10 - The minimum Python version needed to run Ganeti is now 2.6.
11 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
12 to always require authentication, ``--require-authentication``. It can
13 be specified in ``$sysconfdir/default/ganeti``.
14 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
15 be used to manage the disk templates to be used by instances in the cluster.
16 Initially, it will be set to a list that includes plain, drbd, if they were
17 enabled by specifying a volume group name, and file and sharedfile, if those
18 were enabled at configure time. Additionally, it will include all disk
19 templates that are currently used by instances. The order of disk templates
20 will be based on Ganeti's history of supporting them. In the future, the
21 first entry of the list will be used as a default disk template on instance
23 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
24 configuration back to the previous stable version.
30 *(Released Fri, 24 May 2013)*
32 Incompatible/important changes
33 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
35 - Instance policies for disk size were documented to be on a per-disk
36 basis, but hail applied them to the sum of all disks. This has been
38 - ``hbal`` will now exit with status 0 if, during job execution over
39 LUXI, early exit has been requested and all jobs are successful;
40 before, exit status 1 was used, which cannot be differentiated from
42 - Compatibility with newer versions of rbd has been fixed
43 - ``gnt-instance batch-create`` has been changed to use the bulk create
44 opcode from Ganeti. This lead to incompatible changes in the format of
45 the JSON file. It's now not a custom dict anymore but a dict
46 compatible with the ``OpInstanceCreate`` opcode.
47 - Parent directories for file storage need to be listed in
48 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
49 write the file automatically based on old configuration values, but it
50 can not distribute it across all nodes and the file contents should be
51 verified. Use ``gnt-cluster copyfile
52 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
53 upgraded. The reason for requiring this list of paths now is that
54 before it would have been possible to inject new paths via RPC,
55 allowing files to be created in arbitrary locations. The RPC protocol
56 is protected using SSL/X.509 certificates, but as a design principle
57 Ganeti does not permit arbitrary paths to be passed.
58 - The parsing of the variants file for OSes (see
59 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
60 lines and comment lines (starting with ``#``) are ignored for better
62 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
63 longer available. ``gnt-node add`` now invokes a new tool on the
64 destination node, named ``prepare-node-join``, to configure the SSH
65 daemon. Paramiko is no longer necessary to configure nodes' SSH
66 daemons via ``gnt-node add``.
67 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
68 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
69 the flag is not reset after a master failover.
70 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
71 is recommended. 2.6 will be mandatory from the 2.8 series.
77 - New network management functionality to support automatic allocation
78 of IP addresses and managing of network parameters. See
79 :manpage:`gnt-network(8)` for more details.
80 - New external storage backend, to allow managing arbitrary storage
81 systems external to the cluster. See
82 :manpage:`ganeti-extstorage-interface(7)`.
83 - New ``exclusive-storage`` node parameter added, restricted to
84 nodegroup level. When it's set to true, physical disks are assigned in
85 an exclusive fashion to instances, as documented in :doc:`Partitioned
86 Ganeti <design-partitioned>`. Currently, only instances using the
87 ``plain`` disk template are supported.
88 - The KVM hypervisor has been updated with many new hypervisor
89 parameters, including a generic one for passing arbitrary command line
90 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
91 compatible up to qemu 1.4.
92 - A new tool, called ``mon-collector``, is the stand-alone executor of
93 the data collectors for a monitoring system. As of this version, it
94 just includes the DRBD data collector, that can be executed by calling
95 ``mon-collector`` using the ``drbd`` parameter. See
96 :manpage:`mon-collector(7)`.
97 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
98 for RAPI users. It allows granting permissions to query for
99 information to a specific user without giving
100 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
101 - A new tool named ``node-cleanup`` has been added. It cleans remains of
102 a cluster from a machine by stopping all daemons, removing
103 certificates and ssconf files. Unless the ``--no-backup`` option is
104 given, copies of the certificates are made.
105 - Instance creations now support the use of opportunistic locking,
106 potentially speeding up the (parallel) creation of multiple instances.
107 This feature is currently only available via the :doc:`RAPI
108 <rapi>` interface and when an instance allocator is used. If the
109 ``opportunistic_locking`` parameter is set the opcode will try to
110 acquire as many locks as possible, but will not wait for any locks
111 held by other opcodes. If not enough resources can be found to
112 allocate the instance, the temporary error code
113 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
114 retried thereafter, with or without opportunistic locking.
115 - New experimental linux-ha resource scripts.
116 - Restricted-commands support: ganeti can now be asked (via command line
117 or rapi) to perform commands on a node. These are passed via ganeti
118 RPC rather than ssh. This functionality is restricted to commands
119 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
120 reasons. The file is not copied automatically.
126 - Diskless instances are now externally mirrored (Issue 237). This for
127 now has only been tested in conjunction with explicit target nodes for
129 - Queries not needing locks or RPC access to the node can now be
130 performed by the confd daemon, making them independent from jobs, and
131 thus faster to execute. This is selectable at configure time.
132 - The functionality for allocating multiple instances at once has been
133 overhauled and is now also available through :doc:`RAPI <rapi>`.
137 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
139 - Disks added through ``gnt-instance modify`` or created through
140 ``gnt-instance recreate-disks`` are wiped, if the
141 ``prealloc_wipe_disks`` flag is set.
142 - If wiping newly created disks fails, the disks are removed. Also,
143 partial failures in creating disks through ``gnt-instance modify``
144 triggers a cleanup of the partially-created disks.
145 - Removing the master IP address doesn't fail if the address has been
147 - Fix ownership of the OS log dir
148 - Workaround missing SO_PEERCRED constant (Issue 191)
154 *(Released Fri, 3 May 2013)*
156 This was the first release candidate of the 2.7 series. Since beta3:
158 - Fix kvm compatibility with qemu 1.4 (Issue 389)
159 - Documentation updates (admin guide, upgrade notes, install
160 instructions) (Issue 372)
161 - Fix gnt-group list nodes and instances count (Issue 436)
162 - Fix compilation without non-mandatory libraries (Issue 441)
163 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
164 - Make confd logging more verbose at INFO level (Issue 435)
165 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
166 - Fix failure path for instance storage type conversion (Issue 229)
167 - Update htools text backend documentation
168 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
169 - Disable inter-cluster instance move for file-based instances, because
170 it is dependant on instance export, which is not supported for
171 file-based instances. (Issue 414)
172 - Fix gnt-job crashes on non-ascii characters (Issue 427)
173 - Fix volume group checks on non-vm-capable nodes (Issue 432)
179 *(Released Mon, 22 Apr 2013)*
181 This was the third beta release of the 2.7 series. Since beta2:
183 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
184 - Fix data loss on wrong usage of ``gnt-instance move``
185 - Properly export errors in confd-based job queries
186 - Add ``users-setup`` tool
187 - Fix iallocator protocol to report 0 as a disk size for diskless
188 instances. This avoids hail breaking when a diskless instance is
190 - Fix job queue directory permission problem that made confd job queries
191 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
192 for access to archived jobs (Issue 406).
193 - Limit the sizes of networks supported by ``gnt-network`` to something
194 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
195 - Fix bugs in instance disk template conversion
196 - Fix GHC 7 compatibility
197 - Fix ``burnin`` install path (Issue 426).
198 - Allow very small disk grows (Issue 347).
199 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
200 that noded doesn't import masterd code (Issue 419).
201 - Make sure the default metavg at cluster init is the same as the vg, if
202 unspecified (Issue 358).
203 - Fix cleanup of partially created disks (part of Issue 416)
209 *(Released Tue, 2 Apr 2013)*
211 This was the second beta release of the 2.7 series. Since beta1:
213 - Networks no longer have a "type" slot, since this information was
214 unused in Ganeti: instead of it tags should be used.
215 - The rapi client now has a ``target_node`` option to MigrateInstance.
216 - Fix early exit return code for hbal (Issue 386).
217 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
218 - Fix ``rbd showmapped`` output parsing (Issue 312).
219 - Networks are now referenced indexed by UUID, rather than name. This
220 will require running cfgupgrade, from 2.7.0beta1, if networks are in
222 - The OS environment now includes network information.
223 - Deleting of a network is now disallowed if any instance nic is using
224 it, to prevent dangling references.
225 - External storage is now documented in man pages.
226 - The exclusive_storage flag can now only be set at nodegroup level.
227 - Hbal can now submit an explicit priority with its jobs.
228 - Many network related locking fixes.
229 - Bump up the required pylint version to 0.25.1.
230 - Fix the ``no_remember`` option in RAPI client.
231 - Many ipolicy related tests, qa, and fixes.
232 - Many documentation improvements and fixes.
233 - Fix building with ``--disable-file-storage``.
234 - Fix ``-q`` option in htools, which was broken if passed more than
236 - Some haskell/python interaction improvements and fixes.
237 - Fix iallocator in case of missing LVM storage.
238 - Fix confd config load in case of ``--no-lvm-storage``.
239 - The confd/query functionality is now mentioned in the security
246 *(Released Wed, 6 Feb 2013)*
248 This was the first beta release of the 2.7 series. All important changes
249 are listed in the latest 2.7 entry.
255 *(Released Fri, 21 Dec 2012)*
257 Important behaviour change: hbal won't rebalance anymore instances which
258 have the ``auto_balance`` attribute set to false. This was the intention
259 all along, but until now it only skipped those from the N+1 memory
260 reservation (DRBD-specific).
262 A significant number of bug fixes in this release:
264 - Fixed disk adoption interaction with ipolicy checks.
265 - Fixed networking issues when instances are started, stopped or
266 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
267 - Fixed the warning in cluster verify for shared storage instances not
269 - Fixed removal of storage directory on shared file storage (issue 262).
270 - Fixed validation of LVM volume group name in OpClusterSetParams
271 (``gnt-cluster modify``) (issue 285).
272 - Fixed runtime memory increases (``gnt-instance modify -m``).
273 - Fixed live migration under Xen's ``xl`` mode.
274 - Fixed ``gnt-instance console`` with ``xl``.
275 - Fixed building with newer Haskell compiler/libraries.
276 - Fixed PID file writing in Haskell daemons (confd); this prevents
277 restart issues if confd was launched manually (outside of
278 ``daemon-util``) while another copy of it was running
279 - Fixed a type error when doing live migrations with KVM (issue 297) and
280 the error messages for failing migrations have been improved.
281 - Fixed opcode validation for the out-of-band commands (``gnt-node
283 - Fixed a type error when unsetting OS hypervisor parameters (issue
284 311); now it's possible to unset all OS-specific hypervisor
286 - Fixed the ``dry-run`` mode for many operations: verification of
287 results was over-zealous but didn't take into account the ``dry-run``
288 operation, resulting in "wrong" failures.
289 - Fixed bash completion in ``gnt-job list`` when the job queue has
290 hundreds of entries; especially with older ``bash`` versions, this
291 results in significant CPU usage.
293 And lastly, a few other improvements have been made:
295 - Added option to force master-failover without voting (issue 282).
296 - Clarified error message on lock conflict (issue 287).
297 - Logging of newly submitted jobs has been improved (issue 290).
298 - Hostname checks have been made uniform between instance rename and
300 - The ``--submit`` option is now supported by ``gnt-debug delay``.
301 - Shutting down the master daemon by sending SIGTERM now stops it from
302 processing jobs waiting for locks; instead, those jobs will be started
303 once again after the master daemon is started the next time (issue
305 - Support for Xen's ``xl`` program has been improved (besides the fixes
307 - Reduced logging noise in the Haskell confd daemon (only show one log
308 entry for each config reload, instead of two).
309 - Several man page updates and typo fixes.
315 *(Released Fri, 12 Oct 2012)*
317 A small bugfix release. Among the bugs fixed:
319 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
320 made the command unusable.
321 - Commands that issue many jobs don't fail anymore just because some jobs
322 take so long that other jobs are archived.
323 - Failures during ``gnt-instance reinstall`` are reflected by the exit
325 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
327 - When ``always_failover`` is set, ``--allow-failover`` is not required
328 in migrate commands anymore.
329 - ``bash_completion`` works even if extglob is disabled.
330 - Fixed bug with locks that made failover for RDB-based instances fail.
331 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
332 a random node instead of one based on the allocator metric.
333 - Support for newer versions of pylint and pep8.
334 - Hail doesn't fail anymore when trying to add an instance of type
335 ``file``, ``sharedfile`` or ``rbd``.
336 - Added new Makefile target to rebuild the whole distribution, so that
337 all files are included.
343 *(Released Fri, 27 Jul 2012)*
346 .. attention:: The ``LUXI`` protocol has been made more consistent
347 regarding its handling of command arguments. This, however, leads to
348 incompatibility issues with previous versions. Please ensure that you
349 restart Ganeti daemons soon after the upgrade, otherwise most
350 ``LUXI`` calls (job submission, setting/resetting the drain flag,
351 pausing/resuming the watcher, cancelling and archiving jobs, querying
352 the cluster configuration) will fail.
361 The current ``admin_up`` field, which used to denote whether an instance
362 should be running or not, has been removed. Instead, ``admin_state`` is
363 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
365 The rational behind this is that an instance being “down” can have
368 - it could be down during a reboot
369 - it could be temporarily be down for a reinstall
370 - or it could be down because it is deprecated and kept just for its
373 The previous Boolean state was making it difficult to do capacity
374 calculations: should Ganeti reserve memory for a down instance? Now, the
375 tri-state field makes it clear:
377 - in ``up`` and ``down`` state, all resources are reserved for the
378 instance, and it can be at any time brought up if it is down
379 - in ``offline`` state, only disk space is reserved for it, but not
382 The field can have an extra use: since the transition between ``up`` and
383 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
384 transition between ``offline`` and ``down`` is done via ``gnt-instance
385 modify``, it is possible to given different rights to users. For
386 example, owners of an instance could be allowed to start/stop it, but
387 not transition it out of the offline state.
389 Instance policies and specs
390 +++++++++++++++++++++++++++
392 In previous Ganeti versions, an instance creation request was not
393 limited on the minimum size and on the maximum size just by the cluster
394 resources. As such, any policy could be implemented only in third-party
395 clients (RAPI clients, or shell wrappers over ``gnt-*``
396 tools). Furthermore, calculating cluster capacity via ``hspace`` again
397 required external input with regards to instance sizes.
399 In order to improve these workflows and to allow for example better
400 per-node group differentiation, we introduced instance specs, which
403 - minimum instance disk size, disk count, memory size, cpu count
404 - maximum values for the above metrics
405 - and “standard” values (used in ``hspace`` to calculate the standard
408 The minimum/maximum values can be also customised at node-group level,
409 for example allowing more powerful hardware to support bigger instance
412 Beside the instance specs, there are a few other settings belonging to
413 the instance policy framework. It is possible now to customise, per
414 cluster and node-group:
416 - the list of allowed disk templates
417 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
418 - the maximum ratio of instance to spindles (see below for more
419 information) for local storage
421 All these together should allow all tools that talk to Ganeti to know
422 what are the ranges of allowed values for instances and the
423 over-subscription that is allowed.
425 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
426 instance configuration, and the physical CPU configuration from the
427 node. For the spindle ratios however, we didn't track before these
428 values, so new parameters have been added:
430 - a new node parameter ``spindle_count``, defaults to 1, customisable at
431 node group or node level
432 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
434 Note that spindles in this context doesn't need to mean actual
435 mechanical hard-drives; it's just a relative number for both the node
436 I/O capacity and instance I/O consumption.
438 Instance migration behaviour
439 ++++++++++++++++++++++++++++
441 While live-migration is in general desirable over failover, it is
442 possible that for some workloads it is actually worse, due to the
443 variable time of the “suspend” phase during live migration.
445 To allow the tools to work consistently over such instances (without
446 having to hard-code instance names), a new backend parameter
447 ``always_failover`` has been added to control the migration/failover
448 behaviour. When set to True, all migration requests for an instance will
449 instead fall-back to failover.
451 Instance memory ballooning
452 ++++++++++++++++++++++++++
454 Initial support for memory ballooning has been added. The memory for an
455 instance is no longer fixed (backend parameter ``memory``), but instead
456 can vary between minimum and maximum values (backend parameters
457 ``minmem`` and ``maxmem``). Currently we only change an instance's
460 - live migrating or failing over and instance and the target node
461 doesn't have enough memory
462 - user requests changing the memory via ``gnt-instance modify
468 In order to control the use of specific CPUs by instance, support for
469 controlling CPU pinning has been added for the Xen, HVM and LXC
470 hypervisors. This is controlled by a new hypervisor parameter
471 ``cpu_mask``; details about possible values for this are in the
472 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
473 VCPU-to-CPU mapping) form will work well only when all nodes in your
474 cluster have the same amount of CPUs.
479 Another area in which Ganeti was not customisable were the parameters
480 used for storage configuration, e.g. how many stripes to use for LVM,
481 DRBD resync configuration, etc.
483 To improve this area, we've added disks parameters, which are
484 customisable at cluster and node group level, and which allow to
485 specify various parameters for disks (DRBD has the most parameters
486 currently), for example:
488 - DRBD resync algorithm and parameters (e.g. speed)
489 - the default VG for meta-data volumes for DRBD
490 - number of stripes for LVM (plain disk template)
493 These parameters can be modified via ``gnt-cluster modify -D …`` and
494 ``gnt-group modify -D …``, and are used at either instance creation (in
495 case of LVM stripes, for example) or at disk “activation” time
498 Rados block device support
499 ++++++++++++++++++++++++++
501 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
502 denoted by the ``rbd`` disk template type. This is considered
503 experimental, feedback is welcome. For details on configuring it, see
504 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
509 The existing master IP functionality works well only in simple setups (a
510 single network shared by all nodes); however, if nodes belong to
511 different networks, then the ``/32`` setup and lack of routing
512 information is not enough.
514 To allow the master IP to function well in more complex cases, the
515 system was reworked as follows:
517 - a master IP netmask setting has been added
518 - the master IP activation/turn-down code was moved from the node daemon
520 - whether to run the Ganeti-supplied master IP script or a user-supplied
521 on is a ``gnt-cluster init`` setting
523 Details about the location of the standard and custom setup scripts are
524 in the man page :manpage:`gnt-cluster(8)`; for information about the
525 setup script protocol, look at the Ganeti-supplied script.
530 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
533 It is now possible to use TLS-protected connections, and when renewing
534 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
535 it is now possible to specify spice or spice CA certificates. Also, it
536 is possible to configure a password for SPICE sessions via the
537 hypervisor parameter ``spice_password_file``.
539 There are also new parameters to control the compression and streaming
540 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
541 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
542 for the spice parameters.
544 Lastly, it is now possible to see the SPICE connection information via
545 ``gnt-instance console``.
550 A new tool (``tools/ovfconverter``) has been added that supports
551 conversion between Ganeti and the `Open Virtualization Format
552 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
555 This relies on the ``qemu-img`` tool to convert the disk formats, so the
556 actual compatibility with other virtualization solutions depends on it.
561 The configuration query daemon (``ganeti-confd``) is now optional, and
562 has been rewritten in Haskell; whether to use the daemon at all, use the
563 Python (default) or the Haskell version is selectable at configure time
564 via the ``--enable-confd`` parameter, which can take one of the
565 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
566 daemon will result in a smaller footprint; for larger systems, we
567 welcome feedback on the Haskell version which might become the default
570 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
571 daemon running. The Python version doesn't implement the new call.
574 User interface changes
575 ~~~~~~~~~~~~~~~~~~~~~~
577 We have replaced the ``--disks`` option of ``gnt-instance
578 replace-disks`` with a more flexible ``--disk`` option, which allows
579 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
580 disk size and mode can be changed upon recreation (via ``gnt-instance
581 recreate-disks``, which accepts the same ``--disk`` option).
583 As many people are used to a ``show`` command, we have added that as an
584 alias to ``info`` on all ``gnt-*`` commands.
586 The ``gnt-instance grow-disk`` command has a new mode in which it can
587 accept the target size of the disk, instead of the delta; this can be
588 more safe since two runs in absolute mode will be idempotent, and
589 sometimes it's also easier to specify the desired size directly.
591 Also the handling of instances with regard to offline secondaries has
592 been improved. Instance operations should not fail because one of it's
593 secondary nodes is offline, even though it's safe to proceed.
595 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
596 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
597 minors to instance name.
602 RAPI coverage has improved, with (for example) new resources for
603 recreate-disks, node power-cycle, etc.
608 There is partial support for ``xl`` in the Xen hypervisor; feedback is
611 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
612 whether to still support Python 2.4 or move to Python 2.6 as minimum
615 Support for Fedora has been slightly improved; the provided example
616 init.d script should work better on it and the INSTALL file should
617 document the needed dependencies.
622 The deprecated ``QueryLocks`` LUXI request has been removed. Use
623 ``Query(what=QR_LOCK, ...)`` instead.
625 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
626 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
627 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
628 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
629 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
631 RAPI client: ``CertificateError`` now derives from
632 ``GanetiApiError``. This should make it more easy to handle Ganeti
635 Deprecation warnings due to PyCrypto/paramiko import in
636 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
637 make sure to run an up-to-date paramiko version, if you use this tool.
639 The QA scripts now depend on Python 2.5 or above (the main code base
640 still works with Python 2.4).
642 The configuration file (``config.data``) is now written without
643 indentation for performance reasons; if you want to edit it, it can be
644 re-formatted via ``tools/fmtjson``.
646 A number of bugs has been fixed in the cluster merge tool.
648 ``x509`` certification verification (used in import-export) has been
649 changed to allow the same clock skew as permitted by the cluster
650 verification. This will remove some rare but hard to diagnose errors in
657 *(Released Thu, 19 Jul 2012)*
659 Very few changes from rc4 to the final release, only bugfixes:
661 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
662 instance, fix CDROM booting for KVM instances)
663 - fixed node group modification of node parameters
664 - fixed issue in LUClusterVerifyGroup with multi-group clusters
665 - fixed generation of bash completion to ensure a stable ordering
672 *(Released Fri, 13 Jul 2012)*
674 Third release candidate for 2.6. The following changes were done from
677 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
678 - Fixed an inconsistency in the LUXI protocol with the provided
679 arguments (NOT backwards compatible)
680 - Fixed a bug with node groups ipolicy where ``min`` was greater than
681 the cluster ``std`` value
682 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
683 easier instance debugging on nodes (requires ``hconfd`` to work)
689 *(Released Tue, 03 Jul 2012)*
691 Second release candidate for 2.6. The following changes were done from
694 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
696 - Fixed a RAPI regression on missing beparams/memory
697 - Fixed redistribution of files on offline nodes
698 - Added possibility to run activate-disks even though secondaries are
699 offline. With this change it relaxes also the strictness on some other
700 commands which use activate disks internally:
701 * ``gnt-instance start|reboot|rename|backup|export``
702 - Made it possible to remove safely an instance if its secondaries are
704 - Made it possible to reinstall even though secondaries are offline
710 *(Released Mon, 25 Jun 2012)*
712 First release candidate for 2.6. The following changes were done from
715 - Fixed bugs with disk parameters and ``rbd`` templates as well as
717 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
718 behaviour. It supports now the modify operation
719 - ``hcheck`` implemented to analyze cluster health and possibility of
720 improving health by rebalance
721 - ``hbal`` has been improved in dealing with split instances
727 *(Released Mon, 11 Jun 2012)*
729 Second beta release of 2.6. The following changes were done from beta2
732 - Fixed ``daemon-util`` with non-root user models
733 - Fixed creation of plain instances with ``--no-wait-for-sync``
734 - Fix wrong iv_names when running ``cfgupgrade``
735 - Export more information in RAPI group queries
736 - Fixed bug when changing instance network interfaces
737 - Extended burnin to do NIC changes
738 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
739 - Changed default for DRBD barriers
740 - Fixed DRBD error reporting for syncer rate
741 - Verify the options on disk parameters
743 And of course various fixes to documentation and improved unittests and
750 *(Released Wed, 23 May 2012)*
752 First beta release of 2.6. The following changes were done from beta1 to
755 - integrated patch for distributions without ``start-stop-daemon``
756 - adapted example init.d script to work on Fedora
757 - fixed log handling in Haskell daemons
758 - adapted checks in the watcher for pycurl linked against libnss
759 - add partial support for ``xl`` instead of ``xm`` for Xen
760 - fixed a type issue in cluster verification
761 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
764 Plus integrated fixes from the 2.5 branch:
766 - fixed ``kvm-ifup`` to use ``/bin/bash``
767 - fixed parallel build failures
768 - KVM live migration when using a custom keymap
774 *(Released Tue, 24 Jul 2012)*
776 A small bugfix release, with no new features:
778 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
779 different default shell (e.g. Debian, Ubuntu)
780 - fixed KVM startup and live migration with a custom keymap (fixes Issue
781 243 and Debian bug #650664)
782 - fixed compatibility with KVM versions that don't support multiple boot
783 devices (fixes Issue 230 and Debian bug #624256)
785 Additionally, a few fixes were done to the build system (fixed parallel
786 build failures) and to the unittests (fixed race condition in test for
787 FileID functions, and the default enable/disable mode for QA test is now
794 *(Released Fri, 11 May 2012)*
796 A small bugfix release.
798 The main issues solved are on the topic of compatibility with newer LVM
801 - fixed parsing of ``lv_attr`` field
802 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
803 the ``--force`` flag is needed
805 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
806 to accept kernel 3.x too (was hardcoded to 2.6.*).
808 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
809 of hook results and that also made display of other errors suboptimal
810 was fixed; the code behaves now like 2.4 and earlier.
812 Another change in 2.5, the cleanup of the OS scripts environment, is too
813 aggressive: it removed even the ``PATH`` variable, which requires the OS
814 scripts to *always* need to export it. Since this is a bit too strict,
815 we now export a minimal PATH, the same that we export for hooks.
817 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
818 integrated into this release.
820 Finally, a few other miscellaneous changes were done (no new features,
821 just small improvements):
823 - Fix ``gnt-group --help`` display
824 - Fix hardcoded Xen kernel path
825 - Fix grow-disk handling of invalid units
826 - Update synopsis for ``gnt-cluster repair-disk-sizes``
827 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
833 *(Released Thu, 12 Apr 2012)*
835 Incompatible/important changes and bugfixes
836 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
838 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
839 resource's ``ip_check`` parameter changed from ``True`` to ``False``
840 to match the underlying LUXI interface.
841 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
842 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
843 not maintain backwards-compatibility as the underlying operation
844 changed in an incompatible way. The RAPI client can talk to old
845 servers, but it needs to be told so as the return value changed.
846 - When creating file-based instances via RAPI, the ``file_driver``
847 parameter no longer defaults to ``loop`` and must be specified.
848 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
850 - Support for the undocumented and deprecated RAPI instance creation
851 request format version 0 has been dropped. Use version 1, supported
852 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
853 - Pyparsing 1.4.6 or above is required, see :doc:`installation
854 documentation <install>`.
855 - The "cluster-verify" hooks are now executed per group by the
856 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
857 if you just run ``gnt-cluster verify``, which generates one opcode per
859 - The environment as passed to the OS scripts is cleared, and thus no
860 environment variables defined in the node daemon's environment will be
861 inherited by the scripts.
862 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
864 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
865 support operations involving multiple node groups.
866 - Offline nodes are ignored when failing over an instance.
867 - Support for KVM version 1.0, which changed the version reporting format
869 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
871 - ``Makefile`` is now compatible with Automake 1.11.2
872 - Includes all bugfixes made in the 2.4 series
877 - The ganeti-htools project has been merged into the ganeti-core source
878 tree and will be built as part of Ganeti (see :doc:`install-quick`).
879 - Implemented support for :doc:`shared storage <design-shared-storage>`.
880 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
881 GPT-style partition tables (requires `parted
882 <http://www.gnu.org/s/parted/>`_).
883 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
884 - Allowed adding tags on instance creation.
885 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
886 - Allow instances to be started in a paused state, enabling the user to
887 see the complete console output on boot using the console.
888 - Added new hypervisor flag to control default reboot behaviour
889 (``reboot_behavior``).
890 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
891 - Improved out-of-band management support:
893 - Added ``gnt-node health`` command reporting the health status of
895 - Added ``gnt-node power`` command to manage power status of nodes.
896 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
898 - Instance migration can fall back to failover if instance is not
900 - Filters can be used when listing nodes, instances, groups and locks;
901 see :manpage:`ganeti(7)` manpage.
902 - Added post-execution status as variables to :doc:`hooks <hooks>`
904 - Instance tags are exported/imported together with the instance.
905 - When given an explicit job ID, ``gnt-job info`` will work for archived
907 - Jobs can define dependencies on other jobs (not yet supported via
908 RAPI or command line, but used by internal commands and usable via
911 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
914 - Instance failover is now available as a RAPI resource
915 (``/2/instances/[instance_name]/failover``).
916 - ``gnt-instance info`` defaults to static information if primary node
918 - Opcodes have a new ``comment`` attribute.
919 - Added basic SPICE support to KVM hypervisor.
920 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
922 Node group improvements
923 ~~~~~~~~~~~~~~~~~~~~~~~
925 - ``gnt-cluster verify`` has been modified to check groups separately,
926 thereby improving performance.
927 - Node group support has been added to ``gnt-cluster verify-disks``,
928 which now operates per node group.
929 - Watcher has been changed to work better with node groups.
931 - One process and state file per node group.
932 - Slow watcher in one group doesn't block other group's watcher.
934 - Added new command, ``gnt-group evacuate``, to move all instances in a
935 node group to other groups.
936 - Added ``gnt-instance change-group`` to move an instance to another
938 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
939 per-group operations.
940 - Node groups can be tagged.
941 - Some operations switch from an exclusive to a shared lock as soon as
943 - Instance's primary and secondary nodes' groups are now available as
944 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
945 and ``snodes.group.uuid``).
950 - Numerous updates to documentation and manpages.
952 - :doc:`RAPI <rapi>` documentation now has detailed parameter
954 - Some opcode/job results are now also documented, see :doc:`RAPI
957 - A lockset's internal lock is now also visible in lock monitor.
958 - Log messages from job queue workers now contain information about the
959 opcode they're processing.
960 - ``gnt-instance console`` no longer requires the instance lock.
961 - A short delay when waiting for job changes reduces the number of LUXI
962 requests significantly.
963 - DRBD metadata volumes are overwritten with zeros during disk creation.
964 - Out-of-band commands no longer acquire the cluster lock in exclusive
966 - ``devel/upload`` now uses correct permissions for directories.
972 *(Released Fri, 23 Mar 2012)*
974 This was the sixth release candidate of the 2.5 series.
980 *(Released Mon, 9 Jan 2012)*
982 This was the fifth release candidate of the 2.5 series.
988 *(Released Thu, 27 Oct 2011)*
990 This was the fourth release candidate of the 2.5 series.
996 *(Released Wed, 26 Oct 2011)*
998 This was the third release candidate of the 2.5 series.
1004 *(Released Tue, 18 Oct 2011)*
1006 This was the second release candidate of the 2.5 series.
1012 *(Released Tue, 4 Oct 2011)*
1014 This was the first release candidate of the 2.5 series.
1020 *(Released Wed, 31 Aug 2011)*
1022 This was the third beta release of the 2.5 series.
1028 *(Released Mon, 22 Aug 2011)*
1030 This was the second beta release of the 2.5 series.
1036 *(Released Fri, 12 Aug 2011)*
1038 This was the first beta release of the 2.5 series.
1044 *(Released Thu, 27 Oct 2011)*
1046 - Fixed bug when parsing command line parameter values ending in
1048 - Fixed assertion error after unclean master shutdown
1049 - Disable HTTP client pool for RPC, significantly reducing memory usage
1051 - Fixed queue archive creation with wrong permissions
1057 *(Released Tue, 23 Aug 2011)*
1061 - Fixed documentation for importing with ``--src-dir`` option
1062 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1063 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1069 *(Released Fri, 5 Aug 2011)*
1071 Many bug-fixes and a few small features:
1073 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1074 issues when you tried to add an instance with two MAC addresses in one
1076 - KVM: fixed per-instance stored UID value
1077 - KVM: configure bridged NICs at migration start
1078 - KVM: Fix a bug where instance will not start with never KVM versions
1080 - Added OS search path to ``gnt-cluster info``
1081 - Fixed an issue with ``file_storage_dir`` where you were forced to
1082 provide an absolute path, but the documentation states it is a
1083 relative path, the documentation was right
1084 - Added a new parameter to instance stop/start called ``--no-remember``
1085 that will make the state change to not be remembered
1086 - Implemented ``no_remember`` at RAPI level
1087 - Improved the documentation
1088 - Node evacuation: don't call IAllocator if node is already empty
1089 - Fixed bug in DRBD8 replace disks on current nodes
1090 - Fixed bug in recreate-disks for DRBD instances
1091 - Moved assertion checking locks in ``gnt-instance replace-disks``
1092 causing it to abort with not owning the right locks for some situation
1093 - Job queue: Fixed potential race condition when cancelling queued jobs
1094 - Fixed off-by-one bug in job serial generation
1095 - ``gnt-node volumes``: Fix instance names
1096 - Fixed aliases in bash completion
1097 - Fixed a bug in reopening log files after being sent a SIGHUP
1098 - Added a flag to burnin to allow specifying VCPU count
1099 - Bugfixes to non-root Ganeti configuration
1105 *(Released Thu, 12 May 2011)*
1107 Many bug-fixes and a few new small features:
1109 - Fixed a bug related to log opening failures
1110 - Fixed a bug in instance listing with orphan instances
1111 - Fixed a bug which prevented resetting the cluster-level node parameter
1112 ``oob_program`` to the default
1113 - Many fixes related to the ``cluster-merge`` tool
1114 - Fixed a race condition in the lock monitor, which caused failures
1115 during (at least) creation of many instances in parallel
1116 - Improved output for gnt-job info
1117 - Removed the quiet flag on some ssh calls which prevented debugging
1119 - Improved the N+1 failure messages in cluster verify by actually
1120 showing the memory values (needed and available)
1121 - Increased lock attempt timeouts so that when executing long operations
1122 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1123 too early and thus prevent the use of the 'fair' mechanism
1124 - Changed instance query data (``gnt-instance info``) to not acquire
1125 locks unless needed, thus allowing its use on locked instance if only
1126 static information is asked for
1127 - Improved behaviour with filesystems that do not support rename on an
1129 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1130 kept locks on all nodes during the wipe, which is unneeded
1131 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1132 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1133 10GiB) which caused the wipe to fail right at the end in some cases
1134 - Fixed master IP activation when doing master failover with no-voting
1135 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1136 the master node itself
1137 - Fixed potential data-loss in under disk full conditions, where Ganeti
1138 wouldn't check correctly the return code and would consider
1139 partially-written files 'correct'
1140 - Fixed bug related to multiple VGs and DRBD disk replacing
1141 - Added new disk parameter ``metavg`` that allows placement of the meta
1142 device for DRBD in a different volume group
1143 - Fixed error handling in the node daemon when the system libc doesn't
1144 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1145 - Fixed lock release during replace-disks, which kept cluster-wide locks
1146 when doing disk replaces with an iallocator script
1147 - Added check for missing bridges in cluster verify
1148 - Handle EPIPE errors while writing to the terminal better, so that
1149 piping the output to e.g. ``less`` doesn't cause a backtrace
1150 - Fixed rare case where a ^C during Luxi calls could have been
1151 interpreted as server errors, instead of simply terminating
1152 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1154 - Added a few more parameters to the KVM hypervisor, allowing a second
1155 CDROM, custom disk type for CDROMs and a floppy image
1156 - Removed redundant message in instance rename when the name is given
1158 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1159 disks on new nodes, allowing recreation when the original instance
1160 nodes are completely gone
1161 - Added option when converting disk templates to DRBD to skip waiting
1162 for the resync, in order to make the instance available sooner
1163 - Added two new variables to the OS scripts environment (containing the
1165 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1166 to allow use of ``pvgrub`` as bootloader
1167 - Changed the instance memory modifications to only check out-of-memory
1168 conditions on memory increases, and turned the secondary node warnings
1169 into errors (they can still be overridden via ``--force``)
1170 - Fixed the handling of a corner case when the Python installation gets
1171 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1172 to execute a command that doesn't exist
1173 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1174 node of the instance returned failures during instance shutdown; this
1175 adds the option ``--ignore-consistency`` to gnt-instance move
1177 And as usual, various improvements to the error messages, documentation
1184 *(Released Wed, 09 Mar 2011)*
1186 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1187 the RAPI users file if run twice (even with ``--dry-run``).
1189 The release fixes that bug (nothing else changed).
1195 *(Released Mon, 07 Mar 2011)*
1197 Final 2.4.0 release. Just a few small fixes:
1199 - Fixed RAPI node evacuate
1200 - Fixed the kvm-ifup script
1201 - Fixed internal error handling for special job cases
1202 - Updated man page to specify the escaping feature for options
1208 *(Released Mon, 28 Feb 2011)*
1210 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1211 that this feature wiped the disks of the wrong instance, leading to loss
1216 - Fixed title of query field containing instance name
1217 - Expanded the glossary in the documentation
1218 - Fixed one unittest (internal issue)
1224 *(Released Mon, 21 Feb 2011)*
1226 A number of bug fixes plus just a couple functionality changes.
1228 On the user-visible side, the ``gnt-* list`` command output has changed
1229 with respect to "special" field states. The current rc1 style of display
1230 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1231 the default output mode special fields states are displayed as follows:
1233 - Offline resource: ``*``
1234 - Unavailable/not applicable: ``-``
1235 - Data missing (RPC failure): ``?``
1236 - Unknown field: ``??``
1238 Another user-visible change is the addition of ``--force-join`` to
1243 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1244 - Fixed regression in RAPI/instance reinstall where all parameters were
1245 required (instead of optional)
1246 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1247 - Fixed iallocator usage (offline nodes were not considered offline)
1248 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1249 - Fixed hypervisor and OS parameter validation with respect to
1250 non-vm_capable nodes
1251 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1253 - Fixed ``tools/listrunner`` with respect to agent-based usage
1259 *(Released Fri, 4 Feb 2011)*
1261 Many changes and fixes since the beta1 release. While there were some
1262 internal changes, the code has been mostly stabilised for the RC
1265 Note: the dumb allocator was removed in this release, as it was not kept
1266 up-to-date with the IAllocator protocol changes. It is recommended to
1267 use the ``hail`` command from the ganeti-htools package.
1269 Note: the 2.4 and up versions of Ganeti are not compatible with the
1270 0.2.x branch of ganeti-htools. You need to upgrade to
1271 ganeti-htools-0.3.0 (or later).
1273 Regressions fixed from 2.3
1274 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1276 - Fixed the ``gnt-cluster verify-disks`` command
1277 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1279 - Fixed disk adoption breakage
1280 - Fixed wrong headers in instance listing for field aliases
1285 - Fixed corner case in KVM handling of NICs
1286 - Fixed many cases of wrong handling of non-vm_capable nodes
1287 - Fixed a bug where a missing instance symlink was not possible to
1288 recreate with any ``gnt-*`` command (now ``gnt-instance
1289 activate-disks`` does it)
1290 - Fixed the volume group name as reported by ``gnt-cluster
1292 - Increased timeouts for the import-export code, hopefully leading to
1293 fewer aborts due network or instance timeouts
1294 - Fixed bug in ``gnt-node list-storage``
1295 - Fixed bug where not all daemons were started on cluster
1296 initialisation, but only at the first watcher run
1297 - Fixed many bugs in the OOB implementation
1298 - Fixed watcher behaviour in presence of instances with offline
1300 - Fixed instance list output for instances running on the wrong node
1301 - a few fixes to the cluster-merge tool, but it still cannot merge
1302 multi-node groups (currently it is not recommended to use this tool)
1308 - Improved network configuration for the KVM hypervisor
1309 - Added e1000 as a supported NIC for Xen-HVM
1310 - Improved the lvmstrap tool to also be able to use partitions, as
1311 opposed to full disks
1312 - Improved speed of disk wiping (the cluster parameter
1313 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1314 of instance creations
1315 - Added documentation for the OS parameters
1316 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1317 hypervisor is not responding
1318 - Added display of blacklisted and hidden OS information in
1319 ``gnt-cluster info``
1320 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1321 NIC and node parameters, which might create problems with currently
1322 invalid (but undetected) configuration files, but prevents validation
1323 failures when unrelated parameters are modified
1324 - Changed cluster initialisation to wait for the master daemon to become
1326 - Expanded the RAPI interface:
1328 - Added config redistribution resource
1329 - Added activation/deactivation of instance disks
1330 - Added export of console information
1332 - Implemented log file reopening on SIGHUP, which allows using
1333 logrotate(8) for the Ganeti log files
1334 - Added a basic OOB helper script as an example
1340 *(Released Fri, 14 Jan 2011)*
1345 - Fixed timezone issues when formatting timestamps
1346 - Added support for node groups, available via ``gnt-group`` and other
1348 - Added out-of-band framework and management, see :doc:`design
1349 document <design-oob>`
1350 - Removed support for roman numbers from ``gnt-node list`` and
1351 ``gnt-instance list``.
1352 - Allowed modification of master network interface via ``gnt-cluster
1353 modify --master-netdev``
1354 - Accept offline secondaries while shutting down instance disks
1355 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1356 - Added support for multiple LVM volume groups
1357 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1359 - Added commands to list available fields:
1361 - ``gnt-node list-fields``
1362 - ``gnt-group list-fields``
1363 - ``gnt-instance list-fields``
1365 - Updated documentation and man pages
1370 - Moved ``rapi_users`` file into separate directory, now named
1371 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1373 - Added new tool for running commands on many machines,
1374 ``tools/ganeti-listrunner``
1375 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1376 improving the ``gnt-instance console`` output
1377 - Allowed customisation of disk index separator at ``configure`` time
1378 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1379 - Added support for non-partitioned md disks in ``lvmstrap``
1380 - Added script to gracefully power off KVM instances
1381 - Split ``utils`` module into smaller parts
1382 - Changed query operations to return more detailed information, e.g.
1383 whether an information is unavailable due to an offline node. To use
1384 this new functionality, the LUXI call ``Query`` must be used. Field
1385 information is now stored by the master daemon and can be retrieved
1386 using ``QueryFields``. Instances, nodes and groups can also be queried
1387 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1388 exposed via RAPI). The following commands make use of this
1389 infrastructure change:
1391 - ``gnt-group list``
1392 - ``gnt-group list-fields``
1394 - ``gnt-node list-fields``
1395 - ``gnt-instance list``
1396 - ``gnt-instance list-fields``
1397 - ``gnt-debug locks``
1402 - New RAPI resources (see :doc:`rapi`):
1406 - ``/2/groups/[group_name]``
1407 - ``/2/groups/[group_name]/assign-nodes``
1408 - ``/2/groups/[group_name]/modify``
1409 - ``/2/groups/[group_name]/rename``
1410 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1414 - Implemented ``no_install`` for instance creation
1415 - Implemented OS parameters for instance reinstallation, allowing
1416 use of special settings on reinstallation (e.g. for preserving data)
1421 - Added IPv6 support in import/export
1422 - Pause DRBD synchronization while wiping disks on instance creation
1423 - Updated unittests and QA scripts
1424 - Improved network parameters passed to KVM
1425 - Converted man pages from docbook to reStructuredText
1431 *(Released Mon, 20 Dec 2010)*
1433 Released version 2.3.1~rc1 without any changes.
1439 *(Released Wed, 1 Dec 2010)*
1441 - impexpd: Disable OpenSSL compression in socat if possible (backport
1442 from master, commit e90739d625b, see :doc:`installation guide
1443 <install-quick>` for details)
1444 - Changed unittest coverage report to exclude test scripts
1445 - Added script to check version format
1451 *(Released Wed, 1 Dec 2010)*
1453 Released version 2.3.0~rc1 without any changes.
1459 *(Released Fri, 19 Nov 2010)*
1461 A number of bugfixes and documentation updates:
1463 - Update ganeti-os-interface documentation
1464 - Fixed a bug related to duplicate MACs or similar items which should be
1466 - Fix breakage in OS state modify
1467 - Reinstall instance: disallow offline secondaries (fixes bug related to
1468 OS changing but reinstall failing)
1469 - plus all the other fixes between 2.2.1 and 2.2.2
1475 *(Released Tue, 2 Nov 2010)*
1477 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1478 - Fixed master failover race with watcher
1479 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1481 - Accept previously stopped instance for export with instance removal
1482 - Simplify and extend the environment variables for instance OS scripts
1483 - Added new node flags, ``master_capable`` and ``vm_capable``
1484 - Added optional instance disk wiping prior during allocation. This is a
1485 cluster-wide option and can be set/modified using
1486 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1487 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1488 :doc:`install-quick`
1489 - Added a new watcher option (``--ignore-pause``)
1490 - Added option to ignore offline node on instance start/stop
1491 (``--ignore-offline``)
1492 - Allow overriding OS parameters with ``gnt-instance reinstall``
1493 - Added ability to change node's secondary IP address using ``gnt-node
1495 - Implemented privilege separation for all daemons except
1496 ``ganeti-noded``, see ``configure`` options
1497 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1499 - Implemented job priorities (see ``ganeti(7)`` manpage)
1500 - Ignore failures while shutting down instances during failover from
1502 - Exit daemon's bootstrap process only once daemon is ready
1503 - Export more information via ``LUInstanceQuery``/remote API
1504 - Improved documentation, QA and unittests
1505 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1506 restart if the file was created or changed
1507 - Added LUXI protocol version sent with each request and response,
1508 allowing detection of server/client mismatches
1509 - Moved the Python scripts among gnt-* and ganeti-* into modules
1510 - Moved all code related to setting up SSH to an external script,
1512 - Infrastructure changes for node group support in future versions
1518 *(Released Fri, 19 Nov 2010)*
1520 A few small bugs fixed, and some improvements to the build system:
1522 - Fix documentation regarding conversion to drbd
1523 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1525 - Fix error handling in node modify with multiple changes
1526 - Allow remote imports without checked names
1532 *(Released Tue, 19 Oct 2010)*
1534 - Disable SSL session ID cache in RPC client
1540 *(Released Thu, 14 Oct 2010)*
1542 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1543 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1545 - Fix problems with interaction between Curl and Python's HTTP server,
1546 resulting in increased speed in many RPC calls
1547 - Improve our release script to prevent breakage with older aclocal and
1554 *(Released Thu, 7 Oct 2010)*
1556 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1557 value retrieved from master
1558 - Added support for blacklisted or hidden OS definitions
1559 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1560 - Added support for -mem-path in KVM hypervisor abstraction layer
1561 - Allow overriding instance parameters in tool for inter-cluster
1562 instance moves (``tools/move-instance``)
1563 - Improved opcode summaries (e.g. in ``gnt-job list``)
1564 - Improve consistency of OS listing by sorting it
1565 - Documentation updates
1571 *(Released Fri, 8 Oct 2010)*
1573 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1579 *(Released Mon, 4 Oct 2010)*
1581 - Fixed regression in ``gnt-instance rename``
1587 *(Released Wed, 22 Sep 2010)*
1589 - Fixed OS_VARIANT variable for OS scripts
1590 - Fixed cluster tag operations via RAPI
1591 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1592 - Disabled RAPI CA checks in watcher
1598 *(Released Mon, 23 Aug 2010)*
1600 - Support DRBD versions of the format "a.b.c.d"
1602 - Re-introduce support for usage from multiple threads in RAPI client
1603 - Instance renames and modify via RAPI
1604 - Work around race condition between processing and archival in job
1606 - Mark opcodes following failed one as failed, too
1607 - Job field ``lock_status`` was removed due to difficulties making it
1608 work with the changed job queue in Ganeti 2.2; a better way to monitor
1609 locks is expected for a later 2.2.x release
1610 - Fixed dry-run behaviour with many commands
1611 - Support ``ssh-agent`` again when adding nodes
1612 - Many additional bugfixes
1618 *(Released Fri, 30 Jul 2010)*
1620 Important change: the internal RPC mechanism between Ganeti nodes has
1621 changed from using a home-grown http library (based on the Python base
1622 libraries) to use the PycURL library. This requires that PycURL is
1623 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1624 against GnuTLS by default. cURL's support for GnuTLS had known issues
1625 before cURL 7.21.0 and we recommend using the latest cURL release or
1626 linking against OpenSSL. Most other distributions already link PycURL
1627 and cURL against OpenSSL. The command::
1629 python -c 'import pycurl; print pycurl.version'
1631 can be used to determine the libraries PycURL and cURL are linked
1634 Other significant changes:
1636 - Rewrote much of the internals of the job queue, in order to achieve
1637 better parallelism; this decouples job query operations from the job
1638 processing, and it should allow much nicer behaviour of the master
1639 daemon under load, and it also has uncovered some long-standing bugs
1640 related to the job serialisation (now fixed)
1641 - Added a default iallocator setting to the cluster parameters,
1642 eliminating the need to always pass nodes or an iallocator for
1643 operations that require selection of new node(s)
1644 - Added experimental support for the LXC virtualization method
1645 - Added support for OS parameters, which allows the installation of
1646 instances to pass parameter to OS scripts in order to customise the
1648 - Added a hypervisor parameter controlling the migration type (live or
1649 non-live), since hypervisors have various levels of reliability; this
1650 has renamed the 'live' parameter to 'mode'
1651 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1652 logical volumes, meaning that cluster verify will ignore them and not
1653 flag their presence as errors
1654 - The watcher will now reset the error count for failed instances after
1655 8 hours, thus allowing self-healing if the problem that caused the
1656 instances to be down/fail to start has cleared in the meantime
1657 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1658 check for, and warn, if the drbd module parameter ``usermode_helper``
1659 is not consistent with the cluster-wide setting; this is needed to
1660 make diagnose easier of failed drbd creations
1661 - Started adding base IPv6 support, but this is not yet
1662 enabled/available for use
1663 - Rename operations (cluster, instance) will now return the new name,
1664 which is especially useful if a short name was passed in
1665 - Added support for instance migration in RAPI
1666 - Added a tool to pre-configure nodes for the SSH setup, before joining
1667 them to the cluster; this will allow in the future a simplified model
1668 for node joining (but not yet fully enabled in 2.2); this needs the
1669 paramiko python library
1670 - Fixed handling of name-resolving errors
1671 - Fixed consistency of job results on the error path
1672 - Fixed master-failover race condition when executed multiple times in
1674 - Fixed many bugs related to the job queue (mostly introduced during the
1675 2.2 development cycle, so not all are impacting 2.1)
1676 - Fixed instance migration with missing disk symlinks
1677 - Fixed handling of unknown jobs in ``gnt-job archive``
1678 - And many other small fixes/improvements
1682 - Enhanced both the unittest and the QA coverage
1683 - Switched the opcode validation to a generic model, and extended the
1684 validation to all opcode parameters
1685 - Changed more parts of the code that write shell scripts to use the
1687 - Switched the master daemon to use the asyncore library for the Luxi
1694 *(Released Thu, 17 Jun 2010)*
1696 - Added tool (``move-instance``) and infrastructure to move instances
1697 between separate clusters (see :doc:`separate documentation
1698 <move-instance>` and :doc:`design document <design-2.2>`)
1699 - Added per-request RPC timeout
1700 - RAPI now requires a Content-Type header for requests with a body (e.g.
1701 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1702 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1703 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1705 - Implemented initial support for running Ganeti daemons as separate
1706 users, see configure-time flags ``--with-user-prefix`` and
1707 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1709 - Instances can be removed after export (``gnt-backup export
1710 --remove-instance``)
1711 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1712 key (instead of 1024 bit)
1713 - Added new cluster configuration file for cluster domain secret
1714 - Import/export now use SSL instead of SSH
1715 - Added support for showing estimated time when exporting an instance,
1716 see the ``ganeti-os-interface(7)`` manpage and look for
1723 *(Released Tue, 16 Nov 2010)*
1725 Some more bugfixes. Unless critical bugs occur, this will be the last
1728 - Fix case of MAC special-values
1729 - Fix mac checker regex
1730 - backend: Fix typo causing "out of range" error
1731 - Add missing --units in gnt-instance list man page
1737 *(Released Tue, 24 Aug 2010)*
1740 - Don't ignore secondary node silently on non-mirrored disk templates
1742 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1743 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1744 - Properly document the usb_mouse parameter
1745 - Fix path in ganeti-rapi(8) (issue 116)
1746 - Adjust error message when the ganeti user's .ssh directory is
1748 - Add same-node-check when changing the disk template to drbd
1754 *(Released Fri, 16 Jul 2010)*
1757 - Add an option to only select some reboot types during qa/burnin.
1758 (on some hypervisors consequent reboots are not supported)
1759 - Fix infrequent race condition in master failover. Sometimes the old
1760 master ip address would be still detected as up for a short time
1761 after it was removed, causing failover to fail.
1762 - Decrease mlockall warnings when the ctypes module is missing. On
1763 Python 2.4 we support running even if no ctypes module is installed,
1764 but we were too verbose about this issue.
1765 - Fix building on old distributions, on which man doesn't have a
1767 - Fix RAPI not to ignore the MAC address on instance creation
1768 - Implement the old instance creation format in the RAPI client.
1774 *(Released Thu, 01 Jul 2010)*
1776 A small bugfix release:
1777 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1778 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1779 a non-existing option
1780 - Fix instance create: the --force-variant option was ignored
1781 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1782 - Fix modify node storage with non-FQDN arguments
1783 - Fix RAPI client to authenticate under Python 2.6 when used
1784 for more than 5 requests needing authentication
1785 - Fix gnt-instance modify -t (storage) giving a wrong error message
1786 when converting a non-shutdown drbd instance to plain
1792 *(Released Fri, 18 Jun 2010)*
1794 A small bugfix release:
1796 - Fix live migration of KVM instances started with older Ganeti
1797 versions which had fewer hypervisor parameters
1798 - Fix gnt-instance grow-disk on down instances
1799 - Fix an error-reporting bug during instance migration
1800 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1801 silently ignoring broken ones
1802 - Fix an RPC error reporting bug affecting, for example, RAPI client
1804 - Fix bug triggered by different API version os-es on different nodes
1805 - Fix a bug in instance startup with custom hvparams: OS level
1806 parameters would fail to be applied.
1807 - Fix the RAPI client under Python 2.6 (but more work is needed to
1808 make it work completely well with OpenSSL)
1809 - Fix handling of errors when resolving names from DNS
1815 *(Released Thu, 3 Jun 2010)*
1817 A medium sized development cycle. Some new features, and some
1818 fixes/small improvements/cleanups.
1820 Significant features
1821 ~~~~~~~~~~~~~~~~~~~~
1823 The node deamon now tries to mlock itself into memory, unless the
1824 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1825 its logs, and falls back to console logging. This allows emergency
1826 features such as ``gnt-node powercycle`` to work even in the event of a
1827 broken node disk (tested offlining the disk hosting the node's
1828 filesystem and dropping its memory caches; don't try this at home)
1830 KVM: add vhost-net acceleration support. It can be tested with a new
1831 enough version of the kernel and of qemu-kvm.
1833 KVM: Add instance chrooting feature. If you use privilege dropping for
1834 your VMs you can also now force them to chroot to an empty directory,
1835 before starting the emulated guest.
1837 KVM: Add maximum migration bandwith and maximum downtime tweaking
1838 support (requires a new-enough version of qemu-kvm).
1840 Cluster verify will now warn if the master node doesn't have the master
1841 ip configured on it.
1843 Add a new (incompatible) instance creation request format to RAPI which
1844 supports all parameters (previously only a subset was supported, and it
1845 wasn't possible to extend the old format to accomodate all the new
1846 features. The old format is still supported, and a client can check for
1847 this feature, before using it, by checking for its presence in the
1848 ``features`` RAPI resource.
1850 Now with ancient latin support. Try it passing the ``--roman`` option to
1851 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1852 (requires the python-roman module to be installed, in order to work).
1857 As usual many internal code refactorings, documentation updates, and
1860 - Lots of improvements and cleanups to the experimental Remote API
1861 (RAPI) client library.
1862 - A new unit test suite for the core daemon libraries.
1863 - A fix to creating missing directories makes sure the umask is not
1864 applied anymore. This enforces the same directory permissions
1866 - Better handling terminating daemons with ctrl+c (used when running
1867 them in debugging mode).
1868 - Fix a race condition in live migrating a KVM instance, when stat()
1869 on the old proc status file returned EINVAL, which is an unexpected
1871 - Fixed manpage checking with newer man and utf-8 charachters. But now
1872 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1878 *(Released Fri, 7 May 2010)*
1880 Fix a bug which prevented untagged KVM instances from starting.
1886 *(Released Fri, 7 May 2010)*
1888 Another release with a long development cycle, during which many
1889 different features were added.
1891 Significant features
1892 ~~~~~~~~~~~~~~~~~~~~
1894 The KVM hypervisor now can run the individual instances as non-root, to
1895 reduce the impact of a VM being hijacked due to bugs in the
1896 hypervisor. It is possible to run all instances as a single (non-root)
1897 user, to manually specify a user for each instance, or to dynamically
1898 allocate a user out of a cluster-wide pool to each instance, with the
1899 guarantee that no two instances will run under the same user ID on any
1902 An experimental RAPI client library, that can be used standalone
1903 (without the other Ganeti libraries), is provided in the source tree as
1904 ``lib/rapi/client.py``. Note this client might change its interface in
1905 the future, as we iterate on its capabilities.
1907 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1908 replace the cluster's certificates and crypto keys. This might help in
1909 case they have been compromised, or have simply expired.
1911 A new disk option for instance creation has been added that allows one
1912 to "adopt" currently existing logical volumes, with data
1913 preservation. This should allow easier migration to Ganeti from
1914 unmanaged (or managed via other software) instances.
1916 Another disk improvement is the possibility to convert between redundant
1917 (DRBD) and plain (LVM) disk configuration for an instance. This should
1918 allow better scalability (starting with one node and growing the
1919 cluster, or shrinking a two-node cluster to one node).
1921 A new feature that could help with automated node failovers has been
1922 implemented: if a node sees itself as offline (by querying the master
1923 candidates), it will try to shutdown (hard) all instances and any active
1924 DRBD devices. This reduces the risk of duplicate instances if an
1925 external script automatically failovers the instances on such nodes. To
1926 enable this, the cluster parameter ``maintain_node_health`` should be
1927 enabled; in the future this option (per the name) will enable other
1928 automatic maintenance features.
1930 Instance export/import now will reuse the original instance
1931 specifications for all parameters; that means exporting an instance,
1932 deleting it and the importing it back should give an almost identical
1933 instance. Note that the default import behaviour has changed from
1934 before, where it created only one NIC; now it recreates the original
1937 Cluster verify has added a few new checks: SSL certificates validity,
1938 /etc/hosts consistency across the cluster, etc.
1943 As usual, many internal changes were done, documentation fixes,
1946 - Fixed cluster initialization with disabled cluster storage (regression
1947 introduced in 2.1.1)
1948 - File-based storage supports growing the disks
1949 - Fixed behaviour of node role changes
1950 - Fixed cluster verify for some corner cases, plus a general rewrite of
1951 cluster verify to allow future extension with more checks
1952 - Fixed log spamming by watcher and node daemon (regression introduced
1954 - Fixed possible validation issues when changing the list of enabled
1956 - Fixed cleanup of /etc/hosts during node removal
1957 - Fixed RAPI response for invalid methods
1958 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1959 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1960 from ide disks, etc.)
1961 - Allow OS changes without re-installation (to record a changed OS
1962 outside of Ganeti, or to allow OS renames)
1963 - Allow instance creation without OS installation (useful for example if
1964 the OS will be installed manually, or restored from a backup not in
1966 - Implemented option to make cluster ``copyfile`` use the replication
1968 - Added list of enabled hypervisors to ssconf (possibly useful for
1970 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1972 - A partial form of node re-IP is possible via node readd, which now
1973 allows changed node primary IP
1974 - Command line utilities now show an informational message if the job is
1976 - The logs of the master daemon now show the PID/UID/GID of the
1983 *(Released Fri, 12 Mar 2010)*
1985 During the 2.1.0 long release candidate cycle, a lot of improvements and
1986 changes have accumulated with were released later as 2.1.1.
1991 The node evacuate command (``gnt-node evacuate``) was significantly
1992 rewritten, and as such the IAllocator protocol was changed - a new
1993 request type has been added. This unfortunate change during a stable
1994 series is designed to improve performance of node evacuations; on
1995 clusters with more than about five nodes and which are well-balanced,
1996 evacuation should proceed in parallel for all instances of the node
1997 being evacuated. As such, any existing IAllocator scripts need to be
1998 updated, otherwise the above command will fail due to the unknown
1999 request. The provided "dumb" allocator has not been updated; but the
2000 ganeti-htools package supports the new protocol since version 0.2.4.
2002 Another important change is increased validation of node and instance
2003 names. This might create problems in special cases, if invalid host
2004 names are being used.
2006 Also, a new layer of hypervisor parameters has been added, that sits at
2007 OS level between the cluster defaults and the instance ones. This allows
2008 customisation of virtualization parameters depending on the installed
2009 OS. For example instances with OS 'X' may have a different KVM kernel
2010 (or any other parameter) than the cluster defaults. This is intended to
2011 help managing a multiple OSes on the same cluster, without manual
2012 modification of each instance's parameters.
2014 A tool for merging clusters, ``cluster-merge``, has been added in the
2015 tools sub-directory.
2020 - Improved the int/float conversions that should make the code more
2021 robust in face of errors from the node daemons
2022 - Fixed the remove node code in case of internal configuration errors
2023 - Fixed the node daemon behaviour in face of inconsistent queue
2024 directory (e.g. read-only file-system where we can't open the files
2026 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2027 now it either aborts cleanly or, if given the new "auto_promote"
2028 parameter, will automatically promote other nodes as needed
2029 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2030 completely prevent Ganeti from working
2031 - Fixed bug for instance export when not all disks were successfully
2033 - Fixed behaviour of node add when the new node is slow in starting up
2035 - Fixed handling of signals in the LUXI client, which should improve
2036 behaviour of command-line scripts
2037 - Added checks for invalid node/instance names in the configuration (now
2038 flagged during cluster verify)
2039 - Fixed watcher behaviour for disk activation errors
2040 - Fixed two potentially endless loops in http library, which led to the
2041 RAPI daemon hanging and consuming 100% CPU in some cases
2042 - Fixed bug in RAPI daemon related to hashed passwords
2043 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2045 - Enhanced compatibility with non-Debian OSes, but not using absolute
2046 path in some commands and allowing customisation of the ssh
2047 configuration directory
2048 - Fixed possible future issue with new Python versions by abiding to the
2049 proper use of ``__slots__`` attribute on classes
2050 - Added checks that should prevent directory traversal attacks
2051 - Many documentation fixes based on feedback from users
2056 - Added an "early_release" more for instance replace disks and node
2057 evacuate, where we release locks earlier and thus allow higher
2058 parallelism within the cluster
2059 - Added watcher hooks, intended to allow the watcher to restart other
2060 daemons (e.g. from the ganeti-nbma project), but they can be used of
2061 course for any other purpose
2062 - Added a compile-time disable for DRBD barriers, to increase
2063 performance if the administrator trusts the power supply or the
2064 storage system to not lose writes
2065 - Added the option of using syslog for logging instead of, or in
2066 addition to, Ganeti's own log files
2067 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2068 can indeed boot from a paravirtual NIC
2069 - Added a generic debug level for many operations; while this is not
2070 used widely yet, it allows one to pass the debug value all the way to
2072 - Enhanced the hooks environment for instance moves (failovers,
2073 migrations) where the primary/secondary nodes changed during the
2074 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2075 - Enhanced data validations for many user-supplied values; one important
2076 item is the restrictions imposed on instance and node names, which
2077 might reject some (invalid) host names
2078 - Add a configure-time option to disable file-based storage, if it's not
2079 needed; this allows greater security separation between the master
2080 node and the other nodes from the point of view of the inter-node RPC
2082 - Added user notification in interactive tools if job is waiting in the
2083 job queue or trying to acquire locks
2084 - Added log messages when a job is waiting for locks
2085 - Added filtering by node tags in instance operations which admit
2086 multiple instances (start, stop, reboot, reinstall)
2087 - Added a new tool for cluster mergers, ``cluster-merge``
2088 - Parameters from command line which are of the form ``a=b,c=d`` can now
2089 use backslash escapes to pass in values which contain commas,
2090 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2092 - For KVM, the instance name is the first parameter passed to KVM, so
2093 that it's more visible in the process list
2099 *(Released Tue, 2 Mar 2010)*
2101 Ganeti 2.1 brings many improvements with it. Major changes:
2103 - Added infrastructure to ease automated disk repairs
2104 - Added new daemon to export configuration data in a cheaper way than
2105 using the remote API
2106 - Instance NICs can now be routed instead of being associated with a
2108 - Improved job locking logic to reduce impact of jobs acquiring multiple
2109 locks waiting for other long-running jobs
2111 In-depth implementation details can be found in the Ganeti 2.1 design
2117 - Added chroot hypervisor
2118 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2120 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2121 ``bootloader_path`` and ``bootloader_args``)
2122 - Added the ``use_localtime`` option for the xen-hvm and kvm
2123 hypervisors, and the default value for this has changed to false (in
2124 2.0 xen-hvm always enabled it)
2125 - Added luxi call to submit multiple jobs in one go
2126 - Added cluster initialization option to not modify ``/etc/hosts``
2128 - Added network interface parameters
2129 - Added dry run mode to some LUs
2130 - Added RAPI resources:
2132 - ``/2/instances/[instance_name]/info``
2133 - ``/2/instances/[instance_name]/replace-disks``
2134 - ``/2/nodes/[node_name]/evacuate``
2135 - ``/2/nodes/[node_name]/migrate``
2136 - ``/2/nodes/[node_name]/role``
2137 - ``/2/nodes/[node_name]/storage``
2138 - ``/2/nodes/[node_name]/storage/modify``
2139 - ``/2/nodes/[node_name]/storage/repair``
2141 - Added OpCodes to evacuate or migrate all instances on a node
2142 - Added new command to list storage elements on nodes (``gnt-node
2143 list-storage``) and modify them (``gnt-node modify-storage``)
2144 - Added new ssconf files with master candidate IP address
2145 (``ssconf_master_candidates_ips``), node primary IP address
2146 (``ssconf_node_primary_ips``) and node secondary IP address
2147 (``ssconf_node_secondary_ips``)
2148 - Added ``ganeti-confd`` and a client library to query the Ganeti
2149 configuration via UDP
2150 - Added ability to run hooks after cluster initialization and before
2152 - Added automatic mode for disk replace (``gnt-instance replace-disks
2154 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2155 after catastrophic data-loss
2156 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2158 - Added ``gnt-instance move`` command to move instances
2159 - Added ``gnt-cluster watcher`` command to control watcher
2160 - Added ``gnt-node powercycle`` command to powercycle nodes
2161 - Added new job status field ``lock_status``
2162 - Added parseable error codes to cluster verification (``gnt-cluster
2163 verify --error-codes``) and made output less verbose (use
2164 ``--verbose`` to restore previous behaviour)
2165 - Added UUIDs to the main config entities (cluster, nodes, instances)
2166 - Added support for OS variants
2167 - Added support for hashed passwords in the Ganeti remote API users file
2169 - Added option to specify maximum timeout on instance shutdown
2170 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2171 - Added new helper script to start and stop Ganeti daemons
2172 (``daemon-util``), with the intent to reduce the work necessary to
2173 adjust Ganeti for non-Debian distributions and to start/stop daemons
2175 - Added more unittests
2176 - Fixed critical bug in ganeti-masterd startup
2177 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2178 now customisable at the cluster level for both the KVM and Xen
2179 hypervisors using the new ``migration_port`` parameter
2180 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2181 reinstalling an instance
2182 - Allowed ``@`` in tag names
2183 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2184 - Many documentation updates
2185 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2186 - ``gnt-instance reinstall`` can now reinstall multiple instances
2187 - Updated many command line parameters
2188 - Introduced new OS API version 15
2189 - No longer support a default hypervisor
2190 - Treat virtual LVs as inexistent
2191 - Improved job locking logic to reduce lock contention
2192 - Match instance and node names case insensitively
2193 - Reimplemented bash completion script to be more complete
2200 *(Released Thu, 4 Feb 2010)*
2202 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2203 - Fix a string formatting bug
2204 - Improve safety of the code in some error paths
2205 - Improve data validation in the master of values returned from nodes
2211 *(Released Thu, 17 Dec 2009)*
2213 - Fix security issue due to missing validation of iallocator names; this
2214 allows local and remote execution of arbitrary executables
2215 - Fix failure of gnt-node list during instance removal
2216 - Ship the RAPI documentation in the archive
2222 *(Released Wed, 30 Sep 2009)*
2224 - Fixed many wrong messages
2225 - Fixed a few bugs related to the locking library
2226 - Fixed MAC checking at instance creation time
2227 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2228 - Fixed a few issues related to signal handling in both daemons and
2230 - Fixed the example startup script provided
2231 - Fixed insserv dependencies in the example startup script (patch from
2233 - Fixed handling of drained nodes in the iallocator framework
2234 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2236 - Fixed error related to invalid job IDs in job polling
2237 - Fixed job/opcode persistence on unclean master shutdown
2238 - Fixed handling of partial job processing after unclean master
2240 - Fixed error reporting from LUs, previously all errors were converted
2241 into execution errors
2242 - Fixed error reporting from burnin
2243 - Decreased significantly the memory usage of the job queue
2244 - Optimised slightly multi-job submission
2245 - Optimised slightly opcode loading
2246 - Backported the multi-job submit framework from the development
2247 branch; multi-instance start and stop should be faster
2248 - Added script to clean archived jobs after 21 days; this will reduce
2249 the size of the queue directory
2250 - Added some extra checks in disk size tracking
2251 - Added an example ethers hook script
2252 - Added a cluster parameter that prevents Ganeti from modifying of
2254 - Added more node information to RAPI responses
2255 - Added a ``gnt-job watch`` command that allows following the ouput of a
2257 - Added a bind-address option to ganeti-rapi
2258 - Added more checks to the configuration verify
2259 - Enhanced the burnin script such that some operations can be retried
2261 - Converted instance reinstall to multi-instance model
2267 *(Released Fri, 7 Aug 2009)*
2269 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2270 to allow using the pre-2.0.2 behaviour in activation, if any existing
2271 instances have mismatched disk sizes in the configuration
2272 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2273 any configuration mismatches for disk sizes
2274 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2275 failover to work on two-node clusters
2276 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2278 - Fixed detection of OS script errors in ``gnt-backup export``
2279 - Fixed exit code of ``gnt-backup export``
2285 *(Released Fri, 17 Jul 2009)*
2287 - Added experimental support for stripped logical volumes; this should
2288 enhance performance but comes with a higher complexity in the block
2289 device handling; stripping is only enabled when passing
2290 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2291 affected even in the non-stripped mode
2292 - Improved resiliency against transient failures at the end of DRBD
2293 resyncs, and in general of DRBD resync checks
2294 - Fixed a couple of issues with exports and snapshot errors
2295 - Fixed a couple of issues in instance listing
2296 - Added display of the disk size in ``gnt-instance info``
2297 - Fixed checking for valid OSes in instance creation
2298 - Fixed handling of the "vcpus" parameter in instance listing and in
2299 general of invalid parameters
2300 - Fixed http server library, and thus RAPI, to handle invalid
2301 username/password combinations correctly; this means that now they
2302 report unauthorized for queries too, not only for modifications,
2303 allowing earlier detect of configuration problems
2304 - Added a new "role" node list field, equivalent to the master/master
2305 candidate/drained/offline flags combinations
2306 - Fixed cluster modify and changes of candidate pool size
2307 - Fixed cluster verify error messages for wrong files on regular nodes
2308 - Fixed a couple of issues with node demotion from master candidate role
2309 - Fixed node readd issues
2310 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2311 - Added a new ``--no-voting`` option for masterfailover to fix failover
2312 on two-nodes clusters when the former master node is unreachable
2313 - Added instance reinstall over RAPI
2319 *(Released Tue, 16 Jun 2009)*
2321 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2322 allow re-adding the start in single-user option (regression from 1.2)
2323 - the watcher writes the instance status to a file, to allow monitoring
2324 to report the instance status (from the master) based on cached
2325 results of the watcher's queries; while this can get stale if the
2326 watcher is being locked due to other work on the cluster, this is
2327 still an improvement
2328 - the watcher now also restarts the node daemon and the rapi daemon if
2330 - fixed the watcher to handle full and drained queue cases
2331 - hooks export more instance data in the environment, which helps if
2332 hook scripts need to take action based on the instance's properties
2333 (no longer need to query back into ganeti)
2334 - instance failovers when the instance is stopped do not check for free
2335 RAM, so that failing over a stopped instance is possible in low memory
2337 - rapi uses queries for tags instead of jobs (for less job traffic), and
2338 for cluster tags it won't talk to masterd at all but read them from
2340 - a couple of error handling fixes in RAPI
2341 - drbd handling: improved the error handling of inconsistent disks after
2342 resync to reduce the frequency of "there are some degraded disks for
2343 this instance" messages
2344 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2345 error handling path called a wrong function name)
2351 *(Released Wed, 27 May 2009)*
2353 - no changes from rc5
2359 *(Released Wed, 20 May 2009)*
2361 - fix a couple of bugs (validation, argument checks)
2362 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2363 - some small improvements to RAPI and IAllocator
2364 - make watcher automatically start the master daemon if down
2370 *(Released Mon, 27 Apr 2009)*
2372 - change the OS list to not require locks; this helps with big clusters
2373 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2374 volume group is broken
2375 - ``gnt-instance info``, without any arguments, doesn't run for all
2376 instances anymore; either pass ``--all`` or pass the desired
2377 instances; this helps against mistakes on big clusters where listing
2378 the information for all instances takes a long time
2379 - miscellaneous doc and man pages fixes
2385 *(Released Wed, 8 Apr 2009)*
2387 - Change the internal locking model of some ``gnt-node`` commands, in
2388 order to reduce contention (and blocking of master daemon) when
2389 batching many creation/reinstall jobs
2390 - Fixes to Xen soft reboot
2391 - No longer build documentation at build time, instead distribute it in
2392 the archive, in order to reduce the need for the whole docbook/rst
2399 *(Released Fri, 27 Mar 2009)*
2401 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2402 - Fix watcher startup sequence, improves the behaviour of busy clusters
2403 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2404 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2405 bind address and other places
2406 - Some documentation fixes and updates
2412 *(Released Mon, 2 Mar 2009)*
2414 - More documentation updates, now all docs should be more-or-less
2416 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2418 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2420 - Fix an issue related to $libdir/run/ganeti and cluster creation
2426 *(Released Thu, 19 Feb 2009)*
2428 - Xen PVM and KVM have switched the default value for the instance root
2429 disk to the first partition on the first drive, instead of the whole
2430 drive; this means that the OS installation scripts must be changed
2432 - Man pages have been updated
2433 - RAPI has been switched by default to HTTPS, and the exported functions
2434 should all work correctly
2435 - RAPI v1 has been removed
2436 - Many improvements to the KVM hypervisor
2437 - Block device errors are now better reported
2438 - Many other bugfixes and small improvements
2444 *(Released Mon, 26 Jan 2009)*
2446 - Version 2 is a general rewrite of the code and therefore the
2447 differences are too many to list, see the design document for 2.0 in
2448 the ``doc/`` subdirectory for more details
2449 - In this beta version there is not yet a migration path from 1.2 (there
2450 will be one in the final 2.0 release)
2451 - A few significant changes are:
2453 - all commands are executed by a daemon (``ganeti-masterd``) and the
2454 various ``gnt-*`` commands are just front-ends to it
2455 - all the commands are entered into, and executed from a job queue,
2456 see the ``gnt-job(8)`` manpage
2457 - the RAPI daemon supports read-write operations, secured by basic
2458 HTTP authentication on top of HTTPS
2459 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2460 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2461 to migrate to DRBD 8 first while still running Ganeti 1.2)
2462 - DRBD devices are using statically allocated minor numbers, which
2463 will be assigned to existing instances during the migration process
2464 - there is support for both Xen PVM and Xen HVM instances running on
2466 - KVM virtualization is supported too
2467 - file-based storage has been implemented, which means that it is
2468 possible to run the cluster without LVM and DRBD storage, for
2469 example using a shared filesystem exported from shared storage (and
2470 still have live migration)
2476 *(Released Tue, 13 Jan 2009)*
2478 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2479 - Reuse the old instance mac address by default on instance import, if
2480 the instance name is the same.
2481 - Handle situations in which the node info rpc returns incomplete
2483 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2484 - Improved version of batcher:
2486 - state file support
2487 - instance mac address support
2488 - support for HVM clusters/instances
2490 - Add an option to show the number of cpu sockets and nodes in
2492 - Support OSes that handle more than one version of the OS api (but do
2493 not change the current API in any other way)
2494 - Fix ``gnt-node migrate``
2495 - ``gnt-debug`` man page
2496 - Fixes various more typos and small issues
2497 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2503 *(Released Wed, 24 Sep 2008)*
2505 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2506 type of disk exported to fully virtualized instances.
2507 - provide access to the serial console of HVM instances
2508 - instance auto_balance flag, set by default. If turned off it will
2509 avoid warnings on cluster verify if there is not enough memory to fail
2510 over an instance. in the future it will prevent automatically failing
2511 it over when we will support that.
2512 - batcher tool for instance creation, see ``tools/README.batcher``
2513 - ``gnt-instance reinstall --select-os`` to interactively select a new
2514 operating system when reinstalling an instance.
2515 - when changing the memory amount on instance modify a check has been
2516 added that the instance will be able to start. also warnings are
2517 emitted if the instance will not be able to fail over, if auto_balance
2519 - documentation fixes
2520 - sync fields between ``gnt-instance list/modify/add/import``
2521 - fix a race condition in drbd when the sync speed was set after giving
2522 the device a remote peer.
2528 *(Released Tue, 22 Jul 2008)*
2530 - note: the allowed size and number of tags per object were reduced
2531 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2532 - fixed twisted 8.x compatibility
2533 - fixed ``gnt-instance replace-disks`` with iallocator
2534 - add TCP keepalives on twisted connections to detect restarted nodes
2535 - disk increase support, see ``gnt-instance grow-disk``
2536 - implement bulk node/instance query for RAPI
2537 - add tags in node/instance listing (optional)
2538 - experimental migration (and live migration) support, read the man page
2539 for ``gnt-instance migrate``
2540 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2541 has some small improvements in handling its state file
2547 *(Released Fri, 13 Jun 2008)*
2549 - Experimental readonly, REST-based remote API implementation;
2550 automatically started on master node, TCP port 5080, if enabled by
2551 ``--enable-rapi`` parameter to configure script.
2552 - Instance allocator support. Add and import instance accept a
2553 ``--iallocator`` parameter, and call that instance allocator to decide
2554 which node to use for the instance. The iallocator document describes
2555 what's expected from an allocator script.
2556 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2557 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2558 node is lost there is still enough memory to fail over the instances
2560 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2561 ``gnt-cluster verify``, to check for site-specific compliance. All the
2562 hooks will run, and their output, if any, will be displayed. Any
2563 failing hook will make the verification return an error value.
2564 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2565 primary and secondary interfaces
2566 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2567 that is still declared as part of the cluster and has failed.
2568 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2569 specifying output fields, that just adds the chosen fields to the
2571 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2572 export from the filesystem.
2573 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2574 have been added. Using them you can enable/disable acpi and pae
2575 support, and specify a path for a cd image to be exported to the
2576 instance. These parameters as the name suggest only work on HVM
2578 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2579 PAE support will be set to the previously hardcoded values, but the
2580 (previously hardcoded) path to the CDROM ISO image will be unset and
2581 if required, needs to be set manually with ``gnt-instance modify``
2583 - The address to which an instance's VNC console is bound is now
2584 selectable per-instance, rather than being cluster wide. Of course
2585 this only applies to instances controlled via VNC, so currently just
2586 applies to HVM clusters.
2592 *(Released Mon, 18 Feb 2008)*
2594 - more tweaks to the disk activation code (especially helpful for DRBD)
2595 - change the default ``gnt-instance list`` output format, now there is
2596 one combined status field (see the manpage for the exact values this
2598 - some more fixes for the mac export to hooks change
2599 - make Ganeti not break with DRBD 8.2.x (which changed the version
2600 format in ``/proc/drbd``) (issue 24)
2601 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2602 template, allowing migration from DRBD0.7+MD to DRBD8
2608 *(Released Wed, 30 Jan 2008)*
2610 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2612 - add command aliases infrastructure and a few aliases
2613 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2614 man pages and the ``--help`` option of ``gnt-node
2615 list``/``gnt-instance list``
2616 - fix ``gnt-backup list`` with down nodes (issue 21)
2617 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2618 - fix the dist archive and add a check for including svn/git files in
2620 - some developer-related changes: improve the burnin and the QA suite,
2621 add an upload script for testing during development
2627 *(Released Wed, 16 Jan 2008)*
2629 - experimental HVM support, read the install document, section
2630 "Initializing the cluster"
2631 - allow for the PVM hypervisor per-instance kernel and initrd paths
2632 - add a new command ``gnt-cluster verify-disks`` which uses a new
2633 algorithm to improve the reconnection of the DRBD pairs if the device
2634 on the secondary node has gone away
2635 - make logical volume code auto-activate LVs at disk activation time
2636 - slightly improve the speed of activating disks
2637 - allow specification of the MAC address at instance creation time, and
2638 changing it later via ``gnt-instance modify``
2639 - fix handling of external commands that generate lots of output on
2641 - update documentation with regard to minimum version of DRBD8 supported
2647 *(Released Tue, 4 Dec 2007)*
2649 - Log the ``xm create`` output to the node daemon log on failure (to
2650 help diagnosing the error)
2651 - In debug mode, log all external commands output if failed to the logs
2652 - Change parsing of lvm commands to ignore stderr
2658 *(Released Wed, 28 Nov 2007)*
2660 - Another round of updates to the DRBD 8 code to deal with more failures
2661 in the replace secondary node operation
2662 - Some more logging of failures in disk operations (lvm, drbd)
2663 - A few documentation updates
2670 *(Released Tue, 13 Nov 2007)*
2672 - Change configuration file format from Python's Pickle to JSON.
2673 Upgrading is possible using the cfgupgrade utility.
2674 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2675 faster replace disks and is more stable (DRBD 8 has many improvements
2676 compared to DRBD 0.7)
2677 - Added command line tags support (see man pages for ``gnt-instance``,
2678 ``gnt-node``, ``gnt-cluster``)
2679 - Added instance rename support
2680 - Added multi-instance startup/shutdown
2681 - Added cluster rename support
2682 - Added ``gnt-node evacuate`` to simplify some node operations
2683 - Added instance reboot operation that can speedup reboot as compared to
2685 - Soften the requirement that hostnames are in FQDN format
2686 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2688 - Removed dependency on debian's patched fping that uses the
2689 non-standard ``-S`` option
2690 - Now the OS definitions are searched for in multiple, configurable
2691 paths (easier for distros to package)
2692 - Some changes to the hooks infrastructure (especially the new
2693 post-configuration update hook)
2694 - Other small bugfixes
2696 .. vim: set textwidth=72 syntax=rst :