695 compared to DRBD 0.7)
2696 - Added command line tags support (see man pages for ``gnt-instance``,
2697 ``gnt-node``, ``gnt-cluster``)
2698 - Added instance rename support
2699 - Added multi-instance startup/shutdown
2700 - Added cluster rename support
2701 - Added ``gnt-node evacuate`` to simplify some node operations
2702 - Added instance reboot operation that can speedup reboot as compared to
2704 - Soften the requirement that hostnames are in FQDN format
2705 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2707 - Removed dependency on debian's patched fping that uses the
2708 non-standard ``-S`` option
2709 - Now the OS definitions are searched for in multiple, configurable
2710 paths (easier for distros to package)
2711 - Some changes to the hooks infrastructure (especially the new
2712 post-configuration update hook)
2713 - Other small bugfixes
2715 .. vim: set textwidth=72 syntax=rst :