10 - The minimum Python version needed to run Ganeti is now 2.6.
11 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
12 to always require authentication, ``--require-authentication``. It can
13 be specified in ``$sysconfdir/default/ganeti``.
14 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
15 be used to manage the disk templates to be used by instances in the cluster.
16 Initially, it will be set to a list that includes plain, drbd, if they were
17 enabled by specifying a volume group name, and file and sharedfile, if those
18 were enabled at configure time. Additionally, it will include all disk
19 templates that are currently used by instances. The order of disk templates
20 will be based on Ganeti's history of supporting them. In the future, the
21 first entry of the list will be used as a default disk template on instance
23 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
24 configuration back to the previous stable version.
32 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
38 *(Released Tue, 2 Apr 2013)*
40 - Networks no longer have a "type" slot, since this information was
41 unused in Ganeti: instead of it tags should be used.
42 - Diskless instances are now externally mirrored (Issue 237). This for
43 now has only been tested in conjunction with explicit target nodes for
45 - The rapi client now has a ``target_node`` option to MigrateInstance.
46 - Fix early exit return code for hbal (Issue 386).
47 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
48 - Fix ``rbd showmapped`` output parsing (Issue 312).
49 - Networks are now referenced indexed by UUID, rather than name. This
50 will require running cfgupgrade, from 2.7.0beta1, if networks are in
52 - The OS environment now includes network information.
53 - Deleting of a network is now disallowed if any instance nic is using
54 it, to prevent dangling references.
55 - External storage is now documented in man pages.
56 - The exclusive_storage flag can now only be set at nodegroup level.
57 - Hbal can now submit an explicit priority with its jobs.
58 - Many network related locking fixes.
59 - Bump up the required pylint version to 0.25.1.
60 - Fix the ``no_remember`` option in RAPI client.
61 - Many ipolicy related tests, qa, and fixes.
62 - Many documentation improvements and fixes.
63 - Fix building with ``--disable-file-storage``.
64 - Fix ``-q`` option in htools, which was broken if passed more than
66 - Some haskell/python interaction improvements and fixes.
67 - Fix iallocator in case of missing LVM storage.
68 - Fix confd config load in case of ``--no-lvm-storage``.
69 - The confd/query functionality is now mentioned in the security
76 *(Released Wed, 6 Feb 2013)*
78 - ``gnt-instance batch-create`` has been changed to use the bulk create
79 opcode from Ganeti. This lead to incompatible changes in the format of
80 the JSON file. It's now not a custom dict anymore but a dict
81 compatible with the ``OpInstanceCreate`` opcode.
82 - Parent directories for file storage need to be listed in
83 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
84 write the file automatically based on old configuration values, but it
85 can not distribute it across all nodes and the file contents should be
86 verified. Use ``gnt-cluster copyfile
87 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
88 upgraded. The reason for requiring this list of paths now is that
89 before it would have been possible to inject new paths via RPC,
90 allowing files to be created in arbitrary locations. The RPC protocol
91 is protected using SSL/X.509 certificates, but as a design principle
92 Ganeti does not permit arbitrary paths to be passed.
93 - The parsing of the variants file for OSes (see
94 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
95 lines and comment lines (starting with ``#``) are ignored for better
97 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
98 longer available. ``gnt-node add`` now invokes a new tool on the
99 destination node, named ``prepare-node-join``, to configure the SSH
100 daemon. Paramiko is no longer necessary to configure nodes' SSH
101 daemons via ``gnt-node add``.
102 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
103 for RAPI users. It allows granting permissions to query for
104 information to a specific user without giving
105 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
106 - A new tool named ``node-cleanup`` has been added. It cleans remains of
107 a cluster from a machine by stopping all daemons, removing
108 certificates and ssconf files. Unless the ``--no-backup`` option is
109 given, copies of the certificates are made.
110 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
111 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
112 the flag is not reset after a master failover.
113 - Instance creations now support the use of opportunistic locking,
114 potentially speeding up the (parallel) creation of multiple instances.
115 This feature is currently only available via the :doc:`RAPI
116 <rapi>` interface and when an instance allocator is used. If the
117 ``opportunistic_locking`` parameter is set the opcode will try to
118 acquire as many locks as possible, but will not wait for any locks
119 held by other opcodes. If not enough resources can be found to
120 allocate the instance, the temporary error code
121 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
122 retried thereafter, with or without opportunistic locking.
123 - The functionality for allocating multiple instances at once has been
124 overhauled and is now also available through :doc:`RAPI <rapi>`.
125 - Man pages can now be included when the documentation is built, in
126 which case the output is in ``doc/man-html``. The configure-time
127 option is ``--enable-manpages-in-doc``. Sphinx 1.0 or higher is
129 - A new htool called ``hroller`` helps scheduling parallel shutdown of
130 nodes on clusters where all instances are shut down, but without ever
131 rebooting both primary and secondary node for an instance at the same
133 - Queries not needing locks or RPC access to the node can now be
134 performed by the confd daemon, making them independent from jobs, and
135 thus faster to execute. This is selectable at configure time.
136 - The KVM hypervisor has been updated with many new hypervisor
137 parameters, including a generic one for passing arbitrary command line
138 values. See a complete list in :manpage:`gnt-instance(8)`.
139 - A new tool, called ``mon-collector``, is the stand-alone executor of
140 the data collectors for a monitoring system. As of this version, it
141 just includes the DRBD data collector, that can be executed by calling
142 ``mon-collector`` using the ``drbd`` parameter. See
143 :manpage:`mon-collector(7)`.
144 - New ``exclusive-storage`` node parameter added. When it's set to true,
145 physical disks are assigned in an exclusive fashion to instances, as
146 documented in :doc:`Partitioned Ganeti <design-partitioned>`.
147 Currently, only instances using the ``plain`` disk template are
149 - New network management functionality to support automatic allocation
150 of IP addresses and managing of network parameters. See
151 :manpage:`gnt-network(8)` for more details.
152 - New external storage backend, to allow managing arbitrary storage
153 systems external to the cluster. See
154 :manpage:`ganeti-extstorage-interface(7)`.
155 - Instance renames of LVM-based instances will now update the LV tags
156 (which can be used to recover the instance-to-LV mapping in case of
158 - ``hbal`` will now exit with status 0 if, during job execution over
159 LUXI, early exit has been requested and all jobs are successful;
160 before, exit status 1 was used, which cannot be differentiated from
167 *(Released Fri, 21 Dec 2012)*
169 Important behaviour change: hbal won't rebalance anymore instances which
170 have the ``auto_balance`` attribute set to false. This was the intention
171 all along, but until now it only skipped those from the N+1 memory
172 reservation (DRBD-specific).
174 A significant number of bug fixes in this release:
176 - Fixed disk adoption interaction with ipolicy checks.
177 - Fixed networking issues when instances are started, stopped or
178 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
179 - Fixed the warning in cluster verify for shared storage instances not
181 - Fixed removal of storage directory on shared file storage (issue 262).
182 - Fixed validation of LVM volume group name in OpClusterSetParams
183 (``gnt-cluster modify``) (issue 285).
184 - Fixed runtime memory increases (``gnt-instance modify -m``).
185 - Fixed live migration under Xen's ``xl`` mode.
186 - Fixed ``gnt-instance console`` with ``xl``.
187 - Fixed building with newer Haskell compiler/libraries.
188 - Fixed PID file writing in Haskell daemons (confd); this prevents
189 restart issues if confd was launched manually (outside of
190 ``daemon-util``) while another copy of it was running
191 - Fixed a type error when doing live migrations with KVM (issue 297) and
192 the error messages for failing migrations have been improved.
193 - Fixed opcode validation for the out-of-band commands (``gnt-node
195 - Fixed a type error when unsetting OS hypervisor parameters (issue
196 311); now it's possible to unset all OS-specific hypervisor
198 - Fixed the ``dry-run`` mode for many operations: verification of
199 results was over-zealous but didn't take into account the ``dry-run``
200 operation, resulting in "wrong" failures.
201 - Fixed bash completion in ``gnt-job list`` when the job queue has
202 hundreds of entries; especially with older ``bash`` versions, this
203 results in significant CPU usage.
205 And lastly, a few other improvements have been made:
207 - Added option to force master-failover without voting (issue 282).
208 - Clarified error message on lock conflict (issue 287).
209 - Logging of newly submitted jobs has been improved (issue 290).
210 - Hostname checks have been made uniform between instance rename and
212 - The ``--submit`` option is now supported by ``gnt-debug delay``.
213 - Shutting down the master daemon by sending SIGTERM now stops it from
214 processing jobs waiting for locks; instead, those jobs will be started
215 once again after the master daemon is started the next time (issue
217 - Support for Xen's ``xl`` program has been improved (besides the fixes
219 - Reduced logging noise in the Haskell confd daemon (only show one log
220 entry for each config reload, instead of two).
221 - Several man page updates and typo fixes.
227 *(Released Fri, 12 Oct 2012)*
229 A small bugfix release. Among the bugs fixed:
231 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
232 made the command unusable.
233 - Commands that issue many jobs don't fail anymore just because some jobs
234 take so long that other jobs are archived.
235 - Failures during ``gnt-instance reinstall`` are reflected by the exit
237 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
239 - When ``always_failover`` is set, ``--allow-failover`` is not required
240 in migrate commands anymore.
241 - ``bash_completion`` works even if extglob is disabled.
242 - Fixed bug with locks that made failover for RDB-based instances fail.
243 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
244 a random node instead of one based on the allocator metric.
245 - Support for newer versions of pylint and pep8.
246 - Hail doesn't fail anymore when trying to add an instance of type
247 ``file``, ``sharedfile`` or ``rbd``.
248 - Added new Makefile target to rebuild the whole distribution, so that
249 all files are included.
255 *(Released Fri, 27 Jul 2012)*
258 .. attention:: The ``LUXI`` protocol has been made more consistent
259 regarding its handling of command arguments. This, however, leads to
260 incompatibility issues with previous versions. Please ensure that you
261 restart Ganeti daemons soon after the upgrade, otherwise most
262 ``LUXI`` calls (job submission, setting/resetting the drain flag,
263 pausing/resuming the watcher, cancelling and archiving jobs, querying
264 the cluster configuration) will fail.
273 The current ``admin_up`` field, which used to denote whether an instance
274 should be running or not, has been removed. Instead, ``admin_state`` is
275 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
277 The rational behind this is that an instance being “down” can have
280 - it could be down during a reboot
281 - it could be temporarily be down for a reinstall
282 - or it could be down because it is deprecated and kept just for its
285 The previous Boolean state was making it difficult to do capacity
286 calculations: should Ganeti reserve memory for a down instance? Now, the
287 tri-state field makes it clear:
289 - in ``up`` and ``down`` state, all resources are reserved for the
290 instance, and it can be at any time brought up if it is down
291 - in ``offline`` state, only disk space is reserved for it, but not
294 The field can have an extra use: since the transition between ``up`` and
295 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
296 transition between ``offline`` and ``down`` is done via ``gnt-instance
297 modify``, it is possible to given different rights to users. For
298 example, owners of an instance could be allowed to start/stop it, but
299 not transition it out of the offline state.
301 Instance policies and specs
302 +++++++++++++++++++++++++++
304 In previous Ganeti versions, an instance creation request was not
305 limited on the minimum size and on the maximum size just by the cluster
306 resources. As such, any policy could be implemented only in third-party
307 clients (RAPI clients, or shell wrappers over ``gnt-*``
308 tools). Furthermore, calculating cluster capacity via ``hspace`` again
309 required external input with regards to instance sizes.
311 In order to improve these workflows and to allow for example better
312 per-node group differentiation, we introduced instance specs, which
315 - minimum instance disk size, disk count, memory size, cpu count
316 - maximum values for the above metrics
317 - and “standard” values (used in ``hspace`` to calculate the standard
320 The minimum/maximum values can be also customised at node-group level,
321 for example allowing more powerful hardware to support bigger instance
324 Beside the instance specs, there are a few other settings belonging to
325 the instance policy framework. It is possible now to customise, per
326 cluster and node-group:
328 - the list of allowed disk templates
329 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
330 - the maximum ratio of instance to spindles (see below for more
331 information) for local storage
333 All these together should allow all tools that talk to Ganeti to know
334 what are the ranges of allowed values for instances and the
335 over-subscription that is allowed.
337 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
338 instance configuration, and the physical CPU configuration from the
339 node. For the spindle ratios however, we didn't track before these
340 values, so new parameters have been added:
342 - a new node parameter ``spindle_count``, defaults to 1, customisable at
343 node group or node level
344 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
346 Note that spindles in this context doesn't need to mean actual
347 mechanical hard-drives; it's just a relative number for both the node
348 I/O capacity and instance I/O consumption.
350 Instance migration behaviour
351 ++++++++++++++++++++++++++++
353 While live-migration is in general desirable over failover, it is
354 possible that for some workloads it is actually worse, due to the
355 variable time of the “suspend” phase during live migration.
357 To allow the tools to work consistently over such instances (without
358 having to hard-code instance names), a new backend parameter
359 ``always_failover`` has been added to control the migration/failover
360 behaviour. When set to True, all migration requests for an instance will
361 instead fall-back to failover.
363 Instance memory ballooning
364 ++++++++++++++++++++++++++
366 Initial support for memory ballooning has been added. The memory for an
367 instance is no longer fixed (backend parameter ``memory``), but instead
368 can vary between minimum and maximum values (backend parameters
369 ``minmem`` and ``maxmem``). Currently we only change an instance's
372 - live migrating or failing over and instance and the target node
373 doesn't have enough memory
374 - user requests changing the memory via ``gnt-instance modify
380 In order to control the use of specific CPUs by instance, support for
381 controlling CPU pinning has been added for the Xen, HVM and LXC
382 hypervisors. This is controlled by a new hypervisor parameter
383 ``cpu_mask``; details about possible values for this are in the
384 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
385 VCPU-to-CPU mapping) form will work well only when all nodes in your
386 cluster have the same amount of CPUs.
391 Another area in which Ganeti was not customisable were the parameters
392 used for storage configuration, e.g. how many stripes to use for LVM,
393 DRBD resync configuration, etc.
395 To improve this area, we've added disks parameters, which are
396 customisable at cluster and node group level, and which allow to
397 specify various parameters for disks (DRBD has the most parameters
398 currently), for example:
400 - DRBD resync algorithm and parameters (e.g. speed)
401 - the default VG for meta-data volumes for DRBD
402 - number of stripes for LVM (plain disk template)
405 These parameters can be modified via ``gnt-cluster modify -D …`` and
406 ``gnt-group modify -D …``, and are used at either instance creation (in
407 case of LVM stripes, for example) or at disk “activation” time
410 Rados block device support
411 ++++++++++++++++++++++++++
413 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
414 denoted by the ``rbd`` disk template type. This is considered
415 experimental, feedback is welcome. For details on configuring it, see
416 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
421 The existing master IP functionality works well only in simple setups (a
422 single network shared by all nodes); however, if nodes belong to
423 different networks, then the ``/32`` setup and lack of routing
424 information is not enough.
426 To allow the master IP to function well in more complex cases, the
427 system was reworked as follows:
429 - a master IP netmask setting has been added
430 - the master IP activation/turn-down code was moved from the node daemon
432 - whether to run the Ganeti-supplied master IP script or a user-supplied
433 on is a ``gnt-cluster init`` setting
435 Details about the location of the standard and custom setup scripts are
436 in the man page :manpage:`gnt-cluster(8)`; for information about the
437 setup script protocol, look at the Ganeti-supplied script.
442 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
445 It is now possible to use TLS-protected connections, and when renewing
446 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
447 it is now possible to specify spice or spice CA certificates. Also, it
448 is possible to configure a password for SPICE sessions via the
449 hypervisor parameter ``spice_password_file``.
451 There are also new parameters to control the compression and streaming
452 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
453 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
454 for the spice parameters.
456 Lastly, it is now possible to see the SPICE connection information via
457 ``gnt-instance console``.
462 A new tool (``tools/ovfconverter``) has been added that supports
463 conversion between Ganeti and the `Open Virtualization Format
464 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
467 This relies on the ``qemu-img`` tool to convert the disk formats, so the
468 actual compatibility with other virtualization solutions depends on it.
473 The configuration query daemon (``ganeti-confd``) is now optional, and
474 has been rewritten in Haskell; whether to use the daemon at all, use the
475 Python (default) or the Haskell version is selectable at configure time
476 via the ``--enable-confd`` parameter, which can take one of the
477 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
478 daemon will result in a smaller footprint; for larger systems, we
479 welcome feedback on the Haskell version which might become the default
482 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
483 daemon running. The Python version doesn't implement the new call.
486 User interface changes
487 ~~~~~~~~~~~~~~~~~~~~~~
489 We have replaced the ``--disks`` option of ``gnt-instance
490 replace-disks`` with a more flexible ``--disk`` option, which allows
491 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
492 disk size and mode can be changed upon recreation (via ``gnt-instance
493 recreate-disks``, which accepts the same ``--disk`` option).
495 As many people are used to a ``show`` command, we have added that as an
496 alias to ``info`` on all ``gnt-*`` commands.
498 The ``gnt-instance grow-disk`` command has a new mode in which it can
499 accept the target size of the disk, instead of the delta; this can be
500 more safe since two runs in absolute mode will be idempotent, and
501 sometimes it's also easier to specify the desired size directly.
503 Also the handling of instances with regard to offline secondaries has
504 been improved. Instance operations should not fail because one of it's
505 secondary nodes is offline, even though it's safe to proceed.
507 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
508 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
509 minors to instance name.
514 RAPI coverage has improved, with (for example) new resources for
515 recreate-disks, node power-cycle, etc.
520 There is partial support for ``xl`` in the Xen hypervisor; feedback is
523 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
524 whether to still support Python 2.4 or move to Python 2.6 as minimum
527 Support for Fedora has been slightly improved; the provided example
528 init.d script should work better on it and the INSTALL file should
529 document the needed dependencies.
534 The deprecated ``QueryLocks`` LUXI request has been removed. Use
535 ``Query(what=QR_LOCK, ...)`` instead.
537 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
538 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
539 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
540 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
541 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
543 RAPI client: ``CertificateError`` now derives from
544 ``GanetiApiError``. This should make it more easy to handle Ganeti
547 Deprecation warnings due to PyCrypto/paramiko import in
548 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
549 make sure to run an up-to-date paramiko version, if you use this tool.
551 The QA scripts now depend on Python 2.5 or above (the main code base
552 still works with Python 2.4).
554 The configuration file (``config.data``) is now written without
555 indentation for performance reasons; if you want to edit it, it can be
556 re-formatted via ``tools/fmtjson``.
558 A number of bugs has been fixed in the cluster merge tool.
560 ``x509`` certification verification (used in import-export) has been
561 changed to allow the same clock skew as permitted by the cluster
562 verification. This will remove some rare but hard to diagnose errors in
569 *(Released Thu, 19 Jul 2012)*
571 Very few changes from rc4 to the final release, only bugfixes:
573 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
574 instance, fix CDROM booting for KVM instances)
575 - fixed node group modification of node parameters
576 - fixed issue in LUClusterVerifyGroup with multi-group clusters
577 - fixed generation of bash completion to ensure a stable ordering
584 *(Released Fri, 13 Jul 2012)*
586 Third release candidate for 2.6. The following changes were done from
589 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
590 - Fixed an inconsistency in the LUXI protocol with the provided
591 arguments (NOT backwards compatible)
592 - Fixed a bug with node groups ipolicy where ``min`` was greater than
593 the cluster ``std`` value
594 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
595 easier instance debugging on nodes (requires ``hconfd`` to work)
601 *(Released Tue, 03 Jul 2012)*
603 Second release candidate for 2.6. The following changes were done from
606 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
608 - Fixed a RAPI regression on missing beparams/memory
609 - Fixed redistribution of files on offline nodes
610 - Added possibility to run activate-disks even though secondaries are
611 offline. With this change it relaxes also the strictness on some other
612 commands which use activate disks internally:
613 * ``gnt-instance start|reboot|rename|backup|export``
614 - Made it possible to remove safely an instance if its secondaries are
616 - Made it possible to reinstall even though secondaries are offline
622 *(Released Mon, 25 Jun 2012)*
624 First release candidate for 2.6. The following changes were done from
627 - Fixed bugs with disk parameters and ``rbd`` templates as well as
629 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
630 behaviour. It supports now the modify operation
631 - ``hcheck`` implemented to analyze cluster health and possibility of
632 improving health by rebalance
633 - ``hbal`` has been improved in dealing with split instances
639 *(Released Mon, 11 Jun 2012)*
641 Second beta release of 2.6. The following changes were done from beta2
644 - Fixed ``daemon-util`` with non-root user models
645 - Fixed creation of plain instances with ``--no-wait-for-sync``
646 - Fix wrong iv_names when running ``cfgupgrade``
647 - Export more information in RAPI group queries
648 - Fixed bug when changing instance network interfaces
649 - Extended burnin to do NIC changes
650 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
651 - Changed default for DRBD barriers
652 - Fixed DRBD error reporting for syncer rate
653 - Verify the options on disk parameters
655 And of course various fixes to documentation and improved unittests and
662 *(Released Wed, 23 May 2012)*
664 First beta release of 2.6. The following changes were done from beta1 to
667 - integrated patch for distributions without ``start-stop-daemon``
668 - adapted example init.d script to work on Fedora
669 - fixed log handling in Haskell daemons
670 - adapted checks in the watcher for pycurl linked against libnss
671 - add partial support for ``xl`` instead of ``xm`` for Xen
672 - fixed a type issue in cluster verification
673 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
676 Plus integrated fixes from the 2.5 branch:
678 - fixed ``kvm-ifup`` to use ``/bin/bash``
679 - fixed parallel build failures
680 - KVM live migration when using a custom keymap
686 *(Released Tue, 24 Jul 2012)*
688 A small bugfix release, with no new features:
690 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
691 different default shell (e.g. Debian, Ubuntu)
692 - fixed KVM startup and live migration with a custom keymap (fixes Issue
693 243 and Debian bug #650664)
694 - fixed compatibility with KVM versions that don't support multiple boot
695 devices (fixes Issue 230 and Debian bug #624256)
697 Additionally, a few fixes were done to the build system (fixed parallel
698 build failures) and to the unittests (fixed race condition in test for
699 FileID functions, and the default enable/disable mode for QA test is now
706 *(Released Fri, 11 May 2012)*
708 A small bugfix release.
710 The main issues solved are on the topic of compatibility with newer LVM
713 - fixed parsing of ``lv_attr`` field
714 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
715 the ``--force`` flag is needed
717 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
718 to accept kernel 3.x too (was hardcoded to 2.6.*).
720 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
721 of hook results and that also made display of other errors suboptimal
722 was fixed; the code behaves now like 2.4 and earlier.
724 Another change in 2.5, the cleanup of the OS scripts environment, is too
725 aggressive: it removed even the ``PATH`` variable, which requires the OS
726 scripts to *always* need to export it. Since this is a bit too strict,
727 we now export a minimal PATH, the same that we export for hooks.
729 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
730 integrated into this release.
732 Finally, a few other miscellaneous changes were done (no new features,
733 just small improvements):
735 - Fix ``gnt-group --help`` display
736 - Fix hardcoded Xen kernel path
737 - Fix grow-disk handling of invalid units
738 - Update synopsis for ``gnt-cluster repair-disk-sizes``
739 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
745 *(Released Thu, 12 Apr 2012)*
747 Incompatible/important changes and bugfixes
748 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
750 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
751 resource's ``ip_check`` parameter changed from ``True`` to ``False``
752 to match the underlying LUXI interface.
753 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
754 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
755 not maintain backwards-compatibility as the underlying operation
756 changed in an incompatible way. The RAPI client can talk to old
757 servers, but it needs to be told so as the return value changed.
758 - When creating file-based instances via RAPI, the ``file_driver``
759 parameter no longer defaults to ``loop`` and must be specified.
760 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
762 - Support for the undocumented and deprecated RAPI instance creation
763 request format version 0 has been dropped. Use version 1, supported
764 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
765 - Pyparsing 1.4.6 or above is required, see :doc:`installation
766 documentation <install>`.
767 - The "cluster-verify" hooks are now executed per group by the
768 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
769 if you just run ``gnt-cluster verify``, which generates one opcode per
771 - The environment as passed to the OS scripts is cleared, and thus no
772 environment variables defined in the node daemon's environment will be
773 inherited by the scripts.
774 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
776 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
777 support operations involving multiple node groups.
778 - Offline nodes are ignored when failing over an instance.
779 - Support for KVM version 1.0, which changed the version reporting format
781 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
783 - ``Makefile`` is now compatible with Automake 1.11.2
784 - Includes all bugfixes made in the 2.4 series
789 - The ganeti-htools project has been merged into the ganeti-core source
790 tree and will be built as part of Ganeti (see :doc:`install-quick`).
791 - Implemented support for :doc:`shared storage <design-shared-storage>`.
792 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
793 GPT-style partition tables (requires `parted
794 <http://www.gnu.org/s/parted/>`_).
795 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
796 - Allowed adding tags on instance creation.
797 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
798 - Allow instances to be started in a paused state, enabling the user to
799 see the complete console output on boot using the console.
800 - Added new hypervisor flag to control default reboot behaviour
801 (``reboot_behavior``).
802 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
803 - Improved out-of-band management support:
805 - Added ``gnt-node health`` command reporting the health status of
807 - Added ``gnt-node power`` command to manage power status of nodes.
808 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
810 - Instance migration can fall back to failover if instance is not
812 - Filters can be used when listing nodes, instances, groups and locks;
813 see :manpage:`ganeti(7)` manpage.
814 - Added post-execution status as variables to :doc:`hooks <hooks>`
816 - Instance tags are exported/imported together with the instance.
817 - When given an explicit job ID, ``gnt-job info`` will work for archived
819 - Jobs can define dependencies on other jobs (not yet supported via
820 RAPI or command line, but used by internal commands and usable via
823 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
826 - Instance failover is now available as a RAPI resource
827 (``/2/instances/[instance_name]/failover``).
828 - ``gnt-instance info`` defaults to static information if primary node
830 - Opcodes have a new ``comment`` attribute.
831 - Added basic SPICE support to KVM hypervisor.
832 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
834 Node group improvements
835 ~~~~~~~~~~~~~~~~~~~~~~~
837 - ``gnt-cluster verify`` has been modified to check groups separately,
838 thereby improving performance.
839 - Node group support has been added to ``gnt-cluster verify-disks``,
840 which now operates per node group.
841 - Watcher has been changed to work better with node groups.
843 - One process and state file per node group.
844 - Slow watcher in one group doesn't block other group's watcher.
846 - Added new command, ``gnt-group evacuate``, to move all instances in a
847 node group to other groups.
848 - Added ``gnt-instance change-group`` to move an instance to another
850 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
851 per-group operations.
852 - Node groups can be tagged.
853 - Some operations switch from an exclusive to a shared lock as soon as
855 - Instance's primary and secondary nodes' groups are now available as
856 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
857 and ``snodes.group.uuid``).
862 - Numerous updates to documentation and manpages.
864 - :doc:`RAPI <rapi>` documentation now has detailed parameter
866 - Some opcode/job results are now also documented, see :doc:`RAPI
869 - A lockset's internal lock is now also visible in lock monitor.
870 - Log messages from job queue workers now contain information about the
871 opcode they're processing.
872 - ``gnt-instance console`` no longer requires the instance lock.
873 - A short delay when waiting for job changes reduces the number of LUXI
874 requests significantly.
875 - DRBD metadata volumes are overwritten with zeros during disk creation.
876 - Out-of-band commands no longer acquire the cluster lock in exclusive
878 - ``devel/upload`` now uses correct permissions for directories.
884 *(Released Fri, 23 Mar 2012)*
886 This was the sixth release candidate of the 2.5 series.
892 *(Released Mon, 9 Jan 2012)*
894 This was the fifth release candidate of the 2.5 series.
900 *(Released Thu, 27 Oct 2011)*
902 This was the fourth release candidate of the 2.5 series.
908 *(Released Wed, 26 Oct 2011)*
910 This was the third release candidate of the 2.5 series.
916 *(Released Tue, 18 Oct 2011)*
918 This was the second release candidate of the 2.5 series.
924 *(Released Tue, 4 Oct 2011)*
926 This was the first release candidate of the 2.5 series.
932 *(Released Wed, 31 Aug 2011)*
934 This was the third beta release of the 2.5 series.
940 *(Released Mon, 22 Aug 2011)*
942 This was the second beta release of the 2.5 series.
948 *(Released Fri, 12 Aug 2011)*
950 This was the first beta release of the 2.5 series.
956 *(Released Thu, 27 Oct 2011)*
958 - Fixed bug when parsing command line parameter values ending in
960 - Fixed assertion error after unclean master shutdown
961 - Disable HTTP client pool for RPC, significantly reducing memory usage
963 - Fixed queue archive creation with wrong permissions
969 *(Released Tue, 23 Aug 2011)*
973 - Fixed documentation for importing with ``--src-dir`` option
974 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
975 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
981 *(Released Fri, 5 Aug 2011)*
983 Many bug-fixes and a few small features:
985 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
986 issues when you tried to add an instance with two MAC addresses in one
988 - KVM: fixed per-instance stored UID value
989 - KVM: configure bridged NICs at migration start
990 - KVM: Fix a bug where instance will not start with never KVM versions
992 - Added OS search path to ``gnt-cluster info``
993 - Fixed an issue with ``file_storage_dir`` where you were forced to
994 provide an absolute path, but the documentation states it is a
995 relative path, the documentation was right
996 - Added a new parameter to instance stop/start called ``--no-remember``
997 that will make the state change to not be remembered
998 - Implemented ``no_remember`` at RAPI level
999 - Improved the documentation
1000 - Node evacuation: don't call IAllocator if node is already empty
1001 - Fixed bug in DRBD8 replace disks on current nodes
1002 - Fixed bug in recreate-disks for DRBD instances
1003 - Moved assertion checking locks in ``gnt-instance replace-disks``
1004 causing it to abort with not owning the right locks for some situation
1005 - Job queue: Fixed potential race condition when cancelling queued jobs
1006 - Fixed off-by-one bug in job serial generation
1007 - ``gnt-node volumes``: Fix instance names
1008 - Fixed aliases in bash completion
1009 - Fixed a bug in reopening log files after being sent a SIGHUP
1010 - Added a flag to burnin to allow specifying VCPU count
1011 - Bugfixes to non-root Ganeti configuration
1017 *(Released Thu, 12 May 2011)*
1019 Many bug-fixes and a few new small features:
1021 - Fixed a bug related to log opening failures
1022 - Fixed a bug in instance listing with orphan instances
1023 - Fixed a bug which prevented resetting the cluster-level node parameter
1024 ``oob_program`` to the default
1025 - Many fixes related to the ``cluster-merge`` tool
1026 - Fixed a race condition in the lock monitor, which caused failures
1027 during (at least) creation of many instances in parallel
1028 - Improved output for gnt-job info
1029 - Removed the quiet flag on some ssh calls which prevented debugging
1031 - Improved the N+1 failure messages in cluster verify by actually
1032 showing the memory values (needed and available)
1033 - Increased lock attempt timeouts so that when executing long operations
1034 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1035 too early and thus prevent the use of the 'fair' mechanism
1036 - Changed instance query data (``gnt-instance info``) to not acquire
1037 locks unless needed, thus allowing its use on locked instance if only
1038 static information is asked for
1039 - Improved behaviour with filesystems that do not support rename on an
1041 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1042 kept locks on all nodes during the wipe, which is unneeded
1043 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1044 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1045 10GiB) which caused the wipe to fail right at the end in some cases
1046 - Fixed master IP activation when doing master failover with no-voting
1047 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1048 the master node itself
1049 - Fixed potential data-loss in under disk full conditions, where Ganeti
1050 wouldn't check correctly the return code and would consider
1051 partially-written files 'correct'
1052 - Fixed bug related to multiple VGs and DRBD disk replacing
1053 - Added new disk parameter ``metavg`` that allows placement of the meta
1054 device for DRBD in a different volume group
1055 - Fixed error handling in the node daemon when the system libc doesn't
1056 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1057 - Fixed lock release during replace-disks, which kept cluster-wide locks
1058 when doing disk replaces with an iallocator script
1059 - Added check for missing bridges in cluster verify
1060 - Handle EPIPE errors while writing to the terminal better, so that
1061 piping the output to e.g. ``less`` doesn't cause a backtrace
1062 - Fixed rare case where a ^C during Luxi calls could have been
1063 interpreted as server errors, instead of simply terminating
1064 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1066 - Added a few more parameters to the KVM hypervisor, allowing a second
1067 CDROM, custom disk type for CDROMs and a floppy image
1068 - Removed redundant message in instance rename when the name is given
1070 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1071 disks on new nodes, allowing recreation when the original instance
1072 nodes are completely gone
1073 - Added option when converting disk templates to DRBD to skip waiting
1074 for the resync, in order to make the instance available sooner
1075 - Added two new variables to the OS scripts environment (containing the
1077 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1078 to allow use of ``pvgrub`` as bootloader
1079 - Changed the instance memory modifications to only check out-of-memory
1080 conditions on memory increases, and turned the secondary node warnings
1081 into errors (they can still be overridden via ``--force``)
1082 - Fixed the handling of a corner case when the Python installation gets
1083 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1084 to execute a command that doesn't exist
1085 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1086 node of the instance returned failures during instance shutdown; this
1087 adds the option ``--ignore-consistency`` to gnt-instance move
1089 And as usual, various improvements to the error messages, documentation
1096 *(Released Wed, 09 Mar 2011)*
1098 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1099 the RAPI users file if run twice (even with ``--dry-run``).
1101 The release fixes that bug (nothing else changed).
1107 *(Released Mon, 07 Mar 2011)*
1109 Final 2.4.0 release. Just a few small fixes:
1111 - Fixed RAPI node evacuate
1112 - Fixed the kvm-ifup script
1113 - Fixed internal error handling for special job cases
1114 - Updated man page to specify the escaping feature for options
1120 *(Released Mon, 28 Feb 2011)*
1122 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1123 that this feature wiped the disks of the wrong instance, leading to loss
1128 - Fixed title of query field containing instance name
1129 - Expanded the glossary in the documentation
1130 - Fixed one unittest (internal issue)
1136 *(Released Mon, 21 Feb 2011)*
1138 A number of bug fixes plus just a couple functionality changes.
1140 On the user-visible side, the ``gnt-* list`` command output has changed
1141 with respect to "special" field states. The current rc1 style of display
1142 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1143 the default output mode special fields states are displayed as follows:
1145 - Offline resource: ``*``
1146 - Unavailable/not applicable: ``-``
1147 - Data missing (RPC failure): ``?``
1148 - Unknown field: ``??``
1150 Another user-visible change is the addition of ``--force-join`` to
1155 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1156 - Fixed regression in RAPI/instance reinstall where all parameters were
1157 required (instead of optional)
1158 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1159 - Fixed iallocator usage (offline nodes were not considered offline)
1160 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1161 - Fixed hypervisor and OS parameter validation with respect to
1162 non-vm_capable nodes
1163 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1165 - Fixed ``tools/listrunner`` with respect to agent-based usage
1171 *(Released Fri, 4 Feb 2011)*
1173 Many changes and fixes since the beta1 release. While there were some
1174 internal changes, the code has been mostly stabilised for the RC
1177 Note: the dumb allocator was removed in this release, as it was not kept
1178 up-to-date with the IAllocator protocol changes. It is recommended to
1179 use the ``hail`` command from the ganeti-htools package.
1181 Note: the 2.4 and up versions of Ganeti are not compatible with the
1182 0.2.x branch of ganeti-htools. You need to upgrade to
1183 ganeti-htools-0.3.0 (or later).
1185 Regressions fixed from 2.3
1186 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1188 - Fixed the ``gnt-cluster verify-disks`` command
1189 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1191 - Fixed disk adoption breakage
1192 - Fixed wrong headers in instance listing for field aliases
1197 - Fixed corner case in KVM handling of NICs
1198 - Fixed many cases of wrong handling of non-vm_capable nodes
1199 - Fixed a bug where a missing instance symlink was not possible to
1200 recreate with any ``gnt-*`` command (now ``gnt-instance
1201 activate-disks`` does it)
1202 - Fixed the volume group name as reported by ``gnt-cluster
1204 - Increased timeouts for the import-export code, hopefully leading to
1205 fewer aborts due network or instance timeouts
1206 - Fixed bug in ``gnt-node list-storage``
1207 - Fixed bug where not all daemons were started on cluster
1208 initialisation, but only at the first watcher run
1209 - Fixed many bugs in the OOB implementation
1210 - Fixed watcher behaviour in presence of instances with offline
1212 - Fixed instance list output for instances running on the wrong node
1213 - a few fixes to the cluster-merge tool, but it still cannot merge
1214 multi-node groups (currently it is not recommended to use this tool)
1220 - Improved network configuration for the KVM hypervisor
1221 - Added e1000 as a supported NIC for Xen-HVM
1222 - Improved the lvmstrap tool to also be able to use partitions, as
1223 opposed to full disks
1224 - Improved speed of disk wiping (the cluster parameter
1225 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1226 of instance creations
1227 - Added documentation for the OS parameters
1228 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1229 hypervisor is not responding
1230 - Added display of blacklisted and hidden OS information in
1231 ``gnt-cluster info``
1232 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1233 NIC and node parameters, which might create problems with currently
1234 invalid (but undetected) configuration files, but prevents validation
1235 failures when unrelated parameters are modified
1236 - Changed cluster initialisation to wait for the master daemon to become
1238 - Expanded the RAPI interface:
1240 - Added config redistribution resource
1241 - Added activation/deactivation of instance disks
1242 - Added export of console information
1244 - Implemented log file reopening on SIGHUP, which allows using
1245 logrotate(8) for the Ganeti log files
1246 - Added a basic OOB helper script as an example
1252 *(Released Fri, 14 Jan 2011)*
1257 - Fixed timezone issues when formatting timestamps
1258 - Added support for node groups, available via ``gnt-group`` and other
1260 - Added out-of-band framework and management, see :doc:`design
1261 document <design-oob>`
1262 - Removed support for roman numbers from ``gnt-node list`` and
1263 ``gnt-instance list``.
1264 - Allowed modification of master network interface via ``gnt-cluster
1265 modify --master-netdev``
1266 - Accept offline secondaries while shutting down instance disks
1267 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1268 - Added support for multiple LVM volume groups
1269 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1271 - Added commands to list available fields:
1273 - ``gnt-node list-fields``
1274 - ``gnt-group list-fields``
1275 - ``gnt-instance list-fields``
1277 - Updated documentation and man pages
1282 - Moved ``rapi_users`` file into separate directory, now named
1283 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1285 - Added new tool for running commands on many machines,
1286 ``tools/ganeti-listrunner``
1287 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1288 improving the ``gnt-instance console`` output
1289 - Allowed customisation of disk index separator at ``configure`` time
1290 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1291 - Added support for non-partitioned md disks in ``lvmstrap``
1292 - Added script to gracefully power off KVM instances
1293 - Split ``utils`` module into smaller parts
1294 - Changed query operations to return more detailed information, e.g.
1295 whether an information is unavailable due to an offline node. To use
1296 this new functionality, the LUXI call ``Query`` must be used. Field
1297 information is now stored by the master daemon and can be retrieved
1298 using ``QueryFields``. Instances, nodes and groups can also be queried
1299 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1300 exposed via RAPI). The following commands make use of this
1301 infrastructure change:
1303 - ``gnt-group list``
1304 - ``gnt-group list-fields``
1306 - ``gnt-node list-fields``
1307 - ``gnt-instance list``
1308 - ``gnt-instance list-fields``
1309 - ``gnt-debug locks``
1314 - New RAPI resources (see :doc:`rapi`):
1318 - ``/2/groups/[group_name]``
1319 - ``/2/groups/[group_name]/assign-nodes``
1320 - ``/2/groups/[group_name]/modify``
1321 - ``/2/groups/[group_name]/rename``
1322 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1326 - Implemented ``no_install`` for instance creation
1327 - Implemented OS parameters for instance reinstallation, allowing
1328 use of special settings on reinstallation (e.g. for preserving data)
1333 - Added IPv6 support in import/export
1334 - Pause DRBD synchronization while wiping disks on instance creation
1335 - Updated unittests and QA scripts
1336 - Improved network parameters passed to KVM
1337 - Converted man pages from docbook to reStructuredText
1343 *(Released Mon, 20 Dec 2010)*
1345 Released version 2.3.1~rc1 without any changes.
1351 *(Released Wed, 1 Dec 2010)*
1353 - impexpd: Disable OpenSSL compression in socat if possible (backport
1354 from master, commit e90739d625b, see :doc:`installation guide
1355 <install-quick>` for details)
1356 - Changed unittest coverage report to exclude test scripts
1357 - Added script to check version format
1363 *(Released Wed, 1 Dec 2010)*
1365 Released version 2.3.0~rc1 without any changes.
1371 *(Released Fri, 19 Nov 2010)*
1373 A number of bugfixes and documentation updates:
1375 - Update ganeti-os-interface documentation
1376 - Fixed a bug related to duplicate MACs or similar items which should be
1378 - Fix breakage in OS state modify
1379 - Reinstall instance: disallow offline secondaries (fixes bug related to
1380 OS changing but reinstall failing)
1381 - plus all the other fixes between 2.2.1 and 2.2.2
1387 *(Released Tue, 2 Nov 2010)*
1389 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1390 - Fixed master failover race with watcher
1391 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1393 - Accept previously stopped instance for export with instance removal
1394 - Simplify and extend the environment variables for instance OS scripts
1395 - Added new node flags, ``master_capable`` and ``vm_capable``
1396 - Added optional instance disk wiping prior during allocation. This is a
1397 cluster-wide option and can be set/modified using
1398 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1399 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1400 :doc:`install-quick`
1401 - Added a new watcher option (``--ignore-pause``)
1402 - Added option to ignore offline node on instance start/stop
1403 (``--ignore-offline``)
1404 - Allow overriding OS parameters with ``gnt-instance reinstall``
1405 - Added ability to change node's secondary IP address using ``gnt-node
1407 - Implemented privilege separation for all daemons except
1408 ``ganeti-noded``, see ``configure`` options
1409 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1411 - Implemented job priorities (see ``ganeti(7)`` manpage)
1412 - Ignore failures while shutting down instances during failover from
1414 - Exit daemon's bootstrap process only once daemon is ready
1415 - Export more information via ``LUInstanceQuery``/remote API
1416 - Improved documentation, QA and unittests
1417 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1418 restart if the file was created or changed
1419 - Added LUXI protocol version sent with each request and response,
1420 allowing detection of server/client mismatches
1421 - Moved the Python scripts among gnt-* and ganeti-* into modules
1422 - Moved all code related to setting up SSH to an external script,
1424 - Infrastructure changes for node group support in future versions
1430 *(Released Fri, 19 Nov 2010)*
1432 A few small bugs fixed, and some improvements to the build system:
1434 - Fix documentation regarding conversion to drbd
1435 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1437 - Fix error handling in node modify with multiple changes
1438 - Allow remote imports without checked names
1444 *(Released Tue, 19 Oct 2010)*
1446 - Disable SSL session ID cache in RPC client
1452 *(Released Thu, 14 Oct 2010)*
1454 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1455 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1457 - Fix problems with interaction between Curl and Python's HTTP server,
1458 resulting in increased speed in many RPC calls
1459 - Improve our release script to prevent breakage with older aclocal and
1466 *(Released Thu, 7 Oct 2010)*
1468 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1469 value retrieved from master
1470 - Added support for blacklisted or hidden OS definitions
1471 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1472 - Added support for -mem-path in KVM hypervisor abstraction layer
1473 - Allow overriding instance parameters in tool for inter-cluster
1474 instance moves (``tools/move-instance``)
1475 - Improved opcode summaries (e.g. in ``gnt-job list``)
1476 - Improve consistency of OS listing by sorting it
1477 - Documentation updates
1483 *(Released Fri, 8 Oct 2010)*
1485 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1491 *(Released Mon, 4 Oct 2010)*
1493 - Fixed regression in ``gnt-instance rename``
1499 *(Released Wed, 22 Sep 2010)*
1501 - Fixed OS_VARIANT variable for OS scripts
1502 - Fixed cluster tag operations via RAPI
1503 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1504 - Disabled RAPI CA checks in watcher
1510 *(Released Mon, 23 Aug 2010)*
1512 - Support DRBD versions of the format "a.b.c.d"
1514 - Re-introduce support for usage from multiple threads in RAPI client
1515 - Instance renames and modify via RAPI
1516 - Work around race condition between processing and archival in job
1518 - Mark opcodes following failed one as failed, too
1519 - Job field ``lock_status`` was removed due to difficulties making it
1520 work with the changed job queue in Ganeti 2.2; a better way to monitor
1521 locks is expected for a later 2.2.x release
1522 - Fixed dry-run behaviour with many commands
1523 - Support ``ssh-agent`` again when adding nodes
1524 - Many additional bugfixes
1530 *(Released Fri, 30 Jul 2010)*
1532 Important change: the internal RPC mechanism between Ganeti nodes has
1533 changed from using a home-grown http library (based on the Python base
1534 libraries) to use the PycURL library. This requires that PycURL is
1535 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1536 against GnuTLS by default. cURL's support for GnuTLS had known issues
1537 before cURL 7.21.0 and we recommend using the latest cURL release or
1538 linking against OpenSSL. Most other distributions already link PycURL
1539 and cURL against OpenSSL. The command::
1541 python -c 'import pycurl; print pycurl.version'
1543 can be used to determine the libraries PycURL and cURL are linked
1546 Other significant changes:
1548 - Rewrote much of the internals of the job queue, in order to achieve
1549 better parallelism; this decouples job query operations from the job
1550 processing, and it should allow much nicer behaviour of the master
1551 daemon under load, and it also has uncovered some long-standing bugs
1552 related to the job serialisation (now fixed)
1553 - Added a default iallocator setting to the cluster parameters,
1554 eliminating the need to always pass nodes or an iallocator for
1555 operations that require selection of new node(s)
1556 - Added experimental support for the LXC virtualization method
1557 - Added support for OS parameters, which allows the installation of
1558 instances to pass parameter to OS scripts in order to customise the
1560 - Added a hypervisor parameter controlling the migration type (live or
1561 non-live), since hypervisors have various levels of reliability; this
1562 has renamed the 'live' parameter to 'mode'
1563 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1564 logical volumes, meaning that cluster verify will ignore them and not
1565 flag their presence as errors
1566 - The watcher will now reset the error count for failed instances after
1567 8 hours, thus allowing self-healing if the problem that caused the
1568 instances to be down/fail to start has cleared in the meantime
1569 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1570 check for, and warn, if the drbd module parameter ``usermode_helper``
1571 is not consistent with the cluster-wide setting; this is needed to
1572 make diagnose easier of failed drbd creations
1573 - Started adding base IPv6 support, but this is not yet
1574 enabled/available for use
1575 - Rename operations (cluster, instance) will now return the new name,
1576 which is especially useful if a short name was passed in
1577 - Added support for instance migration in RAPI
1578 - Added a tool to pre-configure nodes for the SSH setup, before joining
1579 them to the cluster; this will allow in the future a simplified model
1580 for node joining (but not yet fully enabled in 2.2); this needs the
1581 paramiko python library
1582 - Fixed handling of name-resolving errors
1583 - Fixed consistency of job results on the error path
1584 - Fixed master-failover race condition when executed multiple times in
1586 - Fixed many bugs related to the job queue (mostly introduced during the
1587 2.2 development cycle, so not all are impacting 2.1)
1588 - Fixed instance migration with missing disk symlinks
1589 - Fixed handling of unknown jobs in ``gnt-job archive``
1590 - And many other small fixes/improvements
1594 - Enhanced both the unittest and the QA coverage
1595 - Switched the opcode validation to a generic model, and extended the
1596 validation to all opcode parameters
1597 - Changed more parts of the code that write shell scripts to use the
1599 - Switched the master daemon to use the asyncore library for the Luxi
1606 *(Released Thu, 17 Jun 2010)*
1608 - Added tool (``move-instance``) and infrastructure to move instances
1609 between separate clusters (see :doc:`separate documentation
1610 <move-instance>` and :doc:`design document <design-2.2>`)
1611 - Added per-request RPC timeout
1612 - RAPI now requires a Content-Type header for requests with a body (e.g.
1613 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1614 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1615 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1617 - Implemented initial support for running Ganeti daemons as separate
1618 users, see configure-time flags ``--with-user-prefix`` and
1619 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1621 - Instances can be removed after export (``gnt-backup export
1622 --remove-instance``)
1623 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1624 key (instead of 1024 bit)
1625 - Added new cluster configuration file for cluster domain secret
1626 - Import/export now use SSL instead of SSH
1627 - Added support for showing estimated time when exporting an instance,
1628 see the ``ganeti-os-interface(7)`` manpage and look for
1635 *(Released Tue, 16 Nov 2010)*
1637 Some more bugfixes. Unless critical bugs occur, this will be the last
1640 - Fix case of MAC special-values
1641 - Fix mac checker regex
1642 - backend: Fix typo causing "out of range" error
1643 - Add missing --units in gnt-instance list man page
1649 *(Released Tue, 24 Aug 2010)*
1652 - Don't ignore secondary node silently on non-mirrored disk templates
1654 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1655 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1656 - Properly document the usb_mouse parameter
1657 - Fix path in ganeti-rapi(8) (issue 116)
1658 - Adjust error message when the ganeti user's .ssh directory is
1660 - Add same-node-check when changing the disk template to drbd
1666 *(Released Fri, 16 Jul 2010)*
1669 - Add an option to only select some reboot types during qa/burnin.
1670 (on some hypervisors consequent reboots are not supported)
1671 - Fix infrequent race condition in master failover. Sometimes the old
1672 master ip address would be still detected as up for a short time
1673 after it was removed, causing failover to fail.
1674 - Decrease mlockall warnings when the ctypes module is missing. On
1675 Python 2.4 we support running even if no ctypes module is installed,
1676 but we were too verbose about this issue.
1677 - Fix building on old distributions, on which man doesn't have a
1679 - Fix RAPI not to ignore the MAC address on instance creation
1680 - Implement the old instance creation format in the RAPI client.
1686 *(Released Thu, 01 Jul 2010)*
1688 A small bugfix release:
1689 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1690 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1691 a non-existing option
1692 - Fix instance create: the --force-variant option was ignored
1693 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1694 - Fix modify node storage with non-FQDN arguments
1695 - Fix RAPI client to authenticate under Python 2.6 when used
1696 for more than 5 requests needing authentication
1697 - Fix gnt-instance modify -t (storage) giving a wrong error message
1698 when converting a non-shutdown drbd instance to plain
1704 *(Released Fri, 18 Jun 2010)*
1706 A small bugfix release:
1708 - Fix live migration of KVM instances started with older Ganeti
1709 versions which had fewer hypervisor parameters
1710 - Fix gnt-instance grow-disk on down instances
1711 - Fix an error-reporting bug during instance migration
1712 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1713 silently ignoring broken ones
1714 - Fix an RPC error reporting bug affecting, for example, RAPI client
1716 - Fix bug triggered by different API version os-es on different nodes
1717 - Fix a bug in instance startup with custom hvparams: OS level
1718 parameters would fail to be applied.
1719 - Fix the RAPI client under Python 2.6 (but more work is needed to
1720 make it work completely well with OpenSSL)
1721 - Fix handling of errors when resolving names from DNS
1727 *(Released Thu, 3 Jun 2010)*
1729 A medium sized development cycle. Some new features, and some
1730 fixes/small improvements/cleanups.
1732 Significant features
1733 ~~~~~~~~~~~~~~~~~~~~
1735 The node deamon now tries to mlock itself into memory, unless the
1736 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1737 its logs, and falls back to console logging. This allows emergency
1738 features such as ``gnt-node powercycle`` to work even in the event of a
1739 broken node disk (tested offlining the disk hosting the node's
1740 filesystem and dropping its memory caches; don't try this at home)
1742 KVM: add vhost-net acceleration support. It can be tested with a new
1743 enough version of the kernel and of qemu-kvm.
1745 KVM: Add instance chrooting feature. If you use privilege dropping for
1746 your VMs you can also now force them to chroot to an empty directory,
1747 before starting the emulated guest.
1749 KVM: Add maximum migration bandwith and maximum downtime tweaking
1750 support (requires a new-enough version of qemu-kvm).
1752 Cluster verify will now warn if the master node doesn't have the master
1753 ip configured on it.
1755 Add a new (incompatible) instance creation request format to RAPI which
1756 supports all parameters (previously only a subset was supported, and it
1757 wasn't possible to extend the old format to accomodate all the new
1758 features. The old format is still supported, and a client can check for
1759 this feature, before using it, by checking for its presence in the
1760 ``features`` RAPI resource.
1762 Now with ancient latin support. Try it passing the ``--roman`` option to
1763 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1764 (requires the python-roman module to be installed, in order to work).
1769 As usual many internal code refactorings, documentation updates, and
1772 - Lots of improvements and cleanups to the experimental Remote API
1773 (RAPI) client library.
1774 - A new unit test suite for the core daemon libraries.
1775 - A fix to creating missing directories makes sure the umask is not
1776 applied anymore. This enforces the same directory permissions
1778 - Better handling terminating daemons with ctrl+c (used when running
1779 them in debugging mode).
1780 - Fix a race condition in live migrating a KVM instance, when stat()
1781 on the old proc status file returned EINVAL, which is an unexpected
1783 - Fixed manpage checking with newer man and utf-8 charachters. But now
1784 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1790 *(Released Fri, 7 May 2010)*
1792 Fix a bug which prevented untagged KVM instances from starting.
1798 *(Released Fri, 7 May 2010)*
1800 Another release with a long development cycle, during which many
1801 different features were added.
1803 Significant features
1804 ~~~~~~~~~~~~~~~~~~~~
1806 The KVM hypervisor now can run the individual instances as non-root, to
1807 reduce the impact of a VM being hijacked due to bugs in the
1808 hypervisor. It is possible to run all instances as a single (non-root)
1809 user, to manually specify a user for each instance, or to dynamically
1810 allocate a user out of a cluster-wide pool to each instance, with the
1811 guarantee that no two instances will run under the same user ID on any
1814 An experimental RAPI client library, that can be used standalone
1815 (without the other Ganeti libraries), is provided in the source tree as
1816 ``lib/rapi/client.py``. Note this client might change its interface in
1817 the future, as we iterate on its capabilities.
1819 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1820 replace the cluster's certificates and crypto keys. This might help in
1821 case they have been compromised, or have simply expired.
1823 A new disk option for instance creation has been added that allows one
1824 to "adopt" currently existing logical volumes, with data
1825 preservation. This should allow easier migration to Ganeti from
1826 unmanaged (or managed via other software) instances.
1828 Another disk improvement is the possibility to convert between redundant
1829 (DRBD) and plain (LVM) disk configuration for an instance. This should
1830 allow better scalability (starting with one node and growing the
1831 cluster, or shrinking a two-node cluster to one node).
1833 A new feature that could help with automated node failovers has been
1834 implemented: if a node sees itself as offline (by querying the master
1835 candidates), it will try to shutdown (hard) all instances and any active
1836 DRBD devices. This reduces the risk of duplicate instances if an
1837 external script automatically failovers the instances on such nodes. To
1838 enable this, the cluster parameter ``maintain_node_health`` should be
1839 enabled; in the future this option (per the name) will enable other
1840 automatic maintenance features.
1842 Instance export/import now will reuse the original instance
1843 specifications for all parameters; that means exporting an instance,
1844 deleting it and the importing it back should give an almost identical
1845 instance. Note that the default import behaviour has changed from
1846 before, where it created only one NIC; now it recreates the original
1849 Cluster verify has added a few new checks: SSL certificates validity,
1850 /etc/hosts consistency across the cluster, etc.
1855 As usual, many internal changes were done, documentation fixes,
1858 - Fixed cluster initialization with disabled cluster storage (regression
1859 introduced in 2.1.1)
1860 - File-based storage supports growing the disks
1861 - Fixed behaviour of node role changes
1862 - Fixed cluster verify for some corner cases, plus a general rewrite of
1863 cluster verify to allow future extension with more checks
1864 - Fixed log spamming by watcher and node daemon (regression introduced
1866 - Fixed possible validation issues when changing the list of enabled
1868 - Fixed cleanup of /etc/hosts during node removal
1869 - Fixed RAPI response for invalid methods
1870 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1871 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1872 from ide disks, etc.)
1873 - Allow OS changes without re-installation (to record a changed OS
1874 outside of Ganeti, or to allow OS renames)
1875 - Allow instance creation without OS installation (useful for example if
1876 the OS will be installed manually, or restored from a backup not in
1878 - Implemented option to make cluster ``copyfile`` use the replication
1880 - Added list of enabled hypervisors to ssconf (possibly useful for
1882 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1884 - A partial form of node re-IP is possible via node readd, which now
1885 allows changed node primary IP
1886 - Command line utilities now show an informational message if the job is
1888 - The logs of the master daemon now show the PID/UID/GID of the
1895 *(Released Fri, 12 Mar 2010)*
1897 During the 2.1.0 long release candidate cycle, a lot of improvements and
1898 changes have accumulated with were released later as 2.1.1.
1903 The node evacuate command (``gnt-node evacuate``) was significantly
1904 rewritten, and as such the IAllocator protocol was changed - a new
1905 request type has been added. This unfortunate change during a stable
1906 series is designed to improve performance of node evacuations; on
1907 clusters with more than about five nodes and which are well-balanced,
1908 evacuation should proceed in parallel for all instances of the node
1909 being evacuated. As such, any existing IAllocator scripts need to be
1910 updated, otherwise the above command will fail due to the unknown
1911 request. The provided "dumb" allocator has not been updated; but the
1912 ganeti-htools package supports the new protocol since version 0.2.4.
1914 Another important change is increased validation of node and instance
1915 names. This might create problems in special cases, if invalid host
1916 names are being used.
1918 Also, a new layer of hypervisor parameters has been added, that sits at
1919 OS level between the cluster defaults and the instance ones. This allows
1920 customisation of virtualization parameters depending on the installed
1921 OS. For example instances with OS 'X' may have a different KVM kernel
1922 (or any other parameter) than the cluster defaults. This is intended to
1923 help managing a multiple OSes on the same cluster, without manual
1924 modification of each instance's parameters.
1926 A tool for merging clusters, ``cluster-merge``, has been added in the
1927 tools sub-directory.
1932 - Improved the int/float conversions that should make the code more
1933 robust in face of errors from the node daemons
1934 - Fixed the remove node code in case of internal configuration errors
1935 - Fixed the node daemon behaviour in face of inconsistent queue
1936 directory (e.g. read-only file-system where we can't open the files
1938 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1939 now it either aborts cleanly or, if given the new "auto_promote"
1940 parameter, will automatically promote other nodes as needed
1941 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1942 completely prevent Ganeti from working
1943 - Fixed bug for instance export when not all disks were successfully
1945 - Fixed behaviour of node add when the new node is slow in starting up
1947 - Fixed handling of signals in the LUXI client, which should improve
1948 behaviour of command-line scripts
1949 - Added checks for invalid node/instance names in the configuration (now
1950 flagged during cluster verify)
1951 - Fixed watcher behaviour for disk activation errors
1952 - Fixed two potentially endless loops in http library, which led to the
1953 RAPI daemon hanging and consuming 100% CPU in some cases
1954 - Fixed bug in RAPI daemon related to hashed passwords
1955 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1957 - Enhanced compatibility with non-Debian OSes, but not using absolute
1958 path in some commands and allowing customisation of the ssh
1959 configuration directory
1960 - Fixed possible future issue with new Python versions by abiding to the
1961 proper use of ``__slots__`` attribute on classes
1962 - Added checks that should prevent directory traversal attacks
1963 - Many documentation fixes based on feedback from users
1968 - Added an "early_release" more for instance replace disks and node
1969 evacuate, where we release locks earlier and thus allow higher
1970 parallelism within the cluster
1971 - Added watcher hooks, intended to allow the watcher to restart other
1972 daemons (e.g. from the ganeti-nbma project), but they can be used of
1973 course for any other purpose
1974 - Added a compile-time disable for DRBD barriers, to increase
1975 performance if the administrator trusts the power supply or the
1976 storage system to not lose writes
1977 - Added the option of using syslog for logging instead of, or in
1978 addition to, Ganeti's own log files
1979 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1980 can indeed boot from a paravirtual NIC
1981 - Added a generic debug level for many operations; while this is not
1982 used widely yet, it allows one to pass the debug value all the way to
1984 - Enhanced the hooks environment for instance moves (failovers,
1985 migrations) where the primary/secondary nodes changed during the
1986 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1987 - Enhanced data validations for many user-supplied values; one important
1988 item is the restrictions imposed on instance and node names, which
1989 might reject some (invalid) host names
1990 - Add a configure-time option to disable file-based storage, if it's not
1991 needed; this allows greater security separation between the master
1992 node and the other nodes from the point of view of the inter-node RPC
1994 - Added user notification in interactive tools if job is waiting in the
1995 job queue or trying to acquire locks
1996 - Added log messages when a job is waiting for locks
1997 - Added filtering by node tags in instance operations which admit
1998 multiple instances (start, stop, reboot, reinstall)
1999 - Added a new tool for cluster mergers, ``cluster-merge``
2000 - Parameters from command line which are of the form ``a=b,c=d`` can now
2001 use backslash escapes to pass in values which contain commas,
2002 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2004 - For KVM, the instance name is the first parameter passed to KVM, so
2005 that it's more visible in the process list
2011 *(Released Tue, 2 Mar 2010)*
2013 Ganeti 2.1 brings many improvements with it. Major changes:
2015 - Added infrastructure to ease automated disk repairs
2016 - Added new daemon to export configuration data in a cheaper way than
2017 using the remote API
2018 - Instance NICs can now be routed instead of being associated with a
2020 - Improved job locking logic to reduce impact of jobs acquiring multiple
2021 locks waiting for other long-running jobs
2023 In-depth implementation details can be found in the Ganeti 2.1 design
2029 - Added chroot hypervisor
2030 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2032 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2033 ``bootloader_path`` and ``bootloader_args``)
2034 - Added the ``use_localtime`` option for the xen-hvm and kvm
2035 hypervisors, and the default value for this has changed to false (in
2036 2.0 xen-hvm always enabled it)
2037 - Added luxi call to submit multiple jobs in one go
2038 - Added cluster initialization option to not modify ``/etc/hosts``
2040 - Added network interface parameters
2041 - Added dry run mode to some LUs
2042 - Added RAPI resources:
2044 - ``/2/instances/[instance_name]/info``
2045 - ``/2/instances/[instance_name]/replace-disks``
2046 - ``/2/nodes/[node_name]/evacuate``
2047 - ``/2/nodes/[node_name]/migrate``
2048 - ``/2/nodes/[node_name]/role``
2049 - ``/2/nodes/[node_name]/storage``
2050 - ``/2/nodes/[node_name]/storage/modify``
2051 - ``/2/nodes/[node_name]/storage/repair``
2053 - Added OpCodes to evacuate or migrate all instances on a node
2054 - Added new command to list storage elements on nodes (``gnt-node
2055 list-storage``) and modify them (``gnt-node modify-storage``)
2056 - Added new ssconf files with master candidate IP address
2057 (``ssconf_master_candidates_ips``), node primary IP address
2058 (``ssconf_node_primary_ips``) and node secondary IP address
2059 (``ssconf_node_secondary_ips``)
2060 - Added ``ganeti-confd`` and a client library to query the Ganeti
2061 configuration via UDP
2062 - Added ability to run hooks after cluster initialization and before
2064 - Added automatic mode for disk replace (``gnt-instance replace-disks
2066 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2067 after catastrophic data-loss
2068 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2070 - Added ``gnt-instance move`` command to move instances
2071 - Added ``gnt-cluster watcher`` command to control watcher
2072 - Added ``gnt-node powercycle`` command to powercycle nodes
2073 - Added new job status field ``lock_status``
2074 - Added parseable error codes to cluster verification (``gnt-cluster
2075 verify --error-codes``) and made output less verbose (use
2076 ``--verbose`` to restore previous behaviour)
2077 - Added UUIDs to the main config entities (cluster, nodes, instances)
2078 - Added support for OS variants
2079 - Added support for hashed passwords in the Ganeti remote API users file
2081 - Added option to specify maximum timeout on instance shutdown
2082 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2083 - Added new helper script to start and stop Ganeti daemons
2084 (``daemon-util``), with the intent to reduce the work necessary to
2085 adjust Ganeti for non-Debian distributions and to start/stop daemons
2087 - Added more unittests
2088 - Fixed critical bug in ganeti-masterd startup
2089 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2090 now customisable at the cluster level for both the KVM and Xen
2091 hypervisors using the new ``migration_port`` parameter
2092 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2093 reinstalling an instance
2094 - Allowed ``@`` in tag names
2095 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2096 - Many documentation updates
2097 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2098 - ``gnt-instance reinstall`` can now reinstall multiple instances
2099 - Updated many command line parameters
2100 - Introduced new OS API version 15
2101 - No longer support a default hypervisor
2102 - Treat virtual LVs as inexistent
2103 - Improved job locking logic to reduce lock contention
2104 - Match instance and node names case insensitively
2105 - Reimplemented bash completion script to be more complete
2112 *(Released Thu, 4 Feb 2010)*
2114 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2115 - Fix a string formatting bug
2116 - Improve safety of the code in some error paths
2117 - Improve data validation in the master of values returned from nodes
2123 *(Released Thu, 17 Dec 2009)*
2125 - Fix security issue due to missing validation of iallocator names; this
2126 allows local and remote execution of arbitrary executables
2127 - Fix failure of gnt-node list during instance removal
2128 - Ship the RAPI documentation in the archive
2134 *(Released Wed, 30 Sep 2009)*
2136 - Fixed many wrong messages
2137 - Fixed a few bugs related to the locking library
2138 - Fixed MAC checking at instance creation time
2139 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2140 - Fixed a few issues related to signal handling in both daemons and
2142 - Fixed the example startup script provided
2143 - Fixed insserv dependencies in the example startup script (patch from
2145 - Fixed handling of drained nodes in the iallocator framework
2146 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2148 - Fixed error related to invalid job IDs in job polling
2149 - Fixed job/opcode persistence on unclean master shutdown
2150 - Fixed handling of partial job processing after unclean master
2152 - Fixed error reporting from LUs, previously all errors were converted
2153 into execution errors
2154 - Fixed error reporting from burnin
2155 - Decreased significantly the memory usage of the job queue
2156 - Optimised slightly multi-job submission
2157 - Optimised slightly opcode loading
2158 - Backported the multi-job submit framework from the development
2159 branch; multi-instance start and stop should be faster
2160 - Added script to clean archived jobs after 21 days; this will reduce
2161 the size of the queue directory
2162 - Added some extra checks in disk size tracking
2163 - Added an example ethers hook script
2164 - Added a cluster parameter that prevents Ganeti from modifying of
2166 - Added more node information to RAPI responses
2167 - Added a ``gnt-job watch`` command that allows following the ouput of a
2169 - Added a bind-address option to ganeti-rapi
2170 - Added more checks to the configuration verify
2171 - Enhanced the burnin script such that some operations can be retried
2173 - Converted instance reinstall to multi-instance model
2179 *(Released Fri, 7 Aug 2009)*
2181 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2182 to allow using the pre-2.0.2 behaviour in activation, if any existing
2183 instances have mismatched disk sizes in the configuration
2184 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2185 any configuration mismatches for disk sizes
2186 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2187 failover to work on two-node clusters
2188 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2190 - Fixed detection of OS script errors in ``gnt-backup export``
2191 - Fixed exit code of ``gnt-backup export``
2197 *(Released Fri, 17 Jul 2009)*
2199 - Added experimental support for stripped logical volumes; this should
2200 enhance performance but comes with a higher complexity in the block
2201 device handling; stripping is only enabled when passing
2202 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2203 affected even in the non-stripped mode
2204 - Improved resiliency against transient failures at the end of DRBD
2205 resyncs, and in general of DRBD resync checks
2206 - Fixed a couple of issues with exports and snapshot errors
2207 - Fixed a couple of issues in instance listing
2208 - Added display of the disk size in ``gnt-instance info``
2209 - Fixed checking for valid OSes in instance creation
2210 - Fixed handling of the "vcpus" parameter in instance listing and in
2211 general of invalid parameters
2212 - Fixed http server library, and thus RAPI, to handle invalid
2213 username/password combinations correctly; this means that now they
2214 report unauthorized for queries too, not only for modifications,
2215 allowing earlier detect of configuration problems
2216 - Added a new "role" node list field, equivalent to the master/master
2217 candidate/drained/offline flags combinations
2218 - Fixed cluster modify and changes of candidate pool size
2219 - Fixed cluster verify error messages for wrong files on regular nodes
2220 - Fixed a couple of issues with node demotion from master candidate role
2221 - Fixed node readd issues
2222 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2223 - Added a new ``--no-voting`` option for masterfailover to fix failover
2224 on two-nodes clusters when the former master node is unreachable
2225 - Added instance reinstall over RAPI
2231 *(Released Tue, 16 Jun 2009)*
2233 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2234 allow re-adding the start in single-user option (regression from 1.2)
2235 - the watcher writes the instance status to a file, to allow monitoring
2236 to report the instance status (from the master) based on cached
2237 results of the watcher's queries; while this can get stale if the
2238 watcher is being locked due to other work on the cluster, this is
2239 still an improvement
2240 - the watcher now also restarts the node daemon and the rapi daemon if
2242 - fixed the watcher to handle full and drained queue cases
2243 - hooks export more instance data in the environment, which helps if
2244 hook scripts need to take action based on the instance's properties
2245 (no longer need to query back into ganeti)
2246 - instance failovers when the instance is stopped do not check for free
2247 RAM, so that failing over a stopped instance is possible in low memory
2249 - rapi uses queries for tags instead of jobs (for less job traffic), and
2250 for cluster tags it won't talk to masterd at all but read them from
2252 - a couple of error handling fixes in RAPI
2253 - drbd handling: improved the error handling of inconsistent disks after
2254 resync to reduce the frequency of "there are some degraded disks for
2255 this instance" messages
2256 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2257 error handling path called a wrong function name)
2263 *(Released Wed, 27 May 2009)*
2265 - no changes from rc5
2271 *(Released Wed, 20 May 2009)*
2273 - fix a couple of bugs (validation, argument checks)
2274 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2275 - some small improvements to RAPI and IAllocator
2276 - make watcher automatically start the master daemon if down
2282 *(Released Mon, 27 Apr 2009)*
2284 - change the OS list to not require locks; this helps with big clusters
2285 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2286 volume group is broken
2287 - ``gnt-instance info``, without any arguments, doesn't run for all
2288 instances anymore; either pass ``--all`` or pass the desired
2289 instances; this helps against mistakes on big clusters where listing
2290 the information for all instances takes a long time
2291 - miscellaneous doc and man pages fixes
2297 *(Released Wed, 8 Apr 2009)*
2299 - Change the internal locking model of some ``gnt-node`` commands, in
2300 order to reduce contention (and blocking of master daemon) when
2301 batching many creation/reinstall jobs
2302 - Fixes to Xen soft reboot
2303 - No longer build documentation at build time, instead distribute it in
2304 the archive, in order to reduce the need for the whole docbook/rst
2311 *(Released Fri, 27 Mar 2009)*
2313 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2314 - Fix watcher startup sequence, improves the behaviour of busy clusters
2315 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2316 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2317 bind address and other places
2318 - Some documentation fixes and updates
2324 *(Released Mon, 2 Mar 2009)*
2326 - More documentation updates, now all docs should be more-or-less
2328 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2330 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2332 - Fix an issue related to $libdir/run/ganeti and cluster creation
2338 *(Released Thu, 19 Feb 2009)*
2340 - Xen PVM and KVM have switched the default value for the instance root
2341 disk to the first partition on the first drive, instead of the whole
2342 drive; this means that the OS installation scripts must be changed
2344 - Man pages have been updated
2345 - RAPI has been switched by default to HTTPS, and the exported functions
2346 should all work correctly
2347 - RAPI v1 has been removed
2348 - Many improvements to the KVM hypervisor
2349 - Block device errors are now better reported
2350 - Many other bugfixes and small improvements
2356 *(Released Mon, 26 Jan 2009)*
2358 - Version 2 is a general rewrite of the code and therefore the
2359 differences are too many to list, see the design document for 2.0 in
2360 the ``doc/`` subdirectory for more details
2361 - In this beta version there is not yet a migration path from 1.2 (there
2362 will be one in the final 2.0 release)
2363 - A few significant changes are:
2365 - all commands are executed by a daemon (``ganeti-masterd``) and the
2366 various ``gnt-*`` commands are just front-ends to it
2367 - all the commands are entered into, and executed from a job queue,
2368 see the ``gnt-job(8)`` manpage
2369 - the RAPI daemon supports read-write operations, secured by basic
2370 HTTP authentication on top of HTTPS
2371 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2372 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2373 to migrate to DRBD 8 first while still running Ganeti 1.2)
2374 - DRBD devices are using statically allocated minor numbers, which
2375 will be assigned to existing instances during the migration process
2376 - there is support for both Xen PVM and Xen HVM instances running on
2378 - KVM virtualization is supported too
2379 - file-based storage has been implemented, which means that it is
2380 possible to run the cluster without LVM and DRBD storage, for
2381 example using a shared filesystem exported from shared storage (and
2382 still have live migration)
2388 *(Released Tue, 13 Jan 2009)*
2390 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2391 - Reuse the old instance mac address by default on instance import, if
2392 the instance name is the same.
2393 - Handle situations in which the node info rpc returns incomplete
2395 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2396 - Improved version of batcher:
2398 - state file support
2399 - instance mac address support
2400 - support for HVM clusters/instances
2402 - Add an option to show the number of cpu sockets and nodes in
2404 - Support OSes that handle more than one version of the OS api (but do
2405 not change the current API in any other way)
2406 - Fix ``gnt-node migrate``
2407 - ``gnt-debug`` man page
2408 - Fixes various more typos and small issues
2409 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2415 *(Released Wed, 24 Sep 2008)*
2417 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2418 type of disk exported to fully virtualized instances.
2419 - provide access to the serial console of HVM instances
2420 - instance auto_balance flag, set by default. If turned off it will
2421 avoid warnings on cluster verify if there is not enough memory to fail
2422 over an instance. in the future it will prevent automatically failing
2423 it over when we will support that.
2424 - batcher tool for instance creation, see ``tools/README.batcher``
2425 - ``gnt-instance reinstall --select-os`` to interactively select a new
2426 operating system when reinstalling an instance.
2427 - when changing the memory amount on instance modify a check has been
2428 added that the instance will be able to start. also warnings are
2429 emitted if the instance will not be able to fail over, if auto_balance
2431 - documentation fixes
2432 - sync fields between ``gnt-instance list/modify/add/import``
2433 - fix a race condition in drbd when the sync speed was set after giving
2434 the device a remote peer.
2440 *(Released Tue, 22 Jul 2008)*
2442 - note: the allowed size and number of tags per object were reduced
2443 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2444 - fixed twisted 8.x compatibility
2445 - fixed ``gnt-instance replace-disks`` with iallocator
2446 - add TCP keepalives on twisted connections to detect restarted nodes
2447 - disk increase support, see ``gnt-instance grow-disk``
2448 - implement bulk node/instance query for RAPI
2449 - add tags in node/instance listing (optional)
2450 - experimental migration (and live migration) support, read the man page
2451 for ``gnt-instance migrate``
2452 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2453 has some small improvements in handling its state file
2459 *(Released Fri, 13 Jun 2008)*
2461 - Experimental readonly, REST-based remote API implementation;
2462 automatically started on master node, TCP port 5080, if enabled by
2463 ``--enable-rapi`` parameter to configure script.
2464 - Instance allocator support. Add and import instance accept a
2465 ``--iallocator`` parameter, and call that instance allocator to decide
2466 which node to use for the instance. The iallocator document describes
2467 what's expected from an allocator script.
2468 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2469 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2470 node is lost there is still enough memory to fail over the instances
2472 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2473 ``gnt-cluster verify``, to check for site-specific compliance. All the
2474 hooks will run, and their output, if any, will be displayed. Any
2475 failing hook will make the verification return an error value.
2476 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2477 primary and secondary interfaces
2478 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2479 that is still declared as part of the cluster and has failed.
2480 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2481 specifying output fields, that just adds the chosen fields to the
2483 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2484 export from the filesystem.
2485 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2486 have been added. Using them you can enable/disable acpi and pae
2487 support, and specify a path for a cd image to be exported to the
2488 instance. These parameters as the name suggest only work on HVM
2490 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2491 PAE support will be set to the previously hardcoded values, but the
2492 (previously hardcoded) path to the CDROM ISO image will be unset and
2493 if required, needs to be set manually with ``gnt-instance modify``
2495 - The address to which an instance's VNC console is bound is now
2496 selectable per-instance, rather than being cluster wide. Of course
2497 this only applies to instances controlled via VNC, so currently just
2498 applies to HVM clusters.
2504 *(Released Mon, 18 Feb 2008)*
2506 - more tweaks to the disk activation code (especially helpful for DRBD)
2507 - change the default ``gnt-instance list`` output format, now there is
2508 one combined status field (see the manpage for the exact values this
2510 - some more fixes for the mac export to hooks change
2511 - make Ganeti not break with DRBD 8.2.x (which changed the version
2512 format in ``/proc/drbd``) (issue 24)
2513 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2514 template, allowing migration from DRBD0.7+MD to DRBD8
2520 *(Released Wed, 30 Jan 2008)*
2522 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2524 - add command aliases infrastructure and a few aliases
2525 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2526 man pages and the ``--help`` option of ``gnt-node
2527 list``/``gnt-instance list``
2528 - fix ``gnt-backup list`` with down nodes (issue 21)
2529 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2530 - fix the dist archive and add a check for including svn/git files in
2532 - some developer-related changes: improve the burnin and the QA suite,
2533 add an upload script for testing during development
2539 *(Released Wed, 16 Jan 2008)*
2541 - experimental HVM support, read the install document, section
2542 "Initializing the cluster"
2543 - allow for the PVM hypervisor per-instance kernel and initrd paths
2544 - add a new command ``gnt-cluster verify-disks`` which uses a new
2545 algorithm to improve the reconnection of the DRBD pairs if the device
2546 on the secondary node has gone away
2547 - make logical volume code auto-activate LVs at disk activation time
2548 - slightly improve the speed of activating disks
2549 - allow specification of the MAC address at instance creation time, and
2550 changing it later via ``gnt-instance modify``
2551 - fix handling of external commands that generate lots of output on
2553 - update documentation with regard to minimum version of DRBD8 supported
2559 *(Released Tue, 4 Dec 2007)*
2561 - Log the ``xm create`` output to the node daemon log on failure (to
2562 help diagnosing the error)
2563 - In debug mode, log all external commands output if failed to the logs
2564 - Change parsing of lvm commands to ignore stderr
2570 *(Released Wed, 28 Nov 2007)*
2572 - Another round of updates to the DRBD 8 code to deal with more failures
2573 in the replace secondary node operation
2574 - Some more logging of failures in disk operations (lvm, drbd)
2575 - A few documentation updates
2582 *(Released Tue, 13 Nov 2007)*
2584 - Change configuration file format from Python's Pickle to JSON.
2585 Upgrading is possible using the cfgupgrade utility.
2586 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2587 faster replace disks and is more stable (DRBD 8 has many improvements
2588 compared to DRBD 0.7)
2589 - Added command line tags support (see man pages for ``gnt-instance``,
2590 ``gnt-node``, ``gnt-cluster``)
2591 - Added instance rename support
2592 - Added multi-instance startup/shutdown
2593 - Added cluster rename support
2594 - Added ``gnt-node evacuate`` to simplify some node operations
2595 - Added instance reboot operation that can speedup reboot as compared to
2597 - Soften the requirement that hostnames are in FQDN format
2598 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2600 - Removed dependency on debian's patched fping that uses the
2601 non-standard ``-S`` option
2602 - Now the OS definitions are searched for in multiple, configurable
2603 paths (easier for distros to package)
2604 - Some changes to the hooks infrastructure (especially the new
2605 post-configuration update hook)
2606 - Other small bugfixes
2608 .. vim: set textwidth=72 syntax=rst :