10 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
11 the correct command syntax. It is possible to use different DRBD versions
12 on different nodes as long as they are compatible to each other. This
13 enables rolling upgrades of DRBD with no downtime. As permanent operation
14 of different DRBD versions within a node group is discouraged,
15 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
16 - hroller now also plans for capacity to move non-redundant instances off
17 any node to be rebooted; the old behavior of completely ignoring any
18 non-redundant instances can be restored by adding the --ignore-non-redundant
27 - The minimum Python version needed to run Ganeti is now 2.6.
28 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
29 to always require authentication, ``--require-authentication``. It can
30 be specified in ``$sysconfdir/default/ganeti``.
31 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
32 be used to manage the disk templates to be used by instances in the cluster.
33 Initially, it will be set to a list that includes plain, drbd, if they were
34 enabled by specifying a volume group name, and file and sharedfile, if those
35 were enabled at configure time. Additionally, it will include all disk
36 templates that are currently used by instances. The order of disk templates
37 will be based on Ganeti's history of supporting them. In the future, the
38 first entry of the list will be used as a default disk template on instance
40 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
41 configuration back to the previous stable version.
42 - The cluster option '--no-lvm-storage' was removed in favor of the new option
43 '--enabled-disk-templates'.
51 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
53 - Disks added through ``gnt-instance modify`` or created through
54 ``gnt-instance recreate-disks`` are wiped, if the
55 ``prealloc_wipe_disks`` flag is set.
56 - If wiping newly created disks fails, the disks are removed. Also,
57 partial failures in creating disks through ``gnt-instance modify``
58 triggers a cleanup of the partially-created disks.
59 - Removing the master IP address doesn't fail if the address has been
66 *(Released Fri, 3 May 2013)*
68 Incompatible/important changes
69 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
71 - Instance policies for disk size were documented to be on a per-disk
72 basis, but hail applied them to the sum of all disks. This has been
74 - ``hbal`` will now exit with status 0 if, during job execution over
75 LUXI, early exit has been requested and all jobs are successful;
76 before, exit status 1 was used, which cannot be differentiated from
78 - Compatibility with newer versions of rbd has been fixed
79 - ``gnt-instance batch-create`` has been changed to use the bulk create
80 opcode from Ganeti. This lead to incompatible changes in the format of
81 the JSON file. It's now not a custom dict anymore but a dict
82 compatible with the ``OpInstanceCreate`` opcode.
83 - Parent directories for file storage need to be listed in
84 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
85 write the file automatically based on old configuration values, but it
86 can not distribute it across all nodes and the file contents should be
87 verified. Use ``gnt-cluster copyfile
88 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
89 upgraded. The reason for requiring this list of paths now is that
90 before it would have been possible to inject new paths via RPC,
91 allowing files to be created in arbitrary locations. The RPC protocol
92 is protected using SSL/X.509 certificates, but as a design principle
93 Ganeti does not permit arbitrary paths to be passed.
94 - The parsing of the variants file for OSes (see
95 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
96 lines and comment lines (starting with ``#``) are ignored for better
98 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
99 longer available. ``gnt-node add`` now invokes a new tool on the
100 destination node, named ``prepare-node-join``, to configure the SSH
101 daemon. Paramiko is no longer necessary to configure nodes' SSH
102 daemons via ``gnt-node add``.
103 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
104 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
105 the flag is not reset after a master failover.
106 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
107 is recommended. 2.6 will be mandatory from the 2.8 series.
113 - New network management functionality to support automatic allocation
114 of IP addresses and managing of network parameters. See
115 :manpage:`gnt-network(8)` for more details.
116 - New external storage backend, to allow managing arbitrary storage
117 systems external to the cluster. See
118 :manpage:`ganeti-extstorage-interface(7)`.
119 - New ``exclusive-storage`` node parameter added, restricted to
120 nodegroup level. When it's set to true, physical disks are assigned in
121 an exclusive fashion to instances, as documented in :doc:`Partitioned
122 Ganeti <design-partitioned>`. Currently, only instances using the
123 ``plain`` disk template are supported.
124 - The KVM hypervisor has been updated with many new hypervisor
125 parameters, including a generic one for passing arbitrary command line
126 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
127 compatible up to qemu 1.4.
128 - A new tool, called ``mon-collector``, is the stand-alone executor of
129 the data collectors for a monitoring system. As of this version, it
130 just includes the DRBD data collector, that can be executed by calling
131 ``mon-collector`` using the ``drbd`` parameter. See
132 :manpage:`mon-collector(7)`.
133 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
134 for RAPI users. It allows granting permissions to query for
135 information to a specific user without giving
136 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
137 - A new tool named ``node-cleanup`` has been added. It cleans remains of
138 a cluster from a machine by stopping all daemons, removing
139 certificates and ssconf files. Unless the ``--no-backup`` option is
140 given, copies of the certificates are made.
141 - Instance creations now support the use of opportunistic locking,
142 potentially speeding up the (parallel) creation of multiple instances.
143 This feature is currently only available via the :doc:`RAPI
144 <rapi>` interface and when an instance allocator is used. If the
145 ``opportunistic_locking`` parameter is set the opcode will try to
146 acquire as many locks as possible, but will not wait for any locks
147 held by other opcodes. If not enough resources can be found to
148 allocate the instance, the temporary error code
149 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
150 retried thereafter, with or without opportunistic locking.
151 - New experimental linux-ha resource scripts.
152 - Restricted-commands support: ganeti can now be asked (via command line
153 or rapi) to perform commands on a node. These are passed via ganeti
154 RPC rather than ssh. This functionality is restricted to commands
155 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
156 reasons. The file is not copied automatically.
162 - Diskless instances are now externally mirrored (Issue 237). This for
163 now has only been tested in conjunction with explicit target nodes for
165 - Queries not needing locks or RPC access to the node can now be
166 performed by the confd daemon, making them independent from jobs, and
167 thus faster to execute. This is selectable at configure time.
168 - The functionality for allocating multiple instances at once has been
169 overhauled and is now also available through :doc:`RAPI <rapi>`.
174 - Fix kvm compatibility with qemu 1.4 (Issue 389)
175 - Documentation updates (admin guide, upgrade notes, install
176 instructions) (Issue 372)
177 - Fix gnt-group list nodes and instances count (Issue 436)
178 - Fix compilation without non-mandatory libraries (Issue 441)
179 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
180 - Make confd logging more verbose at INFO level (Issue 435)
181 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
182 - Fix failure path for instance storage type conversion (Issue 229)
183 - Update htools text backend documentation
184 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
185 - Disable inter-cluster instance move for file-based instances, because
186 it is dependant on instance export, which is not supported for
187 file-based instances. (Issue 414)
188 - Fix gnt-job crashes on non-ascii characters (Issue 427)
189 - Fix volume group checks on non-vm-capable nodes (Issue 432)
195 *(Released Mon, 22 Apr 2013)*
197 This was the third beta release of the 2.7 series. Since beta2:
199 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
200 - Fix data loss on wrong usage of ``gnt-instance move``
201 - Properly export errors in confd-based job queries
202 - Add ``users-setup`` tool
203 - Fix iallocator protocol to report 0 as a disk size for diskless
204 instances. This avoids hail breaking when a diskless instance is
206 - Fix job queue directory permission problem that made confd job queries
207 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
208 for access to archived jobs (Issue 406).
209 - Limit the sizes of networks supported by ``gnt-network`` to something
210 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
211 - Fix bugs in instance disk template conversion
212 - Fix GHC 7 compatibility
213 - Fix ``burnin`` install path (Issue 426).
214 - Allow very small disk grows (Issue 347).
215 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
216 that noded doesn't import masterd code (Issue 419).
217 - Make sure the default metavg at cluster init is the same as the vg, if
218 unspecified (Issue 358).
219 - Fix cleanup of partially created disks (part of Issue 416)
225 *(Released Tue, 2 Apr 2013)*
227 This was the second beta release of the 2.7 series. Since beta1:
229 - Networks no longer have a "type" slot, since this information was
230 unused in Ganeti: instead of it tags should be used.
231 - The rapi client now has a ``target_node`` option to MigrateInstance.
232 - Fix early exit return code for hbal (Issue 386).
233 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
234 - Fix ``rbd showmapped`` output parsing (Issue 312).
235 - Networks are now referenced indexed by UUID, rather than name. This
236 will require running cfgupgrade, from 2.7.0beta1, if networks are in
238 - The OS environment now includes network information.
239 - Deleting of a network is now disallowed if any instance nic is using
240 it, to prevent dangling references.
241 - External storage is now documented in man pages.
242 - The exclusive_storage flag can now only be set at nodegroup level.
243 - Hbal can now submit an explicit priority with its jobs.
244 - Many network related locking fixes.
245 - Bump up the required pylint version to 0.25.1.
246 - Fix the ``no_remember`` option in RAPI client.
247 - Many ipolicy related tests, qa, and fixes.
248 - Many documentation improvements and fixes.
249 - Fix building with ``--disable-file-storage``.
250 - Fix ``-q`` option in htools, which was broken if passed more than
252 - Some haskell/python interaction improvements and fixes.
253 - Fix iallocator in case of missing LVM storage.
254 - Fix confd config load in case of ``--no-lvm-storage``.
255 - The confd/query functionality is now mentioned in the security
262 *(Released Wed, 6 Feb 2013)*
264 This was the first beta release of the 2.7 series. All important changes
265 are listed in the latest 2.7 entry.
271 *(Released Fri, 21 Dec 2012)*
273 Important behaviour change: hbal won't rebalance anymore instances which
274 have the ``auto_balance`` attribute set to false. This was the intention
275 all along, but until now it only skipped those from the N+1 memory
276 reservation (DRBD-specific).
278 A significant number of bug fixes in this release:
280 - Fixed disk adoption interaction with ipolicy checks.
281 - Fixed networking issues when instances are started, stopped or
282 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
283 - Fixed the warning in cluster verify for shared storage instances not
285 - Fixed removal of storage directory on shared file storage (issue 262).
286 - Fixed validation of LVM volume group name in OpClusterSetParams
287 (``gnt-cluster modify``) (issue 285).
288 - Fixed runtime memory increases (``gnt-instance modify -m``).
289 - Fixed live migration under Xen's ``xl`` mode.
290 - Fixed ``gnt-instance console`` with ``xl``.
291 - Fixed building with newer Haskell compiler/libraries.
292 - Fixed PID file writing in Haskell daemons (confd); this prevents
293 restart issues if confd was launched manually (outside of
294 ``daemon-util``) while another copy of it was running
295 - Fixed a type error when doing live migrations with KVM (issue 297) and
296 the error messages for failing migrations have been improved.
297 - Fixed opcode validation for the out-of-band commands (``gnt-node
299 - Fixed a type error when unsetting OS hypervisor parameters (issue
300 311); now it's possible to unset all OS-specific hypervisor
302 - Fixed the ``dry-run`` mode for many operations: verification of
303 results was over-zealous but didn't take into account the ``dry-run``
304 operation, resulting in "wrong" failures.
305 - Fixed bash completion in ``gnt-job list`` when the job queue has
306 hundreds of entries; especially with older ``bash`` versions, this
307 results in significant CPU usage.
309 And lastly, a few other improvements have been made:
311 - Added option to force master-failover without voting (issue 282).
312 - Clarified error message on lock conflict (issue 287).
313 - Logging of newly submitted jobs has been improved (issue 290).
314 - Hostname checks have been made uniform between instance rename and
316 - The ``--submit`` option is now supported by ``gnt-debug delay``.
317 - Shutting down the master daemon by sending SIGTERM now stops it from
318 processing jobs waiting for locks; instead, those jobs will be started
319 once again after the master daemon is started the next time (issue
321 - Support for Xen's ``xl`` program has been improved (besides the fixes
323 - Reduced logging noise in the Haskell confd daemon (only show one log
324 entry for each config reload, instead of two).
325 - Several man page updates and typo fixes.
331 *(Released Fri, 12 Oct 2012)*
333 A small bugfix release. Among the bugs fixed:
335 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
336 made the command unusable.
337 - Commands that issue many jobs don't fail anymore just because some jobs
338 take so long that other jobs are archived.
339 - Failures during ``gnt-instance reinstall`` are reflected by the exit
341 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
343 - When ``always_failover`` is set, ``--allow-failover`` is not required
344 in migrate commands anymore.
345 - ``bash_completion`` works even if extglob is disabled.
346 - Fixed bug with locks that made failover for RDB-based instances fail.
347 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
348 a random node instead of one based on the allocator metric.
349 - Support for newer versions of pylint and pep8.
350 - Hail doesn't fail anymore when trying to add an instance of type
351 ``file``, ``sharedfile`` or ``rbd``.
352 - Added new Makefile target to rebuild the whole distribution, so that
353 all files are included.
359 *(Released Fri, 27 Jul 2012)*
362 .. attention:: The ``LUXI`` protocol has been made more consistent
363 regarding its handling of command arguments. This, however, leads to
364 incompatibility issues with previous versions. Please ensure that you
365 restart Ganeti daemons soon after the upgrade, otherwise most
366 ``LUXI`` calls (job submission, setting/resetting the drain flag,
367 pausing/resuming the watcher, cancelling and archiving jobs, querying
368 the cluster configuration) will fail.
377 The current ``admin_up`` field, which used to denote whether an instance
378 should be running or not, has been removed. Instead, ``admin_state`` is
379 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
381 The rational behind this is that an instance being “down” can have
384 - it could be down during a reboot
385 - it could be temporarily be down for a reinstall
386 - or it could be down because it is deprecated and kept just for its
389 The previous Boolean state was making it difficult to do capacity
390 calculations: should Ganeti reserve memory for a down instance? Now, the
391 tri-state field makes it clear:
393 - in ``up`` and ``down`` state, all resources are reserved for the
394 instance, and it can be at any time brought up if it is down
395 - in ``offline`` state, only disk space is reserved for it, but not
398 The field can have an extra use: since the transition between ``up`` and
399 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
400 transition between ``offline`` and ``down`` is done via ``gnt-instance
401 modify``, it is possible to given different rights to users. For
402 example, owners of an instance could be allowed to start/stop it, but
403 not transition it out of the offline state.
405 Instance policies and specs
406 +++++++++++++++++++++++++++
408 In previous Ganeti versions, an instance creation request was not
409 limited on the minimum size and on the maximum size just by the cluster
410 resources. As such, any policy could be implemented only in third-party
411 clients (RAPI clients, or shell wrappers over ``gnt-*``
412 tools). Furthermore, calculating cluster capacity via ``hspace`` again
413 required external input with regards to instance sizes.
415 In order to improve these workflows and to allow for example better
416 per-node group differentiation, we introduced instance specs, which
419 - minimum instance disk size, disk count, memory size, cpu count
420 - maximum values for the above metrics
421 - and “standard” values (used in ``hspace`` to calculate the standard
424 The minimum/maximum values can be also customised at node-group level,
425 for example allowing more powerful hardware to support bigger instance
428 Beside the instance specs, there are a few other settings belonging to
429 the instance policy framework. It is possible now to customise, per
430 cluster and node-group:
432 - the list of allowed disk templates
433 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
434 - the maximum ratio of instance to spindles (see below for more
435 information) for local storage
437 All these together should allow all tools that talk to Ganeti to know
438 what are the ranges of allowed values for instances and the
439 over-subscription that is allowed.
441 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
442 instance configuration, and the physical CPU configuration from the
443 node. For the spindle ratios however, we didn't track before these
444 values, so new parameters have been added:
446 - a new node parameter ``spindle_count``, defaults to 1, customisable at
447 node group or node level
448 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
450 Note that spindles in this context doesn't need to mean actual
451 mechanical hard-drives; it's just a relative number for both the node
452 I/O capacity and instance I/O consumption.
454 Instance migration behaviour
455 ++++++++++++++++++++++++++++
457 While live-migration is in general desirable over failover, it is
458 possible that for some workloads it is actually worse, due to the
459 variable time of the “suspend” phase during live migration.
461 To allow the tools to work consistently over such instances (without
462 having to hard-code instance names), a new backend parameter
463 ``always_failover`` has been added to control the migration/failover
464 behaviour. When set to True, all migration requests for an instance will
465 instead fall-back to failover.
467 Instance memory ballooning
468 ++++++++++++++++++++++++++
470 Initial support for memory ballooning has been added. The memory for an
471 instance is no longer fixed (backend parameter ``memory``), but instead
472 can vary between minimum and maximum values (backend parameters
473 ``minmem`` and ``maxmem``). Currently we only change an instance's
476 - live migrating or failing over and instance and the target node
477 doesn't have enough memory
478 - user requests changing the memory via ``gnt-instance modify
484 In order to control the use of specific CPUs by instance, support for
485 controlling CPU pinning has been added for the Xen, HVM and LXC
486 hypervisors. This is controlled by a new hypervisor parameter
487 ``cpu_mask``; details about possible values for this are in the
488 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
489 VCPU-to-CPU mapping) form will work well only when all nodes in your
490 cluster have the same amount of CPUs.
495 Another area in which Ganeti was not customisable were the parameters
496 used for storage configuration, e.g. how many stripes to use for LVM,
497 DRBD resync configuration, etc.
499 To improve this area, we've added disks parameters, which are
500 customisable at cluster and node group level, and which allow to
501 specify various parameters for disks (DRBD has the most parameters
502 currently), for example:
504 - DRBD resync algorithm and parameters (e.g. speed)
505 - the default VG for meta-data volumes for DRBD
506 - number of stripes for LVM (plain disk template)
509 These parameters can be modified via ``gnt-cluster modify -D …`` and
510 ``gnt-group modify -D …``, and are used at either instance creation (in
511 case of LVM stripes, for example) or at disk “activation” time
514 Rados block device support
515 ++++++++++++++++++++++++++
517 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
518 denoted by the ``rbd`` disk template type. This is considered
519 experimental, feedback is welcome. For details on configuring it, see
520 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
525 The existing master IP functionality works well only in simple setups (a
526 single network shared by all nodes); however, if nodes belong to
527 different networks, then the ``/32`` setup and lack of routing
528 information is not enough.
530 To allow the master IP to function well in more complex cases, the
531 system was reworked as follows:
533 - a master IP netmask setting has been added
534 - the master IP activation/turn-down code was moved from the node daemon
536 - whether to run the Ganeti-supplied master IP script or a user-supplied
537 on is a ``gnt-cluster init`` setting
539 Details about the location of the standard and custom setup scripts are
540 in the man page :manpage:`gnt-cluster(8)`; for information about the
541 setup script protocol, look at the Ganeti-supplied script.
546 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
549 It is now possible to use TLS-protected connections, and when renewing
550 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
551 it is now possible to specify spice or spice CA certificates. Also, it
552 is possible to configure a password for SPICE sessions via the
553 hypervisor parameter ``spice_password_file``.
555 There are also new parameters to control the compression and streaming
556 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
557 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
558 for the spice parameters.
560 Lastly, it is now possible to see the SPICE connection information via
561 ``gnt-instance console``.
566 A new tool (``tools/ovfconverter``) has been added that supports
567 conversion between Ganeti and the `Open Virtualization Format
568 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
571 This relies on the ``qemu-img`` tool to convert the disk formats, so the
572 actual compatibility with other virtualization solutions depends on it.
577 The configuration query daemon (``ganeti-confd``) is now optional, and
578 has been rewritten in Haskell; whether to use the daemon at all, use the
579 Python (default) or the Haskell version is selectable at configure time
580 via the ``--enable-confd`` parameter, which can take one of the
581 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
582 daemon will result in a smaller footprint; for larger systems, we
583 welcome feedback on the Haskell version which might become the default
586 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
587 daemon running. The Python version doesn't implement the new call.
590 User interface changes
591 ~~~~~~~~~~~~~~~~~~~~~~
593 We have replaced the ``--disks`` option of ``gnt-instance
594 replace-disks`` with a more flexible ``--disk`` option, which allows
595 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
596 disk size and mode can be changed upon recreation (via ``gnt-instance
597 recreate-disks``, which accepts the same ``--disk`` option).
599 As many people are used to a ``show`` command, we have added that as an
600 alias to ``info`` on all ``gnt-*`` commands.
602 The ``gnt-instance grow-disk`` command has a new mode in which it can
603 accept the target size of the disk, instead of the delta; this can be
604 more safe since two runs in absolute mode will be idempotent, and
605 sometimes it's also easier to specify the desired size directly.
607 Also the handling of instances with regard to offline secondaries has
608 been improved. Instance operations should not fail because one of it's
609 secondary nodes is offline, even though it's safe to proceed.
611 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
612 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
613 minors to instance name.
618 RAPI coverage has improved, with (for example) new resources for
619 recreate-disks, node power-cycle, etc.
624 There is partial support for ``xl`` in the Xen hypervisor; feedback is
627 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
628 whether to still support Python 2.4 or move to Python 2.6 as minimum
631 Support for Fedora has been slightly improved; the provided example
632 init.d script should work better on it and the INSTALL file should
633 document the needed dependencies.
638 The deprecated ``QueryLocks`` LUXI request has been removed. Use
639 ``Query(what=QR_LOCK, ...)`` instead.
641 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
642 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
643 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
644 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
645 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
647 RAPI client: ``CertificateError`` now derives from
648 ``GanetiApiError``. This should make it more easy to handle Ganeti
651 Deprecation warnings due to PyCrypto/paramiko import in
652 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
653 make sure to run an up-to-date paramiko version, if you use this tool.
655 The QA scripts now depend on Python 2.5 or above (the main code base
656 still works with Python 2.4).
658 The configuration file (``config.data``) is now written without
659 indentation for performance reasons; if you want to edit it, it can be
660 re-formatted via ``tools/fmtjson``.
662 A number of bugs has been fixed in the cluster merge tool.
664 ``x509`` certification verification (used in import-export) has been
665 changed to allow the same clock skew as permitted by the cluster
666 verification. This will remove some rare but hard to diagnose errors in
673 *(Released Thu, 19 Jul 2012)*
675 Very few changes from rc4 to the final release, only bugfixes:
677 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
678 instance, fix CDROM booting for KVM instances)
679 - fixed node group modification of node parameters
680 - fixed issue in LUClusterVerifyGroup with multi-group clusters
681 - fixed generation of bash completion to ensure a stable ordering
688 *(Released Fri, 13 Jul 2012)*
690 Third release candidate for 2.6. The following changes were done from
693 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
694 - Fixed an inconsistency in the LUXI protocol with the provided
695 arguments (NOT backwards compatible)
696 - Fixed a bug with node groups ipolicy where ``min`` was greater than
697 the cluster ``std`` value
698 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
699 easier instance debugging on nodes (requires ``hconfd`` to work)
705 *(Released Tue, 03 Jul 2012)*
707 Second release candidate for 2.6. The following changes were done from
710 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
712 - Fixed a RAPI regression on missing beparams/memory
713 - Fixed redistribution of files on offline nodes
714 - Added possibility to run activate-disks even though secondaries are
715 offline. With this change it relaxes also the strictness on some other
716 commands which use activate disks internally:
717 * ``gnt-instance start|reboot|rename|backup|export``
718 - Made it possible to remove safely an instance if its secondaries are
720 - Made it possible to reinstall even though secondaries are offline
726 *(Released Mon, 25 Jun 2012)*
728 First release candidate for 2.6. The following changes were done from
731 - Fixed bugs with disk parameters and ``rbd`` templates as well as
733 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
734 behaviour. It supports now the modify operation
735 - ``hcheck`` implemented to analyze cluster health and possibility of
736 improving health by rebalance
737 - ``hbal`` has been improved in dealing with split instances
743 *(Released Mon, 11 Jun 2012)*
745 Second beta release of 2.6. The following changes were done from beta2
748 - Fixed ``daemon-util`` with non-root user models
749 - Fixed creation of plain instances with ``--no-wait-for-sync``
750 - Fix wrong iv_names when running ``cfgupgrade``
751 - Export more information in RAPI group queries
752 - Fixed bug when changing instance network interfaces
753 - Extended burnin to do NIC changes
754 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
755 - Changed default for DRBD barriers
756 - Fixed DRBD error reporting for syncer rate
757 - Verify the options on disk parameters
759 And of course various fixes to documentation and improved unittests and
766 *(Released Wed, 23 May 2012)*
768 First beta release of 2.6. The following changes were done from beta1 to
771 - integrated patch for distributions without ``start-stop-daemon``
772 - adapted example init.d script to work on Fedora
773 - fixed log handling in Haskell daemons
774 - adapted checks in the watcher for pycurl linked against libnss
775 - add partial support for ``xl`` instead of ``xm`` for Xen
776 - fixed a type issue in cluster verification
777 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
780 Plus integrated fixes from the 2.5 branch:
782 - fixed ``kvm-ifup`` to use ``/bin/bash``
783 - fixed parallel build failures
784 - KVM live migration when using a custom keymap
790 *(Released Tue, 24 Jul 2012)*
792 A small bugfix release, with no new features:
794 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
795 different default shell (e.g. Debian, Ubuntu)
796 - fixed KVM startup and live migration with a custom keymap (fixes Issue
797 243 and Debian bug #650664)
798 - fixed compatibility with KVM versions that don't support multiple boot
799 devices (fixes Issue 230 and Debian bug #624256)
801 Additionally, a few fixes were done to the build system (fixed parallel
802 build failures) and to the unittests (fixed race condition in test for
803 FileID functions, and the default enable/disable mode for QA test is now
810 *(Released Fri, 11 May 2012)*
812 A small bugfix release.
814 The main issues solved are on the topic of compatibility with newer LVM
817 - fixed parsing of ``lv_attr`` field
818 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
819 the ``--force`` flag is needed
821 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
822 to accept kernel 3.x too (was hardcoded to 2.6.*).
824 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
825 of hook results and that also made display of other errors suboptimal
826 was fixed; the code behaves now like 2.4 and earlier.
828 Another change in 2.5, the cleanup of the OS scripts environment, is too
829 aggressive: it removed even the ``PATH`` variable, which requires the OS
830 scripts to *always* need to export it. Since this is a bit too strict,
831 we now export a minimal PATH, the same that we export for hooks.
833 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
834 integrated into this release.
836 Finally, a few other miscellaneous changes were done (no new features,
837 just small improvements):
839 - Fix ``gnt-group --help`` display
840 - Fix hardcoded Xen kernel path
841 - Fix grow-disk handling of invalid units
842 - Update synopsis for ``gnt-cluster repair-disk-sizes``
843 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
849 *(Released Thu, 12 Apr 2012)*
851 Incompatible/important changes and bugfixes
852 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
854 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
855 resource's ``ip_check`` parameter changed from ``True`` to ``False``
856 to match the underlying LUXI interface.
857 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
858 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
859 not maintain backwards-compatibility as the underlying operation
860 changed in an incompatible way. The RAPI client can talk to old
861 servers, but it needs to be told so as the return value changed.
862 - When creating file-based instances via RAPI, the ``file_driver``
863 parameter no longer defaults to ``loop`` and must be specified.
864 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
866 - Support for the undocumented and deprecated RAPI instance creation
867 request format version 0 has been dropped. Use version 1, supported
868 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
869 - Pyparsing 1.4.6 or above is required, see :doc:`installation
870 documentation <install>`.
871 - The "cluster-verify" hooks are now executed per group by the
872 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
873 if you just run ``gnt-cluster verify``, which generates one opcode per
875 - The environment as passed to the OS scripts is cleared, and thus no
876 environment variables defined in the node daemon's environment will be
877 inherited by the scripts.
878 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
880 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
881 support operations involving multiple node groups.
882 - Offline nodes are ignored when failing over an instance.
883 - Support for KVM version 1.0, which changed the version reporting format
885 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
887 - ``Makefile`` is now compatible with Automake 1.11.2
888 - Includes all bugfixes made in the 2.4 series
893 - The ganeti-htools project has been merged into the ganeti-core source
894 tree and will be built as part of Ganeti (see :doc:`install-quick`).
895 - Implemented support for :doc:`shared storage <design-shared-storage>`.
896 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
897 GPT-style partition tables (requires `parted
898 <http://www.gnu.org/s/parted/>`_).
899 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
900 - Allowed adding tags on instance creation.
901 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
902 - Allow instances to be started in a paused state, enabling the user to
903 see the complete console output on boot using the console.
904 - Added new hypervisor flag to control default reboot behaviour
905 (``reboot_behavior``).
906 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
907 - Improved out-of-band management support:
909 - Added ``gnt-node health`` command reporting the health status of
911 - Added ``gnt-node power`` command to manage power status of nodes.
912 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
914 - Instance migration can fall back to failover if instance is not
916 - Filters can be used when listing nodes, instances, groups and locks;
917 see :manpage:`ganeti(7)` manpage.
918 - Added post-execution status as variables to :doc:`hooks <hooks>`
920 - Instance tags are exported/imported together with the instance.
921 - When given an explicit job ID, ``gnt-job info`` will work for archived
923 - Jobs can define dependencies on other jobs (not yet supported via
924 RAPI or command line, but used by internal commands and usable via
927 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
930 - Instance failover is now available as a RAPI resource
931 (``/2/instances/[instance_name]/failover``).
932 - ``gnt-instance info`` defaults to static information if primary node
934 - Opcodes have a new ``comment`` attribute.
935 - Added basic SPICE support to KVM hypervisor.
936 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
938 Node group improvements
939 ~~~~~~~~~~~~~~~~~~~~~~~
941 - ``gnt-cluster verify`` has been modified to check groups separately,
942 thereby improving performance.
943 - Node group support has been added to ``gnt-cluster verify-disks``,
944 which now operates per node group.
945 - Watcher has been changed to work better with node groups.
947 - One process and state file per node group.
948 - Slow watcher in one group doesn't block other group's watcher.
950 - Added new command, ``gnt-group evacuate``, to move all instances in a
951 node group to other groups.
952 - Added ``gnt-instance change-group`` to move an instance to another
954 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
955 per-group operations.
956 - Node groups can be tagged.
957 - Some operations switch from an exclusive to a shared lock as soon as
959 - Instance's primary and secondary nodes' groups are now available as
960 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
961 and ``snodes.group.uuid``).
966 - Numerous updates to documentation and manpages.
968 - :doc:`RAPI <rapi>` documentation now has detailed parameter
970 - Some opcode/job results are now also documented, see :doc:`RAPI
973 - A lockset's internal lock is now also visible in lock monitor.
974 - Log messages from job queue workers now contain information about the
975 opcode they're processing.
976 - ``gnt-instance console`` no longer requires the instance lock.
977 - A short delay when waiting for job changes reduces the number of LUXI
978 requests significantly.
979 - DRBD metadata volumes are overwritten with zeros during disk creation.
980 - Out-of-band commands no longer acquire the cluster lock in exclusive
982 - ``devel/upload`` now uses correct permissions for directories.
988 *(Released Fri, 23 Mar 2012)*
990 This was the sixth release candidate of the 2.5 series.
996 *(Released Mon, 9 Jan 2012)*
998 This was the fifth release candidate of the 2.5 series.
1004 *(Released Thu, 27 Oct 2011)*
1006 This was the fourth release candidate of the 2.5 series.
1012 *(Released Wed, 26 Oct 2011)*
1014 This was the third release candidate of the 2.5 series.
1020 *(Released Tue, 18 Oct 2011)*
1022 This was the second release candidate of the 2.5 series.
1028 *(Released Tue, 4 Oct 2011)*
1030 This was the first release candidate of the 2.5 series.
1036 *(Released Wed, 31 Aug 2011)*
1038 This was the third beta release of the 2.5 series.
1044 *(Released Mon, 22 Aug 2011)*
1046 This was the second beta release of the 2.5 series.
1052 *(Released Fri, 12 Aug 2011)*
1054 This was the first beta release of the 2.5 series.
1060 *(Released Thu, 27 Oct 2011)*
1062 - Fixed bug when parsing command line parameter values ending in
1064 - Fixed assertion error after unclean master shutdown
1065 - Disable HTTP client pool for RPC, significantly reducing memory usage
1067 - Fixed queue archive creation with wrong permissions
1073 *(Released Tue, 23 Aug 2011)*
1077 - Fixed documentation for importing with ``--src-dir`` option
1078 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1079 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1085 *(Released Fri, 5 Aug 2011)*
1087 Many bug-fixes and a few small features:
1089 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1090 issues when you tried to add an instance with two MAC addresses in one
1092 - KVM: fixed per-instance stored UID value
1093 - KVM: configure bridged NICs at migration start
1094 - KVM: Fix a bug where instance will not start with never KVM versions
1096 - Added OS search path to ``gnt-cluster info``
1097 - Fixed an issue with ``file_storage_dir`` where you were forced to
1098 provide an absolute path, but the documentation states it is a
1099 relative path, the documentation was right
1100 - Added a new parameter to instance stop/start called ``--no-remember``
1101 that will make the state change to not be remembered
1102 - Implemented ``no_remember`` at RAPI level
1103 - Improved the documentation
1104 - Node evacuation: don't call IAllocator if node is already empty
1105 - Fixed bug in DRBD8 replace disks on current nodes
1106 - Fixed bug in recreate-disks for DRBD instances
1107 - Moved assertion checking locks in ``gnt-instance replace-disks``
1108 causing it to abort with not owning the right locks for some situation
1109 - Job queue: Fixed potential race condition when cancelling queued jobs
1110 - Fixed off-by-one bug in job serial generation
1111 - ``gnt-node volumes``: Fix instance names
1112 - Fixed aliases in bash completion
1113 - Fixed a bug in reopening log files after being sent a SIGHUP
1114 - Added a flag to burnin to allow specifying VCPU count
1115 - Bugfixes to non-root Ganeti configuration
1121 *(Released Thu, 12 May 2011)*
1123 Many bug-fixes and a few new small features:
1125 - Fixed a bug related to log opening failures
1126 - Fixed a bug in instance listing with orphan instances
1127 - Fixed a bug which prevented resetting the cluster-level node parameter
1128 ``oob_program`` to the default
1129 - Many fixes related to the ``cluster-merge`` tool
1130 - Fixed a race condition in the lock monitor, which caused failures
1131 during (at least) creation of many instances in parallel
1132 - Improved output for gnt-job info
1133 - Removed the quiet flag on some ssh calls which prevented debugging
1135 - Improved the N+1 failure messages in cluster verify by actually
1136 showing the memory values (needed and available)
1137 - Increased lock attempt timeouts so that when executing long operations
1138 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1139 too early and thus prevent the use of the 'fair' mechanism
1140 - Changed instance query data (``gnt-instance info``) to not acquire
1141 locks unless needed, thus allowing its use on locked instance if only
1142 static information is asked for
1143 - Improved behaviour with filesystems that do not support rename on an
1145 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1146 kept locks on all nodes during the wipe, which is unneeded
1147 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1148 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1149 10GiB) which caused the wipe to fail right at the end in some cases
1150 - Fixed master IP activation when doing master failover with no-voting
1151 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1152 the master node itself
1153 - Fixed potential data-loss in under disk full conditions, where Ganeti
1154 wouldn't check correctly the return code and would consider
1155 partially-written files 'correct'
1156 - Fixed bug related to multiple VGs and DRBD disk replacing
1157 - Added new disk parameter ``metavg`` that allows placement of the meta
1158 device for DRBD in a different volume group
1159 - Fixed error handling in the node daemon when the system libc doesn't
1160 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1161 - Fixed lock release during replace-disks, which kept cluster-wide locks
1162 when doing disk replaces with an iallocator script
1163 - Added check for missing bridges in cluster verify
1164 - Handle EPIPE errors while writing to the terminal better, so that
1165 piping the output to e.g. ``less`` doesn't cause a backtrace
1166 - Fixed rare case where a ^C during Luxi calls could have been
1167 interpreted as server errors, instead of simply terminating
1168 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1170 - Added a few more parameters to the KVM hypervisor, allowing a second
1171 CDROM, custom disk type for CDROMs and a floppy image
1172 - Removed redundant message in instance rename when the name is given
1174 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1175 disks on new nodes, allowing recreation when the original instance
1176 nodes are completely gone
1177 - Added option when converting disk templates to DRBD to skip waiting
1178 for the resync, in order to make the instance available sooner
1179 - Added two new variables to the OS scripts environment (containing the
1181 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1182 to allow use of ``pvgrub`` as bootloader
1183 - Changed the instance memory modifications to only check out-of-memory
1184 conditions on memory increases, and turned the secondary node warnings
1185 into errors (they can still be overridden via ``--force``)
1186 - Fixed the handling of a corner case when the Python installation gets
1187 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1188 to execute a command that doesn't exist
1189 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1190 node of the instance returned failures during instance shutdown; this
1191 adds the option ``--ignore-consistency`` to gnt-instance move
1193 And as usual, various improvements to the error messages, documentation
1200 *(Released Wed, 09 Mar 2011)*
1202 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1203 the RAPI users file if run twice (even with ``--dry-run``).
1205 The release fixes that bug (nothing else changed).
1211 *(Released Mon, 07 Mar 2011)*
1213 Final 2.4.0 release. Just a few small fixes:
1215 - Fixed RAPI node evacuate
1216 - Fixed the kvm-ifup script
1217 - Fixed internal error handling for special job cases
1218 - Updated man page to specify the escaping feature for options
1224 *(Released Mon, 28 Feb 2011)*
1226 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1227 that this feature wiped the disks of the wrong instance, leading to loss
1232 - Fixed title of query field containing instance name
1233 - Expanded the glossary in the documentation
1234 - Fixed one unittest (internal issue)
1240 *(Released Mon, 21 Feb 2011)*
1242 A number of bug fixes plus just a couple functionality changes.
1244 On the user-visible side, the ``gnt-* list`` command output has changed
1245 with respect to "special" field states. The current rc1 style of display
1246 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1247 the default output mode special fields states are displayed as follows:
1249 - Offline resource: ``*``
1250 - Unavailable/not applicable: ``-``
1251 - Data missing (RPC failure): ``?``
1252 - Unknown field: ``??``
1254 Another user-visible change is the addition of ``--force-join`` to
1259 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1260 - Fixed regression in RAPI/instance reinstall where all parameters were
1261 required (instead of optional)
1262 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1263 - Fixed iallocator usage (offline nodes were not considered offline)
1264 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1265 - Fixed hypervisor and OS parameter validation with respect to
1266 non-vm_capable nodes
1267 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1269 - Fixed ``tools/listrunner`` with respect to agent-based usage
1275 *(Released Fri, 4 Feb 2011)*
1277 Many changes and fixes since the beta1 release. While there were some
1278 internal changes, the code has been mostly stabilised for the RC
1281 Note: the dumb allocator was removed in this release, as it was not kept
1282 up-to-date with the IAllocator protocol changes. It is recommended to
1283 use the ``hail`` command from the ganeti-htools package.
1285 Note: the 2.4 and up versions of Ganeti are not compatible with the
1286 0.2.x branch of ganeti-htools. You need to upgrade to
1287 ganeti-htools-0.3.0 (or later).
1289 Regressions fixed from 2.3
1290 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1292 - Fixed the ``gnt-cluster verify-disks`` command
1293 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1295 - Fixed disk adoption breakage
1296 - Fixed wrong headers in instance listing for field aliases
1301 - Fixed corner case in KVM handling of NICs
1302 - Fixed many cases of wrong handling of non-vm_capable nodes
1303 - Fixed a bug where a missing instance symlink was not possible to
1304 recreate with any ``gnt-*`` command (now ``gnt-instance
1305 activate-disks`` does it)
1306 - Fixed the volume group name as reported by ``gnt-cluster
1308 - Increased timeouts for the import-export code, hopefully leading to
1309 fewer aborts due network or instance timeouts
1310 - Fixed bug in ``gnt-node list-storage``
1311 - Fixed bug where not all daemons were started on cluster
1312 initialisation, but only at the first watcher run
1313 - Fixed many bugs in the OOB implementation
1314 - Fixed watcher behaviour in presence of instances with offline
1316 - Fixed instance list output for instances running on the wrong node
1317 - a few fixes to the cluster-merge tool, but it still cannot merge
1318 multi-node groups (currently it is not recommended to use this tool)
1324 - Improved network configuration for the KVM hypervisor
1325 - Added e1000 as a supported NIC for Xen-HVM
1326 - Improved the lvmstrap tool to also be able to use partitions, as
1327 opposed to full disks
1328 - Improved speed of disk wiping (the cluster parameter
1329 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1330 of instance creations
1331 - Added documentation for the OS parameters
1332 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1333 hypervisor is not responding
1334 - Added display of blacklisted and hidden OS information in
1335 ``gnt-cluster info``
1336 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1337 NIC and node parameters, which might create problems with currently
1338 invalid (but undetected) configuration files, but prevents validation
1339 failures when unrelated parameters are modified
1340 - Changed cluster initialisation to wait for the master daemon to become
1342 - Expanded the RAPI interface:
1344 - Added config redistribution resource
1345 - Added activation/deactivation of instance disks
1346 - Added export of console information
1348 - Implemented log file reopening on SIGHUP, which allows using
1349 logrotate(8) for the Ganeti log files
1350 - Added a basic OOB helper script as an example
1356 *(Released Fri, 14 Jan 2011)*
1361 - Fixed timezone issues when formatting timestamps
1362 - Added support for node groups, available via ``gnt-group`` and other
1364 - Added out-of-band framework and management, see :doc:`design
1365 document <design-oob>`
1366 - Removed support for roman numbers from ``gnt-node list`` and
1367 ``gnt-instance list``.
1368 - Allowed modification of master network interface via ``gnt-cluster
1369 modify --master-netdev``
1370 - Accept offline secondaries while shutting down instance disks
1371 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1372 - Added support for multiple LVM volume groups
1373 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1375 - Added commands to list available fields:
1377 - ``gnt-node list-fields``
1378 - ``gnt-group list-fields``
1379 - ``gnt-instance list-fields``
1381 - Updated documentation and man pages
1386 - Moved ``rapi_users`` file into separate directory, now named
1387 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1389 - Added new tool for running commands on many machines,
1390 ``tools/ganeti-listrunner``
1391 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1392 improving the ``gnt-instance console`` output
1393 - Allowed customisation of disk index separator at ``configure`` time
1394 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1395 - Added support for non-partitioned md disks in ``lvmstrap``
1396 - Added script to gracefully power off KVM instances
1397 - Split ``utils`` module into smaller parts
1398 - Changed query operations to return more detailed information, e.g.
1399 whether an information is unavailable due to an offline node. To use
1400 this new functionality, the LUXI call ``Query`` must be used. Field
1401 information is now stored by the master daemon and can be retrieved
1402 using ``QueryFields``. Instances, nodes and groups can also be queried
1403 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1404 exposed via RAPI). The following commands make use of this
1405 infrastructure change:
1407 - ``gnt-group list``
1408 - ``gnt-group list-fields``
1410 - ``gnt-node list-fields``
1411 - ``gnt-instance list``
1412 - ``gnt-instance list-fields``
1413 - ``gnt-debug locks``
1418 - New RAPI resources (see :doc:`rapi`):
1422 - ``/2/groups/[group_name]``
1423 - ``/2/groups/[group_name]/assign-nodes``
1424 - ``/2/groups/[group_name]/modify``
1425 - ``/2/groups/[group_name]/rename``
1426 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1430 - Implemented ``no_install`` for instance creation
1431 - Implemented OS parameters for instance reinstallation, allowing
1432 use of special settings on reinstallation (e.g. for preserving data)
1437 - Added IPv6 support in import/export
1438 - Pause DRBD synchronization while wiping disks on instance creation
1439 - Updated unittests and QA scripts
1440 - Improved network parameters passed to KVM
1441 - Converted man pages from docbook to reStructuredText
1447 *(Released Mon, 20 Dec 2010)*
1449 Released version 2.3.1~rc1 without any changes.
1455 *(Released Wed, 1 Dec 2010)*
1457 - impexpd: Disable OpenSSL compression in socat if possible (backport
1458 from master, commit e90739d625b, see :doc:`installation guide
1459 <install-quick>` for details)
1460 - Changed unittest coverage report to exclude test scripts
1461 - Added script to check version format
1467 *(Released Wed, 1 Dec 2010)*
1469 Released version 2.3.0~rc1 without any changes.
1475 *(Released Fri, 19 Nov 2010)*
1477 A number of bugfixes and documentation updates:
1479 - Update ganeti-os-interface documentation
1480 - Fixed a bug related to duplicate MACs or similar items which should be
1482 - Fix breakage in OS state modify
1483 - Reinstall instance: disallow offline secondaries (fixes bug related to
1484 OS changing but reinstall failing)
1485 - plus all the other fixes between 2.2.1 and 2.2.2
1491 *(Released Tue, 2 Nov 2010)*
1493 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1494 - Fixed master failover race with watcher
1495 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1497 - Accept previously stopped instance for export with instance removal
1498 - Simplify and extend the environment variables for instance OS scripts
1499 - Added new node flags, ``master_capable`` and ``vm_capable``
1500 - Added optional instance disk wiping prior during allocation. This is a
1501 cluster-wide option and can be set/modified using
1502 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1503 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1504 :doc:`install-quick`
1505 - Added a new watcher option (``--ignore-pause``)
1506 - Added option to ignore offline node on instance start/stop
1507 (``--ignore-offline``)
1508 - Allow overriding OS parameters with ``gnt-instance reinstall``
1509 - Added ability to change node's secondary IP address using ``gnt-node
1511 - Implemented privilege separation for all daemons except
1512 ``ganeti-noded``, see ``configure`` options
1513 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1515 - Implemented job priorities (see ``ganeti(7)`` manpage)
1516 - Ignore failures while shutting down instances during failover from
1518 - Exit daemon's bootstrap process only once daemon is ready
1519 - Export more information via ``LUInstanceQuery``/remote API
1520 - Improved documentation, QA and unittests
1521 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1522 restart if the file was created or changed
1523 - Added LUXI protocol version sent with each request and response,
1524 allowing detection of server/client mismatches
1525 - Moved the Python scripts among gnt-* and ganeti-* into modules
1526 - Moved all code related to setting up SSH to an external script,
1528 - Infrastructure changes for node group support in future versions
1534 *(Released Fri, 19 Nov 2010)*
1536 A few small bugs fixed, and some improvements to the build system:
1538 - Fix documentation regarding conversion to drbd
1539 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1541 - Fix error handling in node modify with multiple changes
1542 - Allow remote imports without checked names
1548 *(Released Tue, 19 Oct 2010)*
1550 - Disable SSL session ID cache in RPC client
1556 *(Released Thu, 14 Oct 2010)*
1558 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1559 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1561 - Fix problems with interaction between Curl and Python's HTTP server,
1562 resulting in increased speed in many RPC calls
1563 - Improve our release script to prevent breakage with older aclocal and
1570 *(Released Thu, 7 Oct 2010)*
1572 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1573 value retrieved from master
1574 - Added support for blacklisted or hidden OS definitions
1575 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1576 - Added support for -mem-path in KVM hypervisor abstraction layer
1577 - Allow overriding instance parameters in tool for inter-cluster
1578 instance moves (``tools/move-instance``)
1579 - Improved opcode summaries (e.g. in ``gnt-job list``)
1580 - Improve consistency of OS listing by sorting it
1581 - Documentation updates
1587 *(Released Fri, 8 Oct 2010)*
1589 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1595 *(Released Mon, 4 Oct 2010)*
1597 - Fixed regression in ``gnt-instance rename``
1603 *(Released Wed, 22 Sep 2010)*
1605 - Fixed OS_VARIANT variable for OS scripts
1606 - Fixed cluster tag operations via RAPI
1607 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1608 - Disabled RAPI CA checks in watcher
1614 *(Released Mon, 23 Aug 2010)*
1616 - Support DRBD versions of the format "a.b.c.d"
1618 - Re-introduce support for usage from multiple threads in RAPI client
1619 - Instance renames and modify via RAPI
1620 - Work around race condition between processing and archival in job
1622 - Mark opcodes following failed one as failed, too
1623 - Job field ``lock_status`` was removed due to difficulties making it
1624 work with the changed job queue in Ganeti 2.2; a better way to monitor
1625 locks is expected for a later 2.2.x release
1626 - Fixed dry-run behaviour with many commands
1627 - Support ``ssh-agent`` again when adding nodes
1628 - Many additional bugfixes
1634 *(Released Fri, 30 Jul 2010)*
1636 Important change: the internal RPC mechanism between Ganeti nodes has
1637 changed from using a home-grown http library (based on the Python base
1638 libraries) to use the PycURL library. This requires that PycURL is
1639 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1640 against GnuTLS by default. cURL's support for GnuTLS had known issues
1641 before cURL 7.21.0 and we recommend using the latest cURL release or
1642 linking against OpenSSL. Most other distributions already link PycURL
1643 and cURL against OpenSSL. The command::
1645 python -c 'import pycurl; print pycurl.version'
1647 can be used to determine the libraries PycURL and cURL are linked
1650 Other significant changes:
1652 - Rewrote much of the internals of the job queue, in order to achieve
1653 better parallelism; this decouples job query operations from the job
1654 processing, and it should allow much nicer behaviour of the master
1655 daemon under load, and it also has uncovered some long-standing bugs
1656 related to the job serialisation (now fixed)
1657 - Added a default iallocator setting to the cluster parameters,
1658 eliminating the need to always pass nodes or an iallocator for
1659 operations that require selection of new node(s)
1660 - Added experimental support for the LXC virtualization method
1661 - Added support for OS parameters, which allows the installation of
1662 instances to pass parameter to OS scripts in order to customise the
1664 - Added a hypervisor parameter controlling the migration type (live or
1665 non-live), since hypervisors have various levels of reliability; this
1666 has renamed the 'live' parameter to 'mode'
1667 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1668 logical volumes, meaning that cluster verify will ignore them and not
1669 flag their presence as errors
1670 - The watcher will now reset the error count for failed instances after
1671 8 hours, thus allowing self-healing if the problem that caused the
1672 instances to be down/fail to start has cleared in the meantime
1673 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1674 check for, and warn, if the drbd module parameter ``usermode_helper``
1675 is not consistent with the cluster-wide setting; this is needed to
1676 make diagnose easier of failed drbd creations
1677 - Started adding base IPv6 support, but this is not yet
1678 enabled/available for use
1679 - Rename operations (cluster, instance) will now return the new name,
1680 which is especially useful if a short name was passed in
1681 - Added support for instance migration in RAPI
1682 - Added a tool to pre-configure nodes for the SSH setup, before joining
1683 them to the cluster; this will allow in the future a simplified model
1684 for node joining (but not yet fully enabled in 2.2); this needs the
1685 paramiko python library
1686 - Fixed handling of name-resolving errors
1687 - Fixed consistency of job results on the error path
1688 - Fixed master-failover race condition when executed multiple times in
1690 - Fixed many bugs related to the job queue (mostly introduced during the
1691 2.2 development cycle, so not all are impacting 2.1)
1692 - Fixed instance migration with missing disk symlinks
1693 - Fixed handling of unknown jobs in ``gnt-job archive``
1694 - And many other small fixes/improvements
1698 - Enhanced both the unittest and the QA coverage
1699 - Switched the opcode validation to a generic model, and extended the
1700 validation to all opcode parameters
1701 - Changed more parts of the code that write shell scripts to use the
1703 - Switched the master daemon to use the asyncore library for the Luxi
1710 *(Released Thu, 17 Jun 2010)*
1712 - Added tool (``move-instance``) and infrastructure to move instances
1713 between separate clusters (see :doc:`separate documentation
1714 <move-instance>` and :doc:`design document <design-2.2>`)
1715 - Added per-request RPC timeout
1716 - RAPI now requires a Content-Type header for requests with a body (e.g.
1717 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1718 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1719 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1721 - Implemented initial support for running Ganeti daemons as separate
1722 users, see configure-time flags ``--with-user-prefix`` and
1723 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1725 - Instances can be removed after export (``gnt-backup export
1726 --remove-instance``)
1727 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1728 key (instead of 1024 bit)
1729 - Added new cluster configuration file for cluster domain secret
1730 - Import/export now use SSL instead of SSH
1731 - Added support for showing estimated time when exporting an instance,
1732 see the ``ganeti-os-interface(7)`` manpage and look for
1739 *(Released Tue, 16 Nov 2010)*
1741 Some more bugfixes. Unless critical bugs occur, this will be the last
1744 - Fix case of MAC special-values
1745 - Fix mac checker regex
1746 - backend: Fix typo causing "out of range" error
1747 - Add missing --units in gnt-instance list man page
1753 *(Released Tue, 24 Aug 2010)*
1756 - Don't ignore secondary node silently on non-mirrored disk templates
1758 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1759 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1760 - Properly document the usb_mouse parameter
1761 - Fix path in ganeti-rapi(8) (issue 116)
1762 - Adjust error message when the ganeti user's .ssh directory is
1764 - Add same-node-check when changing the disk template to drbd
1770 *(Released Fri, 16 Jul 2010)*
1773 - Add an option to only select some reboot types during qa/burnin.
1774 (on some hypervisors consequent reboots are not supported)
1775 - Fix infrequent race condition in master failover. Sometimes the old
1776 master ip address would be still detected as up for a short time
1777 after it was removed, causing failover to fail.
1778 - Decrease mlockall warnings when the ctypes module is missing. On
1779 Python 2.4 we support running even if no ctypes module is installed,
1780 but we were too verbose about this issue.
1781 - Fix building on old distributions, on which man doesn't have a
1783 - Fix RAPI not to ignore the MAC address on instance creation
1784 - Implement the old instance creation format in the RAPI client.
1790 *(Released Thu, 01 Jul 2010)*
1792 A small bugfix release:
1793 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1794 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1795 a non-existing option
1796 - Fix instance create: the --force-variant option was ignored
1797 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1798 - Fix modify node storage with non-FQDN arguments
1799 - Fix RAPI client to authenticate under Python 2.6 when used
1800 for more than 5 requests needing authentication
1801 - Fix gnt-instance modify -t (storage) giving a wrong error message
1802 when converting a non-shutdown drbd instance to plain
1808 *(Released Fri, 18 Jun 2010)*
1810 A small bugfix release:
1812 - Fix live migration of KVM instances started with older Ganeti
1813 versions which had fewer hypervisor parameters
1814 - Fix gnt-instance grow-disk on down instances
1815 - Fix an error-reporting bug during instance migration
1816 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1817 silently ignoring broken ones
1818 - Fix an RPC error reporting bug affecting, for example, RAPI client
1820 - Fix bug triggered by different API version os-es on different nodes
1821 - Fix a bug in instance startup with custom hvparams: OS level
1822 parameters would fail to be applied.
1823 - Fix the RAPI client under Python 2.6 (but more work is needed to
1824 make it work completely well with OpenSSL)
1825 - Fix handling of errors when resolving names from DNS
1831 *(Released Thu, 3 Jun 2010)*
1833 A medium sized development cycle. Some new features, and some
1834 fixes/small improvements/cleanups.
1836 Significant features
1837 ~~~~~~~~~~~~~~~~~~~~
1839 The node deamon now tries to mlock itself into memory, unless the
1840 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1841 its logs, and falls back to console logging. This allows emergency
1842 features such as ``gnt-node powercycle`` to work even in the event of a
1843 broken node disk (tested offlining the disk hosting the node's
1844 filesystem and dropping its memory caches; don't try this at home)
1846 KVM: add vhost-net acceleration support. It can be tested with a new
1847 enough version of the kernel and of qemu-kvm.
1849 KVM: Add instance chrooting feature. If you use privilege dropping for
1850 your VMs you can also now force them to chroot to an empty directory,
1851 before starting the emulated guest.
1853 KVM: Add maximum migration bandwith and maximum downtime tweaking
1854 support (requires a new-enough version of qemu-kvm).
1856 Cluster verify will now warn if the master node doesn't have the master
1857 ip configured on it.
1859 Add a new (incompatible) instance creation request format to RAPI which
1860 supports all parameters (previously only a subset was supported, and it
1861 wasn't possible to extend the old format to accomodate all the new
1862 features. The old format is still supported, and a client can check for
1863 this feature, before using it, by checking for its presence in the
1864 ``features`` RAPI resource.
1866 Now with ancient latin support. Try it passing the ``--roman`` option to
1867 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1868 (requires the python-roman module to be installed, in order to work).
1873 As usual many internal code refactorings, documentation updates, and
1876 - Lots of improvements and cleanups to the experimental Remote API
1877 (RAPI) client library.
1878 - A new unit test suite for the core daemon libraries.
1879 - A fix to creating missing directories makes sure the umask is not
1880 applied anymore. This enforces the same directory permissions
1882 - Better handling terminating daemons with ctrl+c (used when running
1883 them in debugging mode).
1884 - Fix a race condition in live migrating a KVM instance, when stat()
1885 on the old proc status file returned EINVAL, which is an unexpected
1887 - Fixed manpage checking with newer man and utf-8 charachters. But now
1888 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1894 *(Released Fri, 7 May 2010)*
1896 Fix a bug which prevented untagged KVM instances from starting.
1902 *(Released Fri, 7 May 2010)*
1904 Another release with a long development cycle, during which many
1905 different features were added.
1907 Significant features
1908 ~~~~~~~~~~~~~~~~~~~~
1910 The KVM hypervisor now can run the individual instances as non-root, to
1911 reduce the impact of a VM being hijacked due to bugs in the
1912 hypervisor. It is possible to run all instances as a single (non-root)
1913 user, to manually specify a user for each instance, or to dynamically
1914 allocate a user out of a cluster-wide pool to each instance, with the
1915 guarantee that no two instances will run under the same user ID on any
1918 An experimental RAPI client library, that can be used standalone
1919 (without the other Ganeti libraries), is provided in the source tree as
1920 ``lib/rapi/client.py``. Note this client might change its interface in
1921 the future, as we iterate on its capabilities.
1923 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1924 replace the cluster's certificates and crypto keys. This might help in
1925 case they have been compromised, or have simply expired.
1927 A new disk option for instance creation has been added that allows one
1928 to "adopt" currently existing logical volumes, with data
1929 preservation. This should allow easier migration to Ganeti from
1930 unmanaged (or managed via other software) instances.
1932 Another disk improvement is the possibility to convert between redundant
1933 (DRBD) and plain (LVM) disk configuration for an instance. This should
1934 allow better scalability (starting with one node and growing the
1935 cluster, or shrinking a two-node cluster to one node).
1937 A new feature that could help with automated node failovers has been
1938 implemented: if a node sees itself as offline (by querying the master
1939 candidates), it will try to shutdown (hard) all instances and any active
1940 DRBD devices. This reduces the risk of duplicate instances if an
1941 external script automatically failovers the instances on such nodes. To
1942 enable this, the cluster parameter ``maintain_node_health`` should be
1943 enabled; in the future this option (per the name) will enable other
1944 automatic maintenance features.
1946 Instance export/import now will reuse the original instance
1947 specifications for all parameters; that means exporting an instance,
1948 deleting it and the importing it back should give an almost identical
1949 instance. Note that the default import behaviour has changed from
1950 before, where it created only one NIC; now it recreates the original
1953 Cluster verify has added a few new checks: SSL certificates validity,
1954 /etc/hosts consistency across the cluster, etc.
1959 As usual, many internal changes were done, documentation fixes,
1962 - Fixed cluster initialization with disabled cluster storage (regression
1963 introduced in 2.1.1)
1964 - File-based storage supports growing the disks
1965 - Fixed behaviour of node role changes
1966 - Fixed cluster verify for some corner cases, plus a general rewrite of
1967 cluster verify to allow future extension with more checks
1968 - Fixed log spamming by watcher and node daemon (regression introduced
1970 - Fixed possible validation issues when changing the list of enabled
1972 - Fixed cleanup of /etc/hosts during node removal
1973 - Fixed RAPI response for invalid methods
1974 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1975 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1976 from ide disks, etc.)
1977 - Allow OS changes without re-installation (to record a changed OS
1978 outside of Ganeti, or to allow OS renames)
1979 - Allow instance creation without OS installation (useful for example if
1980 the OS will be installed manually, or restored from a backup not in
1982 - Implemented option to make cluster ``copyfile`` use the replication
1984 - Added list of enabled hypervisors to ssconf (possibly useful for
1986 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1988 - A partial form of node re-IP is possible via node readd, which now
1989 allows changed node primary IP
1990 - Command line utilities now show an informational message if the job is
1992 - The logs of the master daemon now show the PID/UID/GID of the
1999 *(Released Fri, 12 Mar 2010)*
2001 During the 2.1.0 long release candidate cycle, a lot of improvements and
2002 changes have accumulated with were released later as 2.1.1.
2007 The node evacuate command (``gnt-node evacuate``) was significantly
2008 rewritten, and as such the IAllocator protocol was changed - a new
2009 request type has been added. This unfortunate change during a stable
2010 series is designed to improve performance of node evacuations; on
2011 clusters with more than about five nodes and which are well-balanced,
2012 evacuation should proceed in parallel for all instances of the node
2013 being evacuated. As such, any existing IAllocator scripts need to be
2014 updated, otherwise the above command will fail due to the unknown
2015 request. The provided "dumb" allocator has not been updated; but the
2016 ganeti-htools package supports the new protocol since version 0.2.4.
2018 Another important change is increased validation of node and instance
2019 names. This might create problems in special cases, if invalid host
2020 names are being used.
2022 Also, a new layer of hypervisor parameters has been added, that sits at
2023 OS level between the cluster defaults and the instance ones. This allows
2024 customisation of virtualization parameters depending on the installed
2025 OS. For example instances with OS 'X' may have a different KVM kernel
2026 (or any other parameter) than the cluster defaults. This is intended to
2027 help managing a multiple OSes on the same cluster, without manual
2028 modification of each instance's parameters.
2030 A tool for merging clusters, ``cluster-merge``, has been added in the
2031 tools sub-directory.
2036 - Improved the int/float conversions that should make the code more
2037 robust in face of errors from the node daemons
2038 - Fixed the remove node code in case of internal configuration errors
2039 - Fixed the node daemon behaviour in face of inconsistent queue
2040 directory (e.g. read-only file-system where we can't open the files
2042 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2043 now it either aborts cleanly or, if given the new "auto_promote"
2044 parameter, will automatically promote other nodes as needed
2045 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2046 completely prevent Ganeti from working
2047 - Fixed bug for instance export when not all disks were successfully
2049 - Fixed behaviour of node add when the new node is slow in starting up
2051 - Fixed handling of signals in the LUXI client, which should improve
2052 behaviour of command-line scripts
2053 - Added checks for invalid node/instance names in the configuration (now
2054 flagged during cluster verify)
2055 - Fixed watcher behaviour for disk activation errors
2056 - Fixed two potentially endless loops in http library, which led to the
2057 RAPI daemon hanging and consuming 100% CPU in some cases
2058 - Fixed bug in RAPI daemon related to hashed passwords
2059 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2061 - Enhanced compatibility with non-Debian OSes, but not using absolute
2062 path in some commands and allowing customisation of the ssh
2063 configuration directory
2064 - Fixed possible future issue with new Python versions by abiding to the
2065 proper use of ``__slots__`` attribute on classes
2066 - Added checks that should prevent directory traversal attacks
2067 - Many documentation fixes based on feedback from users
2072 - Added an "early_release" more for instance replace disks and node
2073 evacuate, where we release locks earlier and thus allow higher
2074 parallelism within the cluster
2075 - Added watcher hooks, intended to allow the watcher to restart other
2076 daemons (e.g. from the ganeti-nbma project), but they can be used of
2077 course for any other purpose
2078 - Added a compile-time disable for DRBD barriers, to increase
2079 performance if the administrator trusts the power supply or the
2080 storage system to not lose writes
2081 - Added the option of using syslog for logging instead of, or in
2082 addition to, Ganeti's own log files
2083 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2084 can indeed boot from a paravirtual NIC
2085 - Added a generic debug level for many operations; while this is not
2086 used widely yet, it allows one to pass the debug value all the way to
2088 - Enhanced the hooks environment for instance moves (failovers,
2089 migrations) where the primary/secondary nodes changed during the
2090 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2091 - Enhanced data validations for many user-supplied values; one important
2092 item is the restrictions imposed on instance and node names, which
2093 might reject some (invalid) host names
2094 - Add a configure-time option to disable file-based storage, if it's not
2095 needed; this allows greater security separation between the master
2096 node and the other nodes from the point of view of the inter-node RPC
2098 - Added user notification in interactive tools if job is waiting in the
2099 job queue or trying to acquire locks
2100 - Added log messages when a job is waiting for locks
2101 - Added filtering by node tags in instance operations which admit
2102 multiple instances (start, stop, reboot, reinstall)
2103 - Added a new tool for cluster mergers, ``cluster-merge``
2104 - Parameters from command line which are of the form ``a=b,c=d`` can now
2105 use backslash escapes to pass in values which contain commas,
2106 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2108 - For KVM, the instance name is the first parameter passed to KVM, so
2109 that it's more visible in the process list
2115 *(Released Tue, 2 Mar 2010)*
2117 Ganeti 2.1 brings many improvements with it. Major changes:
2119 - Added infrastructure to ease automated disk repairs
2120 - Added new daemon to export configuration data in a cheaper way than
2121 using the remote API
2122 - Instance NICs can now be routed instead of being associated with a
2124 - Improved job locking logic to reduce impact of jobs acquiring multiple
2125 locks waiting for other long-running jobs
2127 In-depth implementation details can be found in the Ganeti 2.1 design
2133 - Added chroot hypervisor
2134 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2136 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2137 ``bootloader_path`` and ``bootloader_args``)
2138 - Added the ``use_localtime`` option for the xen-hvm and kvm
2139 hypervisors, and the default value for this has changed to false (in
2140 2.0 xen-hvm always enabled it)
2141 - Added luxi call to submit multiple jobs in one go
2142 - Added cluster initialization option to not modify ``/etc/hosts``
2144 - Added network interface parameters
2145 - Added dry run mode to some LUs
2146 - Added RAPI resources:
2148 - ``/2/instances/[instance_name]/info``
2149 - ``/2/instances/[instance_name]/replace-disks``
2150 - ``/2/nodes/[node_name]/evacuate``
2151 - ``/2/nodes/[node_name]/migrate``
2152 - ``/2/nodes/[node_name]/role``
2153 - ``/2/nodes/[node_name]/storage``
2154 - ``/2/nodes/[node_name]/storage/modify``
2155 - ``/2/nodes/[node_name]/storage/repair``
2157 - Added OpCodes to evacuate or migrate all instances on a node
2158 - Added new command to list storage elements on nodes (``gnt-node
2159 list-storage``) and modify them (``gnt-node modify-storage``)
2160 - Added new ssconf files with master candidate IP address
2161 (``ssconf_master_candidates_ips``), node primary IP address
2162 (``ssconf_node_primary_ips``) and node secondary IP address
2163 (``ssconf_node_secondary_ips``)
2164 - Added ``ganeti-confd`` and a client library to query the Ganeti
2165 configuration via UDP
2166 - Added ability to run hooks after cluster initialization and before
2168 - Added automatic mode for disk replace (``gnt-instance replace-disks
2170 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2171 after catastrophic data-loss
2172 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2174 - Added ``gnt-instance move`` command to move instances
2175 - Added ``gnt-cluster watcher`` command to control watcher
2176 - Added ``gnt-node powercycle`` command to powercycle nodes
2177 - Added new job status field ``lock_status``
2178 - Added parseable error codes to cluster verification (``gnt-cluster
2179 verify --error-codes``) and made output less verbose (use
2180 ``--verbose`` to restore previous behaviour)
2181 - Added UUIDs to the main config entities (cluster, nodes, instances)
2182 - Added support for OS variants
2183 - Added support for hashed passwords in the Ganeti remote API users file
2185 - Added option to specify maximum timeout on instance shutdown
2186 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2187 - Added new helper script to start and stop Ganeti daemons
2188 (``daemon-util``), with the intent to reduce the work necessary to
2189 adjust Ganeti for non-Debian distributions and to start/stop daemons
2191 - Added more unittests
2192 - Fixed critical bug in ganeti-masterd startup
2193 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2194 now customisable at the cluster level for both the KVM and Xen
2195 hypervisors using the new ``migration_port`` parameter
2196 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2197 reinstalling an instance
2198 - Allowed ``@`` in tag names
2199 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2200 - Many documentation updates
2201 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2202 - ``gnt-instance reinstall`` can now reinstall multiple instances
2203 - Updated many command line parameters
2204 - Introduced new OS API version 15
2205 - No longer support a default hypervisor
2206 - Treat virtual LVs as inexistent
2207 - Improved job locking logic to reduce lock contention
2208 - Match instance and node names case insensitively
2209 - Reimplemented bash completion script to be more complete
2216 *(Released Thu, 4 Feb 2010)*
2218 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2219 - Fix a string formatting bug
2220 - Improve safety of the code in some error paths
2221 - Improve data validation in the master of values returned from nodes
2227 *(Released Thu, 17 Dec 2009)*
2229 - Fix security issue due to missing validation of iallocator names; this
2230 allows local and remote execution of arbitrary executables
2231 - Fix failure of gnt-node list during instance removal
2232 - Ship the RAPI documentation in the archive
2238 *(Released Wed, 30 Sep 2009)*
2240 - Fixed many wrong messages
2241 - Fixed a few bugs related to the locking library
2242 - Fixed MAC checking at instance creation time
2243 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2244 - Fixed a few issues related to signal handling in both daemons and
2246 - Fixed the example startup script provided
2247 - Fixed insserv dependencies in the example startup script (patch from
2249 - Fixed handling of drained nodes in the iallocator framework
2250 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2252 - Fixed error related to invalid job IDs in job polling
2253 - Fixed job/opcode persistence on unclean master shutdown
2254 - Fixed handling of partial job processing after unclean master
2256 - Fixed error reporting from LUs, previously all errors were converted
2257 into execution errors
2258 - Fixed error reporting from burnin
2259 - Decreased significantly the memory usage of the job queue
2260 - Optimised slightly multi-job submission
2261 - Optimised slightly opcode loading
2262 - Backported the multi-job submit framework from the development
2263 branch; multi-instance start and stop should be faster
2264 - Added script to clean archived jobs after 21 days; this will reduce
2265 the size of the queue directory
2266 - Added some extra checks in disk size tracking
2267 - Added an example ethers hook script
2268 - Added a cluster parameter that prevents Ganeti from modifying of
2270 - Added more node information to RAPI responses
2271 - Added a ``gnt-job watch`` command that allows following the ouput of a
2273 - Added a bind-address option to ganeti-rapi
2274 - Added more checks to the configuration verify
2275 - Enhanced the burnin script such that some operations can be retried
2277 - Converted instance reinstall to multi-instance model
2283 *(Released Fri, 7 Aug 2009)*
2285 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2286 to allow using the pre-2.0.2 behaviour in activation, if any existing
2287 instances have mismatched disk sizes in the configuration
2288 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2289 any configuration mismatches for disk sizes
2290 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2291 failover to work on two-node clusters
2292 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2294 - Fixed detection of OS script errors in ``gnt-backup export``
2295 - Fixed exit code of ``gnt-backup export``
2301 *(Released Fri, 17 Jul 2009)*
2303 - Added experimental support for stripped logical volumes; this should
2304 enhance performance but comes with a higher complexity in the block
2305 device handling; stripping is only enabled when passing
2306 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2307 affected even in the non-stripped mode
2308 - Improved resiliency against transient failures at the end of DRBD
2309 resyncs, and in general of DRBD resync checks
2310 - Fixed a couple of issues with exports and snapshot errors
2311 - Fixed a couple of issues in instance listing
2312 - Added display of the disk size in ``gnt-instance info``
2313 - Fixed checking for valid OSes in instance creation
2314 - Fixed handling of the "vcpus" parameter in instance listing and in
2315 general of invalid parameters
2316 - Fixed http server library, and thus RAPI, to handle invalid
2317 username/password combinations correctly; this means that now they
2318 report unauthorized for queries too, not only for modifications,
2319 allowing earlier detect of configuration problems
2320 - Added a new "role" node list field, equivalent to the master/master
2321 candidate/drained/offline flags combinations
2322 - Fixed cluster modify and changes of candidate pool size
2323 - Fixed cluster verify error messages for wrong files on regular nodes
2324 - Fixed a couple of issues with node demotion from master candidate role
2325 - Fixed node readd issues
2326 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2327 - Added a new ``--no-voting`` option for masterfailover to fix failover
2328 on two-nodes clusters when the former master node is unreachable
2329 - Added instance reinstall over RAPI
2335 *(Released Tue, 16 Jun 2009)*
2337 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2338 allow re-adding the start in single-user option (regression from 1.2)
2339 - the watcher writes the instance status to a file, to allow monitoring
2340 to report the instance status (from the master) based on cached
2341 results of the watcher's queries; while this can get stale if the
2342 watcher is being locked due to other work on the cluster, this is
2343 still an improvement
2344 - the watcher now also restarts the node daemon and the rapi daemon if
2346 - fixed the watcher to handle full and drained queue cases
2347 - hooks export more instance data in the environment, which helps if
2348 hook scripts need to take action based on the instance's properties
2349 (no longer need to query back into ganeti)
2350 - instance failovers when the instance is stopped do not check for free
2351 RAM, so that failing over a stopped instance is possible in low memory
2353 - rapi uses queries for tags instead of jobs (for less job traffic), and
2354 for cluster tags it won't talk to masterd at all but read them from
2356 - a couple of error handling fixes in RAPI
2357 - drbd handling: improved the error handling of inconsistent disks after
2358 resync to reduce the frequency of "there are some degraded disks for
2359 this instance" messages
2360 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2361 error handling path called a wrong function name)
2367 *(Released Wed, 27 May 2009)*
2369 - no changes from rc5
2375 *(Released Wed, 20 May 2009)*
2377 - fix a couple of bugs (validation, argument checks)
2378 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2379 - some small improvements to RAPI and IAllocator
2380 - make watcher automatically start the master daemon if down
2386 *(Released Mon, 27 Apr 2009)*
2388 - change the OS list to not require locks; this helps with big clusters
2389 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2390 volume group is broken
2391 - ``gnt-instance info``, without any arguments, doesn't run for all
2392 instances anymore; either pass ``--all`` or pass the desired
2393 instances; this helps against mistakes on big clusters where listing
2394 the information for all instances takes a long time
2395 - miscellaneous doc and man pages fixes
2401 *(Released Wed, 8 Apr 2009)*
2403 - Change the internal locking model of some ``gnt-node`` commands, in
2404 order to reduce contention (and blocking of master daemon) when
2405 batching many creation/reinstall jobs
2406 - Fixes to Xen soft reboot
2407 - No longer build documentation at build time, instead distribute it in
2408 the archive, in order to reduce the need for the whole docbook/rst
2415 *(Released Fri, 27 Mar 2009)*
2417 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2418 - Fix watcher startup sequence, improves the behaviour of busy clusters
2419 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2420 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2421 bind address and other places
2422 - Some documentation fixes and updates
2428 *(Released Mon, 2 Mar 2009)*
2430 - More documentation updates, now all docs should be more-or-less
2432 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2434 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2436 - Fix an issue related to $libdir/run/ganeti and cluster creation
2442 *(Released Thu, 19 Feb 2009)*
2444 - Xen PVM and KVM have switched the default value for the instance root
2445 disk to the first partition on the first drive, instead of the whole
2446 drive; this means that the OS installation scripts must be changed
2448 - Man pages have been updated
2449 - RAPI has been switched by default to HTTPS, and the exported functions
2450 should all work correctly
2451 - RAPI v1 has been removed
2452 - Many improvements to the KVM hypervisor
2453 - Block device errors are now better reported
2454 - Many other bugfixes and small improvements
2460 *(Released Mon, 26 Jan 2009)*
2462 - Version 2 is a general rewrite of the code and therefore the
2463 differences are too many to list, see the design document for 2.0 in
2464 the ``doc/`` subdirectory for more details
2465 - In this beta version there is not yet a migration path from 1.2 (there
2466 will be one in the final 2.0 release)
2467 - A few significant changes are:
2469 - all commands are executed by a daemon (``ganeti-masterd``) and the
2470 various ``gnt-*`` commands are just front-ends to it
2471 - all the commands are entered into, and executed from a job queue,
2472 see the ``gnt-job(8)`` manpage
2473 - the RAPI daemon supports read-write operations, secured by basic
2474 HTTP authentication on top of HTTPS
2475 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2476 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2477 to migrate to DRBD 8 first while still running Ganeti 1.2)
2478 - DRBD devices are using statically allocated minor numbers, which
2479 will be assigned to existing instances during the migration process
2480 - there is support for both Xen PVM and Xen HVM instances running on
2482 - KVM virtualization is supported too
2483 - file-based storage has been implemented, which means that it is
2484 possible to run the cluster without LVM and DRBD storage, for
2485 example using a shared filesystem exported from shared storage (and
2486 still have live migration)
2492 *(Released Tue, 13 Jan 2009)*
2494 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2495 - Reuse the old instance mac address by default on instance import, if
2496 the instance name is the same.
2497 - Handle situations in which the node info rpc returns incomplete
2499 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2500 - Improved version of batcher:
2502 - state file support
2503 - instance mac address support
2504 - support for HVM clusters/instances
2506 - Add an option to show the number of cpu sockets and nodes in
2508 - Support OSes that handle more than one version of the OS api (but do
2509 not change the current API in any other way)
2510 - Fix ``gnt-node migrate``
2511 - ``gnt-debug`` man page
2512 - Fixes various more typos and small issues
2513 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2519 *(Released Wed, 24 Sep 2008)*
2521 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2522 type of disk exported to fully virtualized instances.
2523 - provide access to the serial console of HVM instances
2524 - instance auto_balance flag, set by default. If turned off it will
2525 avoid warnings on cluster verify if there is not enough memory to fail
2526 over an instance. in the future it will prevent automatically failing
2527 it over when we will support that.
2528 - batcher tool for instance creation, see ``tools/README.batcher``
2529 - ``gnt-instance reinstall --select-os`` to interactively select a new
2530 operating system when reinstalling an instance.
2531 - when changing the memory amount on instance modify a check has been
2532 added that the instance will be able to start. also warnings are
2533 emitted if the instance will not be able to fail over, if auto_balance
2535 - documentation fixes
2536 - sync fields between ``gnt-instance list/modify/add/import``
2537 - fix a race condition in drbd when the sync speed was set after giving
2538 the device a remote peer.
2544 *(Released Tue, 22 Jul 2008)*
2546 - note: the allowed size and number of tags per object were reduced
2547 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2548 - fixed twisted 8.x compatibility
2549 - fixed ``gnt-instance replace-disks`` with iallocator
2550 - add TCP keepalives on twisted connections to detect restarted nodes
2551 - disk increase support, see ``gnt-instance grow-disk``
2552 - implement bulk node/instance query for RAPI
2553 - add tags in node/instance listing (optional)
2554 - experimental migration (and live migration) support, read the man page
2555 for ``gnt-instance migrate``
2556 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2557 has some small improvements in handling its state file
2563 *(Released Fri, 13 Jun 2008)*
2565 - Experimental readonly, REST-based remote API implementation;
2566 automatically started on master node, TCP port 5080, if enabled by
2567 ``--enable-rapi`` parameter to configure script.
2568 - Instance allocator support. Add and import instance accept a
2569 ``--iallocator`` parameter, and call that instance allocator to decide
2570 which node to use for the instance. The iallocator document describes
2571 what's expected from an allocator script.
2572 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2573 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2574 node is lost there is still enough memory to fail over the instances
2576 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2577 ``gnt-cluster verify``, to check for site-specific compliance. All the
2578 hooks will run, and their output, if any, will be displayed. Any
2579 failing hook will make the verification return an error value.
2580 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2581 primary and secondary interfaces
2582 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2583 that is still declared as part of the cluster and has failed.
2584 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2585 specifying output fields, that just adds the chosen fields to the
2587 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2588 export from the filesystem.
2589 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2590 have been added. Using them you can enable/disable acpi and pae
2591 support, and specify a path for a cd image to be exported to the
2592 instance. These parameters as the name suggest only work on HVM
2594 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2595 PAE support will be set to the previously hardcoded values, but the
2596 (previously hardcoded) path to the CDROM ISO image will be unset and
2597 if required, needs to be set manually with ``gnt-instance modify``
2599 - The address to which an instance's VNC console is bound is now
2600 selectable per-instance, rather than being cluster wide. Of course
2601 this only applies to instances controlled via VNC, so currently just
2602 applies to HVM clusters.
2608 *(Released Mon, 18 Feb 2008)*
2610 - more tweaks to the disk activation code (especially helpful for DRBD)
2611 - change the default ``gnt-instance list`` output format, now there is
2612 one combined status field (see the manpage for the exact values this
2614 - some more fixes for the mac export to hooks change
2615 - make Ganeti not break with DRBD 8.2.x (which changed the version
2616 format in ``/proc/drbd``) (issue 24)
2617 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2618 template, allowing migration from DRBD0.7+MD to DRBD8
2624 *(Released Wed, 30 Jan 2008)*
2626 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2628 - add command aliases infrastructure and a few aliases
2629 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2630 man pages and the ``--help`` option of ``gnt-node
2631 list``/``gnt-instance list``
2632 - fix ``gnt-backup list`` with down nodes (issue 21)
2633 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2634 - fix the dist archive and add a check for including svn/git files in
2636 - some developer-related changes: improve the burnin and the QA suite,
2637 add an upload script for testing during development
2643 *(Released Wed, 16 Jan 2008)*
2645 - experimental HVM support, read the install document, section
2646 "Initializing the cluster"
2647 - allow for the PVM hypervisor per-instance kernel and initrd paths
2648 - add a new command ``gnt-cluster verify-disks`` which uses a new
2649 algorithm to improve the reconnection of the DRBD pairs if the device
2650 on the secondary node has gone away
2651 - make logical volume code auto-activate LVs at disk activation time
2652 - slightly improve the speed of activating disks
2653 - allow specification of the MAC address at instance creation time, and
2654 changing it later via ``gnt-instance modify``
2655 - fix handling of external commands that generate lots of output on
2657 - update documentation with regard to minimum version of DRBD8 supported
2663 *(Released Tue, 4 Dec 2007)*
2665 - Log the ``xm create`` output to the node daemon log on failure (to
2666 help diagnosing the error)
2667 - In debug mode, log all external commands output if failed to the logs
2668 - Change parsing of lvm commands to ignore stderr
2674 *(Released Wed, 28 Nov 2007)*
2676 - Another round of updates to the DRBD 8 code to deal with more failures
2677 in the replace secondary node operation
2678 - Some more logging of failures in disk operations (lvm, drbd)
2679 - A few documentation updates
2686 *(Released Tue, 13 Nov 2007)*
2688 - Change configuration file format from Python's Pickle to JSON.
2689 Upgrading is possible using the cfgupgrade utility.
2690 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2691 faster replace disks and is more stable (DRBD 8 has many improvements
2692 compared to DRBD 0.7)
2693 - Added command line tags support (see man pages for ``gnt-instance``,
2694 ``gnt-node``, ``gnt-cluster``)
2695 - Added instance rename support
2696 - Added multi-instance startup/shutdown
2697 - Added cluster rename support
2698 - Added ``gnt-node evacuate`` to simplify some node operations
2699 - Added instance reboot operation that can speedup reboot as compared to
2701 - Soften the requirement that hostnames are in FQDN format
2702 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2704 - Removed dependency on debian's patched fping that uses the
2705 non-standard ``-S`` option
2706 - Now the OS definitions are searched for in multiple, configurable
2707 paths (easier for distros to package)
2708 - Some changes to the hooks infrastructure (especially the new
2709 post-configuration update hook)
2710 - Other small bugfixes
2712 .. vim: set textwidth=72 syntax=rst :