9 - Networks no longer have a "type" slot, since this information was
10 unused in Ganeti: instead of it tags should be used.
11 - Diskless instances are now externally mirrored (Issue 237). This for
12 now has only been tested in conjunction with explicit target nodes for
14 - The rapi client now has a ``target_node`` option to MigrateInstance.
15 - Fix early exit return code for hbal (Issue 386).
16 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
17 - Fix ``rbd showmapped`` output parsing (Issue 312).
18 - Networks are now referenced indexed by UUID, rather than name. This
19 will require running cfgupgrade, from 2.7.0beta1, if networks are in
21 - The OS environment now includes network information.
22 - Deleting of a network is now disallowed if any instance nic is using
23 it, to prevent dangling references.
24 - External storage is now documented in man pages.
25 - The exclusive_storage flag can now only be set at nodegroup level.
26 - Hbal can now submit an explicit priority with its jobs.
27 - Many network related locking fixes.
28 - Bump up the required pylint version to 0.25.1.
29 - Fix the ``no_remember`` option in RAPI client.
30 - Many ipolicy related tests, qa, and fixes.
31 - Many documentation improvements and fixes.
32 - Fix building with ``--disable-file-storage``.
33 - Fix ``-q`` option in htools, which was broken if passed more than
35 - Some haskell/python interaction improvements and fixes.
36 - Fix iallocator in case of missing LVM storage.
37 - Fix confd config load in case of ``--no-lvm-storage``.
38 - The confd/query functionality is now mentioned in the security
45 *(Released Wed, 6 Feb 2013)*
47 - ``gnt-instance batch-create`` has been changed to use the bulk create
48 opcode from Ganeti. This lead to incompatible changes in the format of
49 the JSON file. It's now not a custom dict anymore but a dict
50 compatible with the ``OpInstanceCreate`` opcode.
51 - Parent directories for file storage need to be listed in
52 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
53 write the file automatically based on old configuration values, but it
54 can not distribute it across all nodes and the file contents should be
55 verified. Use ``gnt-cluster copyfile
56 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
57 upgraded. The reason for requiring this list of paths now is that
58 before it would have been possible to inject new paths via RPC,
59 allowing files to be created in arbitrary locations. The RPC protocol
60 is protected using SSL/X.509 certificates, but as a design principle
61 Ganeti does not permit arbitrary paths to be passed.
62 - The parsing of the variants file for OSes (see
63 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
64 lines and comment lines (starting with ``#``) are ignored for better
66 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
67 longer available. ``gnt-node add`` now invokes a new tool on the
68 destination node, named ``prepare-node-join``, to configure the SSH
69 daemon. Paramiko is no longer necessary to configure nodes' SSH
70 daemons via ``gnt-node add``.
71 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
72 for RAPI users. It allows granting permissions to query for
73 information to a specific user without giving
74 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
75 - A new tool named ``node-cleanup`` has been added. It cleans remains of
76 a cluster from a machine by stopping all daemons, removing
77 certificates and ssconf files. Unless the ``--no-backup`` option is
78 given, copies of the certificates are made.
79 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
80 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
81 the flag is not reset after a master failover.
82 - Instance creations now support the use of opportunistic locking,
83 potentially speeding up the (parallel) creation of multiple instances.
84 This feature is currently only available via the :doc:`RAPI
85 <rapi>` interface and when an instance allocator is used. If the
86 ``opportunistic_locking`` parameter is set the opcode will try to
87 acquire as many locks as possible, but will not wait for any locks
88 held by other opcodes. If not enough resources can be found to
89 allocate the instance, the temporary error code
90 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
91 retried thereafter, with or without opportunistic locking.
92 - The functionality for allocating multiple instances at once has been
93 overhauled and is now also available through :doc:`RAPI <rapi>`.
94 - Man pages can now be included when the documentation is built, in
95 which case the output is in ``doc/man-html``. The configure-time
96 option is ``--enable-manpages-in-doc``. Sphinx 1.0 or higher is
98 - A new htool called ``hroller`` helps scheduling parallel shutdown of
99 nodes on clusters where all instances are shut down, but without ever
100 rebooting both primary and secondary node for an instance at the same
102 - Queries not needing locks or RPC access to the node can now be
103 performed by the confd daemon, making them independent from jobs, and
104 thus faster to execute. This is selectable at configure time.
105 - The KVM hypervisor has been updated with many new hypervisor
106 parameters, including a generic one for passing arbitrary command line
107 values. See a complete list in :manpage:`gnt-instance(8)`.
108 - A new tool, called ``mon-collector``, is the stand-alone executor of
109 the data collectors for a monitoring system. As of this version, it
110 just includes the DRBD data collector, that can be executed by calling
111 ``mon-collector`` using the ``drbd`` parameter. See
112 :manpage:`mon-collector(7)`.
113 - New ``exclusive-storage`` node parameter added. When it's set to true,
114 physical disks are assigned in an exclusive fashion to instances, as
115 documented in :doc:`Partitioned Ganeti <design-partitioned>`.
116 Currently, only instances using the ``plain`` disk template are
118 - New network management functionality to support automatic allocation
119 of IP addresses and managing of network parameters. See
120 :manpage:`gnt-network(8)` for more details.
121 - New external storage backend, to allow managing arbitrary storage
122 systems external to the cluster. See
123 :manpage:`ganeti-extstorage-interface(7)`.
124 - Instance renames of LVM-based instances will now update the LV tags
125 (which can be used to recover the instance-to-LV mapping in case of
127 - ``hbal`` will now exit with status 0 if, during job execution over
128 LUXI, early exit has been requested and all jobs are successful;
129 before, exit status 1 was used, which cannot be differentiated from
136 *(Released Fri, 21 Dec 2012)*
138 Important behaviour change: hbal won't rebalance anymore instances which
139 have the ``auto_balance`` attribute set to false. This was the intention
140 all along, but until now it only skipped those from the N+1 memory
141 reservation (DRBD-specific).
143 A significant number of bug fixes in this release:
145 - Fixed disk adoption interaction with ipolicy checks.
146 - Fixed networking issues when instances are started, stopped or
147 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
148 - Fixed the warning in cluster verify for shared storage instances not
150 - Fixed removal of storage directory on shared file storage (issue 262).
151 - Fixed validation of LVM volume group name in OpClusterSetParams
152 (``gnt-cluster modify``) (issue 285).
153 - Fixed runtime memory increases (``gnt-instance modify -m``).
154 - Fixed live migration under Xen's ``xl`` mode.
155 - Fixed ``gnt-instance console`` with ``xl``.
156 - Fixed building with newer Haskell compiler/libraries.
157 - Fixed PID file writing in Haskell daemons (confd); this prevents
158 restart issues if confd was launched manually (outside of
159 ``daemon-util``) while another copy of it was running
160 - Fixed a type error when doing live migrations with KVM (issue 297) and
161 the error messages for failing migrations have been improved.
162 - Fixed opcode validation for the out-of-band commands (``gnt-node
164 - Fixed a type error when unsetting OS hypervisor parameters (issue
165 311); now it's possible to unset all OS-specific hypervisor
167 - Fixed the ``dry-run`` mode for many operations: verification of
168 results was over-zealous but didn't take into account the ``dry-run``
169 operation, resulting in "wrong" failures.
170 - Fixed bash completion in ``gnt-job list`` when the job queue has
171 hundreds of entries; especially with older ``bash`` versions, this
172 results in significant CPU usage.
174 And lastly, a few other improvements have been made:
176 - Added option to force master-failover without voting (issue 282).
177 - Clarified error message on lock conflict (issue 287).
178 - Logging of newly submitted jobs has been improved (issue 290).
179 - Hostname checks have been made uniform between instance rename and
181 - The ``--submit`` option is now supported by ``gnt-debug delay``.
182 - Shutting down the master daemon by sending SIGTERM now stops it from
183 processing jobs waiting for locks; instead, those jobs will be started
184 once again after the master daemon is started the next time (issue
186 - Support for Xen's ``xl`` program has been improved (besides the fixes
188 - Reduced logging noise in the Haskell confd daemon (only show one log
189 entry for each config reload, instead of two).
190 - Several man page updates and typo fixes.
196 *(Released Fri, 12 Oct 2012)*
198 A small bugfix release. Among the bugs fixed:
200 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
201 made the command unusable.
202 - Commands that issue many jobs don't fail anymore just because some jobs
203 take so long that other jobs are archived.
204 - Failures during ``gnt-instance reinstall`` are reflected by the exit
206 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
208 - When ``always_failover`` is set, ``--allow-failover`` is not required
209 in migrate commands anymore.
210 - ``bash_completion`` works even if extglob is disabled.
211 - Fixed bug with locks that made failover for RDB-based instances fail.
212 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
213 a random node instead of one based on the allocator metric.
214 - Support for newer versions of pylint and pep8.
215 - Hail doesn't fail anymore when trying to add an instance of type
216 ``file``, ``sharedfile`` or ``rbd``.
217 - Added new Makefile target to rebuild the whole distribution, so that
218 all files are included.
224 *(Released Fri, 27 Jul 2012)*
227 .. attention:: The ``LUXI`` protocol has been made more consistent
228 regarding its handling of command arguments. This, however, leads to
229 incompatibility issues with previous versions. Please ensure that you
230 restart Ganeti daemons soon after the upgrade, otherwise most
231 ``LUXI`` calls (job submission, setting/resetting the drain flag,
232 pausing/resuming the watcher, cancelling and archiving jobs, querying
233 the cluster configuration) will fail.
242 The current ``admin_up`` field, which used to denote whether an instance
243 should be running or not, has been removed. Instead, ``admin_state`` is
244 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
246 The rational behind this is that an instance being “down” can have
249 - it could be down during a reboot
250 - it could be temporarily be down for a reinstall
251 - or it could be down because it is deprecated and kept just for its
254 The previous Boolean state was making it difficult to do capacity
255 calculations: should Ganeti reserve memory for a down instance? Now, the
256 tri-state field makes it clear:
258 - in ``up`` and ``down`` state, all resources are reserved for the
259 instance, and it can be at any time brought up if it is down
260 - in ``offline`` state, only disk space is reserved for it, but not
263 The field can have an extra use: since the transition between ``up`` and
264 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
265 transition between ``offline`` and ``down`` is done via ``gnt-instance
266 modify``, it is possible to given different rights to users. For
267 example, owners of an instance could be allowed to start/stop it, but
268 not transition it out of the offline state.
270 Instance policies and specs
271 +++++++++++++++++++++++++++
273 In previous Ganeti versions, an instance creation request was not
274 limited on the minimum size and on the maximum size just by the cluster
275 resources. As such, any policy could be implemented only in third-party
276 clients (RAPI clients, or shell wrappers over ``gnt-*``
277 tools). Furthermore, calculating cluster capacity via ``hspace`` again
278 required external input with regards to instance sizes.
280 In order to improve these workflows and to allow for example better
281 per-node group differentiation, we introduced instance specs, which
284 - minimum instance disk size, disk count, memory size, cpu count
285 - maximum values for the above metrics
286 - and “standard” values (used in ``hspace`` to calculate the standard
289 The minimum/maximum values can be also customised at node-group level,
290 for example allowing more powerful hardware to support bigger instance
293 Beside the instance specs, there are a few other settings belonging to
294 the instance policy framework. It is possible now to customise, per
295 cluster and node-group:
297 - the list of allowed disk templates
298 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
299 - the maximum ratio of instance to spindles (see below for more
300 information) for local storage
302 All these together should allow all tools that talk to Ganeti to know
303 what are the ranges of allowed values for instances and the
304 over-subscription that is allowed.
306 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
307 instance configuration, and the physical CPU configuration from the
308 node. For the spindle ratios however, we didn't track before these
309 values, so new parameters have been added:
311 - a new node parameter ``spindle_count``, defaults to 1, customisable at
312 node group or node level
313 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
315 Note that spindles in this context doesn't need to mean actual
316 mechanical hard-drives; it's just a relative number for both the node
317 I/O capacity and instance I/O consumption.
319 Instance migration behaviour
320 ++++++++++++++++++++++++++++
322 While live-migration is in general desirable over failover, it is
323 possible that for some workloads it is actually worse, due to the
324 variable time of the “suspend” phase during live migration.
326 To allow the tools to work consistently over such instances (without
327 having to hard-code instance names), a new backend parameter
328 ``always_failover`` has been added to control the migration/failover
329 behaviour. When set to True, all migration requests for an instance will
330 instead fall-back to failover.
332 Instance memory ballooning
333 ++++++++++++++++++++++++++
335 Initial support for memory ballooning has been added. The memory for an
336 instance is no longer fixed (backend parameter ``memory``), but instead
337 can vary between minimum and maximum values (backend parameters
338 ``minmem`` and ``maxmem``). Currently we only change an instance's
341 - live migrating or failing over and instance and the target node
342 doesn't have enough memory
343 - user requests changing the memory via ``gnt-instance modify
349 In order to control the use of specific CPUs by instance, support for
350 controlling CPU pinning has been added for the Xen, HVM and LXC
351 hypervisors. This is controlled by a new hypervisor parameter
352 ``cpu_mask``; details about possible values for this are in the
353 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
354 VCPU-to-CPU mapping) form will work well only when all nodes in your
355 cluster have the same amount of CPUs.
360 Another area in which Ganeti was not customisable were the parameters
361 used for storage configuration, e.g. how many stripes to use for LVM,
362 DRBD resync configuration, etc.
364 To improve this area, we've added disks parameters, which are
365 customisable at cluster and node group level, and which allow to
366 specify various parameters for disks (DRBD has the most parameters
367 currently), for example:
369 - DRBD resync algorithm and parameters (e.g. speed)
370 - the default VG for meta-data volumes for DRBD
371 - number of stripes for LVM (plain disk template)
374 These parameters can be modified via ``gnt-cluster modify -D …`` and
375 ``gnt-group modify -D …``, and are used at either instance creation (in
376 case of LVM stripes, for example) or at disk “activation” time
379 Rados block device support
380 ++++++++++++++++++++++++++
382 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
383 denoted by the ``rbd`` disk template type. This is considered
384 experimental, feedback is welcome. For details on configuring it, see
385 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
390 The existing master IP functionality works well only in simple setups (a
391 single network shared by all nodes); however, if nodes belong to
392 different networks, then the ``/32`` setup and lack of routing
393 information is not enough.
395 To allow the master IP to function well in more complex cases, the
396 system was reworked as follows:
398 - a master IP netmask setting has been added
399 - the master IP activation/turn-down code was moved from the node daemon
401 - whether to run the Ganeti-supplied master IP script or a user-supplied
402 on is a ``gnt-cluster init`` setting
404 Details about the location of the standard and custom setup scripts are
405 in the man page :manpage:`gnt-cluster(8)`; for information about the
406 setup script protocol, look at the Ganeti-supplied script.
411 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
414 It is now possible to use TLS-protected connections, and when renewing
415 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
416 it is now possible to specify spice or spice CA certificates. Also, it
417 is possible to configure a password for SPICE sessions via the
418 hypervisor parameter ``spice_password_file``.
420 There are also new parameters to control the compression and streaming
421 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
422 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
423 for the spice parameters.
425 Lastly, it is now possible to see the SPICE connection information via
426 ``gnt-instance console``.
431 A new tool (``tools/ovfconverter``) has been added that supports
432 conversion between Ganeti and the `Open Virtualization Format
433 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
436 This relies on the ``qemu-img`` tool to convert the disk formats, so the
437 actual compatibility with other virtualization solutions depends on it.
442 The configuration query daemon (``ganeti-confd``) is now optional, and
443 has been rewritten in Haskell; whether to use the daemon at all, use the
444 Python (default) or the Haskell version is selectable at configure time
445 via the ``--enable-confd`` parameter, which can take one of the
446 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
447 daemon will result in a smaller footprint; for larger systems, we
448 welcome feedback on the Haskell version which might become the default
451 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
452 daemon running. The Python version doesn't implement the new call.
455 User interface changes
456 ~~~~~~~~~~~~~~~~~~~~~~
458 We have replaced the ``--disks`` option of ``gnt-instance
459 replace-disks`` with a more flexible ``--disk`` option, which allows
460 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
461 disk size and mode can be changed upon recreation (via ``gnt-instance
462 recreate-disks``, which accepts the same ``--disk`` option).
464 As many people are used to a ``show`` command, we have added that as an
465 alias to ``info`` on all ``gnt-*`` commands.
467 The ``gnt-instance grow-disk`` command has a new mode in which it can
468 accept the target size of the disk, instead of the delta; this can be
469 more safe since two runs in absolute mode will be idempotent, and
470 sometimes it's also easier to specify the desired size directly.
472 Also the handling of instances with regard to offline secondaries has
473 been improved. Instance operations should not fail because one of it's
474 secondary nodes is offline, even though it's safe to proceed.
476 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
477 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
478 minors to instance name.
483 RAPI coverage has improved, with (for example) new resources for
484 recreate-disks, node power-cycle, etc.
489 There is partial support for ``xl`` in the Xen hypervisor; feedback is
492 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
493 whether to still support Python 2.4 or move to Python 2.6 as minimum
496 Support for Fedora has been slightly improved; the provided example
497 init.d script should work better on it and the INSTALL file should
498 document the needed dependencies.
503 The deprecated ``QueryLocks`` LUXI request has been removed. Use
504 ``Query(what=QR_LOCK, ...)`` instead.
506 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
507 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
508 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
509 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
510 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
512 RAPI client: ``CertificateError`` now derives from
513 ``GanetiApiError``. This should make it more easy to handle Ganeti
516 Deprecation warnings due to PyCrypto/paramiko import in
517 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
518 make sure to run an up-to-date paramiko version, if you use this tool.
520 The QA scripts now depend on Python 2.5 or above (the main code base
521 still works with Python 2.4).
523 The configuration file (``config.data``) is now written without
524 indentation for performance reasons; if you want to edit it, it can be
525 re-formatted via ``tools/fmtjson``.
527 A number of bugs has been fixed in the cluster merge tool.
529 ``x509`` certification verification (used in import-export) has been
530 changed to allow the same clock skew as permitted by the cluster
531 verification. This will remove some rare but hard to diagnose errors in
538 *(Released Thu, 19 Jul 2012)*
540 Very few changes from rc4 to the final release, only bugfixes:
542 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
543 instance, fix CDROM booting for KVM instances)
544 - fixed node group modification of node parameters
545 - fixed issue in LUClusterVerifyGroup with multi-group clusters
546 - fixed generation of bash completion to ensure a stable ordering
553 *(Released Fri, 13 Jul 2012)*
555 Third release candidate for 2.6. The following changes were done from
558 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
559 - Fixed an inconsistency in the LUXI protocol with the provided
560 arguments (NOT backwards compatible)
561 - Fixed a bug with node groups ipolicy where ``min`` was greater than
562 the cluster ``std`` value
563 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
564 easier instance debugging on nodes (requires ``hconfd`` to work)
570 *(Released Tue, 03 Jul 2012)*
572 Second release candidate for 2.6. The following changes were done from
575 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
577 - Fixed a RAPI regression on missing beparams/memory
578 - Fixed redistribution of files on offline nodes
579 - Added possibility to run activate-disks even though secondaries are
580 offline. With this change it relaxes also the strictness on some other
581 commands which use activate disks internally:
582 * ``gnt-instance start|reboot|rename|backup|export``
583 - Made it possible to remove safely an instance if its secondaries are
585 - Made it possible to reinstall even though secondaries are offline
591 *(Released Mon, 25 Jun 2012)*
593 First release candidate for 2.6. The following changes were done from
596 - Fixed bugs with disk parameters and ``rbd`` templates as well as
598 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
599 behaviour. It supports now the modify operation
600 - ``hcheck`` implemented to analyze cluster health and possibility of
601 improving health by rebalance
602 - ``hbal`` has been improved in dealing with split instances
608 *(Released Mon, 11 Jun 2012)*
610 Second beta release of 2.6. The following changes were done from beta2
613 - Fixed ``daemon-util`` with non-root user models
614 - Fixed creation of plain instances with ``--no-wait-for-sync``
615 - Fix wrong iv_names when running ``cfgupgrade``
616 - Export more information in RAPI group queries
617 - Fixed bug when changing instance network interfaces
618 - Extended burnin to do NIC changes
619 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
620 - Changed default for DRBD barriers
621 - Fixed DRBD error reporting for syncer rate
622 - Verify the options on disk parameters
624 And of course various fixes to documentation and improved unittests and
631 *(Released Wed, 23 May 2012)*
633 First beta release of 2.6. The following changes were done from beta1 to
636 - integrated patch for distributions without ``start-stop-daemon``
637 - adapted example init.d script to work on Fedora
638 - fixed log handling in Haskell daemons
639 - adapted checks in the watcher for pycurl linked against libnss
640 - add partial support for ``xl`` instead of ``xm`` for Xen
641 - fixed a type issue in cluster verification
642 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
645 Plus integrated fixes from the 2.5 branch:
647 - fixed ``kvm-ifup`` to use ``/bin/bash``
648 - fixed parallel build failures
649 - KVM live migration when using a custom keymap
655 *(Released Tue, 24 Jul 2012)*
657 A small bugfix release, with no new features:
659 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
660 different default shell (e.g. Debian, Ubuntu)
661 - fixed KVM startup and live migration with a custom keymap (fixes Issue
662 243 and Debian bug #650664)
663 - fixed compatibility with KVM versions that don't support multiple boot
664 devices (fixes Issue 230 and Debian bug #624256)
666 Additionally, a few fixes were done to the build system (fixed parallel
667 build failures) and to the unittests (fixed race condition in test for
668 FileID functions, and the default enable/disable mode for QA test is now
675 *(Released Fri, 11 May 2012)*
677 A small bugfix release.
679 The main issues solved are on the topic of compatibility with newer LVM
682 - fixed parsing of ``lv_attr`` field
683 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
684 the ``--force`` flag is needed
686 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
687 to accept kernel 3.x too (was hardcoded to 2.6.*).
689 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
690 of hook results and that also made display of other errors suboptimal
691 was fixed; the code behaves now like 2.4 and earlier.
693 Another change in 2.5, the cleanup of the OS scripts environment, is too
694 aggressive: it removed even the ``PATH`` variable, which requires the OS
695 scripts to *always* need to export it. Since this is a bit too strict,
696 we now export a minimal PATH, the same that we export for hooks.
698 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
699 integrated into this release.
701 Finally, a few other miscellaneous changes were done (no new features,
702 just small improvements):
704 - Fix ``gnt-group --help`` display
705 - Fix hardcoded Xen kernel path
706 - Fix grow-disk handling of invalid units
707 - Update synopsis for ``gnt-cluster repair-disk-sizes``
708 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
714 *(Released Thu, 12 Apr 2012)*
716 Incompatible/important changes and bugfixes
717 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
719 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
720 resource's ``ip_check`` parameter changed from ``True`` to ``False``
721 to match the underlying LUXI interface.
722 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
723 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
724 not maintain backwards-compatibility as the underlying operation
725 changed in an incompatible way. The RAPI client can talk to old
726 servers, but it needs to be told so as the return value changed.
727 - When creating file-based instances via RAPI, the ``file_driver``
728 parameter no longer defaults to ``loop`` and must be specified.
729 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
731 - Support for the undocumented and deprecated RAPI instance creation
732 request format version 0 has been dropped. Use version 1, supported
733 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
734 - Pyparsing 1.4.6 or above is required, see :doc:`installation
735 documentation <install>`.
736 - The "cluster-verify" hooks are now executed per group by the
737 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
738 if you just run ``gnt-cluster verify``, which generates one opcode per
740 - The environment as passed to the OS scripts is cleared, and thus no
741 environment variables defined in the node daemon's environment will be
742 inherited by the scripts.
743 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
745 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
746 support operations involving multiple node groups.
747 - Offline nodes are ignored when failing over an instance.
748 - Support for KVM version 1.0, which changed the version reporting format
750 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
752 - ``Makefile`` is now compatible with Automake 1.11.2
753 - Includes all bugfixes made in the 2.4 series
758 - The ganeti-htools project has been merged into the ganeti-core source
759 tree and will be built as part of Ganeti (see :doc:`install-quick`).
760 - Implemented support for :doc:`shared storage <design-shared-storage>`.
761 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
762 GPT-style partition tables (requires `parted
763 <http://www.gnu.org/s/parted/>`_).
764 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
765 - Allowed adding tags on instance creation.
766 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
767 - Allow instances to be started in a paused state, enabling the user to
768 see the complete console output on boot using the console.
769 - Added new hypervisor flag to control default reboot behaviour
770 (``reboot_behavior``).
771 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
772 - Improved out-of-band management support:
774 - Added ``gnt-node health`` command reporting the health status of
776 - Added ``gnt-node power`` command to manage power status of nodes.
777 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
779 - Instance migration can fall back to failover if instance is not
781 - Filters can be used when listing nodes, instances, groups and locks;
782 see :manpage:`ganeti(7)` manpage.
783 - Added post-execution status as variables to :doc:`hooks <hooks>`
785 - Instance tags are exported/imported together with the instance.
786 - When given an explicit job ID, ``gnt-job info`` will work for archived
788 - Jobs can define dependencies on other jobs (not yet supported via
789 RAPI or command line, but used by internal commands and usable via
792 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
795 - Instance failover is now available as a RAPI resource
796 (``/2/instances/[instance_name]/failover``).
797 - ``gnt-instance info`` defaults to static information if primary node
799 - Opcodes have a new ``comment`` attribute.
800 - Added basic SPICE support to KVM hypervisor.
801 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
803 Node group improvements
804 ~~~~~~~~~~~~~~~~~~~~~~~
806 - ``gnt-cluster verify`` has been modified to check groups separately,
807 thereby improving performance.
808 - Node group support has been added to ``gnt-cluster verify-disks``,
809 which now operates per node group.
810 - Watcher has been changed to work better with node groups.
812 - One process and state file per node group.
813 - Slow watcher in one group doesn't block other group's watcher.
815 - Added new command, ``gnt-group evacuate``, to move all instances in a
816 node group to other groups.
817 - Added ``gnt-instance change-group`` to move an instance to another
819 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
820 per-group operations.
821 - Node groups can be tagged.
822 - Some operations switch from an exclusive to a shared lock as soon as
824 - Instance's primary and secondary nodes' groups are now available as
825 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
826 and ``snodes.group.uuid``).
831 - Numerous updates to documentation and manpages.
833 - :doc:`RAPI <rapi>` documentation now has detailed parameter
835 - Some opcode/job results are now also documented, see :doc:`RAPI
838 - A lockset's internal lock is now also visible in lock monitor.
839 - Log messages from job queue workers now contain information about the
840 opcode they're processing.
841 - ``gnt-instance console`` no longer requires the instance lock.
842 - A short delay when waiting for job changes reduces the number of LUXI
843 requests significantly.
844 - DRBD metadata volumes are overwritten with zeros during disk creation.
845 - Out-of-band commands no longer acquire the cluster lock in exclusive
847 - ``devel/upload`` now uses correct permissions for directories.
853 *(Released Fri, 23 Mar 2012)*
855 This was the sixth release candidate of the 2.5 series.
861 *(Released Mon, 9 Jan 2012)*
863 This was the fifth release candidate of the 2.5 series.
869 *(Released Thu, 27 Oct 2011)*
871 This was the fourth release candidate of the 2.5 series.
877 *(Released Wed, 26 Oct 2011)*
879 This was the third release candidate of the 2.5 series.
885 *(Released Tue, 18 Oct 2011)*
887 This was the second release candidate of the 2.5 series.
893 *(Released Tue, 4 Oct 2011)*
895 This was the first release candidate of the 2.5 series.
901 *(Released Wed, 31 Aug 2011)*
903 This was the third beta release of the 2.5 series.
909 *(Released Mon, 22 Aug 2011)*
911 This was the second beta release of the 2.5 series.
917 *(Released Fri, 12 Aug 2011)*
919 This was the first beta release of the 2.5 series.
925 *(Released Thu, 27 Oct 2011)*
927 - Fixed bug when parsing command line parameter values ending in
929 - Fixed assertion error after unclean master shutdown
930 - Disable HTTP client pool for RPC, significantly reducing memory usage
932 - Fixed queue archive creation with wrong permissions
938 *(Released Tue, 23 Aug 2011)*
942 - Fixed documentation for importing with ``--src-dir`` option
943 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
944 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
950 *(Released Fri, 5 Aug 2011)*
952 Many bug-fixes and a few small features:
954 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
955 issues when you tried to add an instance with two MAC addresses in one
957 - KVM: fixed per-instance stored UID value
958 - KVM: configure bridged NICs at migration start
959 - KVM: Fix a bug where instance will not start with never KVM versions
961 - Added OS search path to ``gnt-cluster info``
962 - Fixed an issue with ``file_storage_dir`` where you were forced to
963 provide an absolute path, but the documentation states it is a
964 relative path, the documentation was right
965 - Added a new parameter to instance stop/start called ``--no-remember``
966 that will make the state change to not be remembered
967 - Implemented ``no_remember`` at RAPI level
968 - Improved the documentation
969 - Node evacuation: don't call IAllocator if node is already empty
970 - Fixed bug in DRBD8 replace disks on current nodes
971 - Fixed bug in recreate-disks for DRBD instances
972 - Moved assertion checking locks in ``gnt-instance replace-disks``
973 causing it to abort with not owning the right locks for some situation
974 - Job queue: Fixed potential race condition when cancelling queued jobs
975 - Fixed off-by-one bug in job serial generation
976 - ``gnt-node volumes``: Fix instance names
977 - Fixed aliases in bash completion
978 - Fixed a bug in reopening log files after being sent a SIGHUP
979 - Added a flag to burnin to allow specifying VCPU count
980 - Bugfixes to non-root Ganeti configuration
986 *(Released Thu, 12 May 2011)*
988 Many bug-fixes and a few new small features:
990 - Fixed a bug related to log opening failures
991 - Fixed a bug in instance listing with orphan instances
992 - Fixed a bug which prevented resetting the cluster-level node parameter
993 ``oob_program`` to the default
994 - Many fixes related to the ``cluster-merge`` tool
995 - Fixed a race condition in the lock monitor, which caused failures
996 during (at least) creation of many instances in parallel
997 - Improved output for gnt-job info
998 - Removed the quiet flag on some ssh calls which prevented debugging
1000 - Improved the N+1 failure messages in cluster verify by actually
1001 showing the memory values (needed and available)
1002 - Increased lock attempt timeouts so that when executing long operations
1003 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1004 too early and thus prevent the use of the 'fair' mechanism
1005 - Changed instance query data (``gnt-instance info``) to not acquire
1006 locks unless needed, thus allowing its use on locked instance if only
1007 static information is asked for
1008 - Improved behaviour with filesystems that do not support rename on an
1010 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1011 kept locks on all nodes during the wipe, which is unneeded
1012 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1013 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1014 10GiB) which caused the wipe to fail right at the end in some cases
1015 - Fixed master IP activation when doing master failover with no-voting
1016 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1017 the master node itself
1018 - Fixed potential data-loss in under disk full conditions, where Ganeti
1019 wouldn't check correctly the return code and would consider
1020 partially-written files 'correct'
1021 - Fixed bug related to multiple VGs and DRBD disk replacing
1022 - Added new disk parameter ``metavg`` that allows placement of the meta
1023 device for DRBD in a different volume group
1024 - Fixed error handling in the node daemon when the system libc doesn't
1025 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1026 - Fixed lock release during replace-disks, which kept cluster-wide locks
1027 when doing disk replaces with an iallocator script
1028 - Added check for missing bridges in cluster verify
1029 - Handle EPIPE errors while writing to the terminal better, so that
1030 piping the output to e.g. ``less`` doesn't cause a backtrace
1031 - Fixed rare case where a ^C during Luxi calls could have been
1032 interpreted as server errors, instead of simply terminating
1033 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1035 - Added a few more parameters to the KVM hypervisor, allowing a second
1036 CDROM, custom disk type for CDROMs and a floppy image
1037 - Removed redundant message in instance rename when the name is given
1039 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1040 disks on new nodes, allowing recreation when the original instance
1041 nodes are completely gone
1042 - Added option when converting disk templates to DRBD to skip waiting
1043 for the resync, in order to make the instance available sooner
1044 - Added two new variables to the OS scripts environment (containing the
1046 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1047 to allow use of ``pvgrub`` as bootloader
1048 - Changed the instance memory modifications to only check out-of-memory
1049 conditions on memory increases, and turned the secondary node warnings
1050 into errors (they can still be overridden via ``--force``)
1051 - Fixed the handling of a corner case when the Python installation gets
1052 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1053 to execute a command that doesn't exist
1054 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1055 node of the instance returned failures during instance shutdown; this
1056 adds the option ``--ignore-consistency`` to gnt-instance move
1058 And as usual, various improvements to the error messages, documentation
1065 *(Released Wed, 09 Mar 2011)*
1067 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1068 the RAPI users file if run twice (even with ``--dry-run``).
1070 The release fixes that bug (nothing else changed).
1076 *(Released Mon, 07 Mar 2011)*
1078 Final 2.4.0 release. Just a few small fixes:
1080 - Fixed RAPI node evacuate
1081 - Fixed the kvm-ifup script
1082 - Fixed internal error handling for special job cases
1083 - Updated man page to specify the escaping feature for options
1089 *(Released Mon, 28 Feb 2011)*
1091 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1092 that this feature wiped the disks of the wrong instance, leading to loss
1097 - Fixed title of query field containing instance name
1098 - Expanded the glossary in the documentation
1099 - Fixed one unittest (internal issue)
1105 *(Released Mon, 21 Feb 2011)*
1107 A number of bug fixes plus just a couple functionality changes.
1109 On the user-visible side, the ``gnt-* list`` command output has changed
1110 with respect to "special" field states. The current rc1 style of display
1111 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1112 the default output mode special fields states are displayed as follows:
1114 - Offline resource: ``*``
1115 - Unavailable/not applicable: ``-``
1116 - Data missing (RPC failure): ``?``
1117 - Unknown field: ``??``
1119 Another user-visible change is the addition of ``--force-join`` to
1124 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1125 - Fixed regression in RAPI/instance reinstall where all parameters were
1126 required (instead of optional)
1127 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1128 - Fixed iallocator usage (offline nodes were not considered offline)
1129 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1130 - Fixed hypervisor and OS parameter validation with respect to
1131 non-vm_capable nodes
1132 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1134 - Fixed ``tools/listrunner`` with respect to agent-based usage
1140 *(Released Fri, 4 Feb 2011)*
1142 Many changes and fixes since the beta1 release. While there were some
1143 internal changes, the code has been mostly stabilised for the RC
1146 Note: the dumb allocator was removed in this release, as it was not kept
1147 up-to-date with the IAllocator protocol changes. It is recommended to
1148 use the ``hail`` command from the ganeti-htools package.
1150 Note: the 2.4 and up versions of Ganeti are not compatible with the
1151 0.2.x branch of ganeti-htools. You need to upgrade to
1152 ganeti-htools-0.3.0 (or later).
1154 Regressions fixed from 2.3
1155 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1157 - Fixed the ``gnt-cluster verify-disks`` command
1158 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1160 - Fixed disk adoption breakage
1161 - Fixed wrong headers in instance listing for field aliases
1166 - Fixed corner case in KVM handling of NICs
1167 - Fixed many cases of wrong handling of non-vm_capable nodes
1168 - Fixed a bug where a missing instance symlink was not possible to
1169 recreate with any ``gnt-*`` command (now ``gnt-instance
1170 activate-disks`` does it)
1171 - Fixed the volume group name as reported by ``gnt-cluster
1173 - Increased timeouts for the import-export code, hopefully leading to
1174 fewer aborts due network or instance timeouts
1175 - Fixed bug in ``gnt-node list-storage``
1176 - Fixed bug where not all daemons were started on cluster
1177 initialisation, but only at the first watcher run
1178 - Fixed many bugs in the OOB implementation
1179 - Fixed watcher behaviour in presence of instances with offline
1181 - Fixed instance list output for instances running on the wrong node
1182 - a few fixes to the cluster-merge tool, but it still cannot merge
1183 multi-node groups (currently it is not recommended to use this tool)
1189 - Improved network configuration for the KVM hypervisor
1190 - Added e1000 as a supported NIC for Xen-HVM
1191 - Improved the lvmstrap tool to also be able to use partitions, as
1192 opposed to full disks
1193 - Improved speed of disk wiping (the cluster parameter
1194 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1195 of instance creations
1196 - Added documentation for the OS parameters
1197 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1198 hypervisor is not responding
1199 - Added display of blacklisted and hidden OS information in
1200 ``gnt-cluster info``
1201 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1202 NIC and node parameters, which might create problems with currently
1203 invalid (but undetected) configuration files, but prevents validation
1204 failures when unrelated parameters are modified
1205 - Changed cluster initialisation to wait for the master daemon to become
1207 - Expanded the RAPI interface:
1209 - Added config redistribution resource
1210 - Added activation/deactivation of instance disks
1211 - Added export of console information
1213 - Implemented log file reopening on SIGHUP, which allows using
1214 logrotate(8) for the Ganeti log files
1215 - Added a basic OOB helper script as an example
1221 *(Released Fri, 14 Jan 2011)*
1226 - Fixed timezone issues when formatting timestamps
1227 - Added support for node groups, available via ``gnt-group`` and other
1229 - Added out-of-band framework and management, see :doc:`design
1230 document <design-oob>`
1231 - Removed support for roman numbers from ``gnt-node list`` and
1232 ``gnt-instance list``.
1233 - Allowed modification of master network interface via ``gnt-cluster
1234 modify --master-netdev``
1235 - Accept offline secondaries while shutting down instance disks
1236 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1237 - Added support for multiple LVM volume groups
1238 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1240 - Added commands to list available fields:
1242 - ``gnt-node list-fields``
1243 - ``gnt-group list-fields``
1244 - ``gnt-instance list-fields``
1246 - Updated documentation and man pages
1251 - Moved ``rapi_users`` file into separate directory, now named
1252 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1254 - Added new tool for running commands on many machines,
1255 ``tools/ganeti-listrunner``
1256 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1257 improving the ``gnt-instance console`` output
1258 - Allowed customisation of disk index separator at ``configure`` time
1259 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1260 - Added support for non-partitioned md disks in ``lvmstrap``
1261 - Added script to gracefully power off KVM instances
1262 - Split ``utils`` module into smaller parts
1263 - Changed query operations to return more detailed information, e.g.
1264 whether an information is unavailable due to an offline node. To use
1265 this new functionality, the LUXI call ``Query`` must be used. Field
1266 information is now stored by the master daemon and can be retrieved
1267 using ``QueryFields``. Instances, nodes and groups can also be queried
1268 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1269 exposed via RAPI). The following commands make use of this
1270 infrastructure change:
1272 - ``gnt-group list``
1273 - ``gnt-group list-fields``
1275 - ``gnt-node list-fields``
1276 - ``gnt-instance list``
1277 - ``gnt-instance list-fields``
1278 - ``gnt-debug locks``
1283 - New RAPI resources (see :doc:`rapi`):
1287 - ``/2/groups/[group_name]``
1288 - ``/2/groups/[group_name]/assign-nodes``
1289 - ``/2/groups/[group_name]/modify``
1290 - ``/2/groups/[group_name]/rename``
1291 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1295 - Implemented ``no_install`` for instance creation
1296 - Implemented OS parameters for instance reinstallation, allowing
1297 use of special settings on reinstallation (e.g. for preserving data)
1302 - Added IPv6 support in import/export
1303 - Pause DRBD synchronization while wiping disks on instance creation
1304 - Updated unittests and QA scripts
1305 - Improved network parameters passed to KVM
1306 - Converted man pages from docbook to reStructuredText
1312 *(Released Mon, 20 Dec 2010)*
1314 Released version 2.3.1~rc1 without any changes.
1320 *(Released Wed, 1 Dec 2010)*
1322 - impexpd: Disable OpenSSL compression in socat if possible (backport
1323 from master, commit e90739d625b, see :doc:`installation guide
1324 <install-quick>` for details)
1325 - Changed unittest coverage report to exclude test scripts
1326 - Added script to check version format
1332 *(Released Wed, 1 Dec 2010)*
1334 Released version 2.3.0~rc1 without any changes.
1340 *(Released Fri, 19 Nov 2010)*
1342 A number of bugfixes and documentation updates:
1344 - Update ganeti-os-interface documentation
1345 - Fixed a bug related to duplicate MACs or similar items which should be
1347 - Fix breakage in OS state modify
1348 - Reinstall instance: disallow offline secondaries (fixes bug related to
1349 OS changing but reinstall failing)
1350 - plus all the other fixes between 2.2.1 and 2.2.2
1356 *(Released Tue, 2 Nov 2010)*
1358 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1359 - Fixed master failover race with watcher
1360 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1362 - Accept previously stopped instance for export with instance removal
1363 - Simplify and extend the environment variables for instance OS scripts
1364 - Added new node flags, ``master_capable`` and ``vm_capable``
1365 - Added optional instance disk wiping prior during allocation. This is a
1366 cluster-wide option and can be set/modified using
1367 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1368 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1369 :doc:`install-quick`
1370 - Added a new watcher option (``--ignore-pause``)
1371 - Added option to ignore offline node on instance start/stop
1372 (``--ignore-offline``)
1373 - Allow overriding OS parameters with ``gnt-instance reinstall``
1374 - Added ability to change node's secondary IP address using ``gnt-node
1376 - Implemented privilege separation for all daemons except
1377 ``ganeti-noded``, see ``configure`` options
1378 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1380 - Implemented job priorities (see ``ganeti(7)`` manpage)
1381 - Ignore failures while shutting down instances during failover from
1383 - Exit daemon's bootstrap process only once daemon is ready
1384 - Export more information via ``LUInstanceQuery``/remote API
1385 - Improved documentation, QA and unittests
1386 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1387 restart if the file was created or changed
1388 - Added LUXI protocol version sent with each request and response,
1389 allowing detection of server/client mismatches
1390 - Moved the Python scripts among gnt-* and ganeti-* into modules
1391 - Moved all code related to setting up SSH to an external script,
1393 - Infrastructure changes for node group support in future versions
1399 *(Released Fri, 19 Nov 2010)*
1401 A few small bugs fixed, and some improvements to the build system:
1403 - Fix documentation regarding conversion to drbd
1404 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1406 - Fix error handling in node modify with multiple changes
1407 - Allow remote imports without checked names
1413 *(Released Tue, 19 Oct 2010)*
1415 - Disable SSL session ID cache in RPC client
1421 *(Released Thu, 14 Oct 2010)*
1423 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1424 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1426 - Fix problems with interaction between Curl and Python's HTTP server,
1427 resulting in increased speed in many RPC calls
1428 - Improve our release script to prevent breakage with older aclocal and
1435 *(Released Thu, 7 Oct 2010)*
1437 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1438 value retrieved from master
1439 - Added support for blacklisted or hidden OS definitions
1440 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1441 - Added support for -mem-path in KVM hypervisor abstraction layer
1442 - Allow overriding instance parameters in tool for inter-cluster
1443 instance moves (``tools/move-instance``)
1444 - Improved opcode summaries (e.g. in ``gnt-job list``)
1445 - Improve consistency of OS listing by sorting it
1446 - Documentation updates
1452 *(Released Fri, 8 Oct 2010)*
1454 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1460 *(Released Mon, 4 Oct 2010)*
1462 - Fixed regression in ``gnt-instance rename``
1468 *(Released Wed, 22 Sep 2010)*
1470 - Fixed OS_VARIANT variable for OS scripts
1471 - Fixed cluster tag operations via RAPI
1472 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1473 - Disabled RAPI CA checks in watcher
1479 *(Released Mon, 23 Aug 2010)*
1481 - Support DRBD versions of the format "a.b.c.d"
1483 - Re-introduce support for usage from multiple threads in RAPI client
1484 - Instance renames and modify via RAPI
1485 - Work around race condition between processing and archival in job
1487 - Mark opcodes following failed one as failed, too
1488 - Job field ``lock_status`` was removed due to difficulties making it
1489 work with the changed job queue in Ganeti 2.2; a better way to monitor
1490 locks is expected for a later 2.2.x release
1491 - Fixed dry-run behaviour with many commands
1492 - Support ``ssh-agent`` again when adding nodes
1493 - Many additional bugfixes
1499 *(Released Fri, 30 Jul 2010)*
1501 Important change: the internal RPC mechanism between Ganeti nodes has
1502 changed from using a home-grown http library (based on the Python base
1503 libraries) to use the PycURL library. This requires that PycURL is
1504 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1505 against GnuTLS by default. cURL's support for GnuTLS had known issues
1506 before cURL 7.21.0 and we recommend using the latest cURL release or
1507 linking against OpenSSL. Most other distributions already link PycURL
1508 and cURL against OpenSSL. The command::
1510 python -c 'import pycurl; print pycurl.version'
1512 can be used to determine the libraries PycURL and cURL are linked
1515 Other significant changes:
1517 - Rewrote much of the internals of the job queue, in order to achieve
1518 better parallelism; this decouples job query operations from the job
1519 processing, and it should allow much nicer behaviour of the master
1520 daemon under load, and it also has uncovered some long-standing bugs
1521 related to the job serialisation (now fixed)
1522 - Added a default iallocator setting to the cluster parameters,
1523 eliminating the need to always pass nodes or an iallocator for
1524 operations that require selection of new node(s)
1525 - Added experimental support for the LXC virtualization method
1526 - Added support for OS parameters, which allows the installation of
1527 instances to pass parameter to OS scripts in order to customise the
1529 - Added a hypervisor parameter controlling the migration type (live or
1530 non-live), since hypervisors have various levels of reliability; this
1531 has renamed the 'live' parameter to 'mode'
1532 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1533 logical volumes, meaning that cluster verify will ignore them and not
1534 flag their presence as errors
1535 - The watcher will now reset the error count for failed instances after
1536 8 hours, thus allowing self-healing if the problem that caused the
1537 instances to be down/fail to start has cleared in the meantime
1538 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1539 check for, and warn, if the drbd module parameter ``usermode_helper``
1540 is not consistent with the cluster-wide setting; this is needed to
1541 make diagnose easier of failed drbd creations
1542 - Started adding base IPv6 support, but this is not yet
1543 enabled/available for use
1544 - Rename operations (cluster, instance) will now return the new name,
1545 which is especially useful if a short name was passed in
1546 - Added support for instance migration in RAPI
1547 - Added a tool to pre-configure nodes for the SSH setup, before joining
1548 them to the cluster; this will allow in the future a simplified model
1549 for node joining (but not yet fully enabled in 2.2); this needs the
1550 paramiko python library
1551 - Fixed handling of name-resolving errors
1552 - Fixed consistency of job results on the error path
1553 - Fixed master-failover race condition when executed multiple times in
1555 - Fixed many bugs related to the job queue (mostly introduced during the
1556 2.2 development cycle, so not all are impacting 2.1)
1557 - Fixed instance migration with missing disk symlinks
1558 - Fixed handling of unknown jobs in ``gnt-job archive``
1559 - And many other small fixes/improvements
1563 - Enhanced both the unittest and the QA coverage
1564 - Switched the opcode validation to a generic model, and extended the
1565 validation to all opcode parameters
1566 - Changed more parts of the code that write shell scripts to use the
1568 - Switched the master daemon to use the asyncore library for the Luxi
1575 *(Released Thu, 17 Jun 2010)*
1577 - Added tool (``move-instance``) and infrastructure to move instances
1578 between separate clusters (see :doc:`separate documentation
1579 <move-instance>` and :doc:`design document <design-2.2>`)
1580 - Added per-request RPC timeout
1581 - RAPI now requires a Content-Type header for requests with a body (e.g.
1582 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1583 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1584 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1586 - Implemented initial support for running Ganeti daemons as separate
1587 users, see configure-time flags ``--with-user-prefix`` and
1588 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1590 - Instances can be removed after export (``gnt-backup export
1591 --remove-instance``)
1592 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1593 key (instead of 1024 bit)
1594 - Added new cluster configuration file for cluster domain secret
1595 - Import/export now use SSL instead of SSH
1596 - Added support for showing estimated time when exporting an instance,
1597 see the ``ganeti-os-interface(7)`` manpage and look for
1604 *(Released Tue, 16 Nov 2010)*
1606 Some more bugfixes. Unless critical bugs occur, this will be the last
1609 - Fix case of MAC special-values
1610 - Fix mac checker regex
1611 - backend: Fix typo causing "out of range" error
1612 - Add missing --units in gnt-instance list man page
1618 *(Released Tue, 24 Aug 2010)*
1621 - Don't ignore secondary node silently on non-mirrored disk templates
1623 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1624 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1625 - Properly document the usb_mouse parameter
1626 - Fix path in ganeti-rapi(8) (issue 116)
1627 - Adjust error message when the ganeti user's .ssh directory is
1629 - Add same-node-check when changing the disk template to drbd
1635 *(Released Fri, 16 Jul 2010)*
1638 - Add an option to only select some reboot types during qa/burnin.
1639 (on some hypervisors consequent reboots are not supported)
1640 - Fix infrequent race condition in master failover. Sometimes the old
1641 master ip address would be still detected as up for a short time
1642 after it was removed, causing failover to fail.
1643 - Decrease mlockall warnings when the ctypes module is missing. On
1644 Python 2.4 we support running even if no ctypes module is installed,
1645 but we were too verbose about this issue.
1646 - Fix building on old distributions, on which man doesn't have a
1648 - Fix RAPI not to ignore the MAC address on instance creation
1649 - Implement the old instance creation format in the RAPI client.
1655 *(Released Thu, 01 Jul 2010)*
1657 A small bugfix release:
1658 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1659 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1660 a non-existing option
1661 - Fix instance create: the --force-variant option was ignored
1662 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1663 - Fix modify node storage with non-FQDN arguments
1664 - Fix RAPI client to authenticate under Python 2.6 when used
1665 for more than 5 requests needing authentication
1666 - Fix gnt-instance modify -t (storage) giving a wrong error message
1667 when converting a non-shutdown drbd instance to plain
1673 *(Released Fri, 18 Jun 2010)*
1675 A small bugfix release:
1677 - Fix live migration of KVM instances started with older Ganeti
1678 versions which had fewer hypervisor parameters
1679 - Fix gnt-instance grow-disk on down instances
1680 - Fix an error-reporting bug during instance migration
1681 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1682 silently ignoring broken ones
1683 - Fix an RPC error reporting bug affecting, for example, RAPI client
1685 - Fix bug triggered by different API version os-es on different nodes
1686 - Fix a bug in instance startup with custom hvparams: OS level
1687 parameters would fail to be applied.
1688 - Fix the RAPI client under Python 2.6 (but more work is needed to
1689 make it work completely well with OpenSSL)
1690 - Fix handling of errors when resolving names from DNS
1696 *(Released Thu, 3 Jun 2010)*
1698 A medium sized development cycle. Some new features, and some
1699 fixes/small improvements/cleanups.
1701 Significant features
1702 ~~~~~~~~~~~~~~~~~~~~
1704 The node deamon now tries to mlock itself into memory, unless the
1705 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1706 its logs, and falls back to console logging. This allows emergency
1707 features such as ``gnt-node powercycle`` to work even in the event of a
1708 broken node disk (tested offlining the disk hosting the node's
1709 filesystem and dropping its memory caches; don't try this at home)
1711 KVM: add vhost-net acceleration support. It can be tested with a new
1712 enough version of the kernel and of qemu-kvm.
1714 KVM: Add instance chrooting feature. If you use privilege dropping for
1715 your VMs you can also now force them to chroot to an empty directory,
1716 before starting the emulated guest.
1718 KVM: Add maximum migration bandwith and maximum downtime tweaking
1719 support (requires a new-enough version of qemu-kvm).
1721 Cluster verify will now warn if the master node doesn't have the master
1722 ip configured on it.
1724 Add a new (incompatible) instance creation request format to RAPI which
1725 supports all parameters (previously only a subset was supported, and it
1726 wasn't possible to extend the old format to accomodate all the new
1727 features. The old format is still supported, and a client can check for
1728 this feature, before using it, by checking for its presence in the
1729 ``features`` RAPI resource.
1731 Now with ancient latin support. Try it passing the ``--roman`` option to
1732 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1733 (requires the python-roman module to be installed, in order to work).
1738 As usual many internal code refactorings, documentation updates, and
1741 - Lots of improvements and cleanups to the experimental Remote API
1742 (RAPI) client library.
1743 - A new unit test suite for the core daemon libraries.
1744 - A fix to creating missing directories makes sure the umask is not
1745 applied anymore. This enforces the same directory permissions
1747 - Better handling terminating daemons with ctrl+c (used when running
1748 them in debugging mode).
1749 - Fix a race condition in live migrating a KVM instance, when stat()
1750 on the old proc status file returned EINVAL, which is an unexpected
1752 - Fixed manpage checking with newer man and utf-8 charachters. But now
1753 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1759 *(Released Fri, 7 May 2010)*
1761 Fix a bug which prevented untagged KVM instances from starting.
1767 *(Released Fri, 7 May 2010)*
1769 Another release with a long development cycle, during which many
1770 different features were added.
1772 Significant features
1773 ~~~~~~~~~~~~~~~~~~~~
1775 The KVM hypervisor now can run the individual instances as non-root, to
1776 reduce the impact of a VM being hijacked due to bugs in the
1777 hypervisor. It is possible to run all instances as a single (non-root)
1778 user, to manually specify a user for each instance, or to dynamically
1779 allocate a user out of a cluster-wide pool to each instance, with the
1780 guarantee that no two instances will run under the same user ID on any
1783 An experimental RAPI client library, that can be used standalone
1784 (without the other Ganeti libraries), is provided in the source tree as
1785 ``lib/rapi/client.py``. Note this client might change its interface in
1786 the future, as we iterate on its capabilities.
1788 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1789 replace the cluster's certificates and crypto keys. This might help in
1790 case they have been compromised, or have simply expired.
1792 A new disk option for instance creation has been added that allows one
1793 to "adopt" currently existing logical volumes, with data
1794 preservation. This should allow easier migration to Ganeti from
1795 unmanaged (or managed via other software) instances.
1797 Another disk improvement is the possibility to convert between redundant
1798 (DRBD) and plain (LVM) disk configuration for an instance. This should
1799 allow better scalability (starting with one node and growing the
1800 cluster, or shrinking a two-node cluster to one node).
1802 A new feature that could help with automated node failovers has been
1803 implemented: if a node sees itself as offline (by querying the master
1804 candidates), it will try to shutdown (hard) all instances and any active
1805 DRBD devices. This reduces the risk of duplicate instances if an
1806 external script automatically failovers the instances on such nodes. To
1807 enable this, the cluster parameter ``maintain_node_health`` should be
1808 enabled; in the future this option (per the name) will enable other
1809 automatic maintenance features.
1811 Instance export/import now will reuse the original instance
1812 specifications for all parameters; that means exporting an instance,
1813 deleting it and the importing it back should give an almost identical
1814 instance. Note that the default import behaviour has changed from
1815 before, where it created only one NIC; now it recreates the original
1818 Cluster verify has added a few new checks: SSL certificates validity,
1819 /etc/hosts consistency across the cluster, etc.
1824 As usual, many internal changes were done, documentation fixes,
1827 - Fixed cluster initialization with disabled cluster storage (regression
1828 introduced in 2.1.1)
1829 - File-based storage supports growing the disks
1830 - Fixed behaviour of node role changes
1831 - Fixed cluster verify for some corner cases, plus a general rewrite of
1832 cluster verify to allow future extension with more checks
1833 - Fixed log spamming by watcher and node daemon (regression introduced
1835 - Fixed possible validation issues when changing the list of enabled
1837 - Fixed cleanup of /etc/hosts during node removal
1838 - Fixed RAPI response for invalid methods
1839 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1840 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1841 from ide disks, etc.)
1842 - Allow OS changes without re-installation (to record a changed OS
1843 outside of Ganeti, or to allow OS renames)
1844 - Allow instance creation without OS installation (useful for example if
1845 the OS will be installed manually, or restored from a backup not in
1847 - Implemented option to make cluster ``copyfile`` use the replication
1849 - Added list of enabled hypervisors to ssconf (possibly useful for
1851 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1853 - A partial form of node re-IP is possible via node readd, which now
1854 allows changed node primary IP
1855 - Command line utilities now show an informational message if the job is
1857 - The logs of the master daemon now show the PID/UID/GID of the
1864 *(Released Fri, 12 Mar 2010)*
1866 During the 2.1.0 long release candidate cycle, a lot of improvements and
1867 changes have accumulated with were released later as 2.1.1.
1872 The node evacuate command (``gnt-node evacuate``) was significantly
1873 rewritten, and as such the IAllocator protocol was changed - a new
1874 request type has been added. This unfortunate change during a stable
1875 series is designed to improve performance of node evacuations; on
1876 clusters with more than about five nodes and which are well-balanced,
1877 evacuation should proceed in parallel for all instances of the node
1878 being evacuated. As such, any existing IAllocator scripts need to be
1879 updated, otherwise the above command will fail due to the unknown
1880 request. The provided "dumb" allocator has not been updated; but the
1881 ganeti-htools package supports the new protocol since version 0.2.4.
1883 Another important change is increased validation of node and instance
1884 names. This might create problems in special cases, if invalid host
1885 names are being used.
1887 Also, a new layer of hypervisor parameters has been added, that sits at
1888 OS level between the cluster defaults and the instance ones. This allows
1889 customisation of virtualization parameters depending on the installed
1890 OS. For example instances with OS 'X' may have a different KVM kernel
1891 (or any other parameter) than the cluster defaults. This is intended to
1892 help managing a multiple OSes on the same cluster, without manual
1893 modification of each instance's parameters.
1895 A tool for merging clusters, ``cluster-merge``, has been added in the
1896 tools sub-directory.
1901 - Improved the int/float conversions that should make the code more
1902 robust in face of errors from the node daemons
1903 - Fixed the remove node code in case of internal configuration errors
1904 - Fixed the node daemon behaviour in face of inconsistent queue
1905 directory (e.g. read-only file-system where we can't open the files
1907 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1908 now it either aborts cleanly or, if given the new "auto_promote"
1909 parameter, will automatically promote other nodes as needed
1910 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1911 completely prevent Ganeti from working
1912 - Fixed bug for instance export when not all disks were successfully
1914 - Fixed behaviour of node add when the new node is slow in starting up
1916 - Fixed handling of signals in the LUXI client, which should improve
1917 behaviour of command-line scripts
1918 - Added checks for invalid node/instance names in the configuration (now
1919 flagged during cluster verify)
1920 - Fixed watcher behaviour for disk activation errors
1921 - Fixed two potentially endless loops in http library, which led to the
1922 RAPI daemon hanging and consuming 100% CPU in some cases
1923 - Fixed bug in RAPI daemon related to hashed passwords
1924 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1926 - Enhanced compatibility with non-Debian OSes, but not using absolute
1927 path in some commands and allowing customisation of the ssh
1928 configuration directory
1929 - Fixed possible future issue with new Python versions by abiding to the
1930 proper use of ``__slots__`` attribute on classes
1931 - Added checks that should prevent directory traversal attacks
1932 - Many documentation fixes based on feedback from users
1937 - Added an "early_release" more for instance replace disks and node
1938 evacuate, where we release locks earlier and thus allow higher
1939 parallelism within the cluster
1940 - Added watcher hooks, intended to allow the watcher to restart other
1941 daemons (e.g. from the ganeti-nbma project), but they can be used of
1942 course for any other purpose
1943 - Added a compile-time disable for DRBD barriers, to increase
1944 performance if the administrator trusts the power supply or the
1945 storage system to not lose writes
1946 - Added the option of using syslog for logging instead of, or in
1947 addition to, Ganeti's own log files
1948 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1949 can indeed boot from a paravirtual NIC
1950 - Added a generic debug level for many operations; while this is not
1951 used widely yet, it allows one to pass the debug value all the way to
1953 - Enhanced the hooks environment for instance moves (failovers,
1954 migrations) where the primary/secondary nodes changed during the
1955 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1956 - Enhanced data validations for many user-supplied values; one important
1957 item is the restrictions imposed on instance and node names, which
1958 might reject some (invalid) host names
1959 - Add a configure-time option to disable file-based storage, if it's not
1960 needed; this allows greater security separation between the master
1961 node and the other nodes from the point of view of the inter-node RPC
1963 - Added user notification in interactive tools if job is waiting in the
1964 job queue or trying to acquire locks
1965 - Added log messages when a job is waiting for locks
1966 - Added filtering by node tags in instance operations which admit
1967 multiple instances (start, stop, reboot, reinstall)
1968 - Added a new tool for cluster mergers, ``cluster-merge``
1969 - Parameters from command line which are of the form ``a=b,c=d`` can now
1970 use backslash escapes to pass in values which contain commas,
1971 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1973 - For KVM, the instance name is the first parameter passed to KVM, so
1974 that it's more visible in the process list
1980 *(Released Tue, 2 Mar 2010)*
1982 Ganeti 2.1 brings many improvements with it. Major changes:
1984 - Added infrastructure to ease automated disk repairs
1985 - Added new daemon to export configuration data in a cheaper way than
1986 using the remote API
1987 - Instance NICs can now be routed instead of being associated with a
1989 - Improved job locking logic to reduce impact of jobs acquiring multiple
1990 locks waiting for other long-running jobs
1992 In-depth implementation details can be found in the Ganeti 2.1 design
1998 - Added chroot hypervisor
1999 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2001 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2002 ``bootloader_path`` and ``bootloader_args``)
2003 - Added the ``use_localtime`` option for the xen-hvm and kvm
2004 hypervisors, and the default value for this has changed to false (in
2005 2.0 xen-hvm always enabled it)
2006 - Added luxi call to submit multiple jobs in one go
2007 - Added cluster initialization option to not modify ``/etc/hosts``
2009 - Added network interface parameters
2010 - Added dry run mode to some LUs
2011 - Added RAPI resources:
2013 - ``/2/instances/[instance_name]/info``
2014 - ``/2/instances/[instance_name]/replace-disks``
2015 - ``/2/nodes/[node_name]/evacuate``
2016 - ``/2/nodes/[node_name]/migrate``
2017 - ``/2/nodes/[node_name]/role``
2018 - ``/2/nodes/[node_name]/storage``
2019 - ``/2/nodes/[node_name]/storage/modify``
2020 - ``/2/nodes/[node_name]/storage/repair``
2022 - Added OpCodes to evacuate or migrate all instances on a node
2023 - Added new command to list storage elements on nodes (``gnt-node
2024 list-storage``) and modify them (``gnt-node modify-storage``)
2025 - Added new ssconf files with master candidate IP address
2026 (``ssconf_master_candidates_ips``), node primary IP address
2027 (``ssconf_node_primary_ips``) and node secondary IP address
2028 (``ssconf_node_secondary_ips``)
2029 - Added ``ganeti-confd`` and a client library to query the Ganeti
2030 configuration via UDP
2031 - Added ability to run hooks after cluster initialization and before
2033 - Added automatic mode for disk replace (``gnt-instance replace-disks
2035 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2036 after catastrophic data-loss
2037 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2039 - Added ``gnt-instance move`` command to move instances
2040 - Added ``gnt-cluster watcher`` command to control watcher
2041 - Added ``gnt-node powercycle`` command to powercycle nodes
2042 - Added new job status field ``lock_status``
2043 - Added parseable error codes to cluster verification (``gnt-cluster
2044 verify --error-codes``) and made output less verbose (use
2045 ``--verbose`` to restore previous behaviour)
2046 - Added UUIDs to the main config entities (cluster, nodes, instances)
2047 - Added support for OS variants
2048 - Added support for hashed passwords in the Ganeti remote API users file
2050 - Added option to specify maximum timeout on instance shutdown
2051 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2052 - Added new helper script to start and stop Ganeti daemons
2053 (``daemon-util``), with the intent to reduce the work necessary to
2054 adjust Ganeti for non-Debian distributions and to start/stop daemons
2056 - Added more unittests
2057 - Fixed critical bug in ganeti-masterd startup
2058 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2059 now customisable at the cluster level for both the KVM and Xen
2060 hypervisors using the new ``migration_port`` parameter
2061 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2062 reinstalling an instance
2063 - Allowed ``@`` in tag names
2064 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2065 - Many documentation updates
2066 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2067 - ``gnt-instance reinstall`` can now reinstall multiple instances
2068 - Updated many command line parameters
2069 - Introduced new OS API version 15
2070 - No longer support a default hypervisor
2071 - Treat virtual LVs as inexistent
2072 - Improved job locking logic to reduce lock contention
2073 - Match instance and node names case insensitively
2074 - Reimplemented bash completion script to be more complete
2081 *(Released Thu, 4 Feb 2010)*
2083 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2084 - Fix a string formatting bug
2085 - Improve safety of the code in some error paths
2086 - Improve data validation in the master of values returned from nodes
2092 *(Released Thu, 17 Dec 2009)*
2094 - Fix security issue due to missing validation of iallocator names; this
2095 allows local and remote execution of arbitrary executables
2096 - Fix failure of gnt-node list during instance removal
2097 - Ship the RAPI documentation in the archive
2103 *(Released Wed, 30 Sep 2009)*
2105 - Fixed many wrong messages
2106 - Fixed a few bugs related to the locking library
2107 - Fixed MAC checking at instance creation time
2108 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2109 - Fixed a few issues related to signal handling in both daemons and
2111 - Fixed the example startup script provided
2112 - Fixed insserv dependencies in the example startup script (patch from
2114 - Fixed handling of drained nodes in the iallocator framework
2115 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2117 - Fixed error related to invalid job IDs in job polling
2118 - Fixed job/opcode persistence on unclean master shutdown
2119 - Fixed handling of partial job processing after unclean master
2121 - Fixed error reporting from LUs, previously all errors were converted
2122 into execution errors
2123 - Fixed error reporting from burnin
2124 - Decreased significantly the memory usage of the job queue
2125 - Optimised slightly multi-job submission
2126 - Optimised slightly opcode loading
2127 - Backported the multi-job submit framework from the development
2128 branch; multi-instance start and stop should be faster
2129 - Added script to clean archived jobs after 21 days; this will reduce
2130 the size of the queue directory
2131 - Added some extra checks in disk size tracking
2132 - Added an example ethers hook script
2133 - Added a cluster parameter that prevents Ganeti from modifying of
2135 - Added more node information to RAPI responses
2136 - Added a ``gnt-job watch`` command that allows following the ouput of a
2138 - Added a bind-address option to ganeti-rapi
2139 - Added more checks to the configuration verify
2140 - Enhanced the burnin script such that some operations can be retried
2142 - Converted instance reinstall to multi-instance model
2148 *(Released Fri, 7 Aug 2009)*
2150 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2151 to allow using the pre-2.0.2 behaviour in activation, if any existing
2152 instances have mismatched disk sizes in the configuration
2153 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2154 any configuration mismatches for disk sizes
2155 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2156 failover to work on two-node clusters
2157 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2159 - Fixed detection of OS script errors in ``gnt-backup export``
2160 - Fixed exit code of ``gnt-backup export``
2166 *(Released Fri, 17 Jul 2009)*
2168 - Added experimental support for stripped logical volumes; this should
2169 enhance performance but comes with a higher complexity in the block
2170 device handling; stripping is only enabled when passing
2171 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2172 affected even in the non-stripped mode
2173 - Improved resiliency against transient failures at the end of DRBD
2174 resyncs, and in general of DRBD resync checks
2175 - Fixed a couple of issues with exports and snapshot errors
2176 - Fixed a couple of issues in instance listing
2177 - Added display of the disk size in ``gnt-instance info``
2178 - Fixed checking for valid OSes in instance creation
2179 - Fixed handling of the "vcpus" parameter in instance listing and in
2180 general of invalid parameters
2181 - Fixed http server library, and thus RAPI, to handle invalid
2182 username/password combinations correctly; this means that now they
2183 report unauthorized for queries too, not only for modifications,
2184 allowing earlier detect of configuration problems
2185 - Added a new "role" node list field, equivalent to the master/master
2186 candidate/drained/offline flags combinations
2187 - Fixed cluster modify and changes of candidate pool size
2188 - Fixed cluster verify error messages for wrong files on regular nodes
2189 - Fixed a couple of issues with node demotion from master candidate role
2190 - Fixed node readd issues
2191 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2192 - Added a new ``--no-voting`` option for masterfailover to fix failover
2193 on two-nodes clusters when the former master node is unreachable
2194 - Added instance reinstall over RAPI
2200 *(Released Tue, 16 Jun 2009)*
2202 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2203 allow re-adding the start in single-user option (regression from 1.2)
2204 - the watcher writes the instance status to a file, to allow monitoring
2205 to report the instance status (from the master) based on cached
2206 results of the watcher's queries; while this can get stale if the
2207 watcher is being locked due to other work on the cluster, this is
2208 still an improvement
2209 - the watcher now also restarts the node daemon and the rapi daemon if
2211 - fixed the watcher to handle full and drained queue cases
2212 - hooks export more instance data in the environment, which helps if
2213 hook scripts need to take action based on the instance's properties
2214 (no longer need to query back into ganeti)
2215 - instance failovers when the instance is stopped do not check for free
2216 RAM, so that failing over a stopped instance is possible in low memory
2218 - rapi uses queries for tags instead of jobs (for less job traffic), and
2219 for cluster tags it won't talk to masterd at all but read them from
2221 - a couple of error handling fixes in RAPI
2222 - drbd handling: improved the error handling of inconsistent disks after
2223 resync to reduce the frequency of "there are some degraded disks for
2224 this instance" messages
2225 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2226 error handling path called a wrong function name)
2232 *(Released Wed, 27 May 2009)*
2234 - no changes from rc5
2240 *(Released Wed, 20 May 2009)*
2242 - fix a couple of bugs (validation, argument checks)
2243 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2244 - some small improvements to RAPI and IAllocator
2245 - make watcher automatically start the master daemon if down
2251 *(Released Mon, 27 Apr 2009)*
2253 - change the OS list to not require locks; this helps with big clusters
2254 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2255 volume group is broken
2256 - ``gnt-instance info``, without any arguments, doesn't run for all
2257 instances anymore; either pass ``--all`` or pass the desired
2258 instances; this helps against mistakes on big clusters where listing
2259 the information for all instances takes a long time
2260 - miscellaneous doc and man pages fixes
2266 *(Released Wed, 8 Apr 2009)*
2268 - Change the internal locking model of some ``gnt-node`` commands, in
2269 order to reduce contention (and blocking of master daemon) when
2270 batching many creation/reinstall jobs
2271 - Fixes to Xen soft reboot
2272 - No longer build documentation at build time, instead distribute it in
2273 the archive, in order to reduce the need for the whole docbook/rst
2280 *(Released Fri, 27 Mar 2009)*
2282 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2283 - Fix watcher startup sequence, improves the behaviour of busy clusters
2284 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2285 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2286 bind address and other places
2287 - Some documentation fixes and updates
2293 *(Released Mon, 2 Mar 2009)*
2295 - More documentation updates, now all docs should be more-or-less
2297 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2299 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2301 - Fix an issue related to $libdir/run/ganeti and cluster creation
2307 *(Released Thu, 19 Feb 2009)*
2309 - Xen PVM and KVM have switched the default value for the instance root
2310 disk to the first partition on the first drive, instead of the whole
2311 drive; this means that the OS installation scripts must be changed
2313 - Man pages have been updated
2314 - RAPI has been switched by default to HTTPS, and the exported functions
2315 should all work correctly
2316 - RAPI v1 has been removed
2317 - Many improvements to the KVM hypervisor
2318 - Block device errors are now better reported
2319 - Many other bugfixes and small improvements
2325 *(Released Mon, 26 Jan 2009)*
2327 - Version 2 is a general rewrite of the code and therefore the
2328 differences are too many to list, see the design document for 2.0 in
2329 the ``doc/`` subdirectory for more details
2330 - In this beta version there is not yet a migration path from 1.2 (there
2331 will be one in the final 2.0 release)
2332 - A few significant changes are:
2334 - all commands are executed by a daemon (``ganeti-masterd``) and the
2335 various ``gnt-*`` commands are just front-ends to it
2336 - all the commands are entered into, and executed from a job queue,
2337 see the ``gnt-job(8)`` manpage
2338 - the RAPI daemon supports read-write operations, secured by basic
2339 HTTP authentication on top of HTTPS
2340 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2341 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2342 to migrate to DRBD 8 first while still running Ganeti 1.2)
2343 - DRBD devices are using statically allocated minor numbers, which
2344 will be assigned to existing instances during the migration process
2345 - there is support for both Xen PVM and Xen HVM instances running on
2347 - KVM virtualization is supported too
2348 - file-based storage has been implemented, which means that it is
2349 possible to run the cluster without LVM and DRBD storage, for
2350 example using a shared filesystem exported from shared storage (and
2351 still have live migration)
2357 *(Released Tue, 13 Jan 2009)*
2359 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2360 - Reuse the old instance mac address by default on instance import, if
2361 the instance name is the same.
2362 - Handle situations in which the node info rpc returns incomplete
2364 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2365 - Improved version of batcher:
2367 - state file support
2368 - instance mac address support
2369 - support for HVM clusters/instances
2371 - Add an option to show the number of cpu sockets and nodes in
2373 - Support OSes that handle more than one version of the OS api (but do
2374 not change the current API in any other way)
2375 - Fix ``gnt-node migrate``
2376 - ``gnt-debug`` man page
2377 - Fixes various more typos and small issues
2378 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2384 *(Released Wed, 24 Sep 2008)*
2386 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2387 type of disk exported to fully virtualized instances.
2388 - provide access to the serial console of HVM instances
2389 - instance auto_balance flag, set by default. If turned off it will
2390 avoid warnings on cluster verify if there is not enough memory to fail
2391 over an instance. in the future it will prevent automatically failing
2392 it over when we will support that.
2393 - batcher tool for instance creation, see ``tools/README.batcher``
2394 - ``gnt-instance reinstall --select-os`` to interactively select a new
2395 operating system when reinstalling an instance.
2396 - when changing the memory amount on instance modify a check has been
2397 added that the instance will be able to start. also warnings are
2398 emitted if the instance will not be able to fail over, if auto_balance
2400 - documentation fixes
2401 - sync fields between ``gnt-instance list/modify/add/import``
2402 - fix a race condition in drbd when the sync speed was set after giving
2403 the device a remote peer.
2409 *(Released Tue, 22 Jul 2008)*
2411 - note: the allowed size and number of tags per object were reduced
2412 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2413 - fixed twisted 8.x compatibility
2414 - fixed ``gnt-instance replace-disks`` with iallocator
2415 - add TCP keepalives on twisted connections to detect restarted nodes
2416 - disk increase support, see ``gnt-instance grow-disk``
2417 - implement bulk node/instance query for RAPI
2418 - add tags in node/instance listing (optional)
2419 - experimental migration (and live migration) support, read the man page
2420 for ``gnt-instance migrate``
2421 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2422 has some small improvements in handling its state file
2428 *(Released Fri, 13 Jun 2008)*
2430 - Experimental readonly, REST-based remote API implementation;
2431 automatically started on master node, TCP port 5080, if enabled by
2432 ``--enable-rapi`` parameter to configure script.
2433 - Instance allocator support. Add and import instance accept a
2434 ``--iallocator`` parameter, and call that instance allocator to decide
2435 which node to use for the instance. The iallocator document describes
2436 what's expected from an allocator script.
2437 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2438 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2439 node is lost there is still enough memory to fail over the instances
2441 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2442 ``gnt-cluster verify``, to check for site-specific compliance. All the
2443 hooks will run, and their output, if any, will be displayed. Any
2444 failing hook will make the verification return an error value.
2445 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2446 primary and secondary interfaces
2447 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2448 that is still declared as part of the cluster and has failed.
2449 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2450 specifying output fields, that just adds the chosen fields to the
2452 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2453 export from the filesystem.
2454 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2455 have been added. Using them you can enable/disable acpi and pae
2456 support, and specify a path for a cd image to be exported to the
2457 instance. These parameters as the name suggest only work on HVM
2459 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2460 PAE support will be set to the previously hardcoded values, but the
2461 (previously hardcoded) path to the CDROM ISO image will be unset and
2462 if required, needs to be set manually with ``gnt-instance modify``
2464 - The address to which an instance's VNC console is bound is now
2465 selectable per-instance, rather than being cluster wide. Of course
2466 this only applies to instances controlled via VNC, so currently just
2467 applies to HVM clusters.
2473 *(Released Mon, 18 Feb 2008)*
2475 - more tweaks to the disk activation code (especially helpful for DRBD)
2476 - change the default ``gnt-instance list`` output format, now there is
2477 one combined status field (see the manpage for the exact values this
2479 - some more fixes for the mac export to hooks change
2480 - make Ganeti not break with DRBD 8.2.x (which changed the version
2481 format in ``/proc/drbd``) (issue 24)
2482 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2483 template, allowing migration from DRBD0.7+MD to DRBD8
2489 *(Released Wed, 30 Jan 2008)*
2491 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2493 - add command aliases infrastructure and a few aliases
2494 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2495 man pages and the ``--help`` option of ``gnt-node
2496 list``/``gnt-instance list``
2497 - fix ``gnt-backup list`` with down nodes (issue 21)
2498 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2499 - fix the dist archive and add a check for including svn/git files in
2501 - some developer-related changes: improve the burnin and the QA suite,
2502 add an upload script for testing during development
2508 *(Released Wed, 16 Jan 2008)*
2510 - experimental HVM support, read the install document, section
2511 "Initializing the cluster"
2512 - allow for the PVM hypervisor per-instance kernel and initrd paths
2513 - add a new command ``gnt-cluster verify-disks`` which uses a new
2514 algorithm to improve the reconnection of the DRBD pairs if the device
2515 on the secondary node has gone away
2516 - make logical volume code auto-activate LVs at disk activation time
2517 - slightly improve the speed of activating disks
2518 - allow specification of the MAC address at instance creation time, and
2519 changing it later via ``gnt-instance modify``
2520 - fix handling of external commands that generate lots of output on
2522 - update documentation with regard to minimum version of DRBD8 supported
2528 *(Released Tue, 4 Dec 2007)*
2530 - Log the ``xm create`` output to the node daemon log on failure (to
2531 help diagnosing the error)
2532 - In debug mode, log all external commands output if failed to the logs
2533 - Change parsing of lvm commands to ignore stderr
2539 *(Released Wed, 28 Nov 2007)*
2541 - Another round of updates to the DRBD 8 code to deal with more failures
2542 in the replace secondary node operation
2543 - Some more logging of failures in disk operations (lvm, drbd)
2544 - A few documentation updates
2551 *(Released Tue, 13 Nov 2007)*
2553 - Change configuration file format from Python's Pickle to JSON.
2554 Upgrading is possible using the cfgupgrade utility.
2555 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2556 faster replace disks and is more stable (DRBD 8 has many improvements
2557 compared to DRBD 0.7)
2558 - Added command line tags support (see man pages for ``gnt-instance``,
2559 ``gnt-node``, ``gnt-cluster``)
2560 - Added instance rename support
2561 - Added multi-instance startup/shutdown
2562 - Added cluster rename support
2563 - Added ``gnt-node evacuate`` to simplify some node operations
2564 - Added instance reboot operation that can speedup reboot as compared to
2566 - Soften the requirement that hostnames are in FQDN format
2567 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2569 - Removed dependency on debian's patched fping that uses the
2570 non-standard ``-S`` option
2571 - Now the OS definitions are searched for in multiple, configurable
2572 paths (easier for distros to package)
2573 - Some changes to the hooks infrastructure (especially the new
2574 post-configuration update hook)
2575 - Other small bugfixes
2577 .. vim: set textwidth=72 syntax=rst :