10 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
11 the correct command syntax. It is possible to use different DRBD versions
12 on different nodes as long as they are compatible to each other. This
13 enables rolling upgrades of DRBD with no downtime. As permanent operation
14 of different DRBD versions within a node group is discouraged,
15 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
23 - The minimum Python version needed to run Ganeti is now 2.6.
24 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
25 to always require authentication, ``--require-authentication``. It can
26 be specified in ``$sysconfdir/default/ganeti``.
27 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
28 be used to manage the disk templates to be used by instances in the cluster.
29 Initially, it will be set to a list that includes plain, drbd, if they were
30 enabled by specifying a volume group name, and file and sharedfile, if those
31 were enabled at configure time. Additionally, it will include all disk
32 templates that are currently used by instances. The order of disk templates
33 will be based on Ganeti's history of supporting them. In the future, the
34 first entry of the list will be used as a default disk template on instance
36 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
37 configuration back to the previous stable version.
38 - The cluster option '--no-lvm-storage' was removed in favor of the new option
39 '--enabled-disk-templates'.
47 Incompatible/important changes
48 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
50 - Instance policies for disk size were documented to be on a per-disk
51 basis, but hail applied them to the sum of all disks. This has been
53 - ``hbal`` will now exit with status 0 if, during job execution over
54 LUXI, early exit has been requested and all jobs are successful;
55 before, exit status 1 was used, which cannot be differentiated from
57 - Compatibility with newer versions of rbd has been fixed
58 - ``gnt-instance batch-create`` has been changed to use the bulk create
59 opcode from Ganeti. This lead to incompatible changes in the format of
60 the JSON file. It's now not a custom dict anymore but a dict
61 compatible with the ``OpInstanceCreate`` opcode.
62 - Parent directories for file storage need to be listed in
63 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
64 write the file automatically based on old configuration values, but it
65 can not distribute it across all nodes and the file contents should be
66 verified. Use ``gnt-cluster copyfile
67 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
68 upgraded. The reason for requiring this list of paths now is that
69 before it would have been possible to inject new paths via RPC,
70 allowing files to be created in arbitrary locations. The RPC protocol
71 is protected using SSL/X.509 certificates, but as a design principle
72 Ganeti does not permit arbitrary paths to be passed.
73 - The parsing of the variants file for OSes (see
74 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
75 lines and comment lines (starting with ``#``) are ignored for better
77 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
78 longer available. ``gnt-node add`` now invokes a new tool on the
79 destination node, named ``prepare-node-join``, to configure the SSH
80 daemon. Paramiko is no longer necessary to configure nodes' SSH
81 daemons via ``gnt-node add``.
82 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
83 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
84 the flag is not reset after a master failover.
85 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
86 is recommended. 2.6 will be mandatory from the 2.8 series.
92 - New network management functionality to support automatic allocation
93 of IP addresses and managing of network parameters. See
94 :manpage:`gnt-network(8)` for more details.
95 - New external storage backend, to allow managing arbitrary storage
96 systems external to the cluster. See
97 :manpage:`ganeti-extstorage-interface(7)`.
98 - New ``exclusive-storage`` node parameter added, restricted to
99 nodegroup level. When it's set to true, physical disks are assigned in
100 an exclusive fashion to instances, as documented in :doc:`Partitioned
101 Ganeti <design-partitioned>`. Currently, only instances using the
102 ``plain`` disk template are supported.
103 - The KVM hypervisor has been updated with many new hypervisor
104 parameters, including a generic one for passing arbitrary command line
105 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
106 compatible up to qemu 1.4.
107 - A new tool, called ``mon-collector``, is the stand-alone executor of
108 the data collectors for a monitoring system. As of this version, it
109 just includes the DRBD data collector, that can be executed by calling
110 ``mon-collector`` using the ``drbd`` parameter. See
111 :manpage:`mon-collector(7)`.
112 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
113 for RAPI users. It allows granting permissions to query for
114 information to a specific user without giving
115 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
116 - A new tool named ``node-cleanup`` has been added. It cleans remains of
117 a cluster from a machine by stopping all daemons, removing
118 certificates and ssconf files. Unless the ``--no-backup`` option is
119 given, copies of the certificates are made.
120 - Instance creations now support the use of opportunistic locking,
121 potentially speeding up the (parallel) creation of multiple instances.
122 This feature is currently only available via the :doc:`RAPI
123 <rapi>` interface and when an instance allocator is used. If the
124 ``opportunistic_locking`` parameter is set the opcode will try to
125 acquire as many locks as possible, but will not wait for any locks
126 held by other opcodes. If not enough resources can be found to
127 allocate the instance, the temporary error code
128 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
129 retried thereafter, with or without opportunistic locking.
130 - New experimental linux-ha resource scripts.
131 - Restricted-commands support: ganeti can now be asked (via command line
132 or rapi) to perform commands on a node. These are passed via ganeti
133 RPC rather than ssh. This functionality is restricted to commands
134 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
135 reasons. The file is not copied automatically.
141 - Diskless instances are now externally mirrored (Issue 237). This for
142 now has only been tested in conjunction with explicit target nodes for
144 - Queries not needing locks or RPC access to the node can now be
145 performed by the confd daemon, making them independent from jobs, and
146 thus faster to execute. This is selectable at configure time.
147 - The functionality for allocating multiple instances at once has been
148 overhauled and is now also available through :doc:`RAPI <rapi>`.
153 - Fix kvm compatibility with qemu 1.4 (Issue 389)
154 - Documentation updates (admin guide, upgrade notes, install
155 instructions) (Issue 372)
156 - Fix gnt-group list nodes and instances count (Issue 436)
157 - Fix compilation without non-mandatory libraries (Issue 441)
158 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
159 - Make confd logging more verbose at INFO level (Issue 435)
160 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
161 - Fix failure path for instance storage type conversion (Issue 229)
162 - Update htools text backend documentation
163 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
164 - Disable inter-cluster instance move for file-based instances, because
165 it is dependant on instance export, which is not supported for
166 file-based instances. (Issue 414)
172 *(Released Mon, 22 Apr 2013)*
174 This was the third beta release of the 2.7 series. Since beta2:
176 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
177 - Fix data loss on wrong usage of ``gnt-instance move``
178 - Properly export errors in confd-based job queries
179 - Add ``users-setup`` tool
180 - Fix iallocator protocol to report 0 as a disk size for diskless
181 instances. This avoids hail breaking when a diskless instance is
183 - Fix job queue directory permission problem that made confd job queries
184 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
185 for access to archived jobs (Issue 406).
186 - Limit the sizes of networks supported by ``gnt-network`` to something
187 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
188 - Fix bugs in instance disk template conversion
189 - Fix GHC 7 compatibility
190 - Fix ``burnin`` install path (Issue 426).
191 - Allow very small disk grows (Issue 347).
192 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
193 that noded doesn't import masterd code (Issue 419).
194 - Make sure the default metavg at cluster init is the same as the vg, if
195 unspecified (Issue 358).
196 - Fix cleanup of partially created disks (part of Issue 416)
202 *(Released Tue, 2 Apr 2013)*
204 This was the second beta release of the 2.7 series. Since beta1:
206 - Networks no longer have a "type" slot, since this information was
207 unused in Ganeti: instead of it tags should be used.
208 - The rapi client now has a ``target_node`` option to MigrateInstance.
209 - Fix early exit return code for hbal (Issue 386).
210 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
211 - Fix ``rbd showmapped`` output parsing (Issue 312).
212 - Networks are now referenced indexed by UUID, rather than name. This
213 will require running cfgupgrade, from 2.7.0beta1, if networks are in
215 - The OS environment now includes network information.
216 - Deleting of a network is now disallowed if any instance nic is using
217 it, to prevent dangling references.
218 - External storage is now documented in man pages.
219 - The exclusive_storage flag can now only be set at nodegroup level.
220 - Hbal can now submit an explicit priority with its jobs.
221 - Many network related locking fixes.
222 - Bump up the required pylint version to 0.25.1.
223 - Fix the ``no_remember`` option in RAPI client.
224 - Many ipolicy related tests, qa, and fixes.
225 - Many documentation improvements and fixes.
226 - Fix building with ``--disable-file-storage``.
227 - Fix ``-q`` option in htools, which was broken if passed more than
229 - Some haskell/python interaction improvements and fixes.
230 - Fix iallocator in case of missing LVM storage.
231 - Fix confd config load in case of ``--no-lvm-storage``.
232 - The confd/query functionality is now mentioned in the security
239 *(Released Wed, 6 Feb 2013)*
241 This was the first beta release of the 2.7 series. All important changes
242 are listed in the latest 2.7 entry.
248 *(Released Fri, 21 Dec 2012)*
250 Important behaviour change: hbal won't rebalance anymore instances which
251 have the ``auto_balance`` attribute set to false. This was the intention
252 all along, but until now it only skipped those from the N+1 memory
253 reservation (DRBD-specific).
255 A significant number of bug fixes in this release:
257 - Fixed disk adoption interaction with ipolicy checks.
258 - Fixed networking issues when instances are started, stopped or
259 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
260 - Fixed the warning in cluster verify for shared storage instances not
262 - Fixed removal of storage directory on shared file storage (issue 262).
263 - Fixed validation of LVM volume group name in OpClusterSetParams
264 (``gnt-cluster modify``) (issue 285).
265 - Fixed runtime memory increases (``gnt-instance modify -m``).
266 - Fixed live migration under Xen's ``xl`` mode.
267 - Fixed ``gnt-instance console`` with ``xl``.
268 - Fixed building with newer Haskell compiler/libraries.
269 - Fixed PID file writing in Haskell daemons (confd); this prevents
270 restart issues if confd was launched manually (outside of
271 ``daemon-util``) while another copy of it was running
272 - Fixed a type error when doing live migrations with KVM (issue 297) and
273 the error messages for failing migrations have been improved.
274 - Fixed opcode validation for the out-of-band commands (``gnt-node
276 - Fixed a type error when unsetting OS hypervisor parameters (issue
277 311); now it's possible to unset all OS-specific hypervisor
279 - Fixed the ``dry-run`` mode for many operations: verification of
280 results was over-zealous but didn't take into account the ``dry-run``
281 operation, resulting in "wrong" failures.
282 - Fixed bash completion in ``gnt-job list`` when the job queue has
283 hundreds of entries; especially with older ``bash`` versions, this
284 results in significant CPU usage.
286 And lastly, a few other improvements have been made:
288 - Added option to force master-failover without voting (issue 282).
289 - Clarified error message on lock conflict (issue 287).
290 - Logging of newly submitted jobs has been improved (issue 290).
291 - Hostname checks have been made uniform between instance rename and
293 - The ``--submit`` option is now supported by ``gnt-debug delay``.
294 - Shutting down the master daemon by sending SIGTERM now stops it from
295 processing jobs waiting for locks; instead, those jobs will be started
296 once again after the master daemon is started the next time (issue
298 - Support for Xen's ``xl`` program has been improved (besides the fixes
300 - Reduced logging noise in the Haskell confd daemon (only show one log
301 entry for each config reload, instead of two).
302 - Several man page updates and typo fixes.
308 *(Released Fri, 12 Oct 2012)*
310 A small bugfix release. Among the bugs fixed:
312 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
313 made the command unusable.
314 - Commands that issue many jobs don't fail anymore just because some jobs
315 take so long that other jobs are archived.
316 - Failures during ``gnt-instance reinstall`` are reflected by the exit
318 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
320 - When ``always_failover`` is set, ``--allow-failover`` is not required
321 in migrate commands anymore.
322 - ``bash_completion`` works even if extglob is disabled.
323 - Fixed bug with locks that made failover for RDB-based instances fail.
324 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
325 a random node instead of one based on the allocator metric.
326 - Support for newer versions of pylint and pep8.
327 - Hail doesn't fail anymore when trying to add an instance of type
328 ``file``, ``sharedfile`` or ``rbd``.
329 - Added new Makefile target to rebuild the whole distribution, so that
330 all files are included.
336 *(Released Fri, 27 Jul 2012)*
339 .. attention:: The ``LUXI`` protocol has been made more consistent
340 regarding its handling of command arguments. This, however, leads to
341 incompatibility issues with previous versions. Please ensure that you
342 restart Ganeti daemons soon after the upgrade, otherwise most
343 ``LUXI`` calls (job submission, setting/resetting the drain flag,
344 pausing/resuming the watcher, cancelling and archiving jobs, querying
345 the cluster configuration) will fail.
354 The current ``admin_up`` field, which used to denote whether an instance
355 should be running or not, has been removed. Instead, ``admin_state`` is
356 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
358 The rational behind this is that an instance being “down” can have
361 - it could be down during a reboot
362 - it could be temporarily be down for a reinstall
363 - or it could be down because it is deprecated and kept just for its
366 The previous Boolean state was making it difficult to do capacity
367 calculations: should Ganeti reserve memory for a down instance? Now, the
368 tri-state field makes it clear:
370 - in ``up`` and ``down`` state, all resources are reserved for the
371 instance, and it can be at any time brought up if it is down
372 - in ``offline`` state, only disk space is reserved for it, but not
375 The field can have an extra use: since the transition between ``up`` and
376 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
377 transition between ``offline`` and ``down`` is done via ``gnt-instance
378 modify``, it is possible to given different rights to users. For
379 example, owners of an instance could be allowed to start/stop it, but
380 not transition it out of the offline state.
382 Instance policies and specs
383 +++++++++++++++++++++++++++
385 In previous Ganeti versions, an instance creation request was not
386 limited on the minimum size and on the maximum size just by the cluster
387 resources. As such, any policy could be implemented only in third-party
388 clients (RAPI clients, or shell wrappers over ``gnt-*``
389 tools). Furthermore, calculating cluster capacity via ``hspace`` again
390 required external input with regards to instance sizes.
392 In order to improve these workflows and to allow for example better
393 per-node group differentiation, we introduced instance specs, which
396 - minimum instance disk size, disk count, memory size, cpu count
397 - maximum values for the above metrics
398 - and “standard” values (used in ``hspace`` to calculate the standard
401 The minimum/maximum values can be also customised at node-group level,
402 for example allowing more powerful hardware to support bigger instance
405 Beside the instance specs, there are a few other settings belonging to
406 the instance policy framework. It is possible now to customise, per
407 cluster and node-group:
409 - the list of allowed disk templates
410 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
411 - the maximum ratio of instance to spindles (see below for more
412 information) for local storage
414 All these together should allow all tools that talk to Ganeti to know
415 what are the ranges of allowed values for instances and the
416 over-subscription that is allowed.
418 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
419 instance configuration, and the physical CPU configuration from the
420 node. For the spindle ratios however, we didn't track before these
421 values, so new parameters have been added:
423 - a new node parameter ``spindle_count``, defaults to 1, customisable at
424 node group or node level
425 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
427 Note that spindles in this context doesn't need to mean actual
428 mechanical hard-drives; it's just a relative number for both the node
429 I/O capacity and instance I/O consumption.
431 Instance migration behaviour
432 ++++++++++++++++++++++++++++
434 While live-migration is in general desirable over failover, it is
435 possible that for some workloads it is actually worse, due to the
436 variable time of the “suspend” phase during live migration.
438 To allow the tools to work consistently over such instances (without
439 having to hard-code instance names), a new backend parameter
440 ``always_failover`` has been added to control the migration/failover
441 behaviour. When set to True, all migration requests for an instance will
442 instead fall-back to failover.
444 Instance memory ballooning
445 ++++++++++++++++++++++++++
447 Initial support for memory ballooning has been added. The memory for an
448 instance is no longer fixed (backend parameter ``memory``), but instead
449 can vary between minimum and maximum values (backend parameters
450 ``minmem`` and ``maxmem``). Currently we only change an instance's
453 - live migrating or failing over and instance and the target node
454 doesn't have enough memory
455 - user requests changing the memory via ``gnt-instance modify
461 In order to control the use of specific CPUs by instance, support for
462 controlling CPU pinning has been added for the Xen, HVM and LXC
463 hypervisors. This is controlled by a new hypervisor parameter
464 ``cpu_mask``; details about possible values for this are in the
465 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
466 VCPU-to-CPU mapping) form will work well only when all nodes in your
467 cluster have the same amount of CPUs.
472 Another area in which Ganeti was not customisable were the parameters
473 used for storage configuration, e.g. how many stripes to use for LVM,
474 DRBD resync configuration, etc.
476 To improve this area, we've added disks parameters, which are
477 customisable at cluster and node group level, and which allow to
478 specify various parameters for disks (DRBD has the most parameters
479 currently), for example:
481 - DRBD resync algorithm and parameters (e.g. speed)
482 - the default VG for meta-data volumes for DRBD
483 - number of stripes for LVM (plain disk template)
486 These parameters can be modified via ``gnt-cluster modify -D …`` and
487 ``gnt-group modify -D …``, and are used at either instance creation (in
488 case of LVM stripes, for example) or at disk “activation” time
491 Rados block device support
492 ++++++++++++++++++++++++++
494 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
495 denoted by the ``rbd`` disk template type. This is considered
496 experimental, feedback is welcome. For details on configuring it, see
497 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
502 The existing master IP functionality works well only in simple setups (a
503 single network shared by all nodes); however, if nodes belong to
504 different networks, then the ``/32`` setup and lack of routing
505 information is not enough.
507 To allow the master IP to function well in more complex cases, the
508 system was reworked as follows:
510 - a master IP netmask setting has been added
511 - the master IP activation/turn-down code was moved from the node daemon
513 - whether to run the Ganeti-supplied master IP script or a user-supplied
514 on is a ``gnt-cluster init`` setting
516 Details about the location of the standard and custom setup scripts are
517 in the man page :manpage:`gnt-cluster(8)`; for information about the
518 setup script protocol, look at the Ganeti-supplied script.
523 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
526 It is now possible to use TLS-protected connections, and when renewing
527 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
528 it is now possible to specify spice or spice CA certificates. Also, it
529 is possible to configure a password for SPICE sessions via the
530 hypervisor parameter ``spice_password_file``.
532 There are also new parameters to control the compression and streaming
533 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
534 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
535 for the spice parameters.
537 Lastly, it is now possible to see the SPICE connection information via
538 ``gnt-instance console``.
543 A new tool (``tools/ovfconverter``) has been added that supports
544 conversion between Ganeti and the `Open Virtualization Format
545 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
548 This relies on the ``qemu-img`` tool to convert the disk formats, so the
549 actual compatibility with other virtualization solutions depends on it.
554 The configuration query daemon (``ganeti-confd``) is now optional, and
555 has been rewritten in Haskell; whether to use the daemon at all, use the
556 Python (default) or the Haskell version is selectable at configure time
557 via the ``--enable-confd`` parameter, which can take one of the
558 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
559 daemon will result in a smaller footprint; for larger systems, we
560 welcome feedback on the Haskell version which might become the default
563 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
564 daemon running. The Python version doesn't implement the new call.
567 User interface changes
568 ~~~~~~~~~~~~~~~~~~~~~~
570 We have replaced the ``--disks`` option of ``gnt-instance
571 replace-disks`` with a more flexible ``--disk`` option, which allows
572 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
573 disk size and mode can be changed upon recreation (via ``gnt-instance
574 recreate-disks``, which accepts the same ``--disk`` option).
576 As many people are used to a ``show`` command, we have added that as an
577 alias to ``info`` on all ``gnt-*`` commands.
579 The ``gnt-instance grow-disk`` command has a new mode in which it can
580 accept the target size of the disk, instead of the delta; this can be
581 more safe since two runs in absolute mode will be idempotent, and
582 sometimes it's also easier to specify the desired size directly.
584 Also the handling of instances with regard to offline secondaries has
585 been improved. Instance operations should not fail because one of it's
586 secondary nodes is offline, even though it's safe to proceed.
588 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
589 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
590 minors to instance name.
595 RAPI coverage has improved, with (for example) new resources for
596 recreate-disks, node power-cycle, etc.
601 There is partial support for ``xl`` in the Xen hypervisor; feedback is
604 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
605 whether to still support Python 2.4 or move to Python 2.6 as minimum
608 Support for Fedora has been slightly improved; the provided example
609 init.d script should work better on it and the INSTALL file should
610 document the needed dependencies.
615 The deprecated ``QueryLocks`` LUXI request has been removed. Use
616 ``Query(what=QR_LOCK, ...)`` instead.
618 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
619 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
620 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
621 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
622 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
624 RAPI client: ``CertificateError`` now derives from
625 ``GanetiApiError``. This should make it more easy to handle Ganeti
628 Deprecation warnings due to PyCrypto/paramiko import in
629 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
630 make sure to run an up-to-date paramiko version, if you use this tool.
632 The QA scripts now depend on Python 2.5 or above (the main code base
633 still works with Python 2.4).
635 The configuration file (``config.data``) is now written without
636 indentation for performance reasons; if you want to edit it, it can be
637 re-formatted via ``tools/fmtjson``.
639 A number of bugs has been fixed in the cluster merge tool.
641 ``x509`` certification verification (used in import-export) has been
642 changed to allow the same clock skew as permitted by the cluster
643 verification. This will remove some rare but hard to diagnose errors in
650 *(Released Thu, 19 Jul 2012)*
652 Very few changes from rc4 to the final release, only bugfixes:
654 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
655 instance, fix CDROM booting for KVM instances)
656 - fixed node group modification of node parameters
657 - fixed issue in LUClusterVerifyGroup with multi-group clusters
658 - fixed generation of bash completion to ensure a stable ordering
665 *(Released Fri, 13 Jul 2012)*
667 Third release candidate for 2.6. The following changes were done from
670 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
671 - Fixed an inconsistency in the LUXI protocol with the provided
672 arguments (NOT backwards compatible)
673 - Fixed a bug with node groups ipolicy where ``min`` was greater than
674 the cluster ``std`` value
675 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
676 easier instance debugging on nodes (requires ``hconfd`` to work)
682 *(Released Tue, 03 Jul 2012)*
684 Second release candidate for 2.6. The following changes were done from
687 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
689 - Fixed a RAPI regression on missing beparams/memory
690 - Fixed redistribution of files on offline nodes
691 - Added possibility to run activate-disks even though secondaries are
692 offline. With this change it relaxes also the strictness on some other
693 commands which use activate disks internally:
694 * ``gnt-instance start|reboot|rename|backup|export``
695 - Made it possible to remove safely an instance if its secondaries are
697 - Made it possible to reinstall even though secondaries are offline
703 *(Released Mon, 25 Jun 2012)*
705 First release candidate for 2.6. The following changes were done from
708 - Fixed bugs with disk parameters and ``rbd`` templates as well as
710 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
711 behaviour. It supports now the modify operation
712 - ``hcheck`` implemented to analyze cluster health and possibility of
713 improving health by rebalance
714 - ``hbal`` has been improved in dealing with split instances
720 *(Released Mon, 11 Jun 2012)*
722 Second beta release of 2.6. The following changes were done from beta2
725 - Fixed ``daemon-util`` with non-root user models
726 - Fixed creation of plain instances with ``--no-wait-for-sync``
727 - Fix wrong iv_names when running ``cfgupgrade``
728 - Export more information in RAPI group queries
729 - Fixed bug when changing instance network interfaces
730 - Extended burnin to do NIC changes
731 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
732 - Changed default for DRBD barriers
733 - Fixed DRBD error reporting for syncer rate
734 - Verify the options on disk parameters
736 And of course various fixes to documentation and improved unittests and
743 *(Released Wed, 23 May 2012)*
745 First beta release of 2.6. The following changes were done from beta1 to
748 - integrated patch for distributions without ``start-stop-daemon``
749 - adapted example init.d script to work on Fedora
750 - fixed log handling in Haskell daemons
751 - adapted checks in the watcher for pycurl linked against libnss
752 - add partial support for ``xl`` instead of ``xm`` for Xen
753 - fixed a type issue in cluster verification
754 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
757 Plus integrated fixes from the 2.5 branch:
759 - fixed ``kvm-ifup`` to use ``/bin/bash``
760 - fixed parallel build failures
761 - KVM live migration when using a custom keymap
767 *(Released Tue, 24 Jul 2012)*
769 A small bugfix release, with no new features:
771 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
772 different default shell (e.g. Debian, Ubuntu)
773 - fixed KVM startup and live migration with a custom keymap (fixes Issue
774 243 and Debian bug #650664)
775 - fixed compatibility with KVM versions that don't support multiple boot
776 devices (fixes Issue 230 and Debian bug #624256)
778 Additionally, a few fixes were done to the build system (fixed parallel
779 build failures) and to the unittests (fixed race condition in test for
780 FileID functions, and the default enable/disable mode for QA test is now
787 *(Released Fri, 11 May 2012)*
789 A small bugfix release.
791 The main issues solved are on the topic of compatibility with newer LVM
794 - fixed parsing of ``lv_attr`` field
795 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
796 the ``--force`` flag is needed
798 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
799 to accept kernel 3.x too (was hardcoded to 2.6.*).
801 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
802 of hook results and that also made display of other errors suboptimal
803 was fixed; the code behaves now like 2.4 and earlier.
805 Another change in 2.5, the cleanup of the OS scripts environment, is too
806 aggressive: it removed even the ``PATH`` variable, which requires the OS
807 scripts to *always* need to export it. Since this is a bit too strict,
808 we now export a minimal PATH, the same that we export for hooks.
810 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
811 integrated into this release.
813 Finally, a few other miscellaneous changes were done (no new features,
814 just small improvements):
816 - Fix ``gnt-group --help`` display
817 - Fix hardcoded Xen kernel path
818 - Fix grow-disk handling of invalid units
819 - Update synopsis for ``gnt-cluster repair-disk-sizes``
820 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
826 *(Released Thu, 12 Apr 2012)*
828 Incompatible/important changes and bugfixes
829 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
831 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
832 resource's ``ip_check`` parameter changed from ``True`` to ``False``
833 to match the underlying LUXI interface.
834 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
835 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
836 not maintain backwards-compatibility as the underlying operation
837 changed in an incompatible way. The RAPI client can talk to old
838 servers, but it needs to be told so as the return value changed.
839 - When creating file-based instances via RAPI, the ``file_driver``
840 parameter no longer defaults to ``loop`` and must be specified.
841 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
843 - Support for the undocumented and deprecated RAPI instance creation
844 request format version 0 has been dropped. Use version 1, supported
845 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
846 - Pyparsing 1.4.6 or above is required, see :doc:`installation
847 documentation <install>`.
848 - The "cluster-verify" hooks are now executed per group by the
849 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
850 if you just run ``gnt-cluster verify``, which generates one opcode per
852 - The environment as passed to the OS scripts is cleared, and thus no
853 environment variables defined in the node daemon's environment will be
854 inherited by the scripts.
855 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
857 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
858 support operations involving multiple node groups.
859 - Offline nodes are ignored when failing over an instance.
860 - Support for KVM version 1.0, which changed the version reporting format
862 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
864 - ``Makefile`` is now compatible with Automake 1.11.2
865 - Includes all bugfixes made in the 2.4 series
870 - The ganeti-htools project has been merged into the ganeti-core source
871 tree and will be built as part of Ganeti (see :doc:`install-quick`).
872 - Implemented support for :doc:`shared storage <design-shared-storage>`.
873 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
874 GPT-style partition tables (requires `parted
875 <http://www.gnu.org/s/parted/>`_).
876 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
877 - Allowed adding tags on instance creation.
878 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
879 - Allow instances to be started in a paused state, enabling the user to
880 see the complete console output on boot using the console.
881 - Added new hypervisor flag to control default reboot behaviour
882 (``reboot_behavior``).
883 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
884 - Improved out-of-band management support:
886 - Added ``gnt-node health`` command reporting the health status of
888 - Added ``gnt-node power`` command to manage power status of nodes.
889 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
891 - Instance migration can fall back to failover if instance is not
893 - Filters can be used when listing nodes, instances, groups and locks;
894 see :manpage:`ganeti(7)` manpage.
895 - Added post-execution status as variables to :doc:`hooks <hooks>`
897 - Instance tags are exported/imported together with the instance.
898 - When given an explicit job ID, ``gnt-job info`` will work for archived
900 - Jobs can define dependencies on other jobs (not yet supported via
901 RAPI or command line, but used by internal commands and usable via
904 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
907 - Instance failover is now available as a RAPI resource
908 (``/2/instances/[instance_name]/failover``).
909 - ``gnt-instance info`` defaults to static information if primary node
911 - Opcodes have a new ``comment`` attribute.
912 - Added basic SPICE support to KVM hypervisor.
913 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
915 Node group improvements
916 ~~~~~~~~~~~~~~~~~~~~~~~
918 - ``gnt-cluster verify`` has been modified to check groups separately,
919 thereby improving performance.
920 - Node group support has been added to ``gnt-cluster verify-disks``,
921 which now operates per node group.
922 - Watcher has been changed to work better with node groups.
924 - One process and state file per node group.
925 - Slow watcher in one group doesn't block other group's watcher.
927 - Added new command, ``gnt-group evacuate``, to move all instances in a
928 node group to other groups.
929 - Added ``gnt-instance change-group`` to move an instance to another
931 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
932 per-group operations.
933 - Node groups can be tagged.
934 - Some operations switch from an exclusive to a shared lock as soon as
936 - Instance's primary and secondary nodes' groups are now available as
937 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
938 and ``snodes.group.uuid``).
943 - Numerous updates to documentation and manpages.
945 - :doc:`RAPI <rapi>` documentation now has detailed parameter
947 - Some opcode/job results are now also documented, see :doc:`RAPI
950 - A lockset's internal lock is now also visible in lock monitor.
951 - Log messages from job queue workers now contain information about the
952 opcode they're processing.
953 - ``gnt-instance console`` no longer requires the instance lock.
954 - A short delay when waiting for job changes reduces the number of LUXI
955 requests significantly.
956 - DRBD metadata volumes are overwritten with zeros during disk creation.
957 - Out-of-band commands no longer acquire the cluster lock in exclusive
959 - ``devel/upload`` now uses correct permissions for directories.
965 *(Released Fri, 23 Mar 2012)*
967 This was the sixth release candidate of the 2.5 series.
973 *(Released Mon, 9 Jan 2012)*
975 This was the fifth release candidate of the 2.5 series.
981 *(Released Thu, 27 Oct 2011)*
983 This was the fourth release candidate of the 2.5 series.
989 *(Released Wed, 26 Oct 2011)*
991 This was the third release candidate of the 2.5 series.
997 *(Released Tue, 18 Oct 2011)*
999 This was the second release candidate of the 2.5 series.
1005 *(Released Tue, 4 Oct 2011)*
1007 This was the first release candidate of the 2.5 series.
1013 *(Released Wed, 31 Aug 2011)*
1015 This was the third beta release of the 2.5 series.
1021 *(Released Mon, 22 Aug 2011)*
1023 This was the second beta release of the 2.5 series.
1029 *(Released Fri, 12 Aug 2011)*
1031 This was the first beta release of the 2.5 series.
1037 *(Released Thu, 27 Oct 2011)*
1039 - Fixed bug when parsing command line parameter values ending in
1041 - Fixed assertion error after unclean master shutdown
1042 - Disable HTTP client pool for RPC, significantly reducing memory usage
1044 - Fixed queue archive creation with wrong permissions
1050 *(Released Tue, 23 Aug 2011)*
1054 - Fixed documentation for importing with ``--src-dir`` option
1055 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1056 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1062 *(Released Fri, 5 Aug 2011)*
1064 Many bug-fixes and a few small features:
1066 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1067 issues when you tried to add an instance with two MAC addresses in one
1069 - KVM: fixed per-instance stored UID value
1070 - KVM: configure bridged NICs at migration start
1071 - KVM: Fix a bug where instance will not start with never KVM versions
1073 - Added OS search path to ``gnt-cluster info``
1074 - Fixed an issue with ``file_storage_dir`` where you were forced to
1075 provide an absolute path, but the documentation states it is a
1076 relative path, the documentation was right
1077 - Added a new parameter to instance stop/start called ``--no-remember``
1078 that will make the state change to not be remembered
1079 - Implemented ``no_remember`` at RAPI level
1080 - Improved the documentation
1081 - Node evacuation: don't call IAllocator if node is already empty
1082 - Fixed bug in DRBD8 replace disks on current nodes
1083 - Fixed bug in recreate-disks for DRBD instances
1084 - Moved assertion checking locks in ``gnt-instance replace-disks``
1085 causing it to abort with not owning the right locks for some situation
1086 - Job queue: Fixed potential race condition when cancelling queued jobs
1087 - Fixed off-by-one bug in job serial generation
1088 - ``gnt-node volumes``: Fix instance names
1089 - Fixed aliases in bash completion
1090 - Fixed a bug in reopening log files after being sent a SIGHUP
1091 - Added a flag to burnin to allow specifying VCPU count
1092 - Bugfixes to non-root Ganeti configuration
1098 *(Released Thu, 12 May 2011)*
1100 Many bug-fixes and a few new small features:
1102 - Fixed a bug related to log opening failures
1103 - Fixed a bug in instance listing with orphan instances
1104 - Fixed a bug which prevented resetting the cluster-level node parameter
1105 ``oob_program`` to the default
1106 - Many fixes related to the ``cluster-merge`` tool
1107 - Fixed a race condition in the lock monitor, which caused failures
1108 during (at least) creation of many instances in parallel
1109 - Improved output for gnt-job info
1110 - Removed the quiet flag on some ssh calls which prevented debugging
1112 - Improved the N+1 failure messages in cluster verify by actually
1113 showing the memory values (needed and available)
1114 - Increased lock attempt timeouts so that when executing long operations
1115 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1116 too early and thus prevent the use of the 'fair' mechanism
1117 - Changed instance query data (``gnt-instance info``) to not acquire
1118 locks unless needed, thus allowing its use on locked instance if only
1119 static information is asked for
1120 - Improved behaviour with filesystems that do not support rename on an
1122 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1123 kept locks on all nodes during the wipe, which is unneeded
1124 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1125 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1126 10GiB) which caused the wipe to fail right at the end in some cases
1127 - Fixed master IP activation when doing master failover with no-voting
1128 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1129 the master node itself
1130 - Fixed potential data-loss in under disk full conditions, where Ganeti
1131 wouldn't check correctly the return code and would consider
1132 partially-written files 'correct'
1133 - Fixed bug related to multiple VGs and DRBD disk replacing
1134 - Added new disk parameter ``metavg`` that allows placement of the meta
1135 device for DRBD in a different volume group
1136 - Fixed error handling in the node daemon when the system libc doesn't
1137 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1138 - Fixed lock release during replace-disks, which kept cluster-wide locks
1139 when doing disk replaces with an iallocator script
1140 - Added check for missing bridges in cluster verify
1141 - Handle EPIPE errors while writing to the terminal better, so that
1142 piping the output to e.g. ``less`` doesn't cause a backtrace
1143 - Fixed rare case where a ^C during Luxi calls could have been
1144 interpreted as server errors, instead of simply terminating
1145 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1147 - Added a few more parameters to the KVM hypervisor, allowing a second
1148 CDROM, custom disk type for CDROMs and a floppy image
1149 - Removed redundant message in instance rename when the name is given
1151 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1152 disks on new nodes, allowing recreation when the original instance
1153 nodes are completely gone
1154 - Added option when converting disk templates to DRBD to skip waiting
1155 for the resync, in order to make the instance available sooner
1156 - Added two new variables to the OS scripts environment (containing the
1158 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1159 to allow use of ``pvgrub`` as bootloader
1160 - Changed the instance memory modifications to only check out-of-memory
1161 conditions on memory increases, and turned the secondary node warnings
1162 into errors (they can still be overridden via ``--force``)
1163 - Fixed the handling of a corner case when the Python installation gets
1164 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1165 to execute a command that doesn't exist
1166 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1167 node of the instance returned failures during instance shutdown; this
1168 adds the option ``--ignore-consistency`` to gnt-instance move
1170 And as usual, various improvements to the error messages, documentation
1177 *(Released Wed, 09 Mar 2011)*
1179 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1180 the RAPI users file if run twice (even with ``--dry-run``).
1182 The release fixes that bug (nothing else changed).
1188 *(Released Mon, 07 Mar 2011)*
1190 Final 2.4.0 release. Just a few small fixes:
1192 - Fixed RAPI node evacuate
1193 - Fixed the kvm-ifup script
1194 - Fixed internal error handling for special job cases
1195 - Updated man page to specify the escaping feature for options
1201 *(Released Mon, 28 Feb 2011)*
1203 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1204 that this feature wiped the disks of the wrong instance, leading to loss
1209 - Fixed title of query field containing instance name
1210 - Expanded the glossary in the documentation
1211 - Fixed one unittest (internal issue)
1217 *(Released Mon, 21 Feb 2011)*
1219 A number of bug fixes plus just a couple functionality changes.
1221 On the user-visible side, the ``gnt-* list`` command output has changed
1222 with respect to "special" field states. The current rc1 style of display
1223 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1224 the default output mode special fields states are displayed as follows:
1226 - Offline resource: ``*``
1227 - Unavailable/not applicable: ``-``
1228 - Data missing (RPC failure): ``?``
1229 - Unknown field: ``??``
1231 Another user-visible change is the addition of ``--force-join`` to
1236 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1237 - Fixed regression in RAPI/instance reinstall where all parameters were
1238 required (instead of optional)
1239 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1240 - Fixed iallocator usage (offline nodes were not considered offline)
1241 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1242 - Fixed hypervisor and OS parameter validation with respect to
1243 non-vm_capable nodes
1244 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1246 - Fixed ``tools/listrunner`` with respect to agent-based usage
1252 *(Released Fri, 4 Feb 2011)*
1254 Many changes and fixes since the beta1 release. While there were some
1255 internal changes, the code has been mostly stabilised for the RC
1258 Note: the dumb allocator was removed in this release, as it was not kept
1259 up-to-date with the IAllocator protocol changes. It is recommended to
1260 use the ``hail`` command from the ganeti-htools package.
1262 Note: the 2.4 and up versions of Ganeti are not compatible with the
1263 0.2.x branch of ganeti-htools. You need to upgrade to
1264 ganeti-htools-0.3.0 (or later).
1266 Regressions fixed from 2.3
1267 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1269 - Fixed the ``gnt-cluster verify-disks`` command
1270 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1272 - Fixed disk adoption breakage
1273 - Fixed wrong headers in instance listing for field aliases
1278 - Fixed corner case in KVM handling of NICs
1279 - Fixed many cases of wrong handling of non-vm_capable nodes
1280 - Fixed a bug where a missing instance symlink was not possible to
1281 recreate with any ``gnt-*`` command (now ``gnt-instance
1282 activate-disks`` does it)
1283 - Fixed the volume group name as reported by ``gnt-cluster
1285 - Increased timeouts for the import-export code, hopefully leading to
1286 fewer aborts due network or instance timeouts
1287 - Fixed bug in ``gnt-node list-storage``
1288 - Fixed bug where not all daemons were started on cluster
1289 initialisation, but only at the first watcher run
1290 - Fixed many bugs in the OOB implementation
1291 - Fixed watcher behaviour in presence of instances with offline
1293 - Fixed instance list output for instances running on the wrong node
1294 - a few fixes to the cluster-merge tool, but it still cannot merge
1295 multi-node groups (currently it is not recommended to use this tool)
1301 - Improved network configuration for the KVM hypervisor
1302 - Added e1000 as a supported NIC for Xen-HVM
1303 - Improved the lvmstrap tool to also be able to use partitions, as
1304 opposed to full disks
1305 - Improved speed of disk wiping (the cluster parameter
1306 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1307 of instance creations
1308 - Added documentation for the OS parameters
1309 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1310 hypervisor is not responding
1311 - Added display of blacklisted and hidden OS information in
1312 ``gnt-cluster info``
1313 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1314 NIC and node parameters, which might create problems with currently
1315 invalid (but undetected) configuration files, but prevents validation
1316 failures when unrelated parameters are modified
1317 - Changed cluster initialisation to wait for the master daemon to become
1319 - Expanded the RAPI interface:
1321 - Added config redistribution resource
1322 - Added activation/deactivation of instance disks
1323 - Added export of console information
1325 - Implemented log file reopening on SIGHUP, which allows using
1326 logrotate(8) for the Ganeti log files
1327 - Added a basic OOB helper script as an example
1333 *(Released Fri, 14 Jan 2011)*
1338 - Fixed timezone issues when formatting timestamps
1339 - Added support for node groups, available via ``gnt-group`` and other
1341 - Added out-of-band framework and management, see :doc:`design
1342 document <design-oob>`
1343 - Removed support for roman numbers from ``gnt-node list`` and
1344 ``gnt-instance list``.
1345 - Allowed modification of master network interface via ``gnt-cluster
1346 modify --master-netdev``
1347 - Accept offline secondaries while shutting down instance disks
1348 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1349 - Added support for multiple LVM volume groups
1350 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1352 - Added commands to list available fields:
1354 - ``gnt-node list-fields``
1355 - ``gnt-group list-fields``
1356 - ``gnt-instance list-fields``
1358 - Updated documentation and man pages
1363 - Moved ``rapi_users`` file into separate directory, now named
1364 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1366 - Added new tool for running commands on many machines,
1367 ``tools/ganeti-listrunner``
1368 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1369 improving the ``gnt-instance console`` output
1370 - Allowed customisation of disk index separator at ``configure`` time
1371 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1372 - Added support for non-partitioned md disks in ``lvmstrap``
1373 - Added script to gracefully power off KVM instances
1374 - Split ``utils`` module into smaller parts
1375 - Changed query operations to return more detailed information, e.g.
1376 whether an information is unavailable due to an offline node. To use
1377 this new functionality, the LUXI call ``Query`` must be used. Field
1378 information is now stored by the master daemon and can be retrieved
1379 using ``QueryFields``. Instances, nodes and groups can also be queried
1380 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1381 exposed via RAPI). The following commands make use of this
1382 infrastructure change:
1384 - ``gnt-group list``
1385 - ``gnt-group list-fields``
1387 - ``gnt-node list-fields``
1388 - ``gnt-instance list``
1389 - ``gnt-instance list-fields``
1390 - ``gnt-debug locks``
1395 - New RAPI resources (see :doc:`rapi`):
1399 - ``/2/groups/[group_name]``
1400 - ``/2/groups/[group_name]/assign-nodes``
1401 - ``/2/groups/[group_name]/modify``
1402 - ``/2/groups/[group_name]/rename``
1403 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1407 - Implemented ``no_install`` for instance creation
1408 - Implemented OS parameters for instance reinstallation, allowing
1409 use of special settings on reinstallation (e.g. for preserving data)
1414 - Added IPv6 support in import/export
1415 - Pause DRBD synchronization while wiping disks on instance creation
1416 - Updated unittests and QA scripts
1417 - Improved network parameters passed to KVM
1418 - Converted man pages from docbook to reStructuredText
1424 *(Released Mon, 20 Dec 2010)*
1426 Released version 2.3.1~rc1 without any changes.
1432 *(Released Wed, 1 Dec 2010)*
1434 - impexpd: Disable OpenSSL compression in socat if possible (backport
1435 from master, commit e90739d625b, see :doc:`installation guide
1436 <install-quick>` for details)
1437 - Changed unittest coverage report to exclude test scripts
1438 - Added script to check version format
1444 *(Released Wed, 1 Dec 2010)*
1446 Released version 2.3.0~rc1 without any changes.
1452 *(Released Fri, 19 Nov 2010)*
1454 A number of bugfixes and documentation updates:
1456 - Update ganeti-os-interface documentation
1457 - Fixed a bug related to duplicate MACs or similar items which should be
1459 - Fix breakage in OS state modify
1460 - Reinstall instance: disallow offline secondaries (fixes bug related to
1461 OS changing but reinstall failing)
1462 - plus all the other fixes between 2.2.1 and 2.2.2
1468 *(Released Tue, 2 Nov 2010)*
1470 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1471 - Fixed master failover race with watcher
1472 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1474 - Accept previously stopped instance for export with instance removal
1475 - Simplify and extend the environment variables for instance OS scripts
1476 - Added new node flags, ``master_capable`` and ``vm_capable``
1477 - Added optional instance disk wiping prior during allocation. This is a
1478 cluster-wide option and can be set/modified using
1479 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1480 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1481 :doc:`install-quick`
1482 - Added a new watcher option (``--ignore-pause``)
1483 - Added option to ignore offline node on instance start/stop
1484 (``--ignore-offline``)
1485 - Allow overriding OS parameters with ``gnt-instance reinstall``
1486 - Added ability to change node's secondary IP address using ``gnt-node
1488 - Implemented privilege separation for all daemons except
1489 ``ganeti-noded``, see ``configure`` options
1490 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1492 - Implemented job priorities (see ``ganeti(7)`` manpage)
1493 - Ignore failures while shutting down instances during failover from
1495 - Exit daemon's bootstrap process only once daemon is ready
1496 - Export more information via ``LUInstanceQuery``/remote API
1497 - Improved documentation, QA and unittests
1498 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1499 restart if the file was created or changed
1500 - Added LUXI protocol version sent with each request and response,
1501 allowing detection of server/client mismatches
1502 - Moved the Python scripts among gnt-* and ganeti-* into modules
1503 - Moved all code related to setting up SSH to an external script,
1505 - Infrastructure changes for node group support in future versions
1511 *(Released Fri, 19 Nov 2010)*
1513 A few small bugs fixed, and some improvements to the build system:
1515 - Fix documentation regarding conversion to drbd
1516 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1518 - Fix error handling in node modify with multiple changes
1519 - Allow remote imports without checked names
1525 *(Released Tue, 19 Oct 2010)*
1527 - Disable SSL session ID cache in RPC client
1533 *(Released Thu, 14 Oct 2010)*
1535 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1536 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1538 - Fix problems with interaction between Curl and Python's HTTP server,
1539 resulting in increased speed in many RPC calls
1540 - Improve our release script to prevent breakage with older aclocal and
1547 *(Released Thu, 7 Oct 2010)*
1549 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1550 value retrieved from master
1551 - Added support for blacklisted or hidden OS definitions
1552 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1553 - Added support for -mem-path in KVM hypervisor abstraction layer
1554 - Allow overriding instance parameters in tool for inter-cluster
1555 instance moves (``tools/move-instance``)
1556 - Improved opcode summaries (e.g. in ``gnt-job list``)
1557 - Improve consistency of OS listing by sorting it
1558 - Documentation updates
1564 *(Released Fri, 8 Oct 2010)*
1566 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1572 *(Released Mon, 4 Oct 2010)*
1574 - Fixed regression in ``gnt-instance rename``
1580 *(Released Wed, 22 Sep 2010)*
1582 - Fixed OS_VARIANT variable for OS scripts
1583 - Fixed cluster tag operations via RAPI
1584 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1585 - Disabled RAPI CA checks in watcher
1591 *(Released Mon, 23 Aug 2010)*
1593 - Support DRBD versions of the format "a.b.c.d"
1595 - Re-introduce support for usage from multiple threads in RAPI client
1596 - Instance renames and modify via RAPI
1597 - Work around race condition between processing and archival in job
1599 - Mark opcodes following failed one as failed, too
1600 - Job field ``lock_status`` was removed due to difficulties making it
1601 work with the changed job queue in Ganeti 2.2; a better way to monitor
1602 locks is expected for a later 2.2.x release
1603 - Fixed dry-run behaviour with many commands
1604 - Support ``ssh-agent`` again when adding nodes
1605 - Many additional bugfixes
1611 *(Released Fri, 30 Jul 2010)*
1613 Important change: the internal RPC mechanism between Ganeti nodes has
1614 changed from using a home-grown http library (based on the Python base
1615 libraries) to use the PycURL library. This requires that PycURL is
1616 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1617 against GnuTLS by default. cURL's support for GnuTLS had known issues
1618 before cURL 7.21.0 and we recommend using the latest cURL release or
1619 linking against OpenSSL. Most other distributions already link PycURL
1620 and cURL against OpenSSL. The command::
1622 python -c 'import pycurl; print pycurl.version'
1624 can be used to determine the libraries PycURL and cURL are linked
1627 Other significant changes:
1629 - Rewrote much of the internals of the job queue, in order to achieve
1630 better parallelism; this decouples job query operations from the job
1631 processing, and it should allow much nicer behaviour of the master
1632 daemon under load, and it also has uncovered some long-standing bugs
1633 related to the job serialisation (now fixed)
1634 - Added a default iallocator setting to the cluster parameters,
1635 eliminating the need to always pass nodes or an iallocator for
1636 operations that require selection of new node(s)
1637 - Added experimental support for the LXC virtualization method
1638 - Added support for OS parameters, which allows the installation of
1639 instances to pass parameter to OS scripts in order to customise the
1641 - Added a hypervisor parameter controlling the migration type (live or
1642 non-live), since hypervisors have various levels of reliability; this
1643 has renamed the 'live' parameter to 'mode'
1644 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1645 logical volumes, meaning that cluster verify will ignore them and not
1646 flag their presence as errors
1647 - The watcher will now reset the error count for failed instances after
1648 8 hours, thus allowing self-healing if the problem that caused the
1649 instances to be down/fail to start has cleared in the meantime
1650 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1651 check for, and warn, if the drbd module parameter ``usermode_helper``
1652 is not consistent with the cluster-wide setting; this is needed to
1653 make diagnose easier of failed drbd creations
1654 - Started adding base IPv6 support, but this is not yet
1655 enabled/available for use
1656 - Rename operations (cluster, instance) will now return the new name,
1657 which is especially useful if a short name was passed in
1658 - Added support for instance migration in RAPI
1659 - Added a tool to pre-configure nodes for the SSH setup, before joining
1660 them to the cluster; this will allow in the future a simplified model
1661 for node joining (but not yet fully enabled in 2.2); this needs the
1662 paramiko python library
1663 - Fixed handling of name-resolving errors
1664 - Fixed consistency of job results on the error path
1665 - Fixed master-failover race condition when executed multiple times in
1667 - Fixed many bugs related to the job queue (mostly introduced during the
1668 2.2 development cycle, so not all are impacting 2.1)
1669 - Fixed instance migration with missing disk symlinks
1670 - Fixed handling of unknown jobs in ``gnt-job archive``
1671 - And many other small fixes/improvements
1675 - Enhanced both the unittest and the QA coverage
1676 - Switched the opcode validation to a generic model, and extended the
1677 validation to all opcode parameters
1678 - Changed more parts of the code that write shell scripts to use the
1680 - Switched the master daemon to use the asyncore library for the Luxi
1687 *(Released Thu, 17 Jun 2010)*
1689 - Added tool (``move-instance``) and infrastructure to move instances
1690 between separate clusters (see :doc:`separate documentation
1691 <move-instance>` and :doc:`design document <design-2.2>`)
1692 - Added per-request RPC timeout
1693 - RAPI now requires a Content-Type header for requests with a body (e.g.
1694 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1695 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1696 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1698 - Implemented initial support for running Ganeti daemons as separate
1699 users, see configure-time flags ``--with-user-prefix`` and
1700 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1702 - Instances can be removed after export (``gnt-backup export
1703 --remove-instance``)
1704 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1705 key (instead of 1024 bit)
1706 - Added new cluster configuration file for cluster domain secret
1707 - Import/export now use SSL instead of SSH
1708 - Added support for showing estimated time when exporting an instance,
1709 see the ``ganeti-os-interface(7)`` manpage and look for
1716 *(Released Tue, 16 Nov 2010)*
1718 Some more bugfixes. Unless critical bugs occur, this will be the last
1721 - Fix case of MAC special-values
1722 - Fix mac checker regex
1723 - backend: Fix typo causing "out of range" error
1724 - Add missing --units in gnt-instance list man page
1730 *(Released Tue, 24 Aug 2010)*
1733 - Don't ignore secondary node silently on non-mirrored disk templates
1735 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1736 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1737 - Properly document the usb_mouse parameter
1738 - Fix path in ganeti-rapi(8) (issue 116)
1739 - Adjust error message when the ganeti user's .ssh directory is
1741 - Add same-node-check when changing the disk template to drbd
1747 *(Released Fri, 16 Jul 2010)*
1750 - Add an option to only select some reboot types during qa/burnin.
1751 (on some hypervisors consequent reboots are not supported)
1752 - Fix infrequent race condition in master failover. Sometimes the old
1753 master ip address would be still detected as up for a short time
1754 after it was removed, causing failover to fail.
1755 - Decrease mlockall warnings when the ctypes module is missing. On
1756 Python 2.4 we support running even if no ctypes module is installed,
1757 but we were too verbose about this issue.
1758 - Fix building on old distributions, on which man doesn't have a
1760 - Fix RAPI not to ignore the MAC address on instance creation
1761 - Implement the old instance creation format in the RAPI client.
1767 *(Released Thu, 01 Jul 2010)*
1769 A small bugfix release:
1770 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1771 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1772 a non-existing option
1773 - Fix instance create: the --force-variant option was ignored
1774 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1775 - Fix modify node storage with non-FQDN arguments
1776 - Fix RAPI client to authenticate under Python 2.6 when used
1777 for more than 5 requests needing authentication
1778 - Fix gnt-instance modify -t (storage) giving a wrong error message
1779 when converting a non-shutdown drbd instance to plain
1785 *(Released Fri, 18 Jun 2010)*
1787 A small bugfix release:
1789 - Fix live migration of KVM instances started with older Ganeti
1790 versions which had fewer hypervisor parameters
1791 - Fix gnt-instance grow-disk on down instances
1792 - Fix an error-reporting bug during instance migration
1793 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1794 silently ignoring broken ones
1795 - Fix an RPC error reporting bug affecting, for example, RAPI client
1797 - Fix bug triggered by different API version os-es on different nodes
1798 - Fix a bug in instance startup with custom hvparams: OS level
1799 parameters would fail to be applied.
1800 - Fix the RAPI client under Python 2.6 (but more work is needed to
1801 make it work completely well with OpenSSL)
1802 - Fix handling of errors when resolving names from DNS
1808 *(Released Thu, 3 Jun 2010)*
1810 A medium sized development cycle. Some new features, and some
1811 fixes/small improvements/cleanups.
1813 Significant features
1814 ~~~~~~~~~~~~~~~~~~~~
1816 The node deamon now tries to mlock itself into memory, unless the
1817 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1818 its logs, and falls back to console logging. This allows emergency
1819 features such as ``gnt-node powercycle`` to work even in the event of a
1820 broken node disk (tested offlining the disk hosting the node's
1821 filesystem and dropping its memory caches; don't try this at home)
1823 KVM: add vhost-net acceleration support. It can be tested with a new
1824 enough version of the kernel and of qemu-kvm.
1826 KVM: Add instance chrooting feature. If you use privilege dropping for
1827 your VMs you can also now force them to chroot to an empty directory,
1828 before starting the emulated guest.
1830 KVM: Add maximum migration bandwith and maximum downtime tweaking
1831 support (requires a new-enough version of qemu-kvm).
1833 Cluster verify will now warn if the master node doesn't have the master
1834 ip configured on it.
1836 Add a new (incompatible) instance creation request format to RAPI which
1837 supports all parameters (previously only a subset was supported, and it
1838 wasn't possible to extend the old format to accomodate all the new
1839 features. The old format is still supported, and a client can check for
1840 this feature, before using it, by checking for its presence in the
1841 ``features`` RAPI resource.
1843 Now with ancient latin support. Try it passing the ``--roman`` option to
1844 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1845 (requires the python-roman module to be installed, in order to work).
1850 As usual many internal code refactorings, documentation updates, and
1853 - Lots of improvements and cleanups to the experimental Remote API
1854 (RAPI) client library.
1855 - A new unit test suite for the core daemon libraries.
1856 - A fix to creating missing directories makes sure the umask is not
1857 applied anymore. This enforces the same directory permissions
1859 - Better handling terminating daemons with ctrl+c (used when running
1860 them in debugging mode).
1861 - Fix a race condition in live migrating a KVM instance, when stat()
1862 on the old proc status file returned EINVAL, which is an unexpected
1864 - Fixed manpage checking with newer man and utf-8 charachters. But now
1865 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1871 *(Released Fri, 7 May 2010)*
1873 Fix a bug which prevented untagged KVM instances from starting.
1879 *(Released Fri, 7 May 2010)*
1881 Another release with a long development cycle, during which many
1882 different features were added.
1884 Significant features
1885 ~~~~~~~~~~~~~~~~~~~~
1887 The KVM hypervisor now can run the individual instances as non-root, to
1888 reduce the impact of a VM being hijacked due to bugs in the
1889 hypervisor. It is possible to run all instances as a single (non-root)
1890 user, to manually specify a user for each instance, or to dynamically
1891 allocate a user out of a cluster-wide pool to each instance, with the
1892 guarantee that no two instances will run under the same user ID on any
1895 An experimental RAPI client library, that can be used standalone
1896 (without the other Ganeti libraries), is provided in the source tree as
1897 ``lib/rapi/client.py``. Note this client might change its interface in
1898 the future, as we iterate on its capabilities.
1900 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1901 replace the cluster's certificates and crypto keys. This might help in
1902 case they have been compromised, or have simply expired.
1904 A new disk option for instance creation has been added that allows one
1905 to "adopt" currently existing logical volumes, with data
1906 preservation. This should allow easier migration to Ganeti from
1907 unmanaged (or managed via other software) instances.
1909 Another disk improvement is the possibility to convert between redundant
1910 (DRBD) and plain (LVM) disk configuration for an instance. This should
1911 allow better scalability (starting with one node and growing the
1912 cluster, or shrinking a two-node cluster to one node).
1914 A new feature that could help with automated node failovers has been
1915 implemented: if a node sees itself as offline (by querying the master
1916 candidates), it will try to shutdown (hard) all instances and any active
1917 DRBD devices. This reduces the risk of duplicate instances if an
1918 external script automatically failovers the instances on such nodes. To
1919 enable this, the cluster parameter ``maintain_node_health`` should be
1920 enabled; in the future this option (per the name) will enable other
1921 automatic maintenance features.
1923 Instance export/import now will reuse the original instance
1924 specifications for all parameters; that means exporting an instance,
1925 deleting it and the importing it back should give an almost identical
1926 instance. Note that the default import behaviour has changed from
1927 before, where it created only one NIC; now it recreates the original
1930 Cluster verify has added a few new checks: SSL certificates validity,
1931 /etc/hosts consistency across the cluster, etc.
1936 As usual, many internal changes were done, documentation fixes,
1939 - Fixed cluster initialization with disabled cluster storage (regression
1940 introduced in 2.1.1)
1941 - File-based storage supports growing the disks
1942 - Fixed behaviour of node role changes
1943 - Fixed cluster verify for some corner cases, plus a general rewrite of
1944 cluster verify to allow future extension with more checks
1945 - Fixed log spamming by watcher and node daemon (regression introduced
1947 - Fixed possible validation issues when changing the list of enabled
1949 - Fixed cleanup of /etc/hosts during node removal
1950 - Fixed RAPI response for invalid methods
1951 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1952 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1953 from ide disks, etc.)
1954 - Allow OS changes without re-installation (to record a changed OS
1955 outside of Ganeti, or to allow OS renames)
1956 - Allow instance creation without OS installation (useful for example if
1957 the OS will be installed manually, or restored from a backup not in
1959 - Implemented option to make cluster ``copyfile`` use the replication
1961 - Added list of enabled hypervisors to ssconf (possibly useful for
1963 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1965 - A partial form of node re-IP is possible via node readd, which now
1966 allows changed node primary IP
1967 - Command line utilities now show an informational message if the job is
1969 - The logs of the master daemon now show the PID/UID/GID of the
1976 *(Released Fri, 12 Mar 2010)*
1978 During the 2.1.0 long release candidate cycle, a lot of improvements and
1979 changes have accumulated with were released later as 2.1.1.
1984 The node evacuate command (``gnt-node evacuate``) was significantly
1985 rewritten, and as such the IAllocator protocol was changed - a new
1986 request type has been added. This unfortunate change during a stable
1987 series is designed to improve performance of node evacuations; on
1988 clusters with more than about five nodes and which are well-balanced,
1989 evacuation should proceed in parallel for all instances of the node
1990 being evacuated. As such, any existing IAllocator scripts need to be
1991 updated, otherwise the above command will fail due to the unknown
1992 request. The provided "dumb" allocator has not been updated; but the
1993 ganeti-htools package supports the new protocol since version 0.2.4.
1995 Another important change is increased validation of node and instance
1996 names. This might create problems in special cases, if invalid host
1997 names are being used.
1999 Also, a new layer of hypervisor parameters has been added, that sits at
2000 OS level between the cluster defaults and the instance ones. This allows
2001 customisation of virtualization parameters depending on the installed
2002 OS. For example instances with OS 'X' may have a different KVM kernel
2003 (or any other parameter) than the cluster defaults. This is intended to
2004 help managing a multiple OSes on the same cluster, without manual
2005 modification of each instance's parameters.
2007 A tool for merging clusters, ``cluster-merge``, has been added in the
2008 tools sub-directory.
2013 - Improved the int/float conversions that should make the code more
2014 robust in face of errors from the node daemons
2015 - Fixed the remove node code in case of internal configuration errors
2016 - Fixed the node daemon behaviour in face of inconsistent queue
2017 directory (e.g. read-only file-system where we can't open the files
2019 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2020 now it either aborts cleanly or, if given the new "auto_promote"
2021 parameter, will automatically promote other nodes as needed
2022 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2023 completely prevent Ganeti from working
2024 - Fixed bug for instance export when not all disks were successfully
2026 - Fixed behaviour of node add when the new node is slow in starting up
2028 - Fixed handling of signals in the LUXI client, which should improve
2029 behaviour of command-line scripts
2030 - Added checks for invalid node/instance names in the configuration (now
2031 flagged during cluster verify)
2032 - Fixed watcher behaviour for disk activation errors
2033 - Fixed two potentially endless loops in http library, which led to the
2034 RAPI daemon hanging and consuming 100% CPU in some cases
2035 - Fixed bug in RAPI daemon related to hashed passwords
2036 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2038 - Enhanced compatibility with non-Debian OSes, but not using absolute
2039 path in some commands and allowing customisation of the ssh
2040 configuration directory
2041 - Fixed possible future issue with new Python versions by abiding to the
2042 proper use of ``__slots__`` attribute on classes
2043 - Added checks that should prevent directory traversal attacks
2044 - Many documentation fixes based on feedback from users
2049 - Added an "early_release" more for instance replace disks and node
2050 evacuate, where we release locks earlier and thus allow higher
2051 parallelism within the cluster
2052 - Added watcher hooks, intended to allow the watcher to restart other
2053 daemons (e.g. from the ganeti-nbma project), but they can be used of
2054 course for any other purpose
2055 - Added a compile-time disable for DRBD barriers, to increase
2056 performance if the administrator trusts the power supply or the
2057 storage system to not lose writes
2058 - Added the option of using syslog for logging instead of, or in
2059 addition to, Ganeti's own log files
2060 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2061 can indeed boot from a paravirtual NIC
2062 - Added a generic debug level for many operations; while this is not
2063 used widely yet, it allows one to pass the debug value all the way to
2065 - Enhanced the hooks environment for instance moves (failovers,
2066 migrations) where the primary/secondary nodes changed during the
2067 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2068 - Enhanced data validations for many user-supplied values; one important
2069 item is the restrictions imposed on instance and node names, which
2070 might reject some (invalid) host names
2071 - Add a configure-time option to disable file-based storage, if it's not
2072 needed; this allows greater security separation between the master
2073 node and the other nodes from the point of view of the inter-node RPC
2075 - Added user notification in interactive tools if job is waiting in the
2076 job queue or trying to acquire locks
2077 - Added log messages when a job is waiting for locks
2078 - Added filtering by node tags in instance operations which admit
2079 multiple instances (start, stop, reboot, reinstall)
2080 - Added a new tool for cluster mergers, ``cluster-merge``
2081 - Parameters from command line which are of the form ``a=b,c=d`` can now
2082 use backslash escapes to pass in values which contain commas,
2083 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2085 - For KVM, the instance name is the first parameter passed to KVM, so
2086 that it's more visible in the process list
2092 *(Released Tue, 2 Mar 2010)*
2094 Ganeti 2.1 brings many improvements with it. Major changes:
2096 - Added infrastructure to ease automated disk repairs
2097 - Added new daemon to export configuration data in a cheaper way than
2098 using the remote API
2099 - Instance NICs can now be routed instead of being associated with a
2101 - Improved job locking logic to reduce impact of jobs acquiring multiple
2102 locks waiting for other long-running jobs
2104 In-depth implementation details can be found in the Ganeti 2.1 design
2110 - Added chroot hypervisor
2111 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2113 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2114 ``bootloader_path`` and ``bootloader_args``)
2115 - Added the ``use_localtime`` option for the xen-hvm and kvm
2116 hypervisors, and the default value for this has changed to false (in
2117 2.0 xen-hvm always enabled it)
2118 - Added luxi call to submit multiple jobs in one go
2119 - Added cluster initialization option to not modify ``/etc/hosts``
2121 - Added network interface parameters
2122 - Added dry run mode to some LUs
2123 - Added RAPI resources:
2125 - ``/2/instances/[instance_name]/info``
2126 - ``/2/instances/[instance_name]/replace-disks``
2127 - ``/2/nodes/[node_name]/evacuate``
2128 - ``/2/nodes/[node_name]/migrate``
2129 - ``/2/nodes/[node_name]/role``
2130 - ``/2/nodes/[node_name]/storage``
2131 - ``/2/nodes/[node_name]/storage/modify``
2132 - ``/2/nodes/[node_name]/storage/repair``
2134 - Added OpCodes to evacuate or migrate all instances on a node
2135 - Added new command to list storage elements on nodes (``gnt-node
2136 list-storage``) and modify them (``gnt-node modify-storage``)
2137 - Added new ssconf files with master candidate IP address
2138 (``ssconf_master_candidates_ips``), node primary IP address
2139 (``ssconf_node_primary_ips``) and node secondary IP address
2140 (``ssconf_node_secondary_ips``)
2141 - Added ``ganeti-confd`` and a client library to query the Ganeti
2142 configuration via UDP
2143 - Added ability to run hooks after cluster initialization and before
2145 - Added automatic mode for disk replace (``gnt-instance replace-disks
2147 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2148 after catastrophic data-loss
2149 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2151 - Added ``gnt-instance move`` command to move instances
2152 - Added ``gnt-cluster watcher`` command to control watcher
2153 - Added ``gnt-node powercycle`` command to powercycle nodes
2154 - Added new job status field ``lock_status``
2155 - Added parseable error codes to cluster verification (``gnt-cluster
2156 verify --error-codes``) and made output less verbose (use
2157 ``--verbose`` to restore previous behaviour)
2158 - Added UUIDs to the main config entities (cluster, nodes, instances)
2159 - Added support for OS variants
2160 - Added support for hashed passwords in the Ganeti remote API users file
2162 - Added option to specify maximum timeout on instance shutdown
2163 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2164 - Added new helper script to start and stop Ganeti daemons
2165 (``daemon-util``), with the intent to reduce the work necessary to
2166 adjust Ganeti for non-Debian distributions and to start/stop daemons
2168 - Added more unittests
2169 - Fixed critical bug in ganeti-masterd startup
2170 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2171 now customisable at the cluster level for both the KVM and Xen
2172 hypervisors using the new ``migration_port`` parameter
2173 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2174 reinstalling an instance
2175 - Allowed ``@`` in tag names
2176 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2177 - Many documentation updates
2178 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2179 - ``gnt-instance reinstall`` can now reinstall multiple instances
2180 - Updated many command line parameters
2181 - Introduced new OS API version 15
2182 - No longer support a default hypervisor
2183 - Treat virtual LVs as inexistent
2184 - Improved job locking logic to reduce lock contention
2185 - Match instance and node names case insensitively
2186 - Reimplemented bash completion script to be more complete
2193 *(Released Thu, 4 Feb 2010)*
2195 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2196 - Fix a string formatting bug
2197 - Improve safety of the code in some error paths
2198 - Improve data validation in the master of values returned from nodes
2204 *(Released Thu, 17 Dec 2009)*
2206 - Fix security issue due to missing validation of iallocator names; this
2207 allows local and remote execution of arbitrary executables
2208 - Fix failure of gnt-node list during instance removal
2209 - Ship the RAPI documentation in the archive
2215 *(Released Wed, 30 Sep 2009)*
2217 - Fixed many wrong messages
2218 - Fixed a few bugs related to the locking library
2219 - Fixed MAC checking at instance creation time
2220 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2221 - Fixed a few issues related to signal handling in both daemons and
2223 - Fixed the example startup script provided
2224 - Fixed insserv dependencies in the example startup script (patch from
2226 - Fixed handling of drained nodes in the iallocator framework
2227 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2229 - Fixed error related to invalid job IDs in job polling
2230 - Fixed job/opcode persistence on unclean master shutdown
2231 - Fixed handling of partial job processing after unclean master
2233 - Fixed error reporting from LUs, previously all errors were converted
2234 into execution errors
2235 - Fixed error reporting from burnin
2236 - Decreased significantly the memory usage of the job queue
2237 - Optimised slightly multi-job submission
2238 - Optimised slightly opcode loading
2239 - Backported the multi-job submit framework from the development
2240 branch; multi-instance start and stop should be faster
2241 - Added script to clean archived jobs after 21 days; this will reduce
2242 the size of the queue directory
2243 - Added some extra checks in disk size tracking
2244 - Added an example ethers hook script
2245 - Added a cluster parameter that prevents Ganeti from modifying of
2247 - Added more node information to RAPI responses
2248 - Added a ``gnt-job watch`` command that allows following the ouput of a
2250 - Added a bind-address option to ganeti-rapi
2251 - Added more checks to the configuration verify
2252 - Enhanced the burnin script such that some operations can be retried
2254 - Converted instance reinstall to multi-instance model
2260 *(Released Fri, 7 Aug 2009)*
2262 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2263 to allow using the pre-2.0.2 behaviour in activation, if any existing
2264 instances have mismatched disk sizes in the configuration
2265 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2266 any configuration mismatches for disk sizes
2267 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2268 failover to work on two-node clusters
2269 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2271 - Fixed detection of OS script errors in ``gnt-backup export``
2272 - Fixed exit code of ``gnt-backup export``
2278 *(Released Fri, 17 Jul 2009)*
2280 - Added experimental support for stripped logical volumes; this should
2281 enhance performance but comes with a higher complexity in the block
2282 device handling; stripping is only enabled when passing
2283 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2284 affected even in the non-stripped mode
2285 - Improved resiliency against transient failures at the end of DRBD
2286 resyncs, and in general of DRBD resync checks
2287 - Fixed a couple of issues with exports and snapshot errors
2288 - Fixed a couple of issues in instance listing
2289 - Added display of the disk size in ``gnt-instance info``
2290 - Fixed checking for valid OSes in instance creation
2291 - Fixed handling of the "vcpus" parameter in instance listing and in
2292 general of invalid parameters
2293 - Fixed http server library, and thus RAPI, to handle invalid
2294 username/password combinations correctly; this means that now they
2295 report unauthorized for queries too, not only for modifications,
2296 allowing earlier detect of configuration problems
2297 - Added a new "role" node list field, equivalent to the master/master
2298 candidate/drained/offline flags combinations
2299 - Fixed cluster modify and changes of candidate pool size
2300 - Fixed cluster verify error messages for wrong files on regular nodes
2301 - Fixed a couple of issues with node demotion from master candidate role
2302 - Fixed node readd issues
2303 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2304 - Added a new ``--no-voting`` option for masterfailover to fix failover
2305 on two-nodes clusters when the former master node is unreachable
2306 - Added instance reinstall over RAPI
2312 *(Released Tue, 16 Jun 2009)*
2314 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2315 allow re-adding the start in single-user option (regression from 1.2)
2316 - the watcher writes the instance status to a file, to allow monitoring
2317 to report the instance status (from the master) based on cached
2318 results of the watcher's queries; while this can get stale if the
2319 watcher is being locked due to other work on the cluster, this is
2320 still an improvement
2321 - the watcher now also restarts the node daemon and the rapi daemon if
2323 - fixed the watcher to handle full and drained queue cases
2324 - hooks export more instance data in the environment, which helps if
2325 hook scripts need to take action based on the instance's properties
2326 (no longer need to query back into ganeti)
2327 - instance failovers when the instance is stopped do not check for free
2328 RAM, so that failing over a stopped instance is possible in low memory
2330 - rapi uses queries for tags instead of jobs (for less job traffic), and
2331 for cluster tags it won't talk to masterd at all but read them from
2333 - a couple of error handling fixes in RAPI
2334 - drbd handling: improved the error handling of inconsistent disks after
2335 resync to reduce the frequency of "there are some degraded disks for
2336 this instance" messages
2337 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2338 error handling path called a wrong function name)
2344 *(Released Wed, 27 May 2009)*
2346 - no changes from rc5
2352 *(Released Wed, 20 May 2009)*
2354 - fix a couple of bugs (validation, argument checks)
2355 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2356 - some small improvements to RAPI and IAllocator
2357 - make watcher automatically start the master daemon if down
2363 *(Released Mon, 27 Apr 2009)*
2365 - change the OS list to not require locks; this helps with big clusters
2366 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2367 volume group is broken
2368 - ``gnt-instance info``, without any arguments, doesn't run for all
2369 instances anymore; either pass ``--all`` or pass the desired
2370 instances; this helps against mistakes on big clusters where listing
2371 the information for all instances takes a long time
2372 - miscellaneous doc and man pages fixes
2378 *(Released Wed, 8 Apr 2009)*
2380 - Change the internal locking model of some ``gnt-node`` commands, in
2381 order to reduce contention (and blocking of master daemon) when
2382 batching many creation/reinstall jobs
2383 - Fixes to Xen soft reboot
2384 - No longer build documentation at build time, instead distribute it in
2385 the archive, in order to reduce the need for the whole docbook/rst
2392 *(Released Fri, 27 Mar 2009)*
2394 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2395 - Fix watcher startup sequence, improves the behaviour of busy clusters
2396 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2397 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2398 bind address and other places
2399 - Some documentation fixes and updates
2405 *(Released Mon, 2 Mar 2009)*
2407 - More documentation updates, now all docs should be more-or-less
2409 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2411 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2413 - Fix an issue related to $libdir/run/ganeti and cluster creation
2419 *(Released Thu, 19 Feb 2009)*
2421 - Xen PVM and KVM have switched the default value for the instance root
2422 disk to the first partition on the first drive, instead of the whole
2423 drive; this means that the OS installation scripts must be changed
2425 - Man pages have been updated
2426 - RAPI has been switched by default to HTTPS, and the exported functions
2427 should all work correctly
2428 - RAPI v1 has been removed
2429 - Many improvements to the KVM hypervisor
2430 - Block device errors are now better reported
2431 - Many other bugfixes and small improvements
2437 *(Released Mon, 26 Jan 2009)*
2439 - Version 2 is a general rewrite of the code and therefore the
2440 differences are too many to list, see the design document for 2.0 in
2441 the ``doc/`` subdirectory for more details
2442 - In this beta version there is not yet a migration path from 1.2 (there
2443 will be one in the final 2.0 release)
2444 - A few significant changes are:
2446 - all commands are executed by a daemon (``ganeti-masterd``) and the
2447 various ``gnt-*`` commands are just front-ends to it
2448 - all the commands are entered into, and executed from a job queue,
2449 see the ``gnt-job(8)`` manpage
2450 - the RAPI daemon supports read-write operations, secured by basic
2451 HTTP authentication on top of HTTPS
2452 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2453 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2454 to migrate to DRBD 8 first while still running Ganeti 1.2)
2455 - DRBD devices are using statically allocated minor numbers, which
2456 will be assigned to existing instances during the migration process
2457 - there is support for both Xen PVM and Xen HVM instances running on
2459 - KVM virtualization is supported too
2460 - file-based storage has been implemented, which means that it is
2461 possible to run the cluster without LVM and DRBD storage, for
2462 example using a shared filesystem exported from shared storage (and
2463 still have live migration)
2469 *(Released Tue, 13 Jan 2009)*
2471 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2472 - Reuse the old instance mac address by default on instance import, if
2473 the instance name is the same.
2474 - Handle situations in which the node info rpc returns incomplete
2476 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2477 - Improved version of batcher:
2479 - state file support
2480 - instance mac address support
2481 - support for HVM clusters/instances
2483 - Add an option to show the number of cpu sockets and nodes in
2485 - Support OSes that handle more than one version of the OS api (but do
2486 not change the current API in any other way)
2487 - Fix ``gnt-node migrate``
2488 - ``gnt-debug`` man page
2489 - Fixes various more typos and small issues
2490 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2496 *(Released Wed, 24 Sep 2008)*
2498 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2499 type of disk exported to fully virtualized instances.
2500 - provide access to the serial console of HVM instances
2501 - instance auto_balance flag, set by default. If turned off it will
2502 avoid warnings on cluster verify if there is not enough memory to fail
2503 over an instance. in the future it will prevent automatically failing
2504 it over when we will support that.
2505 - batcher tool for instance creation, see ``tools/README.batcher``
2506 - ``gnt-instance reinstall --select-os`` to interactively select a new
2507 operating system when reinstalling an instance.
2508 - when changing the memory amount on instance modify a check has been
2509 added that the instance will be able to start. also warnings are
2510 emitted if the instance will not be able to fail over, if auto_balance
2512 - documentation fixes
2513 - sync fields between ``gnt-instance list/modify/add/import``
2514 - fix a race condition in drbd when the sync speed was set after giving
2515 the device a remote peer.
2521 *(Released Tue, 22 Jul 2008)*
2523 - note: the allowed size and number of tags per object were reduced
2524 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2525 - fixed twisted 8.x compatibility
2526 - fixed ``gnt-instance replace-disks`` with iallocator
2527 - add TCP keepalives on twisted connections to detect restarted nodes
2528 - disk increase support, see ``gnt-instance grow-disk``
2529 - implement bulk node/instance query for RAPI
2530 - add tags in node/instance listing (optional)
2531 - experimental migration (and live migration) support, read the man page
2532 for ``gnt-instance migrate``
2533 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2534 has some small improvements in handling its state file
2540 *(Released Fri, 13 Jun 2008)*
2542 - Experimental readonly, REST-based remote API implementation;
2543 automatically started on master node, TCP port 5080, if enabled by
2544 ``--enable-rapi`` parameter to configure script.
2545 - Instance allocator support. Add and import instance accept a
2546 ``--iallocator`` parameter, and call that instance allocator to decide
2547 which node to use for the instance. The iallocator document describes
2548 what's expected from an allocator script.
2549 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2550 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2551 node is lost there is still enough memory to fail over the instances
2553 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2554 ``gnt-cluster verify``, to check for site-specific compliance. All the
2555 hooks will run, and their output, if any, will be displayed. Any
2556 failing hook will make the verification return an error value.
2557 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2558 primary and secondary interfaces
2559 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2560 that is still declared as part of the cluster and has failed.
2561 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2562 specifying output fields, that just adds the chosen fields to the
2564 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2565 export from the filesystem.
2566 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2567 have been added. Using them you can enable/disable acpi and pae
2568 support, and specify a path for a cd image to be exported to the
2569 instance. These parameters as the name suggest only work on HVM
2571 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2572 PAE support will be set to the previously hardcoded values, but the
2573 (previously hardcoded) path to the CDROM ISO image will be unset and
2574 if required, needs to be set manually with ``gnt-instance modify``
2576 - The address to which an instance's VNC console is bound is now
2577 selectable per-instance, rather than being cluster wide. Of course
2578 this only applies to instances controlled via VNC, so currently just
2579 applies to HVM clusters.
2585 *(Released Mon, 18 Feb 2008)*
2587 - more tweaks to the disk activation code (especially helpful for DRBD)
2588 - change the default ``gnt-instance list`` output format, now there is
2589 one combined status field (see the manpage for the exact values this
2591 - some more fixes for the mac export to hooks change
2592 - make Ganeti not break with DRBD 8.2.x (which changed the version
2593 format in ``/proc/drbd``) (issue 24)
2594 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2595 template, allowing migration from DRBD0.7+MD to DRBD8
2601 *(Released Wed, 30 Jan 2008)*
2603 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2605 - add command aliases infrastructure and a few aliases
2606 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2607 man pages and the ``--help`` option of ``gnt-node
2608 list``/``gnt-instance list``
2609 - fix ``gnt-backup list`` with down nodes (issue 21)
2610 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2611 - fix the dist archive and add a check for including svn/git files in
2613 - some developer-related changes: improve the burnin and the QA suite,
2614 add an upload script for testing during development
2620 *(Released Wed, 16 Jan 2008)*
2622 - experimental HVM support, read the install document, section
2623 "Initializing the cluster"
2624 - allow for the PVM hypervisor per-instance kernel and initrd paths
2625 - add a new command ``gnt-cluster verify-disks`` which uses a new
2626 algorithm to improve the reconnection of the DRBD pairs if the device
2627 on the secondary node has gone away
2628 - make logical volume code auto-activate LVs at disk activation time
2629 - slightly improve the speed of activating disks
2630 - allow specification of the MAC address at instance creation time, and
2631 changing it later via ``gnt-instance modify``
2632 - fix handling of external commands that generate lots of output on
2634 - update documentation with regard to minimum version of DRBD8 supported
2640 *(Released Tue, 4 Dec 2007)*
2642 - Log the ``xm create`` output to the node daemon log on failure (to
2643 help diagnosing the error)
2644 - In debug mode, log all external commands output if failed to the logs
2645 - Change parsing of lvm commands to ignore stderr
2651 *(Released Wed, 28 Nov 2007)*
2653 - Another round of updates to the DRBD 8 code to deal with more failures
2654 in the replace secondary node operation
2655 - Some more logging of failures in disk operations (lvm, drbd)
2656 - A few documentation updates
2663 *(Released Tue, 13 Nov 2007)*
2665 - Change configuration file format from Python's Pickle to JSON.
2666 Upgrading is possible using the cfgupgrade utility.
2667 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2668 faster replace disks and is more stable (DRBD 8 has many improvements
2669 compared to DRBD 0.7)
2670 - Added command line tags support (see man pages for ``gnt-instance``,
2671 ``gnt-node``, ``gnt-cluster``)
2672 - Added instance rename support
2673 - Added multi-instance startup/shutdown
2674 - Added cluster rename support
2675 - Added ``gnt-node evacuate`` to simplify some node operations
2676 - Added instance reboot operation that can speedup reboot as compared to
2678 - Soften the requirement that hostnames are in FQDN format
2679 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2681 - Removed dependency on debian's patched fping that uses the
2682 non-standard ``-S`` option
2683 - Now the OS definitions are searched for in multiple, configurable
2684 paths (easier for distros to package)
2685 - Some changes to the hooks infrastructure (especially the new
2686 post-configuration update hook)
2687 - Other small bugfixes
2689 .. vim: set textwidth=72 syntax=rst :