8 *(Released Thu, 26 Sep 2013)*
10 - Change the connected groups format in ``gnt-network info`` output; it
11 was previously displayed as a raw list by mistake
12 - Check disk template in right dict when copying
13 - Support multi-instance allocs without iallocator
14 - Fix some errors in the documentation
15 - Fix formatting of tuple in an error message
21 *(Released Thu, 25 Jul 2013)*
23 - Add logrotate functionality in daemon-util
24 - Add logrotate example file
25 - Add missing fields to network queries over rapi
26 - Fix network object timestamps
27 - Add support for querying network timestamps
28 - Fix a typo in the example crontab
29 - Fix a documentation typo
35 *(Released Thu, 04 Jul 2013)*
37 Incompatible/important changes
38 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
40 - Instance policies for disk size were documented to be on a per-disk
41 basis, but hail applied them to the sum of all disks. This has been
43 - ``hbal`` will now exit with status 0 if, during job execution over
44 LUXI, early exit has been requested and all jobs are successful;
45 before, exit status 1 was used, which cannot be differentiated from
47 - Compatibility with newer versions of rbd has been fixed
48 - ``gnt-instance batch-create`` has been changed to use the bulk create
49 opcode from Ganeti. This lead to incompatible changes in the format of
50 the JSON file. It's now not a custom dict anymore but a dict
51 compatible with the ``OpInstanceCreate`` opcode.
52 - Parent directories for file storage need to be listed in
53 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
54 write the file automatically based on old configuration values, but it
55 can not distribute it across all nodes and the file contents should be
56 verified. Use ``gnt-cluster copyfile
57 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
58 upgraded. The reason for requiring this list of paths now is that
59 before it would have been possible to inject new paths via RPC,
60 allowing files to be created in arbitrary locations. The RPC protocol
61 is protected using SSL/X.509 certificates, but as a design principle
62 Ganeti does not permit arbitrary paths to be passed.
63 - The parsing of the variants file for OSes (see
64 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
65 lines and comment lines (starting with ``#``) are ignored for better
67 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
68 longer available. ``gnt-node add`` now invokes a new tool on the
69 destination node, named ``prepare-node-join``, to configure the SSH
70 daemon. Paramiko is no longer necessary to configure nodes' SSH
71 daemons via ``gnt-node add``.
72 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
73 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
74 the flag is not reset after a master failover.
75 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
76 is recommended. 2.6 will be mandatory from the 2.8 series.
82 - New network management functionality to support automatic allocation
83 of IP addresses and managing of network parameters. See
84 :manpage:`gnt-network(8)` for more details.
85 - New external storage backend, to allow managing arbitrary storage
86 systems external to the cluster. See
87 :manpage:`ganeti-extstorage-interface(7)`.
88 - New ``exclusive-storage`` node parameter added, restricted to
89 nodegroup level. When it's set to true, physical disks are assigned in
90 an exclusive fashion to instances, as documented in :doc:`Partitioned
91 Ganeti <design-partitioned>`. Currently, only instances using the
92 ``plain`` disk template are supported.
93 - The KVM hypervisor has been updated with many new hypervisor
94 parameters, including a generic one for passing arbitrary command line
95 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
96 compatible up to qemu 1.4.
97 - A new tool, called ``mon-collector``, is the stand-alone executor of
98 the data collectors for a monitoring system. As of this version, it
99 just includes the DRBD data collector, that can be executed by calling
100 ``mon-collector`` using the ``drbd`` parameter. See
101 :manpage:`mon-collector(7)`.
102 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
103 for RAPI users. It allows granting permissions to query for
104 information to a specific user without giving
105 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
106 - A new tool named ``node-cleanup`` has been added. It cleans remains of
107 a cluster from a machine by stopping all daemons, removing
108 certificates and ssconf files. Unless the ``--no-backup`` option is
109 given, copies of the certificates are made.
110 - Instance creations now support the use of opportunistic locking,
111 potentially speeding up the (parallel) creation of multiple instances.
112 This feature is currently only available via the :doc:`RAPI
113 <rapi>` interface and when an instance allocator is used. If the
114 ``opportunistic_locking`` parameter is set the opcode will try to
115 acquire as many locks as possible, but will not wait for any locks
116 held by other opcodes. If not enough resources can be found to
117 allocate the instance, the temporary error code
118 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
119 retried thereafter, with or without opportunistic locking.
120 - New experimental linux-ha resource scripts.
121 - Restricted-commands support: ganeti can now be asked (via command line
122 or rapi) to perform commands on a node. These are passed via ganeti
123 RPC rather than ssh. This functionality is restricted to commands
124 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
125 reasons. The file is not copied automatically.
131 - Diskless instances are now externally mirrored (Issue 237). This for
132 now has only been tested in conjunction with explicit target nodes for
134 - Queries not needing locks or RPC access to the node can now be
135 performed by the confd daemon, making them independent from jobs, and
136 thus faster to execute. This is selectable at configure time.
137 - The functionality for allocating multiple instances at once has been
138 overhauled and is now also available through :doc:`RAPI <rapi>`.
140 There are no significant changes from version 2.7.0~rc3.
146 *(Released Tue, 25 Jun 2013)*
148 - Fix permissions on the confd query socket (Issue 477)
149 - Fix permissions on the job archive dir (Issue 498)
150 - Fix handling of an internal exception in replace-disks (Issue 472)
151 - Fix gnt-node info handling of shortened names (Issue 497)
152 - Fix gnt-instance grow-disk when wiping is enabled
153 - Documentation improvements, and support for newer pandoc
154 - Fix hspace honoring ipolicy for disks (Issue 484)
155 - Improve handling of the ``kvm_extra`` HV parameter
161 *(Released Fri, 24 May 2013)*
163 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
165 - Disks added through ``gnt-instance modify`` or created through
166 ``gnt-instance recreate-disks`` are wiped, if the
167 ``prealloc_wipe_disks`` flag is set.
168 - If wiping newly created disks fails, the disks are removed. Also,
169 partial failures in creating disks through ``gnt-instance modify``
170 triggers a cleanup of the partially-created disks.
171 - Removing the master IP address doesn't fail if the address has been
173 - Fix ownership of the OS log dir
174 - Workaround missing SO_PEERCRED constant (Issue 191)
180 *(Released Fri, 3 May 2013)*
182 This was the first release candidate of the 2.7 series. Since beta3:
184 - Fix kvm compatibility with qemu 1.4 (Issue 389)
185 - Documentation updates (admin guide, upgrade notes, install
186 instructions) (Issue 372)
187 - Fix gnt-group list nodes and instances count (Issue 436)
188 - Fix compilation without non-mandatory libraries (Issue 441)
189 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
190 - Make confd logging more verbose at INFO level (Issue 435)
191 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
192 - Fix failure path for instance storage type conversion (Issue 229)
193 - Update htools text backend documentation
194 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
195 - Disable inter-cluster instance move for file-based instances, because
196 it is dependant on instance export, which is not supported for
197 file-based instances. (Issue 414)
198 - Fix gnt-job crashes on non-ascii characters (Issue 427)
199 - Fix volume group checks on non-vm-capable nodes (Issue 432)
205 *(Released Mon, 22 Apr 2013)*
207 This was the third beta release of the 2.7 series. Since beta2:
209 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
210 - Fix data loss on wrong usage of ``gnt-instance move``
211 - Properly export errors in confd-based job queries
212 - Add ``users-setup`` tool
213 - Fix iallocator protocol to report 0 as a disk size for diskless
214 instances. This avoids hail breaking when a diskless instance is
216 - Fix job queue directory permission problem that made confd job queries
217 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
218 for access to archived jobs (Issue 406).
219 - Limit the sizes of networks supported by ``gnt-network`` to something
220 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
221 - Fix bugs in instance disk template conversion
222 - Fix GHC 7 compatibility
223 - Fix ``burnin`` install path (Issue 426).
224 - Allow very small disk grows (Issue 347).
225 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
226 that noded doesn't import masterd code (Issue 419).
227 - Make sure the default metavg at cluster init is the same as the vg, if
228 unspecified (Issue 358).
229 - Fix cleanup of partially created disks (part of Issue 416)
235 *(Released Tue, 2 Apr 2013)*
237 This was the second beta release of the 2.7 series. Since beta1:
239 - Networks no longer have a "type" slot, since this information was
240 unused in Ganeti: instead of it tags should be used.
241 - The rapi client now has a ``target_node`` option to MigrateInstance.
242 - Fix early exit return code for hbal (Issue 386).
243 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
244 - Fix ``rbd showmapped`` output parsing (Issue 312).
245 - Networks are now referenced indexed by UUID, rather than name. This
246 will require running cfgupgrade, from 2.7.0beta1, if networks are in
248 - The OS environment now includes network information.
249 - Deleting of a network is now disallowed if any instance nic is using
250 it, to prevent dangling references.
251 - External storage is now documented in man pages.
252 - The exclusive_storage flag can now only be set at nodegroup level.
253 - Hbal can now submit an explicit priority with its jobs.
254 - Many network related locking fixes.
255 - Bump up the required pylint version to 0.25.1.
256 - Fix the ``no_remember`` option in RAPI client.
257 - Many ipolicy related tests, qa, and fixes.
258 - Many documentation improvements and fixes.
259 - Fix building with ``--disable-file-storage``.
260 - Fix ``-q`` option in htools, which was broken if passed more than
262 - Some haskell/python interaction improvements and fixes.
263 - Fix iallocator in case of missing LVM storage.
264 - Fix confd config load in case of ``--no-lvm-storage``.
265 - The confd/query functionality is now mentioned in the security
272 *(Released Wed, 6 Feb 2013)*
274 This was the first beta release of the 2.7 series. All important changes
275 are listed in the latest 2.7 entry.
281 *(Released Fri, 21 Dec 2012)*
283 Important behaviour change: hbal won't rebalance anymore instances which
284 have the ``auto_balance`` attribute set to false. This was the intention
285 all along, but until now it only skipped those from the N+1 memory
286 reservation (DRBD-specific).
288 A significant number of bug fixes in this release:
290 - Fixed disk adoption interaction with ipolicy checks.
291 - Fixed networking issues when instances are started, stopped or
292 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
293 - Fixed the warning in cluster verify for shared storage instances not
295 - Fixed removal of storage directory on shared file storage (issue 262).
296 - Fixed validation of LVM volume group name in OpClusterSetParams
297 (``gnt-cluster modify``) (issue 285).
298 - Fixed runtime memory increases (``gnt-instance modify -m``).
299 - Fixed live migration under Xen's ``xl`` mode.
300 - Fixed ``gnt-instance console`` with ``xl``.
301 - Fixed building with newer Haskell compiler/libraries.
302 - Fixed PID file writing in Haskell daemons (confd); this prevents
303 restart issues if confd was launched manually (outside of
304 ``daemon-util``) while another copy of it was running
305 - Fixed a type error when doing live migrations with KVM (issue 297) and
306 the error messages for failing migrations have been improved.
307 - Fixed opcode validation for the out-of-band commands (``gnt-node
309 - Fixed a type error when unsetting OS hypervisor parameters (issue
310 311); now it's possible to unset all OS-specific hypervisor
312 - Fixed the ``dry-run`` mode for many operations: verification of
313 results was over-zealous but didn't take into account the ``dry-run``
314 operation, resulting in "wrong" failures.
315 - Fixed bash completion in ``gnt-job list`` when the job queue has
316 hundreds of entries; especially with older ``bash`` versions, this
317 results in significant CPU usage.
319 And lastly, a few other improvements have been made:
321 - Added option to force master-failover without voting (issue 282).
322 - Clarified error message on lock conflict (issue 287).
323 - Logging of newly submitted jobs has been improved (issue 290).
324 - Hostname checks have been made uniform between instance rename and
326 - The ``--submit`` option is now supported by ``gnt-debug delay``.
327 - Shutting down the master daemon by sending SIGTERM now stops it from
328 processing jobs waiting for locks; instead, those jobs will be started
329 once again after the master daemon is started the next time (issue
331 - Support for Xen's ``xl`` program has been improved (besides the fixes
333 - Reduced logging noise in the Haskell confd daemon (only show one log
334 entry for each config reload, instead of two).
335 - Several man page updates and typo fixes.
341 *(Released Fri, 12 Oct 2012)*
343 A small bugfix release. Among the bugs fixed:
345 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
346 made the command unusable.
347 - Commands that issue many jobs don't fail anymore just because some jobs
348 take so long that other jobs are archived.
349 - Failures during ``gnt-instance reinstall`` are reflected by the exit
351 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
353 - When ``always_failover`` is set, ``--allow-failover`` is not required
354 in migrate commands anymore.
355 - ``bash_completion`` works even if extglob is disabled.
356 - Fixed bug with locks that made failover for RDB-based instances fail.
357 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
358 a random node instead of one based on the allocator metric.
359 - Support for newer versions of pylint and pep8.
360 - Hail doesn't fail anymore when trying to add an instance of type
361 ``file``, ``sharedfile`` or ``rbd``.
362 - Added new Makefile target to rebuild the whole distribution, so that
363 all files are included.
369 *(Released Fri, 27 Jul 2012)*
372 .. attention:: The ``LUXI`` protocol has been made more consistent
373 regarding its handling of command arguments. This, however, leads to
374 incompatibility issues with previous versions. Please ensure that you
375 restart Ganeti daemons soon after the upgrade, otherwise most
376 ``LUXI`` calls (job submission, setting/resetting the drain flag,
377 pausing/resuming the watcher, cancelling and archiving jobs, querying
378 the cluster configuration) will fail.
387 The current ``admin_up`` field, which used to denote whether an instance
388 should be running or not, has been removed. Instead, ``admin_state`` is
389 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
391 The rational behind this is that an instance being “down” can have
394 - it could be down during a reboot
395 - it could be temporarily be down for a reinstall
396 - or it could be down because it is deprecated and kept just for its
399 The previous Boolean state was making it difficult to do capacity
400 calculations: should Ganeti reserve memory for a down instance? Now, the
401 tri-state field makes it clear:
403 - in ``up`` and ``down`` state, all resources are reserved for the
404 instance, and it can be at any time brought up if it is down
405 - in ``offline`` state, only disk space is reserved for it, but not
408 The field can have an extra use: since the transition between ``up`` and
409 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
410 transition between ``offline`` and ``down`` is done via ``gnt-instance
411 modify``, it is possible to given different rights to users. For
412 example, owners of an instance could be allowed to start/stop it, but
413 not transition it out of the offline state.
415 Instance policies and specs
416 +++++++++++++++++++++++++++
418 In previous Ganeti versions, an instance creation request was not
419 limited on the minimum size and on the maximum size just by the cluster
420 resources. As such, any policy could be implemented only in third-party
421 clients (RAPI clients, or shell wrappers over ``gnt-*``
422 tools). Furthermore, calculating cluster capacity via ``hspace`` again
423 required external input with regards to instance sizes.
425 In order to improve these workflows and to allow for example better
426 per-node group differentiation, we introduced instance specs, which
429 - minimum instance disk size, disk count, memory size, cpu count
430 - maximum values for the above metrics
431 - and “standard” values (used in ``hspace`` to calculate the standard
434 The minimum/maximum values can be also customised at node-group level,
435 for example allowing more powerful hardware to support bigger instance
438 Beside the instance specs, there are a few other settings belonging to
439 the instance policy framework. It is possible now to customise, per
440 cluster and node-group:
442 - the list of allowed disk templates
443 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
444 - the maximum ratio of instance to spindles (see below for more
445 information) for local storage
447 All these together should allow all tools that talk to Ganeti to know
448 what are the ranges of allowed values for instances and the
449 over-subscription that is allowed.
451 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
452 instance configuration, and the physical CPU configuration from the
453 node. For the spindle ratios however, we didn't track before these
454 values, so new parameters have been added:
456 - a new node parameter ``spindle_count``, defaults to 1, customisable at
457 node group or node level
458 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
460 Note that spindles in this context doesn't need to mean actual
461 mechanical hard-drives; it's just a relative number for both the node
462 I/O capacity and instance I/O consumption.
464 Instance migration behaviour
465 ++++++++++++++++++++++++++++
467 While live-migration is in general desirable over failover, it is
468 possible that for some workloads it is actually worse, due to the
469 variable time of the “suspend” phase during live migration.
471 To allow the tools to work consistently over such instances (without
472 having to hard-code instance names), a new backend parameter
473 ``always_failover`` has been added to control the migration/failover
474 behaviour. When set to True, all migration requests for an instance will
475 instead fall-back to failover.
477 Instance memory ballooning
478 ++++++++++++++++++++++++++
480 Initial support for memory ballooning has been added. The memory for an
481 instance is no longer fixed (backend parameter ``memory``), but instead
482 can vary between minimum and maximum values (backend parameters
483 ``minmem`` and ``maxmem``). Currently we only change an instance's
486 - live migrating or failing over and instance and the target node
487 doesn't have enough memory
488 - user requests changing the memory via ``gnt-instance modify
494 In order to control the use of specific CPUs by instance, support for
495 controlling CPU pinning has been added for the Xen, HVM and LXC
496 hypervisors. This is controlled by a new hypervisor parameter
497 ``cpu_mask``; details about possible values for this are in the
498 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
499 VCPU-to-CPU mapping) form will work well only when all nodes in your
500 cluster have the same amount of CPUs.
505 Another area in which Ganeti was not customisable were the parameters
506 used for storage configuration, e.g. how many stripes to use for LVM,
507 DRBD resync configuration, etc.
509 To improve this area, we've added disks parameters, which are
510 customisable at cluster and node group level, and which allow to
511 specify various parameters for disks (DRBD has the most parameters
512 currently), for example:
514 - DRBD resync algorithm and parameters (e.g. speed)
515 - the default VG for meta-data volumes for DRBD
516 - number of stripes for LVM (plain disk template)
519 These parameters can be modified via ``gnt-cluster modify -D …`` and
520 ``gnt-group modify -D …``, and are used at either instance creation (in
521 case of LVM stripes, for example) or at disk “activation” time
524 Rados block device support
525 ++++++++++++++++++++++++++
527 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
528 denoted by the ``rbd`` disk template type. This is considered
529 experimental, feedback is welcome. For details on configuring it, see
530 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
535 The existing master IP functionality works well only in simple setups (a
536 single network shared by all nodes); however, if nodes belong to
537 different networks, then the ``/32`` setup and lack of routing
538 information is not enough.
540 To allow the master IP to function well in more complex cases, the
541 system was reworked as follows:
543 - a master IP netmask setting has been added
544 - the master IP activation/turn-down code was moved from the node daemon
546 - whether to run the Ganeti-supplied master IP script or a user-supplied
547 on is a ``gnt-cluster init`` setting
549 Details about the location of the standard and custom setup scripts are
550 in the man page :manpage:`gnt-cluster(8)`; for information about the
551 setup script protocol, look at the Ganeti-supplied script.
556 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
559 It is now possible to use TLS-protected connections, and when renewing
560 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
561 it is now possible to specify spice or spice CA certificates. Also, it
562 is possible to configure a password for SPICE sessions via the
563 hypervisor parameter ``spice_password_file``.
565 There are also new parameters to control the compression and streaming
566 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
567 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
568 for the spice parameters.
570 Lastly, it is now possible to see the SPICE connection information via
571 ``gnt-instance console``.
576 A new tool (``tools/ovfconverter``) has been added that supports
577 conversion between Ganeti and the `Open Virtualization Format
578 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
581 This relies on the ``qemu-img`` tool to convert the disk formats, so the
582 actual compatibility with other virtualization solutions depends on it.
587 The configuration query daemon (``ganeti-confd``) is now optional, and
588 has been rewritten in Haskell; whether to use the daemon at all, use the
589 Python (default) or the Haskell version is selectable at configure time
590 via the ``--enable-confd`` parameter, which can take one of the
591 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
592 daemon will result in a smaller footprint; for larger systems, we
593 welcome feedback on the Haskell version which might become the default
596 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
597 daemon running. The Python version doesn't implement the new call.
600 User interface changes
601 ~~~~~~~~~~~~~~~~~~~~~~
603 We have replaced the ``--disks`` option of ``gnt-instance
604 replace-disks`` with a more flexible ``--disk`` option, which allows
605 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
606 disk size and mode can be changed upon recreation (via ``gnt-instance
607 recreate-disks``, which accepts the same ``--disk`` option).
609 As many people are used to a ``show`` command, we have added that as an
610 alias to ``info`` on all ``gnt-*`` commands.
612 The ``gnt-instance grow-disk`` command has a new mode in which it can
613 accept the target size of the disk, instead of the delta; this can be
614 more safe since two runs in absolute mode will be idempotent, and
615 sometimes it's also easier to specify the desired size directly.
617 Also the handling of instances with regard to offline secondaries has
618 been improved. Instance operations should not fail because one of it's
619 secondary nodes is offline, even though it's safe to proceed.
621 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
622 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
623 minors to instance name.
628 RAPI coverage has improved, with (for example) new resources for
629 recreate-disks, node power-cycle, etc.
634 There is partial support for ``xl`` in the Xen hypervisor; feedback is
637 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
638 whether to still support Python 2.4 or move to Python 2.6 as minimum
641 Support for Fedora has been slightly improved; the provided example
642 init.d script should work better on it and the INSTALL file should
643 document the needed dependencies.
648 The deprecated ``QueryLocks`` LUXI request has been removed. Use
649 ``Query(what=QR_LOCK, ...)`` instead.
651 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
652 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
653 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
654 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
655 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
657 RAPI client: ``CertificateError`` now derives from
658 ``GanetiApiError``. This should make it more easy to handle Ganeti
661 Deprecation warnings due to PyCrypto/paramiko import in
662 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
663 make sure to run an up-to-date paramiko version, if you use this tool.
665 The QA scripts now depend on Python 2.5 or above (the main code base
666 still works with Python 2.4).
668 The configuration file (``config.data``) is now written without
669 indentation for performance reasons; if you want to edit it, it can be
670 re-formatted via ``tools/fmtjson``.
672 A number of bugs has been fixed in the cluster merge tool.
674 ``x509`` certification verification (used in import-export) has been
675 changed to allow the same clock skew as permitted by the cluster
676 verification. This will remove some rare but hard to diagnose errors in
683 *(Released Thu, 19 Jul 2012)*
685 Very few changes from rc4 to the final release, only bugfixes:
687 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
688 instance, fix CDROM booting for KVM instances)
689 - fixed node group modification of node parameters
690 - fixed issue in LUClusterVerifyGroup with multi-group clusters
691 - fixed generation of bash completion to ensure a stable ordering
698 *(Released Fri, 13 Jul 2012)*
700 Third release candidate for 2.6. The following changes were done from
703 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
704 - Fixed an inconsistency in the LUXI protocol with the provided
705 arguments (NOT backwards compatible)
706 - Fixed a bug with node groups ipolicy where ``min`` was greater than
707 the cluster ``std`` value
708 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
709 easier instance debugging on nodes (requires ``hconfd`` to work)
715 *(Released Tue, 03 Jul 2012)*
717 Second release candidate for 2.6. The following changes were done from
720 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
722 - Fixed a RAPI regression on missing beparams/memory
723 - Fixed redistribution of files on offline nodes
724 - Added possibility to run activate-disks even though secondaries are
725 offline. With this change it relaxes also the strictness on some other
726 commands which use activate disks internally:
727 * ``gnt-instance start|reboot|rename|backup|export``
728 - Made it possible to remove safely an instance if its secondaries are
730 - Made it possible to reinstall even though secondaries are offline
736 *(Released Mon, 25 Jun 2012)*
738 First release candidate for 2.6. The following changes were done from
741 - Fixed bugs with disk parameters and ``rbd`` templates as well as
743 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
744 behaviour. It supports now the modify operation
745 - ``hcheck`` implemented to analyze cluster health and possibility of
746 improving health by rebalance
747 - ``hbal`` has been improved in dealing with split instances
753 *(Released Mon, 11 Jun 2012)*
755 Second beta release of 2.6. The following changes were done from beta2
758 - Fixed ``daemon-util`` with non-root user models
759 - Fixed creation of plain instances with ``--no-wait-for-sync``
760 - Fix wrong iv_names when running ``cfgupgrade``
761 - Export more information in RAPI group queries
762 - Fixed bug when changing instance network interfaces
763 - Extended burnin to do NIC changes
764 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
765 - Changed default for DRBD barriers
766 - Fixed DRBD error reporting for syncer rate
767 - Verify the options on disk parameters
769 And of course various fixes to documentation and improved unittests and
776 *(Released Wed, 23 May 2012)*
778 First beta release of 2.6. The following changes were done from beta1 to
781 - integrated patch for distributions without ``start-stop-daemon``
782 - adapted example init.d script to work on Fedora
783 - fixed log handling in Haskell daemons
784 - adapted checks in the watcher for pycurl linked against libnss
785 - add partial support for ``xl`` instead of ``xm`` for Xen
786 - fixed a type issue in cluster verification
787 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
790 Plus integrated fixes from the 2.5 branch:
792 - fixed ``kvm-ifup`` to use ``/bin/bash``
793 - fixed parallel build failures
794 - KVM live migration when using a custom keymap
800 *(Released Tue, 24 Jul 2012)*
802 A small bugfix release, with no new features:
804 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
805 different default shell (e.g. Debian, Ubuntu)
806 - fixed KVM startup and live migration with a custom keymap (fixes Issue
807 243 and Debian bug #650664)
808 - fixed compatibility with KVM versions that don't support multiple boot
809 devices (fixes Issue 230 and Debian bug #624256)
811 Additionally, a few fixes were done to the build system (fixed parallel
812 build failures) and to the unittests (fixed race condition in test for
813 FileID functions, and the default enable/disable mode for QA test is now
820 *(Released Fri, 11 May 2012)*
822 A small bugfix release.
824 The main issues solved are on the topic of compatibility with newer LVM
827 - fixed parsing of ``lv_attr`` field
828 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
829 the ``--force`` flag is needed
831 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
832 to accept kernel 3.x too (was hardcoded to 2.6.*).
834 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
835 of hook results and that also made display of other errors suboptimal
836 was fixed; the code behaves now like 2.4 and earlier.
838 Another change in 2.5, the cleanup of the OS scripts environment, is too
839 aggressive: it removed even the ``PATH`` variable, which requires the OS
840 scripts to *always* need to export it. Since this is a bit too strict,
841 we now export a minimal PATH, the same that we export for hooks.
843 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
844 integrated into this release.
846 Finally, a few other miscellaneous changes were done (no new features,
847 just small improvements):
849 - Fix ``gnt-group --help`` display
850 - Fix hardcoded Xen kernel path
851 - Fix grow-disk handling of invalid units
852 - Update synopsis for ``gnt-cluster repair-disk-sizes``
853 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
859 *(Released Thu, 12 Apr 2012)*
861 Incompatible/important changes and bugfixes
862 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
864 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
865 resource's ``ip_check`` parameter changed from ``True`` to ``False``
866 to match the underlying LUXI interface.
867 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
868 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
869 not maintain backwards-compatibility as the underlying operation
870 changed in an incompatible way. The RAPI client can talk to old
871 servers, but it needs to be told so as the return value changed.
872 - When creating file-based instances via RAPI, the ``file_driver``
873 parameter no longer defaults to ``loop`` and must be specified.
874 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
876 - Support for the undocumented and deprecated RAPI instance creation
877 request format version 0 has been dropped. Use version 1, supported
878 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
879 - Pyparsing 1.4.6 or above is required, see :doc:`installation
880 documentation <install>`.
881 - The "cluster-verify" hooks are now executed per group by the
882 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
883 if you just run ``gnt-cluster verify``, which generates one opcode per
885 - The environment as passed to the OS scripts is cleared, and thus no
886 environment variables defined in the node daemon's environment will be
887 inherited by the scripts.
888 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
890 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
891 support operations involving multiple node groups.
892 - Offline nodes are ignored when failing over an instance.
893 - Support for KVM version 1.0, which changed the version reporting format
895 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
897 - ``Makefile`` is now compatible with Automake 1.11.2
898 - Includes all bugfixes made in the 2.4 series
903 - The ganeti-htools project has been merged into the ganeti-core source
904 tree and will be built as part of Ganeti (see :doc:`install-quick`).
905 - Implemented support for :doc:`shared storage <design-shared-storage>`.
906 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
907 GPT-style partition tables (requires `parted
908 <http://www.gnu.org/s/parted/>`_).
909 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
910 - Allowed adding tags on instance creation.
911 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
912 - Allow instances to be started in a paused state, enabling the user to
913 see the complete console output on boot using the console.
914 - Added new hypervisor flag to control default reboot behaviour
915 (``reboot_behavior``).
916 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
917 - Improved out-of-band management support:
919 - Added ``gnt-node health`` command reporting the health status of
921 - Added ``gnt-node power`` command to manage power status of nodes.
922 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
924 - Instance migration can fall back to failover if instance is not
926 - Filters can be used when listing nodes, instances, groups and locks;
927 see :manpage:`ganeti(7)` manpage.
928 - Added post-execution status as variables to :doc:`hooks <hooks>`
930 - Instance tags are exported/imported together with the instance.
931 - When given an explicit job ID, ``gnt-job info`` will work for archived
933 - Jobs can define dependencies on other jobs (not yet supported via
934 RAPI or command line, but used by internal commands and usable via
937 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
940 - Instance failover is now available as a RAPI resource
941 (``/2/instances/[instance_name]/failover``).
942 - ``gnt-instance info`` defaults to static information if primary node
944 - Opcodes have a new ``comment`` attribute.
945 - Added basic SPICE support to KVM hypervisor.
946 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
948 Node group improvements
949 ~~~~~~~~~~~~~~~~~~~~~~~
951 - ``gnt-cluster verify`` has been modified to check groups separately,
952 thereby improving performance.
953 - Node group support has been added to ``gnt-cluster verify-disks``,
954 which now operates per node group.
955 - Watcher has been changed to work better with node groups.
957 - One process and state file per node group.
958 - Slow watcher in one group doesn't block other group's watcher.
960 - Added new command, ``gnt-group evacuate``, to move all instances in a
961 node group to other groups.
962 - Added ``gnt-instance change-group`` to move an instance to another
964 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
965 per-group operations.
966 - Node groups can be tagged.
967 - Some operations switch from an exclusive to a shared lock as soon as
969 - Instance's primary and secondary nodes' groups are now available as
970 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
971 and ``snodes.group.uuid``).
976 - Numerous updates to documentation and manpages.
978 - :doc:`RAPI <rapi>` documentation now has detailed parameter
980 - Some opcode/job results are now also documented, see :doc:`RAPI
983 - A lockset's internal lock is now also visible in lock monitor.
984 - Log messages from job queue workers now contain information about the
985 opcode they're processing.
986 - ``gnt-instance console`` no longer requires the instance lock.
987 - A short delay when waiting for job changes reduces the number of LUXI
988 requests significantly.
989 - DRBD metadata volumes are overwritten with zeros during disk creation.
990 - Out-of-band commands no longer acquire the cluster lock in exclusive
992 - ``devel/upload`` now uses correct permissions for directories.
998 *(Released Fri, 23 Mar 2012)*
1000 This was the sixth release candidate of the 2.5 series.
1006 *(Released Mon, 9 Jan 2012)*
1008 This was the fifth release candidate of the 2.5 series.
1014 *(Released Thu, 27 Oct 2011)*
1016 This was the fourth release candidate of the 2.5 series.
1022 *(Released Wed, 26 Oct 2011)*
1024 This was the third release candidate of the 2.5 series.
1030 *(Released Tue, 18 Oct 2011)*
1032 This was the second release candidate of the 2.5 series.
1038 *(Released Tue, 4 Oct 2011)*
1040 This was the first release candidate of the 2.5 series.
1046 *(Released Wed, 31 Aug 2011)*
1048 This was the third beta release of the 2.5 series.
1054 *(Released Mon, 22 Aug 2011)*
1056 This was the second beta release of the 2.5 series.
1062 *(Released Fri, 12 Aug 2011)*
1064 This was the first beta release of the 2.5 series.
1070 *(Released Thu, 27 Oct 2011)*
1072 - Fixed bug when parsing command line parameter values ending in
1074 - Fixed assertion error after unclean master shutdown
1075 - Disable HTTP client pool for RPC, significantly reducing memory usage
1077 - Fixed queue archive creation with wrong permissions
1083 *(Released Tue, 23 Aug 2011)*
1087 - Fixed documentation for importing with ``--src-dir`` option
1088 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1089 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1095 *(Released Fri, 5 Aug 2011)*
1097 Many bug-fixes and a few small features:
1099 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1100 issues when you tried to add an instance with two MAC addresses in one
1102 - KVM: fixed per-instance stored UID value
1103 - KVM: configure bridged NICs at migration start
1104 - KVM: Fix a bug where instance will not start with never KVM versions
1106 - Added OS search path to ``gnt-cluster info``
1107 - Fixed an issue with ``file_storage_dir`` where you were forced to
1108 provide an absolute path, but the documentation states it is a
1109 relative path, the documentation was right
1110 - Added a new parameter to instance stop/start called ``--no-remember``
1111 that will make the state change to not be remembered
1112 - Implemented ``no_remember`` at RAPI level
1113 - Improved the documentation
1114 - Node evacuation: don't call IAllocator if node is already empty
1115 - Fixed bug in DRBD8 replace disks on current nodes
1116 - Fixed bug in recreate-disks for DRBD instances
1117 - Moved assertion checking locks in ``gnt-instance replace-disks``
1118 causing it to abort with not owning the right locks for some situation
1119 - Job queue: Fixed potential race condition when cancelling queued jobs
1120 - Fixed off-by-one bug in job serial generation
1121 - ``gnt-node volumes``: Fix instance names
1122 - Fixed aliases in bash completion
1123 - Fixed a bug in reopening log files after being sent a SIGHUP
1124 - Added a flag to burnin to allow specifying VCPU count
1125 - Bugfixes to non-root Ganeti configuration
1131 *(Released Thu, 12 May 2011)*
1133 Many bug-fixes and a few new small features:
1135 - Fixed a bug related to log opening failures
1136 - Fixed a bug in instance listing with orphan instances
1137 - Fixed a bug which prevented resetting the cluster-level node parameter
1138 ``oob_program`` to the default
1139 - Many fixes related to the ``cluster-merge`` tool
1140 - Fixed a race condition in the lock monitor, which caused failures
1141 during (at least) creation of many instances in parallel
1142 - Improved output for gnt-job info
1143 - Removed the quiet flag on some ssh calls which prevented debugging
1145 - Improved the N+1 failure messages in cluster verify by actually
1146 showing the memory values (needed and available)
1147 - Increased lock attempt timeouts so that when executing long operations
1148 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1149 too early and thus prevent the use of the 'fair' mechanism
1150 - Changed instance query data (``gnt-instance info``) to not acquire
1151 locks unless needed, thus allowing its use on locked instance if only
1152 static information is asked for
1153 - Improved behaviour with filesystems that do not support rename on an
1155 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1156 kept locks on all nodes during the wipe, which is unneeded
1157 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1158 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1159 10GiB) which caused the wipe to fail right at the end in some cases
1160 - Fixed master IP activation when doing master failover with no-voting
1161 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1162 the master node itself
1163 - Fixed potential data-loss in under disk full conditions, where Ganeti
1164 wouldn't check correctly the return code and would consider
1165 partially-written files 'correct'
1166 - Fixed bug related to multiple VGs and DRBD disk replacing
1167 - Added new disk parameter ``metavg`` that allows placement of the meta
1168 device for DRBD in a different volume group
1169 - Fixed error handling in the node daemon when the system libc doesn't
1170 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1171 - Fixed lock release during replace-disks, which kept cluster-wide locks
1172 when doing disk replaces with an iallocator script
1173 - Added check for missing bridges in cluster verify
1174 - Handle EPIPE errors while writing to the terminal better, so that
1175 piping the output to e.g. ``less`` doesn't cause a backtrace
1176 - Fixed rare case where a ^C during Luxi calls could have been
1177 interpreted as server errors, instead of simply terminating
1178 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1180 - Added a few more parameters to the KVM hypervisor, allowing a second
1181 CDROM, custom disk type for CDROMs and a floppy image
1182 - Removed redundant message in instance rename when the name is given
1184 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1185 disks on new nodes, allowing recreation when the original instance
1186 nodes are completely gone
1187 - Added option when converting disk templates to DRBD to skip waiting
1188 for the resync, in order to make the instance available sooner
1189 - Added two new variables to the OS scripts environment (containing the
1191 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1192 to allow use of ``pvgrub`` as bootloader
1193 - Changed the instance memory modifications to only check out-of-memory
1194 conditions on memory increases, and turned the secondary node warnings
1195 into errors (they can still be overridden via ``--force``)
1196 - Fixed the handling of a corner case when the Python installation gets
1197 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1198 to execute a command that doesn't exist
1199 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1200 node of the instance returned failures during instance shutdown; this
1201 adds the option ``--ignore-consistency`` to gnt-instance move
1203 And as usual, various improvements to the error messages, documentation
1210 *(Released Wed, 09 Mar 2011)*
1212 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1213 the RAPI users file if run twice (even with ``--dry-run``).
1215 The release fixes that bug (nothing else changed).
1221 *(Released Mon, 07 Mar 2011)*
1223 Final 2.4.0 release. Just a few small fixes:
1225 - Fixed RAPI node evacuate
1226 - Fixed the kvm-ifup script
1227 - Fixed internal error handling for special job cases
1228 - Updated man page to specify the escaping feature for options
1234 *(Released Mon, 28 Feb 2011)*
1236 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1237 that this feature wiped the disks of the wrong instance, leading to loss
1242 - Fixed title of query field containing instance name
1243 - Expanded the glossary in the documentation
1244 - Fixed one unittest (internal issue)
1250 *(Released Mon, 21 Feb 2011)*
1252 A number of bug fixes plus just a couple functionality changes.
1254 On the user-visible side, the ``gnt-* list`` command output has changed
1255 with respect to "special" field states. The current rc1 style of display
1256 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1257 the default output mode special fields states are displayed as follows:
1259 - Offline resource: ``*``
1260 - Unavailable/not applicable: ``-``
1261 - Data missing (RPC failure): ``?``
1262 - Unknown field: ``??``
1264 Another user-visible change is the addition of ``--force-join`` to
1269 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1270 - Fixed regression in RAPI/instance reinstall where all parameters were
1271 required (instead of optional)
1272 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1273 - Fixed iallocator usage (offline nodes were not considered offline)
1274 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1275 - Fixed hypervisor and OS parameter validation with respect to
1276 non-vm_capable nodes
1277 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1279 - Fixed ``tools/listrunner`` with respect to agent-based usage
1285 *(Released Fri, 4 Feb 2011)*
1287 Many changes and fixes since the beta1 release. While there were some
1288 internal changes, the code has been mostly stabilised for the RC
1291 Note: the dumb allocator was removed in this release, as it was not kept
1292 up-to-date with the IAllocator protocol changes. It is recommended to
1293 use the ``hail`` command from the ganeti-htools package.
1295 Note: the 2.4 and up versions of Ganeti are not compatible with the
1296 0.2.x branch of ganeti-htools. You need to upgrade to
1297 ganeti-htools-0.3.0 (or later).
1299 Regressions fixed from 2.3
1300 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1302 - Fixed the ``gnt-cluster verify-disks`` command
1303 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1305 - Fixed disk adoption breakage
1306 - Fixed wrong headers in instance listing for field aliases
1311 - Fixed corner case in KVM handling of NICs
1312 - Fixed many cases of wrong handling of non-vm_capable nodes
1313 - Fixed a bug where a missing instance symlink was not possible to
1314 recreate with any ``gnt-*`` command (now ``gnt-instance
1315 activate-disks`` does it)
1316 - Fixed the volume group name as reported by ``gnt-cluster
1318 - Increased timeouts for the import-export code, hopefully leading to
1319 fewer aborts due network or instance timeouts
1320 - Fixed bug in ``gnt-node list-storage``
1321 - Fixed bug where not all daemons were started on cluster
1322 initialisation, but only at the first watcher run
1323 - Fixed many bugs in the OOB implementation
1324 - Fixed watcher behaviour in presence of instances with offline
1326 - Fixed instance list output for instances running on the wrong node
1327 - a few fixes to the cluster-merge tool, but it still cannot merge
1328 multi-node groups (currently it is not recommended to use this tool)
1334 - Improved network configuration for the KVM hypervisor
1335 - Added e1000 as a supported NIC for Xen-HVM
1336 - Improved the lvmstrap tool to also be able to use partitions, as
1337 opposed to full disks
1338 - Improved speed of disk wiping (the cluster parameter
1339 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1340 of instance creations
1341 - Added documentation for the OS parameters
1342 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1343 hypervisor is not responding
1344 - Added display of blacklisted and hidden OS information in
1345 ``gnt-cluster info``
1346 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1347 NIC and node parameters, which might create problems with currently
1348 invalid (but undetected) configuration files, but prevents validation
1349 failures when unrelated parameters are modified
1350 - Changed cluster initialisation to wait for the master daemon to become
1352 - Expanded the RAPI interface:
1354 - Added config redistribution resource
1355 - Added activation/deactivation of instance disks
1356 - Added export of console information
1358 - Implemented log file reopening on SIGHUP, which allows using
1359 logrotate(8) for the Ganeti log files
1360 - Added a basic OOB helper script as an example
1366 *(Released Fri, 14 Jan 2011)*
1371 - Fixed timezone issues when formatting timestamps
1372 - Added support for node groups, available via ``gnt-group`` and other
1374 - Added out-of-band framework and management, see :doc:`design
1375 document <design-oob>`
1376 - Removed support for roman numbers from ``gnt-node list`` and
1377 ``gnt-instance list``.
1378 - Allowed modification of master network interface via ``gnt-cluster
1379 modify --master-netdev``
1380 - Accept offline secondaries while shutting down instance disks
1381 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1382 - Added support for multiple LVM volume groups
1383 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1385 - Added commands to list available fields:
1387 - ``gnt-node list-fields``
1388 - ``gnt-group list-fields``
1389 - ``gnt-instance list-fields``
1391 - Updated documentation and man pages
1396 - Moved ``rapi_users`` file into separate directory, now named
1397 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1399 - Added new tool for running commands on many machines,
1400 ``tools/ganeti-listrunner``
1401 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1402 improving the ``gnt-instance console`` output
1403 - Allowed customisation of disk index separator at ``configure`` time
1404 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1405 - Added support for non-partitioned md disks in ``lvmstrap``
1406 - Added script to gracefully power off KVM instances
1407 - Split ``utils`` module into smaller parts
1408 - Changed query operations to return more detailed information, e.g.
1409 whether an information is unavailable due to an offline node. To use
1410 this new functionality, the LUXI call ``Query`` must be used. Field
1411 information is now stored by the master daemon and can be retrieved
1412 using ``QueryFields``. Instances, nodes and groups can also be queried
1413 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1414 exposed via RAPI). The following commands make use of this
1415 infrastructure change:
1417 - ``gnt-group list``
1418 - ``gnt-group list-fields``
1420 - ``gnt-node list-fields``
1421 - ``gnt-instance list``
1422 - ``gnt-instance list-fields``
1423 - ``gnt-debug locks``
1428 - New RAPI resources (see :doc:`rapi`):
1432 - ``/2/groups/[group_name]``
1433 - ``/2/groups/[group_name]/assign-nodes``
1434 - ``/2/groups/[group_name]/modify``
1435 - ``/2/groups/[group_name]/rename``
1436 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1440 - Implemented ``no_install`` for instance creation
1441 - Implemented OS parameters for instance reinstallation, allowing
1442 use of special settings on reinstallation (e.g. for preserving data)
1447 - Added IPv6 support in import/export
1448 - Pause DRBD synchronization while wiping disks on instance creation
1449 - Updated unittests and QA scripts
1450 - Improved network parameters passed to KVM
1451 - Converted man pages from docbook to reStructuredText
1457 *(Released Mon, 20 Dec 2010)*
1459 Released version 2.3.1~rc1 without any changes.
1465 *(Released Wed, 1 Dec 2010)*
1467 - impexpd: Disable OpenSSL compression in socat if possible (backport
1468 from master, commit e90739d625b, see :doc:`installation guide
1469 <install-quick>` for details)
1470 - Changed unittest coverage report to exclude test scripts
1471 - Added script to check version format
1477 *(Released Wed, 1 Dec 2010)*
1479 Released version 2.3.0~rc1 without any changes.
1485 *(Released Fri, 19 Nov 2010)*
1487 A number of bugfixes and documentation updates:
1489 - Update ganeti-os-interface documentation
1490 - Fixed a bug related to duplicate MACs or similar items which should be
1492 - Fix breakage in OS state modify
1493 - Reinstall instance: disallow offline secondaries (fixes bug related to
1494 OS changing but reinstall failing)
1495 - plus all the other fixes between 2.2.1 and 2.2.2
1501 *(Released Tue, 2 Nov 2010)*
1503 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1504 - Fixed master failover race with watcher
1505 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1507 - Accept previously stopped instance for export with instance removal
1508 - Simplify and extend the environment variables for instance OS scripts
1509 - Added new node flags, ``master_capable`` and ``vm_capable``
1510 - Added optional instance disk wiping prior during allocation. This is a
1511 cluster-wide option and can be set/modified using
1512 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1513 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1514 :doc:`install-quick`
1515 - Added a new watcher option (``--ignore-pause``)
1516 - Added option to ignore offline node on instance start/stop
1517 (``--ignore-offline``)
1518 - Allow overriding OS parameters with ``gnt-instance reinstall``
1519 - Added ability to change node's secondary IP address using ``gnt-node
1521 - Implemented privilege separation for all daemons except
1522 ``ganeti-noded``, see ``configure`` options
1523 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1525 - Implemented job priorities (see ``ganeti(7)`` manpage)
1526 - Ignore failures while shutting down instances during failover from
1528 - Exit daemon's bootstrap process only once daemon is ready
1529 - Export more information via ``LUInstanceQuery``/remote API
1530 - Improved documentation, QA and unittests
1531 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1532 restart if the file was created or changed
1533 - Added LUXI protocol version sent with each request and response,
1534 allowing detection of server/client mismatches
1535 - Moved the Python scripts among gnt-* and ganeti-* into modules
1536 - Moved all code related to setting up SSH to an external script,
1538 - Infrastructure changes for node group support in future versions
1544 *(Released Fri, 19 Nov 2010)*
1546 A few small bugs fixed, and some improvements to the build system:
1548 - Fix documentation regarding conversion to drbd
1549 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1551 - Fix error handling in node modify with multiple changes
1552 - Allow remote imports without checked names
1558 *(Released Tue, 19 Oct 2010)*
1560 - Disable SSL session ID cache in RPC client
1566 *(Released Thu, 14 Oct 2010)*
1568 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1569 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1571 - Fix problems with interaction between Curl and Python's HTTP server,
1572 resulting in increased speed in many RPC calls
1573 - Improve our release script to prevent breakage with older aclocal and
1580 *(Released Thu, 7 Oct 2010)*
1582 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1583 value retrieved from master
1584 - Added support for blacklisted or hidden OS definitions
1585 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1586 - Added support for -mem-path in KVM hypervisor abstraction layer
1587 - Allow overriding instance parameters in tool for inter-cluster
1588 instance moves (``tools/move-instance``)
1589 - Improved opcode summaries (e.g. in ``gnt-job list``)
1590 - Improve consistency of OS listing by sorting it
1591 - Documentation updates
1597 *(Released Fri, 8 Oct 2010)*
1599 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1605 *(Released Mon, 4 Oct 2010)*
1607 - Fixed regression in ``gnt-instance rename``
1613 *(Released Wed, 22 Sep 2010)*
1615 - Fixed OS_VARIANT variable for OS scripts
1616 - Fixed cluster tag operations via RAPI
1617 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1618 - Disabled RAPI CA checks in watcher
1624 *(Released Mon, 23 Aug 2010)*
1626 - Support DRBD versions of the format "a.b.c.d"
1628 - Re-introduce support for usage from multiple threads in RAPI client
1629 - Instance renames and modify via RAPI
1630 - Work around race condition between processing and archival in job
1632 - Mark opcodes following failed one as failed, too
1633 - Job field ``lock_status`` was removed due to difficulties making it
1634 work with the changed job queue in Ganeti 2.2; a better way to monitor
1635 locks is expected for a later 2.2.x release
1636 - Fixed dry-run behaviour with many commands
1637 - Support ``ssh-agent`` again when adding nodes
1638 - Many additional bugfixes
1644 *(Released Fri, 30 Jul 2010)*
1646 Important change: the internal RPC mechanism between Ganeti nodes has
1647 changed from using a home-grown http library (based on the Python base
1648 libraries) to use the PycURL library. This requires that PycURL is
1649 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1650 against GnuTLS by default. cURL's support for GnuTLS had known issues
1651 before cURL 7.21.0 and we recommend using the latest cURL release or
1652 linking against OpenSSL. Most other distributions already link PycURL
1653 and cURL against OpenSSL. The command::
1655 python -c 'import pycurl; print pycurl.version'
1657 can be used to determine the libraries PycURL and cURL are linked
1660 Other significant changes:
1662 - Rewrote much of the internals of the job queue, in order to achieve
1663 better parallelism; this decouples job query operations from the job
1664 processing, and it should allow much nicer behaviour of the master
1665 daemon under load, and it also has uncovered some long-standing bugs
1666 related to the job serialisation (now fixed)
1667 - Added a default iallocator setting to the cluster parameters,
1668 eliminating the need to always pass nodes or an iallocator for
1669 operations that require selection of new node(s)
1670 - Added experimental support for the LXC virtualization method
1671 - Added support for OS parameters, which allows the installation of
1672 instances to pass parameter to OS scripts in order to customise the
1674 - Added a hypervisor parameter controlling the migration type (live or
1675 non-live), since hypervisors have various levels of reliability; this
1676 has renamed the 'live' parameter to 'mode'
1677 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1678 logical volumes, meaning that cluster verify will ignore them and not
1679 flag their presence as errors
1680 - The watcher will now reset the error count for failed instances after
1681 8 hours, thus allowing self-healing if the problem that caused the
1682 instances to be down/fail to start has cleared in the meantime
1683 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1684 check for, and warn, if the drbd module parameter ``usermode_helper``
1685 is not consistent with the cluster-wide setting; this is needed to
1686 make diagnose easier of failed drbd creations
1687 - Started adding base IPv6 support, but this is not yet
1688 enabled/available for use
1689 - Rename operations (cluster, instance) will now return the new name,
1690 which is especially useful if a short name was passed in
1691 - Added support for instance migration in RAPI
1692 - Added a tool to pre-configure nodes for the SSH setup, before joining
1693 them to the cluster; this will allow in the future a simplified model
1694 for node joining (but not yet fully enabled in 2.2); this needs the
1695 paramiko python library
1696 - Fixed handling of name-resolving errors
1697 - Fixed consistency of job results on the error path
1698 - Fixed master-failover race condition when executed multiple times in
1700 - Fixed many bugs related to the job queue (mostly introduced during the
1701 2.2 development cycle, so not all are impacting 2.1)
1702 - Fixed instance migration with missing disk symlinks
1703 - Fixed handling of unknown jobs in ``gnt-job archive``
1704 - And many other small fixes/improvements
1708 - Enhanced both the unittest and the QA coverage
1709 - Switched the opcode validation to a generic model, and extended the
1710 validation to all opcode parameters
1711 - Changed more parts of the code that write shell scripts to use the
1713 - Switched the master daemon to use the asyncore library for the Luxi
1720 *(Released Thu, 17 Jun 2010)*
1722 - Added tool (``move-instance``) and infrastructure to move instances
1723 between separate clusters (see :doc:`separate documentation
1724 <move-instance>` and :doc:`design document <design-2.2>`)
1725 - Added per-request RPC timeout
1726 - RAPI now requires a Content-Type header for requests with a body (e.g.
1727 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1728 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1729 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1731 - Implemented initial support for running Ganeti daemons as separate
1732 users, see configure-time flags ``--with-user-prefix`` and
1733 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1735 - Instances can be removed after export (``gnt-backup export
1736 --remove-instance``)
1737 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1738 key (instead of 1024 bit)
1739 - Added new cluster configuration file for cluster domain secret
1740 - Import/export now use SSL instead of SSH
1741 - Added support for showing estimated time when exporting an instance,
1742 see the ``ganeti-os-interface(7)`` manpage and look for
1749 *(Released Tue, 16 Nov 2010)*
1751 Some more bugfixes. Unless critical bugs occur, this will be the last
1754 - Fix case of MAC special-values
1755 - Fix mac checker regex
1756 - backend: Fix typo causing "out of range" error
1757 - Add missing --units in gnt-instance list man page
1763 *(Released Tue, 24 Aug 2010)*
1766 - Don't ignore secondary node silently on non-mirrored disk templates
1768 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1769 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1770 - Properly document the usb_mouse parameter
1771 - Fix path in ganeti-rapi(8) (issue 116)
1772 - Adjust error message when the ganeti user's .ssh directory is
1774 - Add same-node-check when changing the disk template to drbd
1780 *(Released Fri, 16 Jul 2010)*
1783 - Add an option to only select some reboot types during qa/burnin.
1784 (on some hypervisors consequent reboots are not supported)
1785 - Fix infrequent race condition in master failover. Sometimes the old
1786 master ip address would be still detected as up for a short time
1787 after it was removed, causing failover to fail.
1788 - Decrease mlockall warnings when the ctypes module is missing. On
1789 Python 2.4 we support running even if no ctypes module is installed,
1790 but we were too verbose about this issue.
1791 - Fix building on old distributions, on which man doesn't have a
1793 - Fix RAPI not to ignore the MAC address on instance creation
1794 - Implement the old instance creation format in the RAPI client.
1800 *(Released Thu, 01 Jul 2010)*
1802 A small bugfix release:
1803 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1804 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1805 a non-existing option
1806 - Fix instance create: the --force-variant option was ignored
1807 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1808 - Fix modify node storage with non-FQDN arguments
1809 - Fix RAPI client to authenticate under Python 2.6 when used
1810 for more than 5 requests needing authentication
1811 - Fix gnt-instance modify -t (storage) giving a wrong error message
1812 when converting a non-shutdown drbd instance to plain
1818 *(Released Fri, 18 Jun 2010)*
1820 A small bugfix release:
1822 - Fix live migration of KVM instances started with older Ganeti
1823 versions which had fewer hypervisor parameters
1824 - Fix gnt-instance grow-disk on down instances
1825 - Fix an error-reporting bug during instance migration
1826 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1827 silently ignoring broken ones
1828 - Fix an RPC error reporting bug affecting, for example, RAPI client
1830 - Fix bug triggered by different API version os-es on different nodes
1831 - Fix a bug in instance startup with custom hvparams: OS level
1832 parameters would fail to be applied.
1833 - Fix the RAPI client under Python 2.6 (but more work is needed to
1834 make it work completely well with OpenSSL)
1835 - Fix handling of errors when resolving names from DNS
1841 *(Released Thu, 3 Jun 2010)*
1843 A medium sized development cycle. Some new features, and some
1844 fixes/small improvements/cleanups.
1846 Significant features
1847 ~~~~~~~~~~~~~~~~~~~~
1849 The node deamon now tries to mlock itself into memory, unless the
1850 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1851 its logs, and falls back to console logging. This allows emergency
1852 features such as ``gnt-node powercycle`` to work even in the event of a
1853 broken node disk (tested offlining the disk hosting the node's
1854 filesystem and dropping its memory caches; don't try this at home)
1856 KVM: add vhost-net acceleration support. It can be tested with a new
1857 enough version of the kernel and of qemu-kvm.
1859 KVM: Add instance chrooting feature. If you use privilege dropping for
1860 your VMs you can also now force them to chroot to an empty directory,
1861 before starting the emulated guest.
1863 KVM: Add maximum migration bandwith and maximum downtime tweaking
1864 support (requires a new-enough version of qemu-kvm).
1866 Cluster verify will now warn if the master node doesn't have the master
1867 ip configured on it.
1869 Add a new (incompatible) instance creation request format to RAPI which
1870 supports all parameters (previously only a subset was supported, and it
1871 wasn't possible to extend the old format to accomodate all the new
1872 features. The old format is still supported, and a client can check for
1873 this feature, before using it, by checking for its presence in the
1874 ``features`` RAPI resource.
1876 Now with ancient latin support. Try it passing the ``--roman`` option to
1877 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1878 (requires the python-roman module to be installed, in order to work).
1883 As usual many internal code refactorings, documentation updates, and
1886 - Lots of improvements and cleanups to the experimental Remote API
1887 (RAPI) client library.
1888 - A new unit test suite for the core daemon libraries.
1889 - A fix to creating missing directories makes sure the umask is not
1890 applied anymore. This enforces the same directory permissions
1892 - Better handling terminating daemons with ctrl+c (used when running
1893 them in debugging mode).
1894 - Fix a race condition in live migrating a KVM instance, when stat()
1895 on the old proc status file returned EINVAL, which is an unexpected
1897 - Fixed manpage checking with newer man and utf-8 charachters. But now
1898 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1904 *(Released Fri, 7 May 2010)*
1906 Fix a bug which prevented untagged KVM instances from starting.
1912 *(Released Fri, 7 May 2010)*
1914 Another release with a long development cycle, during which many
1915 different features were added.
1917 Significant features
1918 ~~~~~~~~~~~~~~~~~~~~
1920 The KVM hypervisor now can run the individual instances as non-root, to
1921 reduce the impact of a VM being hijacked due to bugs in the
1922 hypervisor. It is possible to run all instances as a single (non-root)
1923 user, to manually specify a user for each instance, or to dynamically
1924 allocate a user out of a cluster-wide pool to each instance, with the
1925 guarantee that no two instances will run under the same user ID on any
1928 An experimental RAPI client library, that can be used standalone
1929 (without the other Ganeti libraries), is provided in the source tree as
1930 ``lib/rapi/client.py``. Note this client might change its interface in
1931 the future, as we iterate on its capabilities.
1933 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1934 replace the cluster's certificates and crypto keys. This might help in
1935 case they have been compromised, or have simply expired.
1937 A new disk option for instance creation has been added that allows one
1938 to "adopt" currently existing logical volumes, with data
1939 preservation. This should allow easier migration to Ganeti from
1940 unmanaged (or managed via other software) instances.
1942 Another disk improvement is the possibility to convert between redundant
1943 (DRBD) and plain (LVM) disk configuration for an instance. This should
1944 allow better scalability (starting with one node and growing the
1945 cluster, or shrinking a two-node cluster to one node).
1947 A new feature that could help with automated node failovers has been
1948 implemented: if a node sees itself as offline (by querying the master
1949 candidates), it will try to shutdown (hard) all instances and any active
1950 DRBD devices. This reduces the risk of duplicate instances if an
1951 external script automatically failovers the instances on such nodes. To
1952 enable this, the cluster parameter ``maintain_node_health`` should be
1953 enabled; in the future this option (per the name) will enable other
1954 automatic maintenance features.
1956 Instance export/import now will reuse the original instance
1957 specifications for all parameters; that means exporting an instance,
1958 deleting it and the importing it back should give an almost identical
1959 instance. Note that the default import behaviour has changed from
1960 before, where it created only one NIC; now it recreates the original
1963 Cluster verify has added a few new checks: SSL certificates validity,
1964 /etc/hosts consistency across the cluster, etc.
1969 As usual, many internal changes were done, documentation fixes,
1972 - Fixed cluster initialization with disabled cluster storage (regression
1973 introduced in 2.1.1)
1974 - File-based storage supports growing the disks
1975 - Fixed behaviour of node role changes
1976 - Fixed cluster verify for some corner cases, plus a general rewrite of
1977 cluster verify to allow future extension with more checks
1978 - Fixed log spamming by watcher and node daemon (regression introduced
1980 - Fixed possible validation issues when changing the list of enabled
1982 - Fixed cleanup of /etc/hosts during node removal
1983 - Fixed RAPI response for invalid methods
1984 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1985 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1986 from ide disks, etc.)
1987 - Allow OS changes without re-installation (to record a changed OS
1988 outside of Ganeti, or to allow OS renames)
1989 - Allow instance creation without OS installation (useful for example if
1990 the OS will be installed manually, or restored from a backup not in
1992 - Implemented option to make cluster ``copyfile`` use the replication
1994 - Added list of enabled hypervisors to ssconf (possibly useful for
1996 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1998 - A partial form of node re-IP is possible via node readd, which now
1999 allows changed node primary IP
2000 - Command line utilities now show an informational message if the job is
2002 - The logs of the master daemon now show the PID/UID/GID of the
2009 *(Released Fri, 12 Mar 2010)*
2011 During the 2.1.0 long release candidate cycle, a lot of improvements and
2012 changes have accumulated with were released later as 2.1.1.
2017 The node evacuate command (``gnt-node evacuate``) was significantly
2018 rewritten, and as such the IAllocator protocol was changed - a new
2019 request type has been added. This unfortunate change during a stable
2020 series is designed to improve performance of node evacuations; on
2021 clusters with more than about five nodes and which are well-balanced,
2022 evacuation should proceed in parallel for all instances of the node
2023 being evacuated. As such, any existing IAllocator scripts need to be
2024 updated, otherwise the above command will fail due to the unknown
2025 request. The provided "dumb" allocator has not been updated; but the
2026 ganeti-htools package supports the new protocol since version 0.2.4.
2028 Another important change is increased validation of node and instance
2029 names. This might create problems in special cases, if invalid host
2030 names are being used.
2032 Also, a new layer of hypervisor parameters has been added, that sits at
2033 OS level between the cluster defaults and the instance ones. This allows
2034 customisation of virtualization parameters depending on the installed
2035 OS. For example instances with OS 'X' may have a different KVM kernel
2036 (or any other parameter) than the cluster defaults. This is intended to
2037 help managing a multiple OSes on the same cluster, without manual
2038 modification of each instance's parameters.
2040 A tool for merging clusters, ``cluster-merge``, has been added in the
2041 tools sub-directory.
2046 - Improved the int/float conversions that should make the code more
2047 robust in face of errors from the node daemons
2048 - Fixed the remove node code in case of internal configuration errors
2049 - Fixed the node daemon behaviour in face of inconsistent queue
2050 directory (e.g. read-only file-system where we can't open the files
2052 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2053 now it either aborts cleanly or, if given the new "auto_promote"
2054 parameter, will automatically promote other nodes as needed
2055 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2056 completely prevent Ganeti from working
2057 - Fixed bug for instance export when not all disks were successfully
2059 - Fixed behaviour of node add when the new node is slow in starting up
2061 - Fixed handling of signals in the LUXI client, which should improve
2062 behaviour of command-line scripts
2063 - Added checks for invalid node/instance names in the configuration (now
2064 flagged during cluster verify)
2065 - Fixed watcher behaviour for disk activation errors
2066 - Fixed two potentially endless loops in http library, which led to the
2067 RAPI daemon hanging and consuming 100% CPU in some cases
2068 - Fixed bug in RAPI daemon related to hashed passwords
2069 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2071 - Enhanced compatibility with non-Debian OSes, but not using absolute
2072 path in some commands and allowing customisation of the ssh
2073 configuration directory
2074 - Fixed possible future issue with new Python versions by abiding to the
2075 proper use of ``__slots__`` attribute on classes
2076 - Added checks that should prevent directory traversal attacks
2077 - Many documentation fixes based on feedback from users
2082 - Added an "early_release" more for instance replace disks and node
2083 evacuate, where we release locks earlier and thus allow higher
2084 parallelism within the cluster
2085 - Added watcher hooks, intended to allow the watcher to restart other
2086 daemons (e.g. from the ganeti-nbma project), but they can be used of
2087 course for any other purpose
2088 - Added a compile-time disable for DRBD barriers, to increase
2089 performance if the administrator trusts the power supply or the
2090 storage system to not lose writes
2091 - Added the option of using syslog for logging instead of, or in
2092 addition to, Ganeti's own log files
2093 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2094 can indeed boot from a paravirtual NIC
2095 - Added a generic debug level for many operations; while this is not
2096 used widely yet, it allows one to pass the debug value all the way to
2098 - Enhanced the hooks environment for instance moves (failovers,
2099 migrations) where the primary/secondary nodes changed during the
2100 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2101 - Enhanced data validations for many user-supplied values; one important
2102 item is the restrictions imposed on instance and node names, which
2103 might reject some (invalid) host names
2104 - Add a configure-time option to disable file-based storage, if it's not
2105 needed; this allows greater security separation between the master
2106 node and the other nodes from the point of view of the inter-node RPC
2108 - Added user notification in interactive tools if job is waiting in the
2109 job queue or trying to acquire locks
2110 - Added log messages when a job is waiting for locks
2111 - Added filtering by node tags in instance operations which admit
2112 multiple instances (start, stop, reboot, reinstall)
2113 - Added a new tool for cluster mergers, ``cluster-merge``
2114 - Parameters from command line which are of the form ``a=b,c=d`` can now
2115 use backslash escapes to pass in values which contain commas,
2116 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2118 - For KVM, the instance name is the first parameter passed to KVM, so
2119 that it's more visible in the process list
2125 *(Released Tue, 2 Mar 2010)*
2127 Ganeti 2.1 brings many improvements with it. Major changes:
2129 - Added infrastructure to ease automated disk repairs
2130 - Added new daemon to export configuration data in a cheaper way than
2131 using the remote API
2132 - Instance NICs can now be routed instead of being associated with a
2134 - Improved job locking logic to reduce impact of jobs acquiring multiple
2135 locks waiting for other long-running jobs
2137 In-depth implementation details can be found in the Ganeti 2.1 design
2143 - Added chroot hypervisor
2144 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2146 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2147 ``bootloader_path`` and ``bootloader_args``)
2148 - Added the ``use_localtime`` option for the xen-hvm and kvm
2149 hypervisors, and the default value for this has changed to false (in
2150 2.0 xen-hvm always enabled it)
2151 - Added luxi call to submit multiple jobs in one go
2152 - Added cluster initialization option to not modify ``/etc/hosts``
2154 - Added network interface parameters
2155 - Added dry run mode to some LUs
2156 - Added RAPI resources:
2158 - ``/2/instances/[instance_name]/info``
2159 - ``/2/instances/[instance_name]/replace-disks``
2160 - ``/2/nodes/[node_name]/evacuate``
2161 - ``/2/nodes/[node_name]/migrate``
2162 - ``/2/nodes/[node_name]/role``
2163 - ``/2/nodes/[node_name]/storage``
2164 - ``/2/nodes/[node_name]/storage/modify``
2165 - ``/2/nodes/[node_name]/storage/repair``
2167 - Added OpCodes to evacuate or migrate all instances on a node
2168 - Added new command to list storage elements on nodes (``gnt-node
2169 list-storage``) and modify them (``gnt-node modify-storage``)
2170 - Added new ssconf files with master candidate IP address
2171 (``ssconf_master_candidates_ips``), node primary IP address
2172 (``ssconf_node_primary_ips``) and node secondary IP address
2173 (``ssconf_node_secondary_ips``)
2174 - Added ``ganeti-confd`` and a client library to query the Ganeti
2175 configuration via UDP
2176 - Added ability to run hooks after cluster initialization and before
2178 - Added automatic mode for disk replace (``gnt-instance replace-disks
2180 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2181 after catastrophic data-loss
2182 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2184 - Added ``gnt-instance move`` command to move instances
2185 - Added ``gnt-cluster watcher`` command to control watcher
2186 - Added ``gnt-node powercycle`` command to powercycle nodes
2187 - Added new job status field ``lock_status``
2188 - Added parseable error codes to cluster verification (``gnt-cluster
2189 verify --error-codes``) and made output less verbose (use
2190 ``--verbose`` to restore previous behaviour)
2191 - Added UUIDs to the main config entities (cluster, nodes, instances)
2192 - Added support for OS variants
2193 - Added support for hashed passwords in the Ganeti remote API users file
2195 - Added option to specify maximum timeout on instance shutdown
2196 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2197 - Added new helper script to start and stop Ganeti daemons
2198 (``daemon-util``), with the intent to reduce the work necessary to
2199 adjust Ganeti for non-Debian distributions and to start/stop daemons
2201 - Added more unittests
2202 - Fixed critical bug in ganeti-masterd startup
2203 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2204 now customisable at the cluster level for both the KVM and Xen
2205 hypervisors using the new ``migration_port`` parameter
2206 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2207 reinstalling an instance
2208 - Allowed ``@`` in tag names
2209 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2210 - Many documentation updates
2211 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2212 - ``gnt-instance reinstall`` can now reinstall multiple instances
2213 - Updated many command line parameters
2214 - Introduced new OS API version 15
2215 - No longer support a default hypervisor
2216 - Treat virtual LVs as inexistent
2217 - Improved job locking logic to reduce lock contention
2218 - Match instance and node names case insensitively
2219 - Reimplemented bash completion script to be more complete
2226 *(Released Thu, 4 Feb 2010)*
2228 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2229 - Fix a string formatting bug
2230 - Improve safety of the code in some error paths
2231 - Improve data validation in the master of values returned from nodes
2237 *(Released Thu, 17 Dec 2009)*
2239 - Fix security issue due to missing validation of iallocator names; this
2240 allows local and remote execution of arbitrary executables
2241 - Fix failure of gnt-node list during instance removal
2242 - Ship the RAPI documentation in the archive
2248 *(Released Wed, 30 Sep 2009)*
2250 - Fixed many wrong messages
2251 - Fixed a few bugs related to the locking library
2252 - Fixed MAC checking at instance creation time
2253 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2254 - Fixed a few issues related to signal handling in both daemons and
2256 - Fixed the example startup script provided
2257 - Fixed insserv dependencies in the example startup script (patch from
2259 - Fixed handling of drained nodes in the iallocator framework
2260 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2262 - Fixed error related to invalid job IDs in job polling
2263 - Fixed job/opcode persistence on unclean master shutdown
2264 - Fixed handling of partial job processing after unclean master
2266 - Fixed error reporting from LUs, previously all errors were converted
2267 into execution errors
2268 - Fixed error reporting from burnin
2269 - Decreased significantly the memory usage of the job queue
2270 - Optimised slightly multi-job submission
2271 - Optimised slightly opcode loading
2272 - Backported the multi-job submit framework from the development
2273 branch; multi-instance start and stop should be faster
2274 - Added script to clean archived jobs after 21 days; this will reduce
2275 the size of the queue directory
2276 - Added some extra checks in disk size tracking
2277 - Added an example ethers hook script
2278 - Added a cluster parameter that prevents Ganeti from modifying of
2280 - Added more node information to RAPI responses
2281 - Added a ``gnt-job watch`` command that allows following the ouput of a
2283 - Added a bind-address option to ganeti-rapi
2284 - Added more checks to the configuration verify
2285 - Enhanced the burnin script such that some operations can be retried
2287 - Converted instance reinstall to multi-instance model
2293 *(Released Fri, 7 Aug 2009)*
2295 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2296 to allow using the pre-2.0.2 behaviour in activation, if any existing
2297 instances have mismatched disk sizes in the configuration
2298 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2299 any configuration mismatches for disk sizes
2300 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2301 failover to work on two-node clusters
2302 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2304 - Fixed detection of OS script errors in ``gnt-backup export``
2305 - Fixed exit code of ``gnt-backup export``
2311 *(Released Fri, 17 Jul 2009)*
2313 - Added experimental support for stripped logical volumes; this should
2314 enhance performance but comes with a higher complexity in the block
2315 device handling; stripping is only enabled when passing
2316 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2317 affected even in the non-stripped mode
2318 - Improved resiliency against transient failures at the end of DRBD
2319 resyncs, and in general of DRBD resync checks
2320 - Fixed a couple of issues with exports and snapshot errors
2321 - Fixed a couple of issues in instance listing
2322 - Added display of the disk size in ``gnt-instance info``
2323 - Fixed checking for valid OSes in instance creation
2324 - Fixed handling of the "vcpus" parameter in instance listing and in
2325 general of invalid parameters
2326 - Fixed http server library, and thus RAPI, to handle invalid
2327 username/password combinations correctly; this means that now they
2328 report unauthorized for queries too, not only for modifications,
2329 allowing earlier detect of configuration problems
2330 - Added a new "role" node list field, equivalent to the master/master
2331 candidate/drained/offline flags combinations
2332 - Fixed cluster modify and changes of candidate pool size
2333 - Fixed cluster verify error messages for wrong files on regular nodes
2334 - Fixed a couple of issues with node demotion from master candidate role
2335 - Fixed node readd issues
2336 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2337 - Added a new ``--no-voting`` option for masterfailover to fix failover
2338 on two-nodes clusters when the former master node is unreachable
2339 - Added instance reinstall over RAPI
2345 *(Released Tue, 16 Jun 2009)*
2347 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2348 allow re-adding the start in single-user option (regression from 1.2)
2349 - the watcher writes the instance status to a file, to allow monitoring
2350 to report the instance status (from the master) based on cached
2351 results of the watcher's queries; while this can get stale if the
2352 watcher is being locked due to other work on the cluster, this is
2353 still an improvement
2354 - the watcher now also restarts the node daemon and the rapi daemon if
2356 - fixed the watcher to handle full and drained queue cases
2357 - hooks export more instance data in the environment, which helps if
2358 hook scripts need to take action based on the instance's properties
2359 (no longer need to query back into ganeti)
2360 - instance failovers when the instance is stopped do not check for free
2361 RAM, so that failing over a stopped instance is possible in low memory
2363 - rapi uses queries for tags instead of jobs (for less job traffic), and
2364 for cluster tags it won't talk to masterd at all but read them from
2366 - a couple of error handling fixes in RAPI
2367 - drbd handling: improved the error handling of inconsistent disks after
2368 resync to reduce the frequency of "there are some degraded disks for
2369 this instance" messages
2370 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2371 error handling path called a wrong function name)
2377 *(Released Wed, 27 May 2009)*
2379 - no changes from rc5
2385 *(Released Wed, 20 May 2009)*
2387 - fix a couple of bugs (validation, argument checks)
2388 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2389 - some small improvements to RAPI and IAllocator
2390 - make watcher automatically start the master daemon if down
2396 *(Released Mon, 27 Apr 2009)*
2398 - change the OS list to not require locks; this helps with big clusters
2399 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2400 volume group is broken
2401 - ``gnt-instance info``, without any arguments, doesn't run for all
2402 instances anymore; either pass ``--all`` or pass the desired
2403 instances; this helps against mistakes on big clusters where listing
2404 the information for all instances takes a long time
2405 - miscellaneous doc and man pages fixes
2411 *(Released Wed, 8 Apr 2009)*
2413 - Change the internal locking model of some ``gnt-node`` commands, in
2414 order to reduce contention (and blocking of master daemon) when
2415 batching many creation/reinstall jobs
2416 - Fixes to Xen soft reboot
2417 - No longer build documentation at build time, instead distribute it in
2418 the archive, in order to reduce the need for the whole docbook/rst
2425 *(Released Fri, 27 Mar 2009)*
2427 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2428 - Fix watcher startup sequence, improves the behaviour of busy clusters
2429 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2430 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2431 bind address and other places
2432 - Some documentation fixes and updates
2438 *(Released Mon, 2 Mar 2009)*
2440 - More documentation updates, now all docs should be more-or-less
2442 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2444 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2446 - Fix an issue related to $libdir/run/ganeti and cluster creation
2452 *(Released Thu, 19 Feb 2009)*
2454 - Xen PVM and KVM have switched the default value for the instance root
2455 disk to the first partition on the first drive, instead of the whole
2456 drive; this means that the OS installation scripts must be changed
2458 - Man pages have been updated
2459 - RAPI has been switched by default to HTTPS, and the exported functions
2460 should all work correctly
2461 - RAPI v1 has been removed
2462 - Many improvements to the KVM hypervisor
2463 - Block device errors are now better reported
2464 - Many other bugfixes and small improvements
2470 *(Released Mon, 26 Jan 2009)*
2472 - Version 2 is a general rewrite of the code and therefore the
2473 differences are too many to list, see the design document for 2.0 in
2474 the ``doc/`` subdirectory for more details
2475 - In this beta version there is not yet a migration path from 1.2 (there
2476 will be one in the final 2.0 release)
2477 - A few significant changes are:
2479 - all commands are executed by a daemon (``ganeti-masterd``) and the
2480 various ``gnt-*`` commands are just front-ends to it
2481 - all the commands are entered into, and executed from a job queue,
2482 see the ``gnt-job(8)`` manpage
2483 - the RAPI daemon supports read-write operations, secured by basic
2484 HTTP authentication on top of HTTPS
2485 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2486 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2487 to migrate to DRBD 8 first while still running Ganeti 1.2)
2488 - DRBD devices are using statically allocated minor numbers, which
2489 will be assigned to existing instances during the migration process
2490 - there is support for both Xen PVM and Xen HVM instances running on
2492 - KVM virtualization is supported too
2493 - file-based storage has been implemented, which means that it is
2494 possible to run the cluster without LVM and DRBD storage, for
2495 example using a shared filesystem exported from shared storage (and
2496 still have live migration)
2502 *(Released Tue, 13 Jan 2009)*
2504 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2505 - Reuse the old instance mac address by default on instance import, if
2506 the instance name is the same.
2507 - Handle situations in which the node info rpc returns incomplete
2509 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2510 - Improved version of batcher:
2512 - state file support
2513 - instance mac address support
2514 - support for HVM clusters/instances
2516 - Add an option to show the number of cpu sockets and nodes in
2518 - Support OSes that handle more than one version of the OS api (but do
2519 not change the current API in any other way)
2520 - Fix ``gnt-node migrate``
2521 - ``gnt-debug`` man page
2522 - Fixes various more typos and small issues
2523 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2529 *(Released Wed, 24 Sep 2008)*
2531 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2532 type of disk exported to fully virtualized instances.
2533 - provide access to the serial console of HVM instances
2534 - instance auto_balance flag, set by default. If turned off it will
2535 avoid warnings on cluster verify if there is not enough memory to fail
2536 over an instance. in the future it will prevent automatically failing
2537 it over when we will support that.
2538 - batcher tool for instance creation, see ``tools/README.batcher``
2539 - ``gnt-instance reinstall --select-os`` to interactively select a new
2540 operating system when reinstalling an instance.
2541 - when changing the memory amount on instance modify a check has been
2542 added that the instance will be able to start. also warnings are
2543 emitted if the instance will not be able to fail over, if auto_balance
2545 - documentation fixes
2546 - sync fields between ``gnt-instance list/modify/add/import``
2547 - fix a race condition in drbd when the sync speed was set after giving
2548 the device a remote peer.
2554 *(Released Tue, 22 Jul 2008)*
2556 - note: the allowed size and number of tags per object were reduced
2557 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2558 - fixed twisted 8.x compatibility
2559 - fixed ``gnt-instance replace-disks`` with iallocator
2560 - add TCP keepalives on twisted connections to detect restarted nodes
2561 - disk increase support, see ``gnt-instance grow-disk``
2562 - implement bulk node/instance query for RAPI
2563 - add tags in node/instance listing (optional)
2564 - experimental migration (and live migration) support, read the man page
2565 for ``gnt-instance migrate``
2566 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2567 has some small improvements in handling its state file
2573 *(Released Fri, 13 Jun 2008)*
2575 - Experimental readonly, REST-based remote API implementation;
2576 automatically started on master node, TCP port 5080, if enabled by
2577 ``--enable-rapi`` parameter to configure script.
2578 - Instance allocator support. Add and import instance accept a
2579 ``--iallocator`` parameter, and call that instance allocator to decide
2580 which node to use for the instance. The iallocator document describes
2581 what's expected from an allocator script.
2582 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2583 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2584 node is lost there is still enough memory to fail over the instances
2586 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2587 ``gnt-cluster verify``, to check for site-specific compliance. All the
2588 hooks will run, and their output, if any, will be displayed. Any
2589 failing hook will make the verification return an error value.
2590 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2591 primary and secondary interfaces
2592 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2593 that is still declared as part of the cluster and has failed.
2594 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2595 specifying output fields, that just adds the chosen fields to the
2597 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2598 export from the filesystem.
2599 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2600 have been added. Using them you can enable/disable acpi and pae
2601 support, and specify a path for a cd image to be exported to the
2602 instance. These parameters as the name suggest only work on HVM
2604 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2605 PAE support will be set to the previously hardcoded values, but the
2606 (previously hardcoded) path to the CDROM ISO image will be unset and
2607 if required, needs to be set manually with ``gnt-instance modify``
2609 - The address to which an instance's VNC console is bound is now
2610 selectable per-instance, rather than being cluster wide. Of course
2611 this only applies to instances controlled via VNC, so currently just
2612 applies to HVM clusters.
2618 *(Released Mon, 18 Feb 2008)*
2620 - more tweaks to the disk activation code (especially helpful for DRBD)
2621 - change the default ``gnt-instance list`` output format, now there is
2622 one combined status field (see the manpage for the exact values this
2624 - some more fixes for the mac export to hooks change
2625 - make Ganeti not break with DRBD 8.2.x (which changed the version
2626 format in ``/proc/drbd``) (issue 24)
2627 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2628 template, allowing migration from DRBD0.7+MD to DRBD8
2634 *(Released Wed, 30 Jan 2008)*
2636 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2638 - add command aliases infrastructure and a few aliases
2639 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2640 man pages and the ``--help`` option of ``gnt-node
2641 list``/``gnt-instance list``
2642 - fix ``gnt-backup list`` with down nodes (issue 21)
2643 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2644 - fix the dist archive and add a check for including svn/git files in
2646 - some developer-related changes: improve the burnin and the QA suite,
2647 add an upload script for testing during development
2653 *(Released Wed, 16 Jan 2008)*
2655 - experimental HVM support, read the install document, section
2656 "Initializing the cluster"
2657 - allow for the PVM hypervisor per-instance kernel and initrd paths
2658 - add a new command ``gnt-cluster verify-disks`` which uses a new
2659 algorithm to improve the reconnection of the DRBD pairs if the device
2660 on the secondary node has gone away
2661 - make logical volume code auto-activate LVs at disk activation time
2662 - slightly improve the speed of activating disks
2663 - allow specification of the MAC address at instance creation time, and
2664 changing it later via ``gnt-instance modify``
2665 - fix handling of external commands that generate lots of output on
2667 - update documentation with regard to minimum version of DRBD8 supported
2673 *(Released Tue, 4 Dec 2007)*
2675 - Log the ``xm create`` output to the node daemon log on failure (to
2676 help diagnosing the error)
2677 - In debug mode, log all external commands output if failed to the logs
2678 - Change parsing of lvm commands to ignore stderr
2684 *(Released Wed, 28 Nov 2007)*
2686 - Another round of updates to the DRBD 8 code to deal with more failures
2687 in the replace secondary node operation
2688 - Some more logging of failures in disk operations (lvm, drbd)
2689 - A few documentation updates
2696 *(Released Tue, 13 Nov 2007)*
2698 - Change configuration file format from Python's Pickle to JSON.
2699 Upgrading is possible using the cfgupgrade utility.
2700 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2701 faster replace disks and is more stable (DRBD 8 has many improvements
2702 compared to DRBD 0.7)
2703 - Added command line tags support (see man pages for ``gnt-instance``,
2704 ``gnt-node``, ``gnt-cluster``)
2705 - Added instance rename support
2706 - Added multi-instance startup/shutdown
2707 - Added cluster rename support
2708 - Added ``gnt-node evacuate`` to simplify some node operations
2709 - Added instance reboot operation that can speedup reboot as compared to
2711 - Soften the requirement that hostnames are in FQDN format
2712 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2714 - Removed dependency on debian's patched fping that uses the
2715 non-standard ``-S`` option
2716 - Now the OS definitions are searched for in multiple, configurable
2717 paths (easier for distros to package)
2718 - Some changes to the hooks infrastructure (especially the new
2719 post-configuration update hook)
2720 - Other small bugfixes
2722 .. vim: set textwidth=72 syntax=rst :