10 - The minimum Python version needed to run Ganeti is now 2.6.
11 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
12 to always require authentication, ``--require-authentication``. It can
13 be specified in ``$sysconfdir/default/ganeti``.
14 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
15 be used to manage the disk templates to be used by instances in the cluster.
16 Initially, it will be set to a list that includes plain, drbd, if they were
17 enabled by specifying a volume group name, and file and sharedfile, if those
18 were enabled at configure time. Additionally, it will include all disk
19 templates that are currently used by instances. The order of disk templates
20 will be based on Ganeti's history of supporting them. In the future, the
21 first entry of the list will be used as a default disk template on instance
23 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
24 configuration back to the previous stable version.
30 *(Released Wed, 6 Feb 2013)*
32 - ``gnt-instance batch-create`` has been changed to use the bulk create
33 opcode from Ganeti. This lead to incompatible changes in the format of
34 the JSON file. It's now not a custom dict anymore but a dict
35 compatible with the ``OpInstanceCreate`` opcode.
36 - Parent directories for file storage need to be listed in
37 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
38 write the file automatically based on old configuration values, but it
39 can not distribute it across all nodes and the file contents should be
40 verified. Use ``gnt-cluster copyfile
41 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
42 upgraded. The reason for requiring this list of paths now is that
43 before it would have been possible to inject new paths via RPC,
44 allowing files to be created in arbitrary locations. The RPC protocol
45 is protected using SSL/X.509 certificates, but as a design principle
46 Ganeti does not permit arbitrary paths to be passed.
47 - The parsing of the variants file for OSes (see
48 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
49 lines and comment lines (starting with ``#``) are ignored for better
51 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
52 longer available. ``gnt-node add`` now invokes a new tool on the
53 destination node, named ``prepare-node-join``, to configure the SSH
54 daemon. Paramiko is no longer necessary to configure nodes' SSH
55 daemons via ``gnt-node add``.
56 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
57 for RAPI users. It allows granting permissions to query for
58 information to a specific user without giving
59 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
60 - A new tool named ``node-cleanup`` has been added. It cleans remains of
61 a cluster from a machine by stopping all daemons, removing
62 certificates and ssconf files. Unless the ``--no-backup`` option is
63 given, copies of the certificates are made.
64 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
65 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
66 the flag is not reset after a master failover.
67 - Instance creations now support the use of opportunistic locking,
68 potentially speeding up the (parallel) creation of multiple instances.
69 This feature is currently only available via the :doc:`RAPI
70 <rapi>` interface and when an instance allocator is used. If the
71 ``opportunistic_locking`` parameter is set the opcode will try to
72 acquire as many locks as possible, but will not wait for any locks
73 held by other opcodes. If not enough resources can be found to
74 allocate the instance, the temporary error code
75 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
76 retried thereafter, with or without opportunistic locking.
77 - The functionality for allocating multiple instances at once has been
78 overhauled and is now also available through :doc:`RAPI <rapi>`.
79 - Man pages can now be included when the documentation is built, in
80 which case the output is in ``doc/man-html``. The configure-time
81 option is ``--enable-manpages-in-doc``. Sphinx 1.0 or higher is
83 - A new htool called ``hroller`` helps scheduling parallel shutdown of
84 nodes on clusters where all instances are shut down, but without ever
85 rebooting both primary and secondary node for an instance at the same
87 - Queries not needing locks or RPC access to the node can now be
88 performed by the confd daemon, making them independent from jobs, and
89 thus faster to execute. This is selectable at configure time.
90 - The KVM hypervisor has been updated with many new hypervisor
91 parameters, including a generic one for passing arbitrary command line
92 values. See a complete list in :manpage:`gnt-instance(8)`.
93 - A new tool, called ``mon-collector``, is the stand-alone executor of
94 the data collectors for a monitoring system. As of this version, it
95 just includes the DRBD data collector, that can be executed by calling
96 ``mon-collector`` using the ``drbd`` parameter. See
97 :manpage:`mon-collector(7)`.
98 - New ``exclusive-storage`` node parameter added. When it's set to true,
99 physical disks are assigned in an exclusive fashion to instances, as
100 documented in :doc:`Partitioned Ganeti <design-partitioned>`.
101 Currently, only instances using the ``plain`` disk template are
103 - New network management functionality to support automatic allocation
104 of IP addresses and managing of network parameters. See
105 :manpage:`gnt-network(8)` for more details.
106 - New external storage backend, to allow managing arbitrary storage
107 systems external to the cluster. See
108 :manpage:`ganeti-extstorage-interface(7)`.
109 - Instance renames of LVM-based instances will now update the LV tags
110 (which can be used to recover the instance-to-LV mapping in case of
112 - ``hbal`` will now exit with status 0 if, during job execution over
113 LUXI, early exit has been requested and all jobs are successful;
114 before, exit status 1 was used, which cannot be differentiated from
121 *(Released Fri, 21 Dec 2012)*
123 Important behaviour change: hbal won't rebalance anymore instances which
124 have the ``auto_balance`` attribute set to false. This was the intention
125 all along, but until now it only skipped those from the N+1 memory
126 reservation (DRBD-specific).
128 A significant number of bug fixes in this release:
130 - Fixed disk adoption interaction with ipolicy checks.
131 - Fixed networking issues when instances are started, stopped or
132 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
133 - Fixed the warning in cluster verify for shared storage instances not
135 - Fixed removal of storage directory on shared file storage (issue 262).
136 - Fixed validation of LVM volume group name in OpClusterSetParams
137 (``gnt-cluster modify``) (issue 285).
138 - Fixed runtime memory increases (``gnt-instance modify -m``).
139 - Fixed live migration under Xen's ``xl`` mode.
140 - Fixed ``gnt-instance console`` with ``xl``.
141 - Fixed building with newer Haskell compiler/libraries.
142 - Fixed PID file writing in Haskell daemons (confd); this prevents
143 restart issues if confd was launched manually (outside of
144 ``daemon-util``) while another copy of it was running
145 - Fixed a type error when doing live migrations with KVM (issue 297) and
146 the error messages for failing migrations have been improved.
147 - Fixed opcode validation for the out-of-band commands (``gnt-node
149 - Fixed a type error when unsetting OS hypervisor parameters (issue
150 311); now it's possible to unset all OS-specific hypervisor
152 - Fixed the ``dry-run`` mode for many operations: verification of
153 results was over-zealous but didn't take into account the ``dry-run``
154 operation, resulting in "wrong" failures.
155 - Fixed bash completion in ``gnt-job list`` when the job queue has
156 hundreds of entries; especially with older ``bash`` versions, this
157 results in significant CPU usage.
159 And lastly, a few other improvements have been made:
161 - Added option to force master-failover without voting (issue 282).
162 - Clarified error message on lock conflict (issue 287).
163 - Logging of newly submitted jobs has been improved (issue 290).
164 - Hostname checks have been made uniform between instance rename and
166 - The ``--submit`` option is now supported by ``gnt-debug delay``.
167 - Shutting down the master daemon by sending SIGTERM now stops it from
168 processing jobs waiting for locks; instead, those jobs will be started
169 once again after the master daemon is started the next time (issue
171 - Support for Xen's ``xl`` program has been improved (besides the fixes
173 - Reduced logging noise in the Haskell confd daemon (only show one log
174 entry for each config reload, instead of two).
175 - Several man page updates and typo fixes.
181 *(Released Fri, 12 Oct 2012)*
183 A small bugfix release. Among the bugs fixed:
185 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
186 made the command unusable.
187 - Commands that issue many jobs don't fail anymore just because some jobs
188 take so long that other jobs are archived.
189 - Failures during ``gnt-instance reinstall`` are reflected by the exit
191 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
193 - When ``always_failover`` is set, ``--allow-failover`` is not required
194 in migrate commands anymore.
195 - ``bash_completion`` works even if extglob is disabled.
196 - Fixed bug with locks that made failover for RDB-based instances fail.
197 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
198 a random node instead of one based on the allocator metric.
199 - Support for newer versions of pylint and pep8.
200 - Hail doesn't fail anymore when trying to add an instance of type
201 ``file``, ``sharedfile`` or ``rbd``.
202 - Added new Makefile target to rebuild the whole distribution, so that
203 all files are included.
209 *(Released Fri, 27 Jul 2012)*
212 .. attention:: The ``LUXI`` protocol has been made more consistent
213 regarding its handling of command arguments. This, however, leads to
214 incompatibility issues with previous versions. Please ensure that you
215 restart Ganeti daemons soon after the upgrade, otherwise most
216 ``LUXI`` calls (job submission, setting/resetting the drain flag,
217 pausing/resuming the watcher, cancelling and archiving jobs, querying
218 the cluster configuration) will fail.
227 The current ``admin_up`` field, which used to denote whether an instance
228 should be running or not, has been removed. Instead, ``admin_state`` is
229 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
231 The rational behind this is that an instance being “down” can have
234 - it could be down during a reboot
235 - it could be temporarily be down for a reinstall
236 - or it could be down because it is deprecated and kept just for its
239 The previous Boolean state was making it difficult to do capacity
240 calculations: should Ganeti reserve memory for a down instance? Now, the
241 tri-state field makes it clear:
243 - in ``up`` and ``down`` state, all resources are reserved for the
244 instance, and it can be at any time brought up if it is down
245 - in ``offline`` state, only disk space is reserved for it, but not
248 The field can have an extra use: since the transition between ``up`` and
249 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
250 transition between ``offline`` and ``down`` is done via ``gnt-instance
251 modify``, it is possible to given different rights to users. For
252 example, owners of an instance could be allowed to start/stop it, but
253 not transition it out of the offline state.
255 Instance policies and specs
256 +++++++++++++++++++++++++++
258 In previous Ganeti versions, an instance creation request was not
259 limited on the minimum size and on the maximum size just by the cluster
260 resources. As such, any policy could be implemented only in third-party
261 clients (RAPI clients, or shell wrappers over ``gnt-*``
262 tools). Furthermore, calculating cluster capacity via ``hspace`` again
263 required external input with regards to instance sizes.
265 In order to improve these workflows and to allow for example better
266 per-node group differentiation, we introduced instance specs, which
269 - minimum instance disk size, disk count, memory size, cpu count
270 - maximum values for the above metrics
271 - and “standard” values (used in ``hspace`` to calculate the standard
274 The minimum/maximum values can be also customised at node-group level,
275 for example allowing more powerful hardware to support bigger instance
278 Beside the instance specs, there are a few other settings belonging to
279 the instance policy framework. It is possible now to customise, per
280 cluster and node-group:
282 - the list of allowed disk templates
283 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
284 - the maximum ratio of instance to spindles (see below for more
285 information) for local storage
287 All these together should allow all tools that talk to Ganeti to know
288 what are the ranges of allowed values for instances and the
289 over-subscription that is allowed.
291 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
292 instance configuration, and the physical CPU configuration from the
293 node. For the spindle ratios however, we didn't track before these
294 values, so new parameters have been added:
296 - a new node parameter ``spindle_count``, defaults to 1, customisable at
297 node group or node level
298 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
300 Note that spindles in this context doesn't need to mean actual
301 mechanical hard-drives; it's just a relative number for both the node
302 I/O capacity and instance I/O consumption.
304 Instance migration behaviour
305 ++++++++++++++++++++++++++++
307 While live-migration is in general desirable over failover, it is
308 possible that for some workloads it is actually worse, due to the
309 variable time of the “suspend” phase during live migration.
311 To allow the tools to work consistently over such instances (without
312 having to hard-code instance names), a new backend parameter
313 ``always_failover`` has been added to control the migration/failover
314 behaviour. When set to True, all migration requests for an instance will
315 instead fall-back to failover.
317 Instance memory ballooning
318 ++++++++++++++++++++++++++
320 Initial support for memory ballooning has been added. The memory for an
321 instance is no longer fixed (backend parameter ``memory``), but instead
322 can vary between minimum and maximum values (backend parameters
323 ``minmem`` and ``maxmem``). Currently we only change an instance's
326 - live migrating or failing over and instance and the target node
327 doesn't have enough memory
328 - user requests changing the memory via ``gnt-instance modify
334 In order to control the use of specific CPUs by instance, support for
335 controlling CPU pinning has been added for the Xen, HVM and LXC
336 hypervisors. This is controlled by a new hypervisor parameter
337 ``cpu_mask``; details about possible values for this are in the
338 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
339 VCPU-to-CPU mapping) form will work well only when all nodes in your
340 cluster have the same amount of CPUs.
345 Another area in which Ganeti was not customisable were the parameters
346 used for storage configuration, e.g. how many stripes to use for LVM,
347 DRBD resync configuration, etc.
349 To improve this area, we've added disks parameters, which are
350 customisable at cluster and node group level, and which allow to
351 specify various parameters for disks (DRBD has the most parameters
352 currently), for example:
354 - DRBD resync algorithm and parameters (e.g. speed)
355 - the default VG for meta-data volumes for DRBD
356 - number of stripes for LVM (plain disk template)
359 These parameters can be modified via ``gnt-cluster modify -D …`` and
360 ``gnt-group modify -D …``, and are used at either instance creation (in
361 case of LVM stripes, for example) or at disk “activation” time
364 Rados block device support
365 ++++++++++++++++++++++++++
367 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
368 denoted by the ``rbd`` disk template type. This is considered
369 experimental, feedback is welcome. For details on configuring it, see
370 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
375 The existing master IP functionality works well only in simple setups (a
376 single network shared by all nodes); however, if nodes belong to
377 different networks, then the ``/32`` setup and lack of routing
378 information is not enough.
380 To allow the master IP to function well in more complex cases, the
381 system was reworked as follows:
383 - a master IP netmask setting has been added
384 - the master IP activation/turn-down code was moved from the node daemon
386 - whether to run the Ganeti-supplied master IP script or a user-supplied
387 on is a ``gnt-cluster init`` setting
389 Details about the location of the standard and custom setup scripts are
390 in the man page :manpage:`gnt-cluster(8)`; for information about the
391 setup script protocol, look at the Ganeti-supplied script.
396 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
399 It is now possible to use TLS-protected connections, and when renewing
400 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
401 it is now possible to specify spice or spice CA certificates. Also, it
402 is possible to configure a password for SPICE sessions via the
403 hypervisor parameter ``spice_password_file``.
405 There are also new parameters to control the compression and streaming
406 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
407 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
408 for the spice parameters.
410 Lastly, it is now possible to see the SPICE connection information via
411 ``gnt-instance console``.
416 A new tool (``tools/ovfconverter``) has been added that supports
417 conversion between Ganeti and the `Open Virtualization Format
418 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
421 This relies on the ``qemu-img`` tool to convert the disk formats, so the
422 actual compatibility with other virtualization solutions depends on it.
427 The configuration query daemon (``ganeti-confd``) is now optional, and
428 has been rewritten in Haskell; whether to use the daemon at all, use the
429 Python (default) or the Haskell version is selectable at configure time
430 via the ``--enable-confd`` parameter, which can take one of the
431 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
432 daemon will result in a smaller footprint; for larger systems, we
433 welcome feedback on the Haskell version which might become the default
436 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
437 daemon running. The Python version doesn't implement the new call.
440 User interface changes
441 ~~~~~~~~~~~~~~~~~~~~~~
443 We have replaced the ``--disks`` option of ``gnt-instance
444 replace-disks`` with a more flexible ``--disk`` option, which allows
445 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
446 disk size and mode can be changed upon recreation (via ``gnt-instance
447 recreate-disks``, which accepts the same ``--disk`` option).
449 As many people are used to a ``show`` command, we have added that as an
450 alias to ``info`` on all ``gnt-*`` commands.
452 The ``gnt-instance grow-disk`` command has a new mode in which it can
453 accept the target size of the disk, instead of the delta; this can be
454 more safe since two runs in absolute mode will be idempotent, and
455 sometimes it's also easier to specify the desired size directly.
457 Also the handling of instances with regard to offline secondaries has
458 been improved. Instance operations should not fail because one of it's
459 secondary nodes is offline, even though it's safe to proceed.
461 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
462 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
463 minors to instance name.
468 RAPI coverage has improved, with (for example) new resources for
469 recreate-disks, node power-cycle, etc.
474 There is partial support for ``xl`` in the Xen hypervisor; feedback is
477 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
478 whether to still support Python 2.4 or move to Python 2.6 as minimum
481 Support for Fedora has been slightly improved; the provided example
482 init.d script should work better on it and the INSTALL file should
483 document the needed dependencies.
488 The deprecated ``QueryLocks`` LUXI request has been removed. Use
489 ``Query(what=QR_LOCK, ...)`` instead.
491 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
492 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
493 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
494 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
495 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
497 RAPI client: ``CertificateError`` now derives from
498 ``GanetiApiError``. This should make it more easy to handle Ganeti
501 Deprecation warnings due to PyCrypto/paramiko import in
502 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
503 make sure to run an up-to-date paramiko version, if you use this tool.
505 The QA scripts now depend on Python 2.5 or above (the main code base
506 still works with Python 2.4).
508 The configuration file (``config.data``) is now written without
509 indentation for performance reasons; if you want to edit it, it can be
510 re-formatted via ``tools/fmtjson``.
512 A number of bugs has been fixed in the cluster merge tool.
514 ``x509`` certification verification (used in import-export) has been
515 changed to allow the same clock skew as permitted by the cluster
516 verification. This will remove some rare but hard to diagnose errors in
523 *(Released Thu, 19 Jul 2012)*
525 Very few changes from rc4 to the final release, only bugfixes:
527 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
528 instance, fix CDROM booting for KVM instances)
529 - fixed node group modification of node parameters
530 - fixed issue in LUClusterVerifyGroup with multi-group clusters
531 - fixed generation of bash completion to ensure a stable ordering
538 *(Released Fri, 13 Jul 2012)*
540 Third release candidate for 2.6. The following changes were done from
543 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
544 - Fixed an inconsistency in the LUXI protocol with the provided
545 arguments (NOT backwards compatible)
546 - Fixed a bug with node groups ipolicy where ``min`` was greater than
547 the cluster ``std`` value
548 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
549 easier instance debugging on nodes (requires ``hconfd`` to work)
555 *(Released Tue, 03 Jul 2012)*
557 Second release candidate for 2.6. The following changes were done from
560 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
562 - Fixed a RAPI regression on missing beparams/memory
563 - Fixed redistribution of files on offline nodes
564 - Added possibility to run activate-disks even though secondaries are
565 offline. With this change it relaxes also the strictness on some other
566 commands which use activate disks internally:
567 * ``gnt-instance start|reboot|rename|backup|export``
568 - Made it possible to remove safely an instance if its secondaries are
570 - Made it possible to reinstall even though secondaries are offline
576 *(Released Mon, 25 Jun 2012)*
578 First release candidate for 2.6. The following changes were done from
581 - Fixed bugs with disk parameters and ``rbd`` templates as well as
583 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
584 behaviour. It supports now the modify operation
585 - ``hcheck`` implemented to analyze cluster health and possibility of
586 improving health by rebalance
587 - ``hbal`` has been improved in dealing with split instances
593 *(Released Mon, 11 Jun 2012)*
595 Second beta release of 2.6. The following changes were done from beta2
598 - Fixed ``daemon-util`` with non-root user models
599 - Fixed creation of plain instances with ``--no-wait-for-sync``
600 - Fix wrong iv_names when running ``cfgupgrade``
601 - Export more information in RAPI group queries
602 - Fixed bug when changing instance network interfaces
603 - Extended burnin to do NIC changes
604 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
605 - Changed default for DRBD barriers
606 - Fixed DRBD error reporting for syncer rate
607 - Verify the options on disk parameters
609 And of course various fixes to documentation and improved unittests and
616 *(Released Wed, 23 May 2012)*
618 First beta release of 2.6. The following changes were done from beta1 to
621 - integrated patch for distributions without ``start-stop-daemon``
622 - adapted example init.d script to work on Fedora
623 - fixed log handling in Haskell daemons
624 - adapted checks in the watcher for pycurl linked against libnss
625 - add partial support for ``xl`` instead of ``xm`` for Xen
626 - fixed a type issue in cluster verification
627 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
630 Plus integrated fixes from the 2.5 branch:
632 - fixed ``kvm-ifup`` to use ``/bin/bash``
633 - fixed parallel build failures
634 - KVM live migration when using a custom keymap
640 *(Released Tue, 24 Jul 2012)*
642 A small bugfix release, with no new features:
644 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
645 different default shell (e.g. Debian, Ubuntu)
646 - fixed KVM startup and live migration with a custom keymap (fixes Issue
647 243 and Debian bug #650664)
648 - fixed compatibility with KVM versions that don't support multiple boot
649 devices (fixes Issue 230 and Debian bug #624256)
651 Additionally, a few fixes were done to the build system (fixed parallel
652 build failures) and to the unittests (fixed race condition in test for
653 FileID functions, and the default enable/disable mode for QA test is now
660 *(Released Fri, 11 May 2012)*
662 A small bugfix release.
664 The main issues solved are on the topic of compatibility with newer LVM
667 - fixed parsing of ``lv_attr`` field
668 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
669 the ``--force`` flag is needed
671 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
672 to accept kernel 3.x too (was hardcoded to 2.6.*).
674 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
675 of hook results and that also made display of other errors suboptimal
676 was fixed; the code behaves now like 2.4 and earlier.
678 Another change in 2.5, the cleanup of the OS scripts environment, is too
679 aggressive: it removed even the ``PATH`` variable, which requires the OS
680 scripts to *always* need to export it. Since this is a bit too strict,
681 we now export a minimal PATH, the same that we export for hooks.
683 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
684 integrated into this release.
686 Finally, a few other miscellaneous changes were done (no new features,
687 just small improvements):
689 - Fix ``gnt-group --help`` display
690 - Fix hardcoded Xen kernel path
691 - Fix grow-disk handling of invalid units
692 - Update synopsis for ``gnt-cluster repair-disk-sizes``
693 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
699 *(Released Thu, 12 Apr 2012)*
701 Incompatible/important changes and bugfixes
702 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
704 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
705 resource's ``ip_check`` parameter changed from ``True`` to ``False``
706 to match the underlying LUXI interface.
707 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
708 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
709 not maintain backwards-compatibility as the underlying operation
710 changed in an incompatible way. The RAPI client can talk to old
711 servers, but it needs to be told so as the return value changed.
712 - When creating file-based instances via RAPI, the ``file_driver``
713 parameter no longer defaults to ``loop`` and must be specified.
714 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
716 - Support for the undocumented and deprecated RAPI instance creation
717 request format version 0 has been dropped. Use version 1, supported
718 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
719 - Pyparsing 1.4.6 or above is required, see :doc:`installation
720 documentation <install>`.
721 - The "cluster-verify" hooks are now executed per group by the
722 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
723 if you just run ``gnt-cluster verify``, which generates one opcode per
725 - The environment as passed to the OS scripts is cleared, and thus no
726 environment variables defined in the node daemon's environment will be
727 inherited by the scripts.
728 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
730 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
731 support operations involving multiple node groups.
732 - Offline nodes are ignored when failing over an instance.
733 - Support for KVM version 1.0, which changed the version reporting format
735 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
737 - ``Makefile`` is now compatible with Automake 1.11.2
738 - Includes all bugfixes made in the 2.4 series
743 - The ganeti-htools project has been merged into the ganeti-core source
744 tree and will be built as part of Ganeti (see :doc:`install-quick`).
745 - Implemented support for :doc:`shared storage <design-shared-storage>`.
746 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
747 GPT-style partition tables (requires `parted
748 <http://www.gnu.org/s/parted/>`_).
749 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
750 - Allowed adding tags on instance creation.
751 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
752 - Allow instances to be started in a paused state, enabling the user to
753 see the complete console output on boot using the console.
754 - Added new hypervisor flag to control default reboot behaviour
755 (``reboot_behavior``).
756 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
757 - Improved out-of-band management support:
759 - Added ``gnt-node health`` command reporting the health status of
761 - Added ``gnt-node power`` command to manage power status of nodes.
762 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
764 - Instance migration can fall back to failover if instance is not
766 - Filters can be used when listing nodes, instances, groups and locks;
767 see :manpage:`ganeti(7)` manpage.
768 - Added post-execution status as variables to :doc:`hooks <hooks>`
770 - Instance tags are exported/imported together with the instance.
771 - When given an explicit job ID, ``gnt-job info`` will work for archived
773 - Jobs can define dependencies on other jobs (not yet supported via
774 RAPI or command line, but used by internal commands and usable via
777 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
780 - Instance failover is now available as a RAPI resource
781 (``/2/instances/[instance_name]/failover``).
782 - ``gnt-instance info`` defaults to static information if primary node
784 - Opcodes have a new ``comment`` attribute.
785 - Added basic SPICE support to KVM hypervisor.
786 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
788 Node group improvements
789 ~~~~~~~~~~~~~~~~~~~~~~~
791 - ``gnt-cluster verify`` has been modified to check groups separately,
792 thereby improving performance.
793 - Node group support has been added to ``gnt-cluster verify-disks``,
794 which now operates per node group.
795 - Watcher has been changed to work better with node groups.
797 - One process and state file per node group.
798 - Slow watcher in one group doesn't block other group's watcher.
800 - Added new command, ``gnt-group evacuate``, to move all instances in a
801 node group to other groups.
802 - Added ``gnt-instance change-group`` to move an instance to another
804 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
805 per-group operations.
806 - Node groups can be tagged.
807 - Some operations switch from an exclusive to a shared lock as soon as
809 - Instance's primary and secondary nodes' groups are now available as
810 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
811 and ``snodes.group.uuid``).
816 - Numerous updates to documentation and manpages.
818 - :doc:`RAPI <rapi>` documentation now has detailed parameter
820 - Some opcode/job results are now also documented, see :doc:`RAPI
823 - A lockset's internal lock is now also visible in lock monitor.
824 - Log messages from job queue workers now contain information about the
825 opcode they're processing.
826 - ``gnt-instance console`` no longer requires the instance lock.
827 - A short delay when waiting for job changes reduces the number of LUXI
828 requests significantly.
829 - DRBD metadata volumes are overwritten with zeros during disk creation.
830 - Out-of-band commands no longer acquire the cluster lock in exclusive
832 - ``devel/upload`` now uses correct permissions for directories.
838 *(Released Fri, 23 Mar 2012)*
840 This was the sixth release candidate of the 2.5 series.
846 *(Released Mon, 9 Jan 2012)*
848 This was the fifth release candidate of the 2.5 series.
854 *(Released Thu, 27 Oct 2011)*
856 This was the fourth release candidate of the 2.5 series.
862 *(Released Wed, 26 Oct 2011)*
864 This was the third release candidate of the 2.5 series.
870 *(Released Tue, 18 Oct 2011)*
872 This was the second release candidate of the 2.5 series.
878 *(Released Tue, 4 Oct 2011)*
880 This was the first release candidate of the 2.5 series.
886 *(Released Wed, 31 Aug 2011)*
888 This was the third beta release of the 2.5 series.
894 *(Released Mon, 22 Aug 2011)*
896 This was the second beta release of the 2.5 series.
902 *(Released Fri, 12 Aug 2011)*
904 This was the first beta release of the 2.5 series.
910 *(Released Thu, 27 Oct 2011)*
912 - Fixed bug when parsing command line parameter values ending in
914 - Fixed assertion error after unclean master shutdown
915 - Disable HTTP client pool for RPC, significantly reducing memory usage
917 - Fixed queue archive creation with wrong permissions
923 *(Released Tue, 23 Aug 2011)*
927 - Fixed documentation for importing with ``--src-dir`` option
928 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
929 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
935 *(Released Fri, 5 Aug 2011)*
937 Many bug-fixes and a few small features:
939 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
940 issues when you tried to add an instance with two MAC addresses in one
942 - KVM: fixed per-instance stored UID value
943 - KVM: configure bridged NICs at migration start
944 - KVM: Fix a bug where instance will not start with never KVM versions
946 - Added OS search path to ``gnt-cluster info``
947 - Fixed an issue with ``file_storage_dir`` where you were forced to
948 provide an absolute path, but the documentation states it is a
949 relative path, the documentation was right
950 - Added a new parameter to instance stop/start called ``--no-remember``
951 that will make the state change to not be remembered
952 - Implemented ``no_remember`` at RAPI level
953 - Improved the documentation
954 - Node evacuation: don't call IAllocator if node is already empty
955 - Fixed bug in DRBD8 replace disks on current nodes
956 - Fixed bug in recreate-disks for DRBD instances
957 - Moved assertion checking locks in ``gnt-instance replace-disks``
958 causing it to abort with not owning the right locks for some situation
959 - Job queue: Fixed potential race condition when cancelling queued jobs
960 - Fixed off-by-one bug in job serial generation
961 - ``gnt-node volumes``: Fix instance names
962 - Fixed aliases in bash completion
963 - Fixed a bug in reopening log files after being sent a SIGHUP
964 - Added a flag to burnin to allow specifying VCPU count
965 - Bugfixes to non-root Ganeti configuration
971 *(Released Thu, 12 May 2011)*
973 Many bug-fixes and a few new small features:
975 - Fixed a bug related to log opening failures
976 - Fixed a bug in instance listing with orphan instances
977 - Fixed a bug which prevented resetting the cluster-level node parameter
978 ``oob_program`` to the default
979 - Many fixes related to the ``cluster-merge`` tool
980 - Fixed a race condition in the lock monitor, which caused failures
981 during (at least) creation of many instances in parallel
982 - Improved output for gnt-job info
983 - Removed the quiet flag on some ssh calls which prevented debugging
985 - Improved the N+1 failure messages in cluster verify by actually
986 showing the memory values (needed and available)
987 - Increased lock attempt timeouts so that when executing long operations
988 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
989 too early and thus prevent the use of the 'fair' mechanism
990 - Changed instance query data (``gnt-instance info``) to not acquire
991 locks unless needed, thus allowing its use on locked instance if only
992 static information is asked for
993 - Improved behaviour with filesystems that do not support rename on an
995 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
996 kept locks on all nodes during the wipe, which is unneeded
997 - Fixed ``gnt-watcher`` handling of errors during hooks execution
998 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
999 10GiB) which caused the wipe to fail right at the end in some cases
1000 - Fixed master IP activation when doing master failover with no-voting
1001 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1002 the master node itself
1003 - Fixed potential data-loss in under disk full conditions, where Ganeti
1004 wouldn't check correctly the return code and would consider
1005 partially-written files 'correct'
1006 - Fixed bug related to multiple VGs and DRBD disk replacing
1007 - Added new disk parameter ``metavg`` that allows placement of the meta
1008 device for DRBD in a different volume group
1009 - Fixed error handling in the node daemon when the system libc doesn't
1010 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1011 - Fixed lock release during replace-disks, which kept cluster-wide locks
1012 when doing disk replaces with an iallocator script
1013 - Added check for missing bridges in cluster verify
1014 - Handle EPIPE errors while writing to the terminal better, so that
1015 piping the output to e.g. ``less`` doesn't cause a backtrace
1016 - Fixed rare case where a ^C during Luxi calls could have been
1017 interpreted as server errors, instead of simply terminating
1018 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1020 - Added a few more parameters to the KVM hypervisor, allowing a second
1021 CDROM, custom disk type for CDROMs and a floppy image
1022 - Removed redundant message in instance rename when the name is given
1024 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1025 disks on new nodes, allowing recreation when the original instance
1026 nodes are completely gone
1027 - Added option when converting disk templates to DRBD to skip waiting
1028 for the resync, in order to make the instance available sooner
1029 - Added two new variables to the OS scripts environment (containing the
1031 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1032 to allow use of ``pvgrub`` as bootloader
1033 - Changed the instance memory modifications to only check out-of-memory
1034 conditions on memory increases, and turned the secondary node warnings
1035 into errors (they can still be overridden via ``--force``)
1036 - Fixed the handling of a corner case when the Python installation gets
1037 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1038 to execute a command that doesn't exist
1039 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1040 node of the instance returned failures during instance shutdown; this
1041 adds the option ``--ignore-consistency`` to gnt-instance move
1043 And as usual, various improvements to the error messages, documentation
1050 *(Released Wed, 09 Mar 2011)*
1052 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1053 the RAPI users file if run twice (even with ``--dry-run``).
1055 The release fixes that bug (nothing else changed).
1061 *(Released Mon, 07 Mar 2011)*
1063 Final 2.4.0 release. Just a few small fixes:
1065 - Fixed RAPI node evacuate
1066 - Fixed the kvm-ifup script
1067 - Fixed internal error handling for special job cases
1068 - Updated man page to specify the escaping feature for options
1074 *(Released Mon, 28 Feb 2011)*
1076 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1077 that this feature wiped the disks of the wrong instance, leading to loss
1082 - Fixed title of query field containing instance name
1083 - Expanded the glossary in the documentation
1084 - Fixed one unittest (internal issue)
1090 *(Released Mon, 21 Feb 2011)*
1092 A number of bug fixes plus just a couple functionality changes.
1094 On the user-visible side, the ``gnt-* list`` command output has changed
1095 with respect to "special" field states. The current rc1 style of display
1096 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1097 the default output mode special fields states are displayed as follows:
1099 - Offline resource: ``*``
1100 - Unavailable/not applicable: ``-``
1101 - Data missing (RPC failure): ``?``
1102 - Unknown field: ``??``
1104 Another user-visible change is the addition of ``--force-join`` to
1109 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1110 - Fixed regression in RAPI/instance reinstall where all parameters were
1111 required (instead of optional)
1112 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1113 - Fixed iallocator usage (offline nodes were not considered offline)
1114 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1115 - Fixed hypervisor and OS parameter validation with respect to
1116 non-vm_capable nodes
1117 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1119 - Fixed ``tools/listrunner`` with respect to agent-based usage
1125 *(Released Fri, 4 Feb 2011)*
1127 Many changes and fixes since the beta1 release. While there were some
1128 internal changes, the code has been mostly stabilised for the RC
1131 Note: the dumb allocator was removed in this release, as it was not kept
1132 up-to-date with the IAllocator protocol changes. It is recommended to
1133 use the ``hail`` command from the ganeti-htools package.
1135 Note: the 2.4 and up versions of Ganeti are not compatible with the
1136 0.2.x branch of ganeti-htools. You need to upgrade to
1137 ganeti-htools-0.3.0 (or later).
1139 Regressions fixed from 2.3
1140 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1142 - Fixed the ``gnt-cluster verify-disks`` command
1143 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1145 - Fixed disk adoption breakage
1146 - Fixed wrong headers in instance listing for field aliases
1151 - Fixed corner case in KVM handling of NICs
1152 - Fixed many cases of wrong handling of non-vm_capable nodes
1153 - Fixed a bug where a missing instance symlink was not possible to
1154 recreate with any ``gnt-*`` command (now ``gnt-instance
1155 activate-disks`` does it)
1156 - Fixed the volume group name as reported by ``gnt-cluster
1158 - Increased timeouts for the import-export code, hopefully leading to
1159 fewer aborts due network or instance timeouts
1160 - Fixed bug in ``gnt-node list-storage``
1161 - Fixed bug where not all daemons were started on cluster
1162 initialisation, but only at the first watcher run
1163 - Fixed many bugs in the OOB implementation
1164 - Fixed watcher behaviour in presence of instances with offline
1166 - Fixed instance list output for instances running on the wrong node
1167 - a few fixes to the cluster-merge tool, but it still cannot merge
1168 multi-node groups (currently it is not recommended to use this tool)
1174 - Improved network configuration for the KVM hypervisor
1175 - Added e1000 as a supported NIC for Xen-HVM
1176 - Improved the lvmstrap tool to also be able to use partitions, as
1177 opposed to full disks
1178 - Improved speed of disk wiping (the cluster parameter
1179 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1180 of instance creations
1181 - Added documentation for the OS parameters
1182 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1183 hypervisor is not responding
1184 - Added display of blacklisted and hidden OS information in
1185 ``gnt-cluster info``
1186 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1187 NIC and node parameters, which might create problems with currently
1188 invalid (but undetected) configuration files, but prevents validation
1189 failures when unrelated parameters are modified
1190 - Changed cluster initialisation to wait for the master daemon to become
1192 - Expanded the RAPI interface:
1194 - Added config redistribution resource
1195 - Added activation/deactivation of instance disks
1196 - Added export of console information
1198 - Implemented log file reopening on SIGHUP, which allows using
1199 logrotate(8) for the Ganeti log files
1200 - Added a basic OOB helper script as an example
1206 *(Released Fri, 14 Jan 2011)*
1211 - Fixed timezone issues when formatting timestamps
1212 - Added support for node groups, available via ``gnt-group`` and other
1214 - Added out-of-band framework and management, see :doc:`design
1215 document <design-oob>`
1216 - Removed support for roman numbers from ``gnt-node list`` and
1217 ``gnt-instance list``.
1218 - Allowed modification of master network interface via ``gnt-cluster
1219 modify --master-netdev``
1220 - Accept offline secondaries while shutting down instance disks
1221 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1222 - Added support for multiple LVM volume groups
1223 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1225 - Added commands to list available fields:
1227 - ``gnt-node list-fields``
1228 - ``gnt-group list-fields``
1229 - ``gnt-instance list-fields``
1231 - Updated documentation and man pages
1236 - Moved ``rapi_users`` file into separate directory, now named
1237 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1239 - Added new tool for running commands on many machines,
1240 ``tools/ganeti-listrunner``
1241 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1242 improving the ``gnt-instance console`` output
1243 - Allowed customisation of disk index separator at ``configure`` time
1244 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1245 - Added support for non-partitioned md disks in ``lvmstrap``
1246 - Added script to gracefully power off KVM instances
1247 - Split ``utils`` module into smaller parts
1248 - Changed query operations to return more detailed information, e.g.
1249 whether an information is unavailable due to an offline node. To use
1250 this new functionality, the LUXI call ``Query`` must be used. Field
1251 information is now stored by the master daemon and can be retrieved
1252 using ``QueryFields``. Instances, nodes and groups can also be queried
1253 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1254 exposed via RAPI). The following commands make use of this
1255 infrastructure change:
1257 - ``gnt-group list``
1258 - ``gnt-group list-fields``
1260 - ``gnt-node list-fields``
1261 - ``gnt-instance list``
1262 - ``gnt-instance list-fields``
1263 - ``gnt-debug locks``
1268 - New RAPI resources (see :doc:`rapi`):
1272 - ``/2/groups/[group_name]``
1273 - ``/2/groups/[group_name]/assign-nodes``
1274 - ``/2/groups/[group_name]/modify``
1275 - ``/2/groups/[group_name]/rename``
1276 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1280 - Implemented ``no_install`` for instance creation
1281 - Implemented OS parameters for instance reinstallation, allowing
1282 use of special settings on reinstallation (e.g. for preserving data)
1287 - Added IPv6 support in import/export
1288 - Pause DRBD synchronization while wiping disks on instance creation
1289 - Updated unittests and QA scripts
1290 - Improved network parameters passed to KVM
1291 - Converted man pages from docbook to reStructuredText
1297 *(Released Mon, 20 Dec 2010)*
1299 Released version 2.3.1~rc1 without any changes.
1305 *(Released Wed, 1 Dec 2010)*
1307 - impexpd: Disable OpenSSL compression in socat if possible (backport
1308 from master, commit e90739d625b, see :doc:`installation guide
1309 <install-quick>` for details)
1310 - Changed unittest coverage report to exclude test scripts
1311 - Added script to check version format
1317 *(Released Wed, 1 Dec 2010)*
1319 Released version 2.3.0~rc1 without any changes.
1325 *(Released Fri, 19 Nov 2010)*
1327 A number of bugfixes and documentation updates:
1329 - Update ganeti-os-interface documentation
1330 - Fixed a bug related to duplicate MACs or similar items which should be
1332 - Fix breakage in OS state modify
1333 - Reinstall instance: disallow offline secondaries (fixes bug related to
1334 OS changing but reinstall failing)
1335 - plus all the other fixes between 2.2.1 and 2.2.2
1341 *(Released Tue, 2 Nov 2010)*
1343 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1344 - Fixed master failover race with watcher
1345 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1347 - Accept previously stopped instance for export with instance removal
1348 - Simplify and extend the environment variables for instance OS scripts
1349 - Added new node flags, ``master_capable`` and ``vm_capable``
1350 - Added optional instance disk wiping prior during allocation. This is a
1351 cluster-wide option and can be set/modified using
1352 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1353 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1354 :doc:`install-quick`
1355 - Added a new watcher option (``--ignore-pause``)
1356 - Added option to ignore offline node on instance start/stop
1357 (``--ignore-offline``)
1358 - Allow overriding OS parameters with ``gnt-instance reinstall``
1359 - Added ability to change node's secondary IP address using ``gnt-node
1361 - Implemented privilege separation for all daemons except
1362 ``ganeti-noded``, see ``configure`` options
1363 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1365 - Implemented job priorities (see ``ganeti(7)`` manpage)
1366 - Ignore failures while shutting down instances during failover from
1368 - Exit daemon's bootstrap process only once daemon is ready
1369 - Export more information via ``LUInstanceQuery``/remote API
1370 - Improved documentation, QA and unittests
1371 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1372 restart if the file was created or changed
1373 - Added LUXI protocol version sent with each request and response,
1374 allowing detection of server/client mismatches
1375 - Moved the Python scripts among gnt-* and ganeti-* into modules
1376 - Moved all code related to setting up SSH to an external script,
1378 - Infrastructure changes for node group support in future versions
1384 *(Released Fri, 19 Nov 2010)*
1386 A few small bugs fixed, and some improvements to the build system:
1388 - Fix documentation regarding conversion to drbd
1389 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1391 - Fix error handling in node modify with multiple changes
1392 - Allow remote imports without checked names
1398 *(Released Tue, 19 Oct 2010)*
1400 - Disable SSL session ID cache in RPC client
1406 *(Released Thu, 14 Oct 2010)*
1408 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1409 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1411 - Fix problems with interaction between Curl and Python's HTTP server,
1412 resulting in increased speed in many RPC calls
1413 - Improve our release script to prevent breakage with older aclocal and
1420 *(Released Thu, 7 Oct 2010)*
1422 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1423 value retrieved from master
1424 - Added support for blacklisted or hidden OS definitions
1425 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1426 - Added support for -mem-path in KVM hypervisor abstraction layer
1427 - Allow overriding instance parameters in tool for inter-cluster
1428 instance moves (``tools/move-instance``)
1429 - Improved opcode summaries (e.g. in ``gnt-job list``)
1430 - Improve consistency of OS listing by sorting it
1431 - Documentation updates
1437 *(Released Fri, 8 Oct 2010)*
1439 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1445 *(Released Mon, 4 Oct 2010)*
1447 - Fixed regression in ``gnt-instance rename``
1453 *(Released Wed, 22 Sep 2010)*
1455 - Fixed OS_VARIANT variable for OS scripts
1456 - Fixed cluster tag operations via RAPI
1457 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1458 - Disabled RAPI CA checks in watcher
1464 *(Released Mon, 23 Aug 2010)*
1466 - Support DRBD versions of the format "a.b.c.d"
1468 - Re-introduce support for usage from multiple threads in RAPI client
1469 - Instance renames and modify via RAPI
1470 - Work around race condition between processing and archival in job
1472 - Mark opcodes following failed one as failed, too
1473 - Job field ``lock_status`` was removed due to difficulties making it
1474 work with the changed job queue in Ganeti 2.2; a better way to monitor
1475 locks is expected for a later 2.2.x release
1476 - Fixed dry-run behaviour with many commands
1477 - Support ``ssh-agent`` again when adding nodes
1478 - Many additional bugfixes
1484 *(Released Fri, 30 Jul 2010)*
1486 Important change: the internal RPC mechanism between Ganeti nodes has
1487 changed from using a home-grown http library (based on the Python base
1488 libraries) to use the PycURL library. This requires that PycURL is
1489 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1490 against GnuTLS by default. cURL's support for GnuTLS had known issues
1491 before cURL 7.21.0 and we recommend using the latest cURL release or
1492 linking against OpenSSL. Most other distributions already link PycURL
1493 and cURL against OpenSSL. The command::
1495 python -c 'import pycurl; print pycurl.version'
1497 can be used to determine the libraries PycURL and cURL are linked
1500 Other significant changes:
1502 - Rewrote much of the internals of the job queue, in order to achieve
1503 better parallelism; this decouples job query operations from the job
1504 processing, and it should allow much nicer behaviour of the master
1505 daemon under load, and it also has uncovered some long-standing bugs
1506 related to the job serialisation (now fixed)
1507 - Added a default iallocator setting to the cluster parameters,
1508 eliminating the need to always pass nodes or an iallocator for
1509 operations that require selection of new node(s)
1510 - Added experimental support for the LXC virtualization method
1511 - Added support for OS parameters, which allows the installation of
1512 instances to pass parameter to OS scripts in order to customise the
1514 - Added a hypervisor parameter controlling the migration type (live or
1515 non-live), since hypervisors have various levels of reliability; this
1516 has renamed the 'live' parameter to 'mode'
1517 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1518 logical volumes, meaning that cluster verify will ignore them and not
1519 flag their presence as errors
1520 - The watcher will now reset the error count for failed instances after
1521 8 hours, thus allowing self-healing if the problem that caused the
1522 instances to be down/fail to start has cleared in the meantime
1523 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1524 check for, and warn, if the drbd module parameter ``usermode_helper``
1525 is not consistent with the cluster-wide setting; this is needed to
1526 make diagnose easier of failed drbd creations
1527 - Started adding base IPv6 support, but this is not yet
1528 enabled/available for use
1529 - Rename operations (cluster, instance) will now return the new name,
1530 which is especially useful if a short name was passed in
1531 - Added support for instance migration in RAPI
1532 - Added a tool to pre-configure nodes for the SSH setup, before joining
1533 them to the cluster; this will allow in the future a simplified model
1534 for node joining (but not yet fully enabled in 2.2); this needs the
1535 paramiko python library
1536 - Fixed handling of name-resolving errors
1537 - Fixed consistency of job results on the error path
1538 - Fixed master-failover race condition when executed multiple times in
1540 - Fixed many bugs related to the job queue (mostly introduced during the
1541 2.2 development cycle, so not all are impacting 2.1)
1542 - Fixed instance migration with missing disk symlinks
1543 - Fixed handling of unknown jobs in ``gnt-job archive``
1544 - And many other small fixes/improvements
1548 - Enhanced both the unittest and the QA coverage
1549 - Switched the opcode validation to a generic model, and extended the
1550 validation to all opcode parameters
1551 - Changed more parts of the code that write shell scripts to use the
1553 - Switched the master daemon to use the asyncore library for the Luxi
1560 *(Released Thu, 17 Jun 2010)*
1562 - Added tool (``move-instance``) and infrastructure to move instances
1563 between separate clusters (see :doc:`separate documentation
1564 <move-instance>` and :doc:`design document <design-2.2>`)
1565 - Added per-request RPC timeout
1566 - RAPI now requires a Content-Type header for requests with a body (e.g.
1567 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1568 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1569 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1571 - Implemented initial support for running Ganeti daemons as separate
1572 users, see configure-time flags ``--with-user-prefix`` and
1573 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1575 - Instances can be removed after export (``gnt-backup export
1576 --remove-instance``)
1577 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1578 key (instead of 1024 bit)
1579 - Added new cluster configuration file for cluster domain secret
1580 - Import/export now use SSL instead of SSH
1581 - Added support for showing estimated time when exporting an instance,
1582 see the ``ganeti-os-interface(7)`` manpage and look for
1589 *(Released Tue, 16 Nov 2010)*
1591 Some more bugfixes. Unless critical bugs occur, this will be the last
1594 - Fix case of MAC special-values
1595 - Fix mac checker regex
1596 - backend: Fix typo causing "out of range" error
1597 - Add missing --units in gnt-instance list man page
1603 *(Released Tue, 24 Aug 2010)*
1606 - Don't ignore secondary node silently on non-mirrored disk templates
1608 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1609 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1610 - Properly document the usb_mouse parameter
1611 - Fix path in ganeti-rapi(8) (issue 116)
1612 - Adjust error message when the ganeti user's .ssh directory is
1614 - Add same-node-check when changing the disk template to drbd
1620 *(Released Fri, 16 Jul 2010)*
1623 - Add an option to only select some reboot types during qa/burnin.
1624 (on some hypervisors consequent reboots are not supported)
1625 - Fix infrequent race condition in master failover. Sometimes the old
1626 master ip address would be still detected as up for a short time
1627 after it was removed, causing failover to fail.
1628 - Decrease mlockall warnings when the ctypes module is missing. On
1629 Python 2.4 we support running even if no ctypes module is installed,
1630 but we were too verbose about this issue.
1631 - Fix building on old distributions, on which man doesn't have a
1633 - Fix RAPI not to ignore the MAC address on instance creation
1634 - Implement the old instance creation format in the RAPI client.
1640 *(Released Thu, 01 Jul 2010)*
1642 A small bugfix release:
1643 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1644 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1645 a non-existing option
1646 - Fix instance create: the --force-variant option was ignored
1647 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1648 - Fix modify node storage with non-FQDN arguments
1649 - Fix RAPI client to authenticate under Python 2.6 when used
1650 for more than 5 requests needing authentication
1651 - Fix gnt-instance modify -t (storage) giving a wrong error message
1652 when converting a non-shutdown drbd instance to plain
1658 *(Released Fri, 18 Jun 2010)*
1660 A small bugfix release:
1662 - Fix live migration of KVM instances started with older Ganeti
1663 versions which had fewer hypervisor parameters
1664 - Fix gnt-instance grow-disk on down instances
1665 - Fix an error-reporting bug during instance migration
1666 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1667 silently ignoring broken ones
1668 - Fix an RPC error reporting bug affecting, for example, RAPI client
1670 - Fix bug triggered by different API version os-es on different nodes
1671 - Fix a bug in instance startup with custom hvparams: OS level
1672 parameters would fail to be applied.
1673 - Fix the RAPI client under Python 2.6 (but more work is needed to
1674 make it work completely well with OpenSSL)
1675 - Fix handling of errors when resolving names from DNS
1681 *(Released Thu, 3 Jun 2010)*
1683 A medium sized development cycle. Some new features, and some
1684 fixes/small improvements/cleanups.
1686 Significant features
1687 ~~~~~~~~~~~~~~~~~~~~
1689 The node deamon now tries to mlock itself into memory, unless the
1690 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1691 its logs, and falls back to console logging. This allows emergency
1692 features such as ``gnt-node powercycle`` to work even in the event of a
1693 broken node disk (tested offlining the disk hosting the node's
1694 filesystem and dropping its memory caches; don't try this at home)
1696 KVM: add vhost-net acceleration support. It can be tested with a new
1697 enough version of the kernel and of qemu-kvm.
1699 KVM: Add instance chrooting feature. If you use privilege dropping for
1700 your VMs you can also now force them to chroot to an empty directory,
1701 before starting the emulated guest.
1703 KVM: Add maximum migration bandwith and maximum downtime tweaking
1704 support (requires a new-enough version of qemu-kvm).
1706 Cluster verify will now warn if the master node doesn't have the master
1707 ip configured on it.
1709 Add a new (incompatible) instance creation request format to RAPI which
1710 supports all parameters (previously only a subset was supported, and it
1711 wasn't possible to extend the old format to accomodate all the new
1712 features. The old format is still supported, and a client can check for
1713 this feature, before using it, by checking for its presence in the
1714 ``features`` RAPI resource.
1716 Now with ancient latin support. Try it passing the ``--roman`` option to
1717 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1718 (requires the python-roman module to be installed, in order to work).
1723 As usual many internal code refactorings, documentation updates, and
1726 - Lots of improvements and cleanups to the experimental Remote API
1727 (RAPI) client library.
1728 - A new unit test suite for the core daemon libraries.
1729 - A fix to creating missing directories makes sure the umask is not
1730 applied anymore. This enforces the same directory permissions
1732 - Better handling terminating daemons with ctrl+c (used when running
1733 them in debugging mode).
1734 - Fix a race condition in live migrating a KVM instance, when stat()
1735 on the old proc status file returned EINVAL, which is an unexpected
1737 - Fixed manpage checking with newer man and utf-8 charachters. But now
1738 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1744 *(Released Fri, 7 May 2010)*
1746 Fix a bug which prevented untagged KVM instances from starting.
1752 *(Released Fri, 7 May 2010)*
1754 Another release with a long development cycle, during which many
1755 different features were added.
1757 Significant features
1758 ~~~~~~~~~~~~~~~~~~~~
1760 The KVM hypervisor now can run the individual instances as non-root, to
1761 reduce the impact of a VM being hijacked due to bugs in the
1762 hypervisor. It is possible to run all instances as a single (non-root)
1763 user, to manually specify a user for each instance, or to dynamically
1764 allocate a user out of a cluster-wide pool to each instance, with the
1765 guarantee that no two instances will run under the same user ID on any
1768 An experimental RAPI client library, that can be used standalone
1769 (without the other Ganeti libraries), is provided in the source tree as
1770 ``lib/rapi/client.py``. Note this client might change its interface in
1771 the future, as we iterate on its capabilities.
1773 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1774 replace the cluster's certificates and crypto keys. This might help in
1775 case they have been compromised, or have simply expired.
1777 A new disk option for instance creation has been added that allows one
1778 to "adopt" currently existing logical volumes, with data
1779 preservation. This should allow easier migration to Ganeti from
1780 unmanaged (or managed via other software) instances.
1782 Another disk improvement is the possibility to convert between redundant
1783 (DRBD) and plain (LVM) disk configuration for an instance. This should
1784 allow better scalability (starting with one node and growing the
1785 cluster, or shrinking a two-node cluster to one node).
1787 A new feature that could help with automated node failovers has been
1788 implemented: if a node sees itself as offline (by querying the master
1789 candidates), it will try to shutdown (hard) all instances and any active
1790 DRBD devices. This reduces the risk of duplicate instances if an
1791 external script automatically failovers the instances on such nodes. To
1792 enable this, the cluster parameter ``maintain_node_health`` should be
1793 enabled; in the future this option (per the name) will enable other
1794 automatic maintenance features.
1796 Instance export/import now will reuse the original instance
1797 specifications for all parameters; that means exporting an instance,
1798 deleting it and the importing it back should give an almost identical
1799 instance. Note that the default import behaviour has changed from
1800 before, where it created only one NIC; now it recreates the original
1803 Cluster verify has added a few new checks: SSL certificates validity,
1804 /etc/hosts consistency across the cluster, etc.
1809 As usual, many internal changes were done, documentation fixes,
1812 - Fixed cluster initialization with disabled cluster storage (regression
1813 introduced in 2.1.1)
1814 - File-based storage supports growing the disks
1815 - Fixed behaviour of node role changes
1816 - Fixed cluster verify for some corner cases, plus a general rewrite of
1817 cluster verify to allow future extension with more checks
1818 - Fixed log spamming by watcher and node daemon (regression introduced
1820 - Fixed possible validation issues when changing the list of enabled
1822 - Fixed cleanup of /etc/hosts during node removal
1823 - Fixed RAPI response for invalid methods
1824 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1825 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1826 from ide disks, etc.)
1827 - Allow OS changes without re-installation (to record a changed OS
1828 outside of Ganeti, or to allow OS renames)
1829 - Allow instance creation without OS installation (useful for example if
1830 the OS will be installed manually, or restored from a backup not in
1832 - Implemented option to make cluster ``copyfile`` use the replication
1834 - Added list of enabled hypervisors to ssconf (possibly useful for
1836 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1838 - A partial form of node re-IP is possible via node readd, which now
1839 allows changed node primary IP
1840 - Command line utilities now show an informational message if the job is
1842 - The logs of the master daemon now show the PID/UID/GID of the
1849 *(Released Fri, 12 Mar 2010)*
1851 During the 2.1.0 long release candidate cycle, a lot of improvements and
1852 changes have accumulated with were released later as 2.1.1.
1857 The node evacuate command (``gnt-node evacuate``) was significantly
1858 rewritten, and as such the IAllocator protocol was changed - a new
1859 request type has been added. This unfortunate change during a stable
1860 series is designed to improve performance of node evacuations; on
1861 clusters with more than about five nodes and which are well-balanced,
1862 evacuation should proceed in parallel for all instances of the node
1863 being evacuated. As such, any existing IAllocator scripts need to be
1864 updated, otherwise the above command will fail due to the unknown
1865 request. The provided "dumb" allocator has not been updated; but the
1866 ganeti-htools package supports the new protocol since version 0.2.4.
1868 Another important change is increased validation of node and instance
1869 names. This might create problems in special cases, if invalid host
1870 names are being used.
1872 Also, a new layer of hypervisor parameters has been added, that sits at
1873 OS level between the cluster defaults and the instance ones. This allows
1874 customisation of virtualization parameters depending on the installed
1875 OS. For example instances with OS 'X' may have a different KVM kernel
1876 (or any other parameter) than the cluster defaults. This is intended to
1877 help managing a multiple OSes on the same cluster, without manual
1878 modification of each instance's parameters.
1880 A tool for merging clusters, ``cluster-merge``, has been added in the
1881 tools sub-directory.
1886 - Improved the int/float conversions that should make the code more
1887 robust in face of errors from the node daemons
1888 - Fixed the remove node code in case of internal configuration errors
1889 - Fixed the node daemon behaviour in face of inconsistent queue
1890 directory (e.g. read-only file-system where we can't open the files
1892 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1893 now it either aborts cleanly or, if given the new "auto_promote"
1894 parameter, will automatically promote other nodes as needed
1895 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1896 completely prevent Ganeti from working
1897 - Fixed bug for instance export when not all disks were successfully
1899 - Fixed behaviour of node add when the new node is slow in starting up
1901 - Fixed handling of signals in the LUXI client, which should improve
1902 behaviour of command-line scripts
1903 - Added checks for invalid node/instance names in the configuration (now
1904 flagged during cluster verify)
1905 - Fixed watcher behaviour for disk activation errors
1906 - Fixed two potentially endless loops in http library, which led to the
1907 RAPI daemon hanging and consuming 100% CPU in some cases
1908 - Fixed bug in RAPI daemon related to hashed passwords
1909 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1911 - Enhanced compatibility with non-Debian OSes, but not using absolute
1912 path in some commands and allowing customisation of the ssh
1913 configuration directory
1914 - Fixed possible future issue with new Python versions by abiding to the
1915 proper use of ``__slots__`` attribute on classes
1916 - Added checks that should prevent directory traversal attacks
1917 - Many documentation fixes based on feedback from users
1922 - Added an "early_release" more for instance replace disks and node
1923 evacuate, where we release locks earlier and thus allow higher
1924 parallelism within the cluster
1925 - Added watcher hooks, intended to allow the watcher to restart other
1926 daemons (e.g. from the ganeti-nbma project), but they can be used of
1927 course for any other purpose
1928 - Added a compile-time disable for DRBD barriers, to increase
1929 performance if the administrator trusts the power supply or the
1930 storage system to not lose writes
1931 - Added the option of using syslog for logging instead of, or in
1932 addition to, Ganeti's own log files
1933 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1934 can indeed boot from a paravirtual NIC
1935 - Added a generic debug level for many operations; while this is not
1936 used widely yet, it allows one to pass the debug value all the way to
1938 - Enhanced the hooks environment for instance moves (failovers,
1939 migrations) where the primary/secondary nodes changed during the
1940 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1941 - Enhanced data validations for many user-supplied values; one important
1942 item is the restrictions imposed on instance and node names, which
1943 might reject some (invalid) host names
1944 - Add a configure-time option to disable file-based storage, if it's not
1945 needed; this allows greater security separation between the master
1946 node and the other nodes from the point of view of the inter-node RPC
1948 - Added user notification in interactive tools if job is waiting in the
1949 job queue or trying to acquire locks
1950 - Added log messages when a job is waiting for locks
1951 - Added filtering by node tags in instance operations which admit
1952 multiple instances (start, stop, reboot, reinstall)
1953 - Added a new tool for cluster mergers, ``cluster-merge``
1954 - Parameters from command line which are of the form ``a=b,c=d`` can now
1955 use backslash escapes to pass in values which contain commas,
1956 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1958 - For KVM, the instance name is the first parameter passed to KVM, so
1959 that it's more visible in the process list
1965 *(Released Tue, 2 Mar 2010)*
1967 Ganeti 2.1 brings many improvements with it. Major changes:
1969 - Added infrastructure to ease automated disk repairs
1970 - Added new daemon to export configuration data in a cheaper way than
1971 using the remote API
1972 - Instance NICs can now be routed instead of being associated with a
1974 - Improved job locking logic to reduce impact of jobs acquiring multiple
1975 locks waiting for other long-running jobs
1977 In-depth implementation details can be found in the Ganeti 2.1 design
1983 - Added chroot hypervisor
1984 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1986 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1987 ``bootloader_path`` and ``bootloader_args``)
1988 - Added the ``use_localtime`` option for the xen-hvm and kvm
1989 hypervisors, and the default value for this has changed to false (in
1990 2.0 xen-hvm always enabled it)
1991 - Added luxi call to submit multiple jobs in one go
1992 - Added cluster initialization option to not modify ``/etc/hosts``
1994 - Added network interface parameters
1995 - Added dry run mode to some LUs
1996 - Added RAPI resources:
1998 - ``/2/instances/[instance_name]/info``
1999 - ``/2/instances/[instance_name]/replace-disks``
2000 - ``/2/nodes/[node_name]/evacuate``
2001 - ``/2/nodes/[node_name]/migrate``
2002 - ``/2/nodes/[node_name]/role``
2003 - ``/2/nodes/[node_name]/storage``
2004 - ``/2/nodes/[node_name]/storage/modify``
2005 - ``/2/nodes/[node_name]/storage/repair``
2007 - Added OpCodes to evacuate or migrate all instances on a node
2008 - Added new command to list storage elements on nodes (``gnt-node
2009 list-storage``) and modify them (``gnt-node modify-storage``)
2010 - Added new ssconf files with master candidate IP address
2011 (``ssconf_master_candidates_ips``), node primary IP address
2012 (``ssconf_node_primary_ips``) and node secondary IP address
2013 (``ssconf_node_secondary_ips``)
2014 - Added ``ganeti-confd`` and a client library to query the Ganeti
2015 configuration via UDP
2016 - Added ability to run hooks after cluster initialization and before
2018 - Added automatic mode for disk replace (``gnt-instance replace-disks
2020 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2021 after catastrophic data-loss
2022 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2024 - Added ``gnt-instance move`` command to move instances
2025 - Added ``gnt-cluster watcher`` command to control watcher
2026 - Added ``gnt-node powercycle`` command to powercycle nodes
2027 - Added new job status field ``lock_status``
2028 - Added parseable error codes to cluster verification (``gnt-cluster
2029 verify --error-codes``) and made output less verbose (use
2030 ``--verbose`` to restore previous behaviour)
2031 - Added UUIDs to the main config entities (cluster, nodes, instances)
2032 - Added support for OS variants
2033 - Added support for hashed passwords in the Ganeti remote API users file
2035 - Added option to specify maximum timeout on instance shutdown
2036 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2037 - Added new helper script to start and stop Ganeti daemons
2038 (``daemon-util``), with the intent to reduce the work necessary to
2039 adjust Ganeti for non-Debian distributions and to start/stop daemons
2041 - Added more unittests
2042 - Fixed critical bug in ganeti-masterd startup
2043 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2044 now customisable at the cluster level for both the KVM and Xen
2045 hypervisors using the new ``migration_port`` parameter
2046 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2047 reinstalling an instance
2048 - Allowed ``@`` in tag names
2049 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2050 - Many documentation updates
2051 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2052 - ``gnt-instance reinstall`` can now reinstall multiple instances
2053 - Updated many command line parameters
2054 - Introduced new OS API version 15
2055 - No longer support a default hypervisor
2056 - Treat virtual LVs as inexistent
2057 - Improved job locking logic to reduce lock contention
2058 - Match instance and node names case insensitively
2059 - Reimplemented bash completion script to be more complete
2066 *(Released Thu, 4 Feb 2010)*
2068 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2069 - Fix a string formatting bug
2070 - Improve safety of the code in some error paths
2071 - Improve data validation in the master of values returned from nodes
2077 *(Released Thu, 17 Dec 2009)*
2079 - Fix security issue due to missing validation of iallocator names; this
2080 allows local and remote execution of arbitrary executables
2081 - Fix failure of gnt-node list during instance removal
2082 - Ship the RAPI documentation in the archive
2088 *(Released Wed, 30 Sep 2009)*
2090 - Fixed many wrong messages
2091 - Fixed a few bugs related to the locking library
2092 - Fixed MAC checking at instance creation time
2093 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2094 - Fixed a few issues related to signal handling in both daemons and
2096 - Fixed the example startup script provided
2097 - Fixed insserv dependencies in the example startup script (patch from
2099 - Fixed handling of drained nodes in the iallocator framework
2100 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2102 - Fixed error related to invalid job IDs in job polling
2103 - Fixed job/opcode persistence on unclean master shutdown
2104 - Fixed handling of partial job processing after unclean master
2106 - Fixed error reporting from LUs, previously all errors were converted
2107 into execution errors
2108 - Fixed error reporting from burnin
2109 - Decreased significantly the memory usage of the job queue
2110 - Optimised slightly multi-job submission
2111 - Optimised slightly opcode loading
2112 - Backported the multi-job submit framework from the development
2113 branch; multi-instance start and stop should be faster
2114 - Added script to clean archived jobs after 21 days; this will reduce
2115 the size of the queue directory
2116 - Added some extra checks in disk size tracking
2117 - Added an example ethers hook script
2118 - Added a cluster parameter that prevents Ganeti from modifying of
2120 - Added more node information to RAPI responses
2121 - Added a ``gnt-job watch`` command that allows following the ouput of a
2123 - Added a bind-address option to ganeti-rapi
2124 - Added more checks to the configuration verify
2125 - Enhanced the burnin script such that some operations can be retried
2127 - Converted instance reinstall to multi-instance model
2133 *(Released Fri, 7 Aug 2009)*
2135 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2136 to allow using the pre-2.0.2 behaviour in activation, if any existing
2137 instances have mismatched disk sizes in the configuration
2138 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2139 any configuration mismatches for disk sizes
2140 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2141 failover to work on two-node clusters
2142 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2144 - Fixed detection of OS script errors in ``gnt-backup export``
2145 - Fixed exit code of ``gnt-backup export``
2151 *(Released Fri, 17 Jul 2009)*
2153 - Added experimental support for stripped logical volumes; this should
2154 enhance performance but comes with a higher complexity in the block
2155 device handling; stripping is only enabled when passing
2156 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2157 affected even in the non-stripped mode
2158 - Improved resiliency against transient failures at the end of DRBD
2159 resyncs, and in general of DRBD resync checks
2160 - Fixed a couple of issues with exports and snapshot errors
2161 - Fixed a couple of issues in instance listing
2162 - Added display of the disk size in ``gnt-instance info``
2163 - Fixed checking for valid OSes in instance creation
2164 - Fixed handling of the "vcpus" parameter in instance listing and in
2165 general of invalid parameters
2166 - Fixed http server library, and thus RAPI, to handle invalid
2167 username/password combinations correctly; this means that now they
2168 report unauthorized for queries too, not only for modifications,
2169 allowing earlier detect of configuration problems
2170 - Added a new "role" node list field, equivalent to the master/master
2171 candidate/drained/offline flags combinations
2172 - Fixed cluster modify and changes of candidate pool size
2173 - Fixed cluster verify error messages for wrong files on regular nodes
2174 - Fixed a couple of issues with node demotion from master candidate role
2175 - Fixed node readd issues
2176 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2177 - Added a new ``--no-voting`` option for masterfailover to fix failover
2178 on two-nodes clusters when the former master node is unreachable
2179 - Added instance reinstall over RAPI
2185 *(Released Tue, 16 Jun 2009)*
2187 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2188 allow re-adding the start in single-user option (regression from 1.2)
2189 - the watcher writes the instance status to a file, to allow monitoring
2190 to report the instance status (from the master) based on cached
2191 results of the watcher's queries; while this can get stale if the
2192 watcher is being locked due to other work on the cluster, this is
2193 still an improvement
2194 - the watcher now also restarts the node daemon and the rapi daemon if
2196 - fixed the watcher to handle full and drained queue cases
2197 - hooks export more instance data in the environment, which helps if
2198 hook scripts need to take action based on the instance's properties
2199 (no longer need to query back into ganeti)
2200 - instance failovers when the instance is stopped do not check for free
2201 RAM, so that failing over a stopped instance is possible in low memory
2203 - rapi uses queries for tags instead of jobs (for less job traffic), and
2204 for cluster tags it won't talk to masterd at all but read them from
2206 - a couple of error handling fixes in RAPI
2207 - drbd handling: improved the error handling of inconsistent disks after
2208 resync to reduce the frequency of "there are some degraded disks for
2209 this instance" messages
2210 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2211 error handling path called a wrong function name)
2217 *(Released Wed, 27 May 2009)*
2219 - no changes from rc5
2225 *(Released Wed, 20 May 2009)*
2227 - fix a couple of bugs (validation, argument checks)
2228 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2229 - some small improvements to RAPI and IAllocator
2230 - make watcher automatically start the master daemon if down
2236 *(Released Mon, 27 Apr 2009)*
2238 - change the OS list to not require locks; this helps with big clusters
2239 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2240 volume group is broken
2241 - ``gnt-instance info``, without any arguments, doesn't run for all
2242 instances anymore; either pass ``--all`` or pass the desired
2243 instances; this helps against mistakes on big clusters where listing
2244 the information for all instances takes a long time
2245 - miscellaneous doc and man pages fixes
2251 *(Released Wed, 8 Apr 2009)*
2253 - Change the internal locking model of some ``gnt-node`` commands, in
2254 order to reduce contention (and blocking of master daemon) when
2255 batching many creation/reinstall jobs
2256 - Fixes to Xen soft reboot
2257 - No longer build documentation at build time, instead distribute it in
2258 the archive, in order to reduce the need for the whole docbook/rst
2265 *(Released Fri, 27 Mar 2009)*
2267 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2268 - Fix watcher startup sequence, improves the behaviour of busy clusters
2269 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2270 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2271 bind address and other places
2272 - Some documentation fixes and updates
2278 *(Released Mon, 2 Mar 2009)*
2280 - More documentation updates, now all docs should be more-or-less
2282 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2284 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2286 - Fix an issue related to $libdir/run/ganeti and cluster creation
2292 *(Released Thu, 19 Feb 2009)*
2294 - Xen PVM and KVM have switched the default value for the instance root
2295 disk to the first partition on the first drive, instead of the whole
2296 drive; this means that the OS installation scripts must be changed
2298 - Man pages have been updated
2299 - RAPI has been switched by default to HTTPS, and the exported functions
2300 should all work correctly
2301 - RAPI v1 has been removed
2302 - Many improvements to the KVM hypervisor
2303 - Block device errors are now better reported
2304 - Many other bugfixes and small improvements
2310 *(Released Mon, 26 Jan 2009)*
2312 - Version 2 is a general rewrite of the code and therefore the
2313 differences are too many to list, see the design document for 2.0 in
2314 the ``doc/`` subdirectory for more details
2315 - In this beta version there is not yet a migration path from 1.2 (there
2316 will be one in the final 2.0 release)
2317 - A few significant changes are:
2319 - all commands are executed by a daemon (``ganeti-masterd``) and the
2320 various ``gnt-*`` commands are just front-ends to it
2321 - all the commands are entered into, and executed from a job queue,
2322 see the ``gnt-job(8)`` manpage
2323 - the RAPI daemon supports read-write operations, secured by basic
2324 HTTP authentication on top of HTTPS
2325 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2326 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2327 to migrate to DRBD 8 first while still running Ganeti 1.2)
2328 - DRBD devices are using statically allocated minor numbers, which
2329 will be assigned to existing instances during the migration process
2330 - there is support for both Xen PVM and Xen HVM instances running on
2332 - KVM virtualization is supported too
2333 - file-based storage has been implemented, which means that it is
2334 possible to run the cluster without LVM and DRBD storage, for
2335 example using a shared filesystem exported from shared storage (and
2336 still have live migration)
2342 *(Released Tue, 13 Jan 2009)*
2344 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2345 - Reuse the old instance mac address by default on instance import, if
2346 the instance name is the same.
2347 - Handle situations in which the node info rpc returns incomplete
2349 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2350 - Improved version of batcher:
2352 - state file support
2353 - instance mac address support
2354 - support for HVM clusters/instances
2356 - Add an option to show the number of cpu sockets and nodes in
2358 - Support OSes that handle more than one version of the OS api (but do
2359 not change the current API in any other way)
2360 - Fix ``gnt-node migrate``
2361 - ``gnt-debug`` man page
2362 - Fixes various more typos and small issues
2363 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2369 *(Released Wed, 24 Sep 2008)*
2371 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2372 type of disk exported to fully virtualized instances.
2373 - provide access to the serial console of HVM instances
2374 - instance auto_balance flag, set by default. If turned off it will
2375 avoid warnings on cluster verify if there is not enough memory to fail
2376 over an instance. in the future it will prevent automatically failing
2377 it over when we will support that.
2378 - batcher tool for instance creation, see ``tools/README.batcher``
2379 - ``gnt-instance reinstall --select-os`` to interactively select a new
2380 operating system when reinstalling an instance.
2381 - when changing the memory amount on instance modify a check has been
2382 added that the instance will be able to start. also warnings are
2383 emitted if the instance will not be able to fail over, if auto_balance
2385 - documentation fixes
2386 - sync fields between ``gnt-instance list/modify/add/import``
2387 - fix a race condition in drbd when the sync speed was set after giving
2388 the device a remote peer.
2394 *(Released Tue, 22 Jul 2008)*
2396 - note: the allowed size and number of tags per object were reduced
2397 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2398 - fixed twisted 8.x compatibility
2399 - fixed ``gnt-instance replace-disks`` with iallocator
2400 - add TCP keepalives on twisted connections to detect restarted nodes
2401 - disk increase support, see ``gnt-instance grow-disk``
2402 - implement bulk node/instance query for RAPI
2403 - add tags in node/instance listing (optional)
2404 - experimental migration (and live migration) support, read the man page
2405 for ``gnt-instance migrate``
2406 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2407 has some small improvements in handling its state file
2413 *(Released Fri, 13 Jun 2008)*
2415 - Experimental readonly, REST-based remote API implementation;
2416 automatically started on master node, TCP port 5080, if enabled by
2417 ``--enable-rapi`` parameter to configure script.
2418 - Instance allocator support. Add and import instance accept a
2419 ``--iallocator`` parameter, and call that instance allocator to decide
2420 which node to use for the instance. The iallocator document describes
2421 what's expected from an allocator script.
2422 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2423 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2424 node is lost there is still enough memory to fail over the instances
2426 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2427 ``gnt-cluster verify``, to check for site-specific compliance. All the
2428 hooks will run, and their output, if any, will be displayed. Any
2429 failing hook will make the verification return an error value.
2430 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2431 primary and secondary interfaces
2432 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2433 that is still declared as part of the cluster and has failed.
2434 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2435 specifying output fields, that just adds the chosen fields to the
2437 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2438 export from the filesystem.
2439 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2440 have been added. Using them you can enable/disable acpi and pae
2441 support, and specify a path for a cd image to be exported to the
2442 instance. These parameters as the name suggest only work on HVM
2444 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2445 PAE support will be set to the previously hardcoded values, but the
2446 (previously hardcoded) path to the CDROM ISO image will be unset and
2447 if required, needs to be set manually with ``gnt-instance modify``
2449 - The address to which an instance's VNC console is bound is now
2450 selectable per-instance, rather than being cluster wide. Of course
2451 this only applies to instances controlled via VNC, so currently just
2452 applies to HVM clusters.
2458 *(Released Mon, 18 Feb 2008)*
2460 - more tweaks to the disk activation code (especially helpful for DRBD)
2461 - change the default ``gnt-instance list`` output format, now there is
2462 one combined status field (see the manpage for the exact values this
2464 - some more fixes for the mac export to hooks change
2465 - make Ganeti not break with DRBD 8.2.x (which changed the version
2466 format in ``/proc/drbd``) (issue 24)
2467 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2468 template, allowing migration from DRBD0.7+MD to DRBD8
2474 *(Released Wed, 30 Jan 2008)*
2476 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2478 - add command aliases infrastructure and a few aliases
2479 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2480 man pages and the ``--help`` option of ``gnt-node
2481 list``/``gnt-instance list``
2482 - fix ``gnt-backup list`` with down nodes (issue 21)
2483 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2484 - fix the dist archive and add a check for including svn/git files in
2486 - some developer-related changes: improve the burnin and the QA suite,
2487 add an upload script for testing during development
2493 *(Released Wed, 16 Jan 2008)*
2495 - experimental HVM support, read the install document, section
2496 "Initializing the cluster"
2497 - allow for the PVM hypervisor per-instance kernel and initrd paths
2498 - add a new command ``gnt-cluster verify-disks`` which uses a new
2499 algorithm to improve the reconnection of the DRBD pairs if the device
2500 on the secondary node has gone away
2501 - make logical volume code auto-activate LVs at disk activation time
2502 - slightly improve the speed of activating disks
2503 - allow specification of the MAC address at instance creation time, and
2504 changing it later via ``gnt-instance modify``
2505 - fix handling of external commands that generate lots of output on
2507 - update documentation with regard to minimum version of DRBD8 supported
2513 *(Released Tue, 4 Dec 2007)*
2515 - Log the ``xm create`` output to the node daemon log on failure (to
2516 help diagnosing the error)
2517 - In debug mode, log all external commands output if failed to the logs
2518 - Change parsing of lvm commands to ignore stderr
2524 *(Released Wed, 28 Nov 2007)*
2526 - Another round of updates to the DRBD 8 code to deal with more failures
2527 in the replace secondary node operation
2528 - Some more logging of failures in disk operations (lvm, drbd)
2529 - A few documentation updates
2536 *(Released Tue, 13 Nov 2007)*
2538 - Change configuration file format from Python's Pickle to JSON.
2539 Upgrading is possible using the cfgupgrade utility.
2540 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2541 faster replace disks and is more stable (DRBD 8 has many improvements
2542 compared to DRBD 0.7)
2543 - Added command line tags support (see man pages for ``gnt-instance``,
2544 ``gnt-node``, ``gnt-cluster``)
2545 - Added instance rename support
2546 - Added multi-instance startup/shutdown
2547 - Added cluster rename support
2548 - Added ``gnt-node evacuate`` to simplify some node operations
2549 - Added instance reboot operation that can speedup reboot as compared to
2551 - Soften the requirement that hostnames are in FQDN format
2552 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2554 - Removed dependency on debian's patched fping that uses the
2555 non-standard ``-S`` option
2556 - Now the OS definitions are searched for in multiple, configurable
2557 paths (easier for distros to package)
2558 - Some changes to the hooks infrastructure (especially the new
2559 post-configuration update hook)
2560 - Other small bugfixes
2562 .. vim: set textwidth=72 syntax=rst :