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 ``$sysconfdir/default/ganeti``.
14 - A new cluster attribute 'enabled_storage_types' is introduced. It will
15 be used to manage the storage types to be used by instances in the cluster.
16 Initially, it will be set to a list that includes lvm, file and sharedfile
17 if those are enabled. Additionally, it will include all storage types that
18 are currently used by instances. The order of storage types will be based
19 on Ganeti's history of supporting them. In the future, the first entry of
20 the list will be used as a default storage type on instance creation.
26 *(Released Wed, 6 Feb 2013)*
28 - ``gnt-instance batch-create`` has been changed to use the bulk create
29 opcode from Ganeti. This lead to incompatible changes in the format of
30 the JSON file. It's now not a custom dict anymore but a dict
31 compatible with the ``OpInstanceCreate`` opcode.
32 - Parent directories for file storage need to be listed in
33 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
34 write the file automatically based on old configuration values, but it
35 can not distribute it across all nodes and the file contents should be
36 verified. Use ``gnt-cluster copyfile
37 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
38 upgraded. The reason for requiring this list of paths now is that
39 before it would have been possible to inject new paths via RPC,
40 allowing files to be created in arbitrary locations. The RPC protocol
41 is protected using SSL/X.509 certificates, but as a design principle
42 Ganeti does not permit arbitrary paths to be passed.
43 - The parsing of the variants file for OSes (see
44 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
45 lines and comment lines (starting with ``#``) are ignored for better
47 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
48 longer available. ``gnt-node add`` now invokes a new tool on the
49 destination node, named ``prepare-node-join``, to configure the SSH
50 daemon. Paramiko is no longer necessary to configure nodes' SSH
51 daemons via ``gnt-node add``.
52 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
53 for RAPI users. It allows granting permissions to query for
54 information to a specific user without giving
55 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
56 - A new tool named ``node-cleanup`` has been added. It cleans remains of
57 a cluster from a machine by stopping all daemons, removing
58 certificates and ssconf files. Unless the ``--no-backup`` option is
59 given, copies of the certificates are made.
60 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
61 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
62 the flag is not reset after a master failover.
63 - Instance creations now support the use of opportunistic locking,
64 potentially speeding up the (parallel) creation of multiple instances.
65 This feature is currently only available via the :doc:`RAPI
66 <rapi>` interface and when an instance allocator is used. If the
67 ``opportunistic_locking`` parameter is set the opcode will try to
68 acquire as many locks as possible, but will not wait for any locks
69 held by other opcodes. If not enough resources can be found to
70 allocate the instance, the temporary error code
71 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
72 retried thereafter, with or without opportunistic locking.
73 - The functionality for allocating multiple instances at once has been
74 overhauled and is now also available through :doc:`RAPI <rapi>`.
75 - Man pages can now be included when the documentation is built, in
76 which case the output is in ``doc/man-html``. The configure-time
77 option is ``--enable-manpages-in-doc``. Sphinx 1.0 or higher is
79 - A new htool called ``hroller`` helps scheduling parallel shutdown of
80 nodes on clusters where all instances are shut down, but without ever
81 rebooting both primary and secondary node for an instance at the same
83 - Queries not needing locks or RPC access to the node can now be
84 performed by the confd daemon, making them independent from jobs, and
85 thus faster to execute. This is selectable at configure time.
86 - The KVM hypervisor has been updated with many new hypervisor
87 parameters, including a generic one for passing arbitrary command line
88 values. See a complete list in :manpage:`gnt-instance(8)`.
89 - A new tool, called ``mon-collector``, is the stand-alone executor of
90 the data collectors for a monitoring system. As of this version, it
91 just includes the DRBD data collector, that can be executed by calling
92 ``mon-collector`` using the ``drbd`` parameter. See
93 :manpage:`mon-collector(7)`.
94 - New ``exclusive-storage`` node parameter added. When it's set to true,
95 physical disks are assigned in an exclusive fashion to instances, as
96 documented in :doc:`Partitioned Ganeti <design-partitioned>`.
97 Currently, only instances using the ``plain`` disk template are
99 - New network management functionality to support automatic allocation
100 of IP addresses and managing of network parameters. See
101 :manpage:`gnt-network(8)` for more details.
102 - New external storage backend, to allow managing arbitrary storage
103 systems external to the cluster. See
104 :manpage:`ganeti-extstorage-interface(7)`.
105 - Instance renames of LVM-based instances will now update the LV tags
106 (which can be used to recover the instance-to-LV mapping in case of
108 - ``hbal`` will now exit with status 0 if, during job execution over
109 LUXI, early exit has been requested and all jobs are successful;
110 before, exit status 1 was used, which cannot be differentiated from
117 *(Released Fri, 21 Dec 2012)*
119 Important behaviour change: hbal won't rebalance anymore instances which
120 have the ``auto_balance`` attribute set to false. This was the intention
121 all along, but until now it only skipped those from the N+1 memory
122 reservation (DRBD-specific).
124 A significant number of bug fixes in this release:
126 - Fixed disk adoption interaction with ipolicy checks.
127 - Fixed networking issues when instances are started, stopped or
128 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
129 - Fixed the warning in cluster verify for shared storage instances not
131 - Fixed removal of storage directory on shared file storage (issue 262).
132 - Fixed validation of LVM volume group name in OpClusterSetParams
133 (``gnt-cluster modify``) (issue 285).
134 - Fixed runtime memory increases (``gnt-instance modify -m``).
135 - Fixed live migration under Xen's ``xl`` mode.
136 - Fixed ``gnt-instance console`` with ``xl``.
137 - Fixed building with newer Haskell compiler/libraries.
138 - Fixed PID file writing in Haskell daemons (confd); this prevents
139 restart issues if confd was launched manually (outside of
140 ``daemon-util``) while another copy of it was running
141 - Fixed a type error when doing live migrations with KVM (issue 297) and
142 the error messages for failing migrations have been improved.
143 - Fixed opcode validation for the out-of-band commands (``gnt-node
145 - Fixed a type error when unsetting OS hypervisor parameters (issue
146 311); now it's possible to unset all OS-specific hypervisor
148 - Fixed the ``dry-run`` mode for many operations: verification of
149 results was over-zealous but didn't take into account the ``dry-run``
150 operation, resulting in "wrong" failures.
151 - Fixed bash completion in ``gnt-job list`` when the job queue has
152 hundreds of entries; especially with older ``bash`` versions, this
153 results in significant CPU usage.
155 And lastly, a few other improvements have been made:
157 - Added option to force master-failover without voting (issue 282).
158 - Clarified error message on lock conflict (issue 287).
159 - Logging of newly submitted jobs has been improved (issue 290).
160 - Hostname checks have been made uniform between instance rename and
162 - The ``--submit`` option is now supported by ``gnt-debug delay``.
163 - Shutting down the master daemon by sending SIGTERM now stops it from
164 processing jobs waiting for locks; instead, those jobs will be started
165 once again after the master daemon is started the next time (issue
167 - Support for Xen's ``xl`` program has been improved (besides the fixes
169 - Reduced logging noise in the Haskell confd daemon (only show one log
170 entry for each config reload, instead of two).
171 - Several man page updates and typo fixes.
177 *(Released Fri, 12 Oct 2012)*
179 A small bugfix release. Among the bugs fixed:
181 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
182 made the command unusable.
183 - Commands that issue many jobs don't fail anymore just because some jobs
184 take so long that other jobs are archived.
185 - Failures during ``gnt-instance reinstall`` are reflected by the exit
187 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
189 - When ``always_failover`` is set, ``--allow-failover`` is not required
190 in migrate commands anymore.
191 - ``bash_completion`` works even if extglob is disabled.
192 - Fixed bug with locks that made failover for RDB-based instances fail.
193 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
194 a random node instead of one based on the allocator metric.
195 - Support for newer versions of pylint and pep8.
196 - Hail doesn't fail anymore when trying to add an instance of type
197 ``file``, ``sharedfile`` or ``rbd``.
198 - Added new Makefile target to rebuild the whole distribution, so that
199 all files are included.
205 *(Released Fri, 27 Jul 2012)*
208 .. attention:: The ``LUXI`` protocol has been made more consistent
209 regarding its handling of command arguments. This, however, leads to
210 incompatibility issues with previous versions. Please ensure that you
211 restart Ganeti daemons soon after the upgrade, otherwise most
212 ``LUXI`` calls (job submission, setting/resetting the drain flag,
213 pausing/resuming the watcher, cancelling and archiving jobs, querying
214 the cluster configuration) will fail.
223 The current ``admin_up`` field, which used to denote whether an instance
224 should be running or not, has been removed. Instead, ``admin_state`` is
225 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
227 The rational behind this is that an instance being “down” can have
230 - it could be down during a reboot
231 - it could be temporarily be down for a reinstall
232 - or it could be down because it is deprecated and kept just for its
235 The previous Boolean state was making it difficult to do capacity
236 calculations: should Ganeti reserve memory for a down instance? Now, the
237 tri-state field makes it clear:
239 - in ``up`` and ``down`` state, all resources are reserved for the
240 instance, and it can be at any time brought up if it is down
241 - in ``offline`` state, only disk space is reserved for it, but not
244 The field can have an extra use: since the transition between ``up`` and
245 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
246 transition between ``offline`` and ``down`` is done via ``gnt-instance
247 modify``, it is possible to given different rights to users. For
248 example, owners of an instance could be allowed to start/stop it, but
249 not transition it out of the offline state.
251 Instance policies and specs
252 +++++++++++++++++++++++++++
254 In previous Ganeti versions, an instance creation request was not
255 limited on the minimum size and on the maximum size just by the cluster
256 resources. As such, any policy could be implemented only in third-party
257 clients (RAPI clients, or shell wrappers over ``gnt-*``
258 tools). Furthermore, calculating cluster capacity via ``hspace`` again
259 required external input with regards to instance sizes.
261 In order to improve these workflows and to allow for example better
262 per-node group differentiation, we introduced instance specs, which
265 - minimum instance disk size, disk count, memory size, cpu count
266 - maximum values for the above metrics
267 - and “standard” values (used in ``hspace`` to calculate the standard
270 The minimum/maximum values can be also customised at node-group level,
271 for example allowing more powerful hardware to support bigger instance
274 Beside the instance specs, there are a few other settings belonging to
275 the instance policy framework. It is possible now to customise, per
276 cluster and node-group:
278 - the list of allowed disk templates
279 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
280 - the maximum ratio of instance to spindles (see below for more
281 information) for local storage
283 All these together should allow all tools that talk to Ganeti to know
284 what are the ranges of allowed values for instances and the
285 over-subscription that is allowed.
287 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
288 instance configuration, and the physical CPU configuration from the
289 node. For the spindle ratios however, we didn't track before these
290 values, so new parameters have been added:
292 - a new node parameter ``spindle_count``, defaults to 1, customisable at
293 node group or node level
294 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
296 Note that spindles in this context doesn't need to mean actual
297 mechanical hard-drives; it's just a relative number for both the node
298 I/O capacity and instance I/O consumption.
300 Instance migration behaviour
301 ++++++++++++++++++++++++++++
303 While live-migration is in general desirable over failover, it is
304 possible that for some workloads it is actually worse, due to the
305 variable time of the “suspend” phase during live migration.
307 To allow the tools to work consistently over such instances (without
308 having to hard-code instance names), a new backend parameter
309 ``always_failover`` has been added to control the migration/failover
310 behaviour. When set to True, all migration requests for an instance will
311 instead fall-back to failover.
313 Instance memory ballooning
314 ++++++++++++++++++++++++++
316 Initial support for memory ballooning has been added. The memory for an
317 instance is no longer fixed (backend parameter ``memory``), but instead
318 can vary between minimum and maximum values (backend parameters
319 ``minmem`` and ``maxmem``). Currently we only change an instance's
322 - live migrating or failing over and instance and the target node
323 doesn't have enough memory
324 - user requests changing the memory via ``gnt-instance modify
330 In order to control the use of specific CPUs by instance, support for
331 controlling CPU pinning has been added for the Xen, HVM and LXC
332 hypervisors. This is controlled by a new hypervisor parameter
333 ``cpu_mask``; details about possible values for this are in the
334 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
335 VCPU-to-CPU mapping) form will work well only when all nodes in your
336 cluster have the same amount of CPUs.
341 Another area in which Ganeti was not customisable were the parameters
342 used for storage configuration, e.g. how many stripes to use for LVM,
343 DRBD resync configuration, etc.
345 To improve this area, we've added disks parameters, which are
346 customisable at cluster and node group level, and which allow to
347 specify various parameters for disks (DRBD has the most parameters
348 currently), for example:
350 - DRBD resync algorithm and parameters (e.g. speed)
351 - the default VG for meta-data volumes for DRBD
352 - number of stripes for LVM (plain disk template)
355 These parameters can be modified via ``gnt-cluster modify -D …`` and
356 ``gnt-group modify -D …``, and are used at either instance creation (in
357 case of LVM stripes, for example) or at disk “activation” time
360 Rados block device support
361 ++++++++++++++++++++++++++
363 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
364 denoted by the ``rbd`` disk template type. This is considered
365 experimental, feedback is welcome. For details on configuring it, see
366 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
371 The existing master IP functionality works well only in simple setups (a
372 single network shared by all nodes); however, if nodes belong to
373 different networks, then the ``/32`` setup and lack of routing
374 information is not enough.
376 To allow the master IP to function well in more complex cases, the
377 system was reworked as follows:
379 - a master IP netmask setting has been added
380 - the master IP activation/turn-down code was moved from the node daemon
382 - whether to run the Ganeti-supplied master IP script or a user-supplied
383 on is a ``gnt-cluster init`` setting
385 Details about the location of the standard and custom setup scripts are
386 in the man page :manpage:`gnt-cluster(8)`; for information about the
387 setup script protocol, look at the Ganeti-supplied script.
392 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
395 It is now possible to use TLS-protected connections, and when renewing
396 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
397 it is now possible to specify spice or spice CA certificates. Also, it
398 is possible to configure a password for SPICE sessions via the
399 hypervisor parameter ``spice_password_file``.
401 There are also new parameters to control the compression and streaming
402 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
403 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
404 for the spice parameters.
406 Lastly, it is now possible to see the SPICE connection information via
407 ``gnt-instance console``.
412 A new tool (``tools/ovfconverter``) has been added that supports
413 conversion between Ganeti and the `Open Virtualization Format
414 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
417 This relies on the ``qemu-img`` tool to convert the disk formats, so the
418 actual compatibility with other virtualization solutions depends on it.
423 The configuration query daemon (``ganeti-confd``) is now optional, and
424 has been rewritten in Haskell; whether to use the daemon at all, use the
425 Python (default) or the Haskell version is selectable at configure time
426 via the ``--enable-confd`` parameter, which can take one of the
427 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
428 daemon will result in a smaller footprint; for larger systems, we
429 welcome feedback on the Haskell version which might become the default
432 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
433 daemon running. The Python version doesn't implement the new call.
436 User interface changes
437 ~~~~~~~~~~~~~~~~~~~~~~
439 We have replaced the ``--disks`` option of ``gnt-instance
440 replace-disks`` with a more flexible ``--disk`` option, which allows
441 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
442 disk size and mode can be changed upon recreation (via ``gnt-instance
443 recreate-disks``, which accepts the same ``--disk`` option).
445 As many people are used to a ``show`` command, we have added that as an
446 alias to ``info`` on all ``gnt-*`` commands.
448 The ``gnt-instance grow-disk`` command has a new mode in which it can
449 accept the target size of the disk, instead of the delta; this can be
450 more safe since two runs in absolute mode will be idempotent, and
451 sometimes it's also easier to specify the desired size directly.
453 Also the handling of instances with regard to offline secondaries has
454 been improved. Instance operations should not fail because one of it's
455 secondary nodes is offline, even though it's safe to proceed.
457 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
458 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
459 minors to instance name.
464 RAPI coverage has improved, with (for example) new resources for
465 recreate-disks, node power-cycle, etc.
470 There is partial support for ``xl`` in the Xen hypervisor; feedback is
473 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
474 whether to still support Python 2.4 or move to Python 2.6 as minimum
477 Support for Fedora has been slightly improved; the provided example
478 init.d script should work better on it and the INSTALL file should
479 document the needed dependencies.
484 The deprecated ``QueryLocks`` LUXI request has been removed. Use
485 ``Query(what=QR_LOCK, ...)`` instead.
487 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
488 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
489 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
490 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
491 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
493 RAPI client: ``CertificateError`` now derives from
494 ``GanetiApiError``. This should make it more easy to handle Ganeti
497 Deprecation warnings due to PyCrypto/paramiko import in
498 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
499 make sure to run an up-to-date paramiko version, if you use this tool.
501 The QA scripts now depend on Python 2.5 or above (the main code base
502 still works with Python 2.4).
504 The configuration file (``config.data``) is now written without
505 indentation for performance reasons; if you want to edit it, it can be
506 re-formatted via ``tools/fmtjson``.
508 A number of bugs has been fixed in the cluster merge tool.
510 ``x509`` certification verification (used in import-export) has been
511 changed to allow the same clock skew as permitted by the cluster
512 verification. This will remove some rare but hard to diagnose errors in
519 *(Released Thu, 19 Jul 2012)*
521 Very few changes from rc4 to the final release, only bugfixes:
523 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
524 instance, fix CDROM booting for KVM instances)
525 - fixed node group modification of node parameters
526 - fixed issue in LUClusterVerifyGroup with multi-group clusters
527 - fixed generation of bash completion to ensure a stable ordering
534 *(Released Fri, 13 Jul 2012)*
536 Third release candidate for 2.6. The following changes were done from
539 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
540 - Fixed an inconsistency in the LUXI protocol with the provided
541 arguments (NOT backwards compatible)
542 - Fixed a bug with node groups ipolicy where ``min`` was greater than
543 the cluster ``std`` value
544 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
545 easier instance debugging on nodes (requires ``hconfd`` to work)
551 *(Released Tue, 03 Jul 2012)*
553 Second release candidate for 2.6. The following changes were done from
556 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
558 - Fixed a RAPI regression on missing beparams/memory
559 - Fixed redistribution of files on offline nodes
560 - Added possibility to run activate-disks even though secondaries are
561 offline. With this change it relaxes also the strictness on some other
562 commands which use activate disks internally:
563 * ``gnt-instance start|reboot|rename|backup|export``
564 - Made it possible to remove safely an instance if its secondaries are
566 - Made it possible to reinstall even though secondaries are offline
572 *(Released Mon, 25 Jun 2012)*
574 First release candidate for 2.6. The following changes were done from
577 - Fixed bugs with disk parameters and ``rbd`` templates as well as
579 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
580 behaviour. It supports now the modify operation
581 - ``hcheck`` implemented to analyze cluster health and possibility of
582 improving health by rebalance
583 - ``hbal`` has been improved in dealing with split instances
589 *(Released Mon, 11 Jun 2012)*
591 Second beta release of 2.6. The following changes were done from beta2
594 - Fixed ``daemon-util`` with non-root user models
595 - Fixed creation of plain instances with ``--no-wait-for-sync``
596 - Fix wrong iv_names when running ``cfgupgrade``
597 - Export more information in RAPI group queries
598 - Fixed bug when changing instance network interfaces
599 - Extended burnin to do NIC changes
600 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
601 - Changed default for DRBD barriers
602 - Fixed DRBD error reporting for syncer rate
603 - Verify the options on disk parameters
605 And of course various fixes to documentation and improved unittests and
612 *(Released Wed, 23 May 2012)*
614 First beta release of 2.6. The following changes were done from beta1 to
617 - integrated patch for distributions without ``start-stop-daemon``
618 - adapted example init.d script to work on Fedora
619 - fixed log handling in Haskell daemons
620 - adapted checks in the watcher for pycurl linked against libnss
621 - add partial support for ``xl`` instead of ``xm`` for Xen
622 - fixed a type issue in cluster verification
623 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
626 Plus integrated fixes from the 2.5 branch:
628 - fixed ``kvm-ifup`` to use ``/bin/bash``
629 - fixed parallel build failures
630 - KVM live migration when using a custom keymap
636 *(Released Tue, 24 Jul 2012)*
638 A small bugfix release, with no new features:
640 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
641 different default shell (e.g. Debian, Ubuntu)
642 - fixed KVM startup and live migration with a custom keymap (fixes Issue
643 243 and Debian bug #650664)
644 - fixed compatibility with KVM versions that don't support multiple boot
645 devices (fixes Issue 230 and Debian bug #624256)
647 Additionally, a few fixes were done to the build system (fixed parallel
648 build failures) and to the unittests (fixed race condition in test for
649 FileID functions, and the default enable/disable mode for QA test is now
656 *(Released Fri, 11 May 2012)*
658 A small bugfix release.
660 The main issues solved are on the topic of compatibility with newer LVM
663 - fixed parsing of ``lv_attr`` field
664 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
665 the ``--force`` flag is needed
667 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
668 to accept kernel 3.x too (was hardcoded to 2.6.*).
670 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
671 of hook results and that also made display of other errors suboptimal
672 was fixed; the code behaves now like 2.4 and earlier.
674 Another change in 2.5, the cleanup of the OS scripts environment, is too
675 aggressive: it removed even the ``PATH`` variable, which requires the OS
676 scripts to *always* need to export it. Since this is a bit too strict,
677 we now export a minimal PATH, the same that we export for hooks.
679 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
680 integrated into this release.
682 Finally, a few other miscellaneous changes were done (no new features,
683 just small improvements):
685 - Fix ``gnt-group --help`` display
686 - Fix hardcoded Xen kernel path
687 - Fix grow-disk handling of invalid units
688 - Update synopsis for ``gnt-cluster repair-disk-sizes``
689 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
695 *(Released Thu, 12 Apr 2012)*
697 Incompatible/important changes and bugfixes
698 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
700 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
701 resource's ``ip_check`` parameter changed from ``True`` to ``False``
702 to match the underlying LUXI interface.
703 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
704 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
705 not maintain backwards-compatibility as the underlying operation
706 changed in an incompatible way. The RAPI client can talk to old
707 servers, but it needs to be told so as the return value changed.
708 - When creating file-based instances via RAPI, the ``file_driver``
709 parameter no longer defaults to ``loop`` and must be specified.
710 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
712 - Support for the undocumented and deprecated RAPI instance creation
713 request format version 0 has been dropped. Use version 1, supported
714 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
715 - Pyparsing 1.4.6 or above is required, see :doc:`installation
716 documentation <install>`.
717 - The "cluster-verify" hooks are now executed per group by the
718 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
719 if you just run ``gnt-cluster verify``, which generates one opcode per
721 - The environment as passed to the OS scripts is cleared, and thus no
722 environment variables defined in the node daemon's environment will be
723 inherited by the scripts.
724 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
726 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
727 support operations involving multiple node groups.
728 - Offline nodes are ignored when failing over an instance.
729 - Support for KVM version 1.0, which changed the version reporting format
731 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
733 - ``Makefile`` is now compatible with Automake 1.11.2
734 - Includes all bugfixes made in the 2.4 series
739 - The ganeti-htools project has been merged into the ganeti-core source
740 tree and will be built as part of Ganeti (see :doc:`install-quick`).
741 - Implemented support for :doc:`shared storage <design-shared-storage>`.
742 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
743 GPT-style partition tables (requires `parted
744 <http://www.gnu.org/s/parted/>`_).
745 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
746 - Allowed adding tags on instance creation.
747 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
748 - Allow instances to be started in a paused state, enabling the user to
749 see the complete console output on boot using the console.
750 - Added new hypervisor flag to control default reboot behaviour
751 (``reboot_behavior``).
752 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
753 - Improved out-of-band management support:
755 - Added ``gnt-node health`` command reporting the health status of
757 - Added ``gnt-node power`` command to manage power status of nodes.
758 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
760 - Instance migration can fall back to failover if instance is not
762 - Filters can be used when listing nodes, instances, groups and locks;
763 see :manpage:`ganeti(7)` manpage.
764 - Added post-execution status as variables to :doc:`hooks <hooks>`
766 - Instance tags are exported/imported together with the instance.
767 - When given an explicit job ID, ``gnt-job info`` will work for archived
769 - Jobs can define dependencies on other jobs (not yet supported via
770 RAPI or command line, but used by internal commands and usable via
773 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
776 - Instance failover is now available as a RAPI resource
777 (``/2/instances/[instance_name]/failover``).
778 - ``gnt-instance info`` defaults to static information if primary node
780 - Opcodes have a new ``comment`` attribute.
781 - Added basic SPICE support to KVM hypervisor.
782 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
784 Node group improvements
785 ~~~~~~~~~~~~~~~~~~~~~~~
787 - ``gnt-cluster verify`` has been modified to check groups separately,
788 thereby improving performance.
789 - Node group support has been added to ``gnt-cluster verify-disks``,
790 which now operates per node group.
791 - Watcher has been changed to work better with node groups.
793 - One process and state file per node group.
794 - Slow watcher in one group doesn't block other group's watcher.
796 - Added new command, ``gnt-group evacuate``, to move all instances in a
797 node group to other groups.
798 - Added ``gnt-instance change-group`` to move an instance to another
800 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
801 per-group operations.
802 - Node groups can be tagged.
803 - Some operations switch from an exclusive to a shared lock as soon as
805 - Instance's primary and secondary nodes' groups are now available as
806 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
807 and ``snodes.group.uuid``).
812 - Numerous updates to documentation and manpages.
814 - :doc:`RAPI <rapi>` documentation now has detailed parameter
816 - Some opcode/job results are now also documented, see :doc:`RAPI
819 - A lockset's internal lock is now also visible in lock monitor.
820 - Log messages from job queue workers now contain information about the
821 opcode they're processing.
822 - ``gnt-instance console`` no longer requires the instance lock.
823 - A short delay when waiting for job changes reduces the number of LUXI
824 requests significantly.
825 - DRBD metadata volumes are overwritten with zeros during disk creation.
826 - Out-of-band commands no longer acquire the cluster lock in exclusive
828 - ``devel/upload`` now uses correct permissions for directories.
834 *(Released Fri, 23 Mar 2012)*
836 This was the sixth release candidate of the 2.5 series.
842 *(Released Mon, 9 Jan 2012)*
844 This was the fifth release candidate of the 2.5 series.
850 *(Released Thu, 27 Oct 2011)*
852 This was the fourth release candidate of the 2.5 series.
858 *(Released Wed, 26 Oct 2011)*
860 This was the third release candidate of the 2.5 series.
866 *(Released Tue, 18 Oct 2011)*
868 This was the second release candidate of the 2.5 series.
874 *(Released Tue, 4 Oct 2011)*
876 This was the first release candidate of the 2.5 series.
882 *(Released Wed, 31 Aug 2011)*
884 This was the third beta release of the 2.5 series.
890 *(Released Mon, 22 Aug 2011)*
892 This was the second beta release of the 2.5 series.
898 *(Released Fri, 12 Aug 2011)*
900 This was the first beta release of the 2.5 series.
906 *(Released Thu, 27 Oct 2011)*
908 - Fixed bug when parsing command line parameter values ending in
910 - Fixed assertion error after unclean master shutdown
911 - Disable HTTP client pool for RPC, significantly reducing memory usage
913 - Fixed queue archive creation with wrong permissions
919 *(Released Tue, 23 Aug 2011)*
923 - Fixed documentation for importing with ``--src-dir`` option
924 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
925 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
931 *(Released Fri, 5 Aug 2011)*
933 Many bug-fixes and a few small features:
935 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
936 issues when you tried to add an instance with two MAC addresses in one
938 - KVM: fixed per-instance stored UID value
939 - KVM: configure bridged NICs at migration start
940 - KVM: Fix a bug where instance will not start with never KVM versions
942 - Added OS search path to ``gnt-cluster info``
943 - Fixed an issue with ``file_storage_dir`` where you were forced to
944 provide an absolute path, but the documentation states it is a
945 relative path, the documentation was right
946 - Added a new parameter to instance stop/start called ``--no-remember``
947 that will make the state change to not be remembered
948 - Implemented ``no_remember`` at RAPI level
949 - Improved the documentation
950 - Node evacuation: don't call IAllocator if node is already empty
951 - Fixed bug in DRBD8 replace disks on current nodes
952 - Fixed bug in recreate-disks for DRBD instances
953 - Moved assertion checking locks in ``gnt-instance replace-disks``
954 causing it to abort with not owning the right locks for some situation
955 - Job queue: Fixed potential race condition when cancelling queued jobs
956 - Fixed off-by-one bug in job serial generation
957 - ``gnt-node volumes``: Fix instance names
958 - Fixed aliases in bash completion
959 - Fixed a bug in reopening log files after being sent a SIGHUP
960 - Added a flag to burnin to allow specifying VCPU count
961 - Bugfixes to non-root Ganeti configuration
967 *(Released Thu, 12 May 2011)*
969 Many bug-fixes and a few new small features:
971 - Fixed a bug related to log opening failures
972 - Fixed a bug in instance listing with orphan instances
973 - Fixed a bug which prevented resetting the cluster-level node parameter
974 ``oob_program`` to the default
975 - Many fixes related to the ``cluster-merge`` tool
976 - Fixed a race condition in the lock monitor, which caused failures
977 during (at least) creation of many instances in parallel
978 - Improved output for gnt-job info
979 - Removed the quiet flag on some ssh calls which prevented debugging
981 - Improved the N+1 failure messages in cluster verify by actually
982 showing the memory values (needed and available)
983 - Increased lock attempt timeouts so that when executing long operations
984 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
985 too early and thus prevent the use of the 'fair' mechanism
986 - Changed instance query data (``gnt-instance info``) to not acquire
987 locks unless needed, thus allowing its use on locked instance if only
988 static information is asked for
989 - Improved behaviour with filesystems that do not support rename on an
991 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
992 kept locks on all nodes during the wipe, which is unneeded
993 - Fixed ``gnt-watcher`` handling of errors during hooks execution
994 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
995 10GiB) which caused the wipe to fail right at the end in some cases
996 - Fixed master IP activation when doing master failover with no-voting
997 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
998 the master node itself
999 - Fixed potential data-loss in under disk full conditions, where Ganeti
1000 wouldn't check correctly the return code and would consider
1001 partially-written files 'correct'
1002 - Fixed bug related to multiple VGs and DRBD disk replacing
1003 - Added new disk parameter ``metavg`` that allows placement of the meta
1004 device for DRBD in a different volume group
1005 - Fixed error handling in the node daemon when the system libc doesn't
1006 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1007 - Fixed lock release during replace-disks, which kept cluster-wide locks
1008 when doing disk replaces with an iallocator script
1009 - Added check for missing bridges in cluster verify
1010 - Handle EPIPE errors while writing to the terminal better, so that
1011 piping the output to e.g. ``less`` doesn't cause a backtrace
1012 - Fixed rare case where a ^C during Luxi calls could have been
1013 interpreted as server errors, instead of simply terminating
1014 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1016 - Added a few more parameters to the KVM hypervisor, allowing a second
1017 CDROM, custom disk type for CDROMs and a floppy image
1018 - Removed redundant message in instance rename when the name is given
1020 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1021 disks on new nodes, allowing recreation when the original instance
1022 nodes are completely gone
1023 - Added option when converting disk templates to DRBD to skip waiting
1024 for the resync, in order to make the instance available sooner
1025 - Added two new variables to the OS scripts environment (containing the
1027 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1028 to allow use of ``pvgrub`` as bootloader
1029 - Changed the instance memory modifications to only check out-of-memory
1030 conditions on memory increases, and turned the secondary node warnings
1031 into errors (they can still be overridden via ``--force``)
1032 - Fixed the handling of a corner case when the Python installation gets
1033 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1034 to execute a command that doesn't exist
1035 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1036 node of the instance returned failures during instance shutdown; this
1037 adds the option ``--ignore-consistency`` to gnt-instance move
1039 And as usual, various improvements to the error messages, documentation
1046 *(Released Wed, 09 Mar 2011)*
1048 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1049 the RAPI users file if run twice (even with ``--dry-run``).
1051 The release fixes that bug (nothing else changed).
1057 *(Released Mon, 07 Mar 2011)*
1059 Final 2.4.0 release. Just a few small fixes:
1061 - Fixed RAPI node evacuate
1062 - Fixed the kvm-ifup script
1063 - Fixed internal error handling for special job cases
1064 - Updated man page to specify the escaping feature for options
1070 *(Released Mon, 28 Feb 2011)*
1072 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1073 that this feature wiped the disks of the wrong instance, leading to loss
1078 - Fixed title of query field containing instance name
1079 - Expanded the glossary in the documentation
1080 - Fixed one unittest (internal issue)
1086 *(Released Mon, 21 Feb 2011)*
1088 A number of bug fixes plus just a couple functionality changes.
1090 On the user-visible side, the ``gnt-* list`` command output has changed
1091 with respect to "special" field states. The current rc1 style of display
1092 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1093 the default output mode special fields states are displayed as follows:
1095 - Offline resource: ``*``
1096 - Unavailable/not applicable: ``-``
1097 - Data missing (RPC failure): ``?``
1098 - Unknown field: ``??``
1100 Another user-visible change is the addition of ``--force-join`` to
1105 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1106 - Fixed regression in RAPI/instance reinstall where all parameters were
1107 required (instead of optional)
1108 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1109 - Fixed iallocator usage (offline nodes were not considered offline)
1110 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1111 - Fixed hypervisor and OS parameter validation with respect to
1112 non-vm_capable nodes
1113 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1115 - Fixed ``tools/listrunner`` with respect to agent-based usage
1121 *(Released Fri, 4 Feb 2011)*
1123 Many changes and fixes since the beta1 release. While there were some
1124 internal changes, the code has been mostly stabilised for the RC
1127 Note: the dumb allocator was removed in this release, as it was not kept
1128 up-to-date with the IAllocator protocol changes. It is recommended to
1129 use the ``hail`` command from the ganeti-htools package.
1131 Note: the 2.4 and up versions of Ganeti are not compatible with the
1132 0.2.x branch of ganeti-htools. You need to upgrade to
1133 ganeti-htools-0.3.0 (or later).
1135 Regressions fixed from 2.3
1136 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1138 - Fixed the ``gnt-cluster verify-disks`` command
1139 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1141 - Fixed disk adoption breakage
1142 - Fixed wrong headers in instance listing for field aliases
1147 - Fixed corner case in KVM handling of NICs
1148 - Fixed many cases of wrong handling of non-vm_capable nodes
1149 - Fixed a bug where a missing instance symlink was not possible to
1150 recreate with any ``gnt-*`` command (now ``gnt-instance
1151 activate-disks`` does it)
1152 - Fixed the volume group name as reported by ``gnt-cluster
1154 - Increased timeouts for the import-export code, hopefully leading to
1155 fewer aborts due network or instance timeouts
1156 - Fixed bug in ``gnt-node list-storage``
1157 - Fixed bug where not all daemons were started on cluster
1158 initialisation, but only at the first watcher run
1159 - Fixed many bugs in the OOB implementation
1160 - Fixed watcher behaviour in presence of instances with offline
1162 - Fixed instance list output for instances running on the wrong node
1163 - a few fixes to the cluster-merge tool, but it still cannot merge
1164 multi-node groups (currently it is not recommended to use this tool)
1170 - Improved network configuration for the KVM hypervisor
1171 - Added e1000 as a supported NIC for Xen-HVM
1172 - Improved the lvmstrap tool to also be able to use partitions, as
1173 opposed to full disks
1174 - Improved speed of disk wiping (the cluster parameter
1175 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1176 of instance creations
1177 - Added documentation for the OS parameters
1178 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1179 hypervisor is not responding
1180 - Added display of blacklisted and hidden OS information in
1181 ``gnt-cluster info``
1182 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1183 NIC and node parameters, which might create problems with currently
1184 invalid (but undetected) configuration files, but prevents validation
1185 failures when unrelated parameters are modified
1186 - Changed cluster initialisation to wait for the master daemon to become
1188 - Expanded the RAPI interface:
1190 - Added config redistribution resource
1191 - Added activation/deactivation of instance disks
1192 - Added export of console information
1194 - Implemented log file reopening on SIGHUP, which allows using
1195 logrotate(8) for the Ganeti log files
1196 - Added a basic OOB helper script as an example
1202 *(Released Fri, 14 Jan 2011)*
1207 - Fixed timezone issues when formatting timestamps
1208 - Added support for node groups, available via ``gnt-group`` and other
1210 - Added out-of-band framework and management, see :doc:`design
1211 document <design-oob>`
1212 - Removed support for roman numbers from ``gnt-node list`` and
1213 ``gnt-instance list``.
1214 - Allowed modification of master network interface via ``gnt-cluster
1215 modify --master-netdev``
1216 - Accept offline secondaries while shutting down instance disks
1217 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1218 - Added support for multiple LVM volume groups
1219 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1221 - Added commands to list available fields:
1223 - ``gnt-node list-fields``
1224 - ``gnt-group list-fields``
1225 - ``gnt-instance list-fields``
1227 - Updated documentation and man pages
1232 - Moved ``rapi_users`` file into separate directory, now named
1233 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1235 - Added new tool for running commands on many machines,
1236 ``tools/ganeti-listrunner``
1237 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1238 improving the ``gnt-instance console`` output
1239 - Allowed customisation of disk index separator at ``configure`` time
1240 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1241 - Added support for non-partitioned md disks in ``lvmstrap``
1242 - Added script to gracefully power off KVM instances
1243 - Split ``utils`` module into smaller parts
1244 - Changed query operations to return more detailed information, e.g.
1245 whether an information is unavailable due to an offline node. To use
1246 this new functionality, the LUXI call ``Query`` must be used. Field
1247 information is now stored by the master daemon and can be retrieved
1248 using ``QueryFields``. Instances, nodes and groups can also be queried
1249 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1250 exposed via RAPI). The following commands make use of this
1251 infrastructure change:
1253 - ``gnt-group list``
1254 - ``gnt-group list-fields``
1256 - ``gnt-node list-fields``
1257 - ``gnt-instance list``
1258 - ``gnt-instance list-fields``
1259 - ``gnt-debug locks``
1264 - New RAPI resources (see :doc:`rapi`):
1268 - ``/2/groups/[group_name]``
1269 - ``/2/groups/[group_name]/assign-nodes``
1270 - ``/2/groups/[group_name]/modify``
1271 - ``/2/groups/[group_name]/rename``
1272 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1276 - Implemented ``no_install`` for instance creation
1277 - Implemented OS parameters for instance reinstallation, allowing
1278 use of special settings on reinstallation (e.g. for preserving data)
1283 - Added IPv6 support in import/export
1284 - Pause DRBD synchronization while wiping disks on instance creation
1285 - Updated unittests and QA scripts
1286 - Improved network parameters passed to KVM
1287 - Converted man pages from docbook to reStructuredText
1293 *(Released Mon, 20 Dec 2010)*
1295 Released version 2.3.1~rc1 without any changes.
1301 *(Released Wed, 1 Dec 2010)*
1303 - impexpd: Disable OpenSSL compression in socat if possible (backport
1304 from master, commit e90739d625b, see :doc:`installation guide
1305 <install-quick>` for details)
1306 - Changed unittest coverage report to exclude test scripts
1307 - Added script to check version format
1313 *(Released Wed, 1 Dec 2010)*
1315 Released version 2.3.0~rc1 without any changes.
1321 *(Released Fri, 19 Nov 2010)*
1323 A number of bugfixes and documentation updates:
1325 - Update ganeti-os-interface documentation
1326 - Fixed a bug related to duplicate MACs or similar items which should be
1328 - Fix breakage in OS state modify
1329 - Reinstall instance: disallow offline secondaries (fixes bug related to
1330 OS changing but reinstall failing)
1331 - plus all the other fixes between 2.2.1 and 2.2.2
1337 *(Released Tue, 2 Nov 2010)*
1339 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1340 - Fixed master failover race with watcher
1341 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1343 - Accept previously stopped instance for export with instance removal
1344 - Simplify and extend the environment variables for instance OS scripts
1345 - Added new node flags, ``master_capable`` and ``vm_capable``
1346 - Added optional instance disk wiping prior during allocation. This is a
1347 cluster-wide option and can be set/modified using
1348 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1349 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1350 :doc:`install-quick`
1351 - Added a new watcher option (``--ignore-pause``)
1352 - Added option to ignore offline node on instance start/stop
1353 (``--ignore-offline``)
1354 - Allow overriding OS parameters with ``gnt-instance reinstall``
1355 - Added ability to change node's secondary IP address using ``gnt-node
1357 - Implemented privilege separation for all daemons except
1358 ``ganeti-noded``, see ``configure`` options
1359 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1361 - Implemented job priorities (see ``ganeti(7)`` manpage)
1362 - Ignore failures while shutting down instances during failover from
1364 - Exit daemon's bootstrap process only once daemon is ready
1365 - Export more information via ``LUInstanceQuery``/remote API
1366 - Improved documentation, QA and unittests
1367 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1368 restart if the file was created or changed
1369 - Added LUXI protocol version sent with each request and response,
1370 allowing detection of server/client mismatches
1371 - Moved the Python scripts among gnt-* and ganeti-* into modules
1372 - Moved all code related to setting up SSH to an external script,
1374 - Infrastructure changes for node group support in future versions
1380 *(Released Fri, 19 Nov 2010)*
1382 A few small bugs fixed, and some improvements to the build system:
1384 - Fix documentation regarding conversion to drbd
1385 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1387 - Fix error handling in node modify with multiple changes
1388 - Allow remote imports without checked names
1394 *(Released Tue, 19 Oct 2010)*
1396 - Disable SSL session ID cache in RPC client
1402 *(Released Thu, 14 Oct 2010)*
1404 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1405 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1407 - Fix problems with interaction between Curl and Python's HTTP server,
1408 resulting in increased speed in many RPC calls
1409 - Improve our release script to prevent breakage with older aclocal and
1416 *(Released Thu, 7 Oct 2010)*
1418 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1419 value retrieved from master
1420 - Added support for blacklisted or hidden OS definitions
1421 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1422 - Added support for -mem-path in KVM hypervisor abstraction layer
1423 - Allow overriding instance parameters in tool for inter-cluster
1424 instance moves (``tools/move-instance``)
1425 - Improved opcode summaries (e.g. in ``gnt-job list``)
1426 - Improve consistency of OS listing by sorting it
1427 - Documentation updates
1433 *(Released Fri, 8 Oct 2010)*
1435 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1441 *(Released Mon, 4 Oct 2010)*
1443 - Fixed regression in ``gnt-instance rename``
1449 *(Released Wed, 22 Sep 2010)*
1451 - Fixed OS_VARIANT variable for OS scripts
1452 - Fixed cluster tag operations via RAPI
1453 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1454 - Disabled RAPI CA checks in watcher
1460 *(Released Mon, 23 Aug 2010)*
1462 - Support DRBD versions of the format "a.b.c.d"
1464 - Re-introduce support for usage from multiple threads in RAPI client
1465 - Instance renames and modify via RAPI
1466 - Work around race condition between processing and archival in job
1468 - Mark opcodes following failed one as failed, too
1469 - Job field ``lock_status`` was removed due to difficulties making it
1470 work with the changed job queue in Ganeti 2.2; a better way to monitor
1471 locks is expected for a later 2.2.x release
1472 - Fixed dry-run behaviour with many commands
1473 - Support ``ssh-agent`` again when adding nodes
1474 - Many additional bugfixes
1480 *(Released Fri, 30 Jul 2010)*
1482 Important change: the internal RPC mechanism between Ganeti nodes has
1483 changed from using a home-grown http library (based on the Python base
1484 libraries) to use the PycURL library. This requires that PycURL is
1485 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1486 against GnuTLS by default. cURL's support for GnuTLS had known issues
1487 before cURL 7.21.0 and we recommend using the latest cURL release or
1488 linking against OpenSSL. Most other distributions already link PycURL
1489 and cURL against OpenSSL. The command::
1491 python -c 'import pycurl; print pycurl.version'
1493 can be used to determine the libraries PycURL and cURL are linked
1496 Other significant changes:
1498 - Rewrote much of the internals of the job queue, in order to achieve
1499 better parallelism; this decouples job query operations from the job
1500 processing, and it should allow much nicer behaviour of the master
1501 daemon under load, and it also has uncovered some long-standing bugs
1502 related to the job serialisation (now fixed)
1503 - Added a default iallocator setting to the cluster parameters,
1504 eliminating the need to always pass nodes or an iallocator for
1505 operations that require selection of new node(s)
1506 - Added experimental support for the LXC virtualization method
1507 - Added support for OS parameters, which allows the installation of
1508 instances to pass parameter to OS scripts in order to customise the
1510 - Added a hypervisor parameter controlling the migration type (live or
1511 non-live), since hypervisors have various levels of reliability; this
1512 has renamed the 'live' parameter to 'mode'
1513 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1514 logical volumes, meaning that cluster verify will ignore them and not
1515 flag their presence as errors
1516 - The watcher will now reset the error count for failed instances after
1517 8 hours, thus allowing self-healing if the problem that caused the
1518 instances to be down/fail to start has cleared in the meantime
1519 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1520 check for, and warn, if the drbd module parameter ``usermode_helper``
1521 is not consistent with the cluster-wide setting; this is needed to
1522 make diagnose easier of failed drbd creations
1523 - Started adding base IPv6 support, but this is not yet
1524 enabled/available for use
1525 - Rename operations (cluster, instance) will now return the new name,
1526 which is especially useful if a short name was passed in
1527 - Added support for instance migration in RAPI
1528 - Added a tool to pre-configure nodes for the SSH setup, before joining
1529 them to the cluster; this will allow in the future a simplified model
1530 for node joining (but not yet fully enabled in 2.2); this needs the
1531 paramiko python library
1532 - Fixed handling of name-resolving errors
1533 - Fixed consistency of job results on the error path
1534 - Fixed master-failover race condition when executed multiple times in
1536 - Fixed many bugs related to the job queue (mostly introduced during the
1537 2.2 development cycle, so not all are impacting 2.1)
1538 - Fixed instance migration with missing disk symlinks
1539 - Fixed handling of unknown jobs in ``gnt-job archive``
1540 - And many other small fixes/improvements
1544 - Enhanced both the unittest and the QA coverage
1545 - Switched the opcode validation to a generic model, and extended the
1546 validation to all opcode parameters
1547 - Changed more parts of the code that write shell scripts to use the
1549 - Switched the master daemon to use the asyncore library for the Luxi
1556 *(Released Thu, 17 Jun 2010)*
1558 - Added tool (``move-instance``) and infrastructure to move instances
1559 between separate clusters (see :doc:`separate documentation
1560 <move-instance>` and :doc:`design document <design-2.2>`)
1561 - Added per-request RPC timeout
1562 - RAPI now requires a Content-Type header for requests with a body (e.g.
1563 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1564 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1565 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1567 - Implemented initial support for running Ganeti daemons as separate
1568 users, see configure-time flags ``--with-user-prefix`` and
1569 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1571 - Instances can be removed after export (``gnt-backup export
1572 --remove-instance``)
1573 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1574 key (instead of 1024 bit)
1575 - Added new cluster configuration file for cluster domain secret
1576 - Import/export now use SSL instead of SSH
1577 - Added support for showing estimated time when exporting an instance,
1578 see the ``ganeti-os-interface(7)`` manpage and look for
1585 *(Released Tue, 16 Nov 2010)*
1587 Some more bugfixes. Unless critical bugs occur, this will be the last
1590 - Fix case of MAC special-values
1591 - Fix mac checker regex
1592 - backend: Fix typo causing "out of range" error
1593 - Add missing --units in gnt-instance list man page
1599 *(Released Tue, 24 Aug 2010)*
1602 - Don't ignore secondary node silently on non-mirrored disk templates
1604 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1605 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1606 - Properly document the usb_mouse parameter
1607 - Fix path in ganeti-rapi(8) (issue 116)
1608 - Adjust error message when the ganeti user's .ssh directory is
1610 - Add same-node-check when changing the disk template to drbd
1616 *(Released Fri, 16 Jul 2010)*
1619 - Add an option to only select some reboot types during qa/burnin.
1620 (on some hypervisors consequent reboots are not supported)
1621 - Fix infrequent race condition in master failover. Sometimes the old
1622 master ip address would be still detected as up for a short time
1623 after it was removed, causing failover to fail.
1624 - Decrease mlockall warnings when the ctypes module is missing. On
1625 Python 2.4 we support running even if no ctypes module is installed,
1626 but we were too verbose about this issue.
1627 - Fix building on old distributions, on which man doesn't have a
1629 - Fix RAPI not to ignore the MAC address on instance creation
1630 - Implement the old instance creation format in the RAPI client.
1636 *(Released Thu, 01 Jul 2010)*
1638 A small bugfix release:
1639 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1640 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1641 a non-existing option
1642 - Fix instance create: the --force-variant option was ignored
1643 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1644 - Fix modify node storage with non-FQDN arguments
1645 - Fix RAPI client to authenticate under Python 2.6 when used
1646 for more than 5 requests needing authentication
1647 - Fix gnt-instance modify -t (storage) giving a wrong error message
1648 when converting a non-shutdown drbd instance to plain
1654 *(Released Fri, 18 Jun 2010)*
1656 A small bugfix release:
1658 - Fix live migration of KVM instances started with older Ganeti
1659 versions which had fewer hypervisor parameters
1660 - Fix gnt-instance grow-disk on down instances
1661 - Fix an error-reporting bug during instance migration
1662 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1663 silently ignoring broken ones
1664 - Fix an RPC error reporting bug affecting, for example, RAPI client
1666 - Fix bug triggered by different API version os-es on different nodes
1667 - Fix a bug in instance startup with custom hvparams: OS level
1668 parameters would fail to be applied.
1669 - Fix the RAPI client under Python 2.6 (but more work is needed to
1670 make it work completely well with OpenSSL)
1671 - Fix handling of errors when resolving names from DNS
1677 *(Released Thu, 3 Jun 2010)*
1679 A medium sized development cycle. Some new features, and some
1680 fixes/small improvements/cleanups.
1682 Significant features
1683 ~~~~~~~~~~~~~~~~~~~~
1685 The node deamon now tries to mlock itself into memory, unless the
1686 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1687 its logs, and falls back to console logging. This allows emergency
1688 features such as ``gnt-node powercycle`` to work even in the event of a
1689 broken node disk (tested offlining the disk hosting the node's
1690 filesystem and dropping its memory caches; don't try this at home)
1692 KVM: add vhost-net acceleration support. It can be tested with a new
1693 enough version of the kernel and of qemu-kvm.
1695 KVM: Add instance chrooting feature. If you use privilege dropping for
1696 your VMs you can also now force them to chroot to an empty directory,
1697 before starting the emulated guest.
1699 KVM: Add maximum migration bandwith and maximum downtime tweaking
1700 support (requires a new-enough version of qemu-kvm).
1702 Cluster verify will now warn if the master node doesn't have the master
1703 ip configured on it.
1705 Add a new (incompatible) instance creation request format to RAPI which
1706 supports all parameters (previously only a subset was supported, and it
1707 wasn't possible to extend the old format to accomodate all the new
1708 features. The old format is still supported, and a client can check for
1709 this feature, before using it, by checking for its presence in the
1710 ``features`` RAPI resource.
1712 Now with ancient latin support. Try it passing the ``--roman`` option to
1713 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1714 (requires the python-roman module to be installed, in order to work).
1719 As usual many internal code refactorings, documentation updates, and
1722 - Lots of improvements and cleanups to the experimental Remote API
1723 (RAPI) client library.
1724 - A new unit test suite for the core daemon libraries.
1725 - A fix to creating missing directories makes sure the umask is not
1726 applied anymore. This enforces the same directory permissions
1728 - Better handling terminating daemons with ctrl+c (used when running
1729 them in debugging mode).
1730 - Fix a race condition in live migrating a KVM instance, when stat()
1731 on the old proc status file returned EINVAL, which is an unexpected
1733 - Fixed manpage checking with newer man and utf-8 charachters. But now
1734 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1740 *(Released Fri, 7 May 2010)*
1742 Fix a bug which prevented untagged KVM instances from starting.
1748 *(Released Fri, 7 May 2010)*
1750 Another release with a long development cycle, during which many
1751 different features were added.
1753 Significant features
1754 ~~~~~~~~~~~~~~~~~~~~
1756 The KVM hypervisor now can run the individual instances as non-root, to
1757 reduce the impact of a VM being hijacked due to bugs in the
1758 hypervisor. It is possible to run all instances as a single (non-root)
1759 user, to manually specify a user for each instance, or to dynamically
1760 allocate a user out of a cluster-wide pool to each instance, with the
1761 guarantee that no two instances will run under the same user ID on any
1764 An experimental RAPI client library, that can be used standalone
1765 (without the other Ganeti libraries), is provided in the source tree as
1766 ``lib/rapi/client.py``. Note this client might change its interface in
1767 the future, as we iterate on its capabilities.
1769 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1770 replace the cluster's certificates and crypto keys. This might help in
1771 case they have been compromised, or have simply expired.
1773 A new disk option for instance creation has been added that allows one
1774 to "adopt" currently existing logical volumes, with data
1775 preservation. This should allow easier migration to Ganeti from
1776 unmanaged (or managed via other software) instances.
1778 Another disk improvement is the possibility to convert between redundant
1779 (DRBD) and plain (LVM) disk configuration for an instance. This should
1780 allow better scalability (starting with one node and growing the
1781 cluster, or shrinking a two-node cluster to one node).
1783 A new feature that could help with automated node failovers has been
1784 implemented: if a node sees itself as offline (by querying the master
1785 candidates), it will try to shutdown (hard) all instances and any active
1786 DRBD devices. This reduces the risk of duplicate instances if an
1787 external script automatically failovers the instances on such nodes. To
1788 enable this, the cluster parameter ``maintain_node_health`` should be
1789 enabled; in the future this option (per the name) will enable other
1790 automatic maintenance features.
1792 Instance export/import now will reuse the original instance
1793 specifications for all parameters; that means exporting an instance,
1794 deleting it and the importing it back should give an almost identical
1795 instance. Note that the default import behaviour has changed from
1796 before, where it created only one NIC; now it recreates the original
1799 Cluster verify has added a few new checks: SSL certificates validity,
1800 /etc/hosts consistency across the cluster, etc.
1805 As usual, many internal changes were done, documentation fixes,
1808 - Fixed cluster initialization with disabled cluster storage (regression
1809 introduced in 2.1.1)
1810 - File-based storage supports growing the disks
1811 - Fixed behaviour of node role changes
1812 - Fixed cluster verify for some corner cases, plus a general rewrite of
1813 cluster verify to allow future extension with more checks
1814 - Fixed log spamming by watcher and node daemon (regression introduced
1816 - Fixed possible validation issues when changing the list of enabled
1818 - Fixed cleanup of /etc/hosts during node removal
1819 - Fixed RAPI response for invalid methods
1820 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1821 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1822 from ide disks, etc.)
1823 - Allow OS changes without re-installation (to record a changed OS
1824 outside of Ganeti, or to allow OS renames)
1825 - Allow instance creation without OS installation (useful for example if
1826 the OS will be installed manually, or restored from a backup not in
1828 - Implemented option to make cluster ``copyfile`` use the replication
1830 - Added list of enabled hypervisors to ssconf (possibly useful for
1832 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1834 - A partial form of node re-IP is possible via node readd, which now
1835 allows changed node primary IP
1836 - Command line utilities now show an informational message if the job is
1838 - The logs of the master daemon now show the PID/UID/GID of the
1845 *(Released Fri, 12 Mar 2010)*
1847 During the 2.1.0 long release candidate cycle, a lot of improvements and
1848 changes have accumulated with were released later as 2.1.1.
1853 The node evacuate command (``gnt-node evacuate``) was significantly
1854 rewritten, and as such the IAllocator protocol was changed - a new
1855 request type has been added. This unfortunate change during a stable
1856 series is designed to improve performance of node evacuations; on
1857 clusters with more than about five nodes and which are well-balanced,
1858 evacuation should proceed in parallel for all instances of the node
1859 being evacuated. As such, any existing IAllocator scripts need to be
1860 updated, otherwise the above command will fail due to the unknown
1861 request. The provided "dumb" allocator has not been updated; but the
1862 ganeti-htools package supports the new protocol since version 0.2.4.
1864 Another important change is increased validation of node and instance
1865 names. This might create problems in special cases, if invalid host
1866 names are being used.
1868 Also, a new layer of hypervisor parameters has been added, that sits at
1869 OS level between the cluster defaults and the instance ones. This allows
1870 customisation of virtualization parameters depending on the installed
1871 OS. For example instances with OS 'X' may have a different KVM kernel
1872 (or any other parameter) than the cluster defaults. This is intended to
1873 help managing a multiple OSes on the same cluster, without manual
1874 modification of each instance's parameters.
1876 A tool for merging clusters, ``cluster-merge``, has been added in the
1877 tools sub-directory.
1882 - Improved the int/float conversions that should make the code more
1883 robust in face of errors from the node daemons
1884 - Fixed the remove node code in case of internal configuration errors
1885 - Fixed the node daemon behaviour in face of inconsistent queue
1886 directory (e.g. read-only file-system where we can't open the files
1888 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1889 now it either aborts cleanly or, if given the new "auto_promote"
1890 parameter, will automatically promote other nodes as needed
1891 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1892 completely prevent Ganeti from working
1893 - Fixed bug for instance export when not all disks were successfully
1895 - Fixed behaviour of node add when the new node is slow in starting up
1897 - Fixed handling of signals in the LUXI client, which should improve
1898 behaviour of command-line scripts
1899 - Added checks for invalid node/instance names in the configuration (now
1900 flagged during cluster verify)
1901 - Fixed watcher behaviour for disk activation errors
1902 - Fixed two potentially endless loops in http library, which led to the
1903 RAPI daemon hanging and consuming 100% CPU in some cases
1904 - Fixed bug in RAPI daemon related to hashed passwords
1905 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1907 - Enhanced compatibility with non-Debian OSes, but not using absolute
1908 path in some commands and allowing customisation of the ssh
1909 configuration directory
1910 - Fixed possible future issue with new Python versions by abiding to the
1911 proper use of ``__slots__`` attribute on classes
1912 - Added checks that should prevent directory traversal attacks
1913 - Many documentation fixes based on feedback from users
1918 - Added an "early_release" more for instance replace disks and node
1919 evacuate, where we release locks earlier and thus allow higher
1920 parallelism within the cluster
1921 - Added watcher hooks, intended to allow the watcher to restart other
1922 daemons (e.g. from the ganeti-nbma project), but they can be used of
1923 course for any other purpose
1924 - Added a compile-time disable for DRBD barriers, to increase
1925 performance if the administrator trusts the power supply or the
1926 storage system to not lose writes
1927 - Added the option of using syslog for logging instead of, or in
1928 addition to, Ganeti's own log files
1929 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1930 can indeed boot from a paravirtual NIC
1931 - Added a generic debug level for many operations; while this is not
1932 used widely yet, it allows one to pass the debug value all the way to
1934 - Enhanced the hooks environment for instance moves (failovers,
1935 migrations) where the primary/secondary nodes changed during the
1936 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1937 - Enhanced data validations for many user-supplied values; one important
1938 item is the restrictions imposed on instance and node names, which
1939 might reject some (invalid) host names
1940 - Add a configure-time option to disable file-based storage, if it's not
1941 needed; this allows greater security separation between the master
1942 node and the other nodes from the point of view of the inter-node RPC
1944 - Added user notification in interactive tools if job is waiting in the
1945 job queue or trying to acquire locks
1946 - Added log messages when a job is waiting for locks
1947 - Added filtering by node tags in instance operations which admit
1948 multiple instances (start, stop, reboot, reinstall)
1949 - Added a new tool for cluster mergers, ``cluster-merge``
1950 - Parameters from command line which are of the form ``a=b,c=d`` can now
1951 use backslash escapes to pass in values which contain commas,
1952 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1954 - For KVM, the instance name is the first parameter passed to KVM, so
1955 that it's more visible in the process list
1961 *(Released Tue, 2 Mar 2010)*
1963 Ganeti 2.1 brings many improvements with it. Major changes:
1965 - Added infrastructure to ease automated disk repairs
1966 - Added new daemon to export configuration data in a cheaper way than
1967 using the remote API
1968 - Instance NICs can now be routed instead of being associated with a
1970 - Improved job locking logic to reduce impact of jobs acquiring multiple
1971 locks waiting for other long-running jobs
1973 In-depth implementation details can be found in the Ganeti 2.1 design
1979 - Added chroot hypervisor
1980 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1982 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1983 ``bootloader_path`` and ``bootloader_args``)
1984 - Added the ``use_localtime`` option for the xen-hvm and kvm
1985 hypervisors, and the default value for this has changed to false (in
1986 2.0 xen-hvm always enabled it)
1987 - Added luxi call to submit multiple jobs in one go
1988 - Added cluster initialization option to not modify ``/etc/hosts``
1990 - Added network interface parameters
1991 - Added dry run mode to some LUs
1992 - Added RAPI resources:
1994 - ``/2/instances/[instance_name]/info``
1995 - ``/2/instances/[instance_name]/replace-disks``
1996 - ``/2/nodes/[node_name]/evacuate``
1997 - ``/2/nodes/[node_name]/migrate``
1998 - ``/2/nodes/[node_name]/role``
1999 - ``/2/nodes/[node_name]/storage``
2000 - ``/2/nodes/[node_name]/storage/modify``
2001 - ``/2/nodes/[node_name]/storage/repair``
2003 - Added OpCodes to evacuate or migrate all instances on a node
2004 - Added new command to list storage elements on nodes (``gnt-node
2005 list-storage``) and modify them (``gnt-node modify-storage``)
2006 - Added new ssconf files with master candidate IP address
2007 (``ssconf_master_candidates_ips``), node primary IP address
2008 (``ssconf_node_primary_ips``) and node secondary IP address
2009 (``ssconf_node_secondary_ips``)
2010 - Added ``ganeti-confd`` and a client library to query the Ganeti
2011 configuration via UDP
2012 - Added ability to run hooks after cluster initialization and before
2014 - Added automatic mode for disk replace (``gnt-instance replace-disks
2016 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2017 after catastrophic data-loss
2018 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2020 - Added ``gnt-instance move`` command to move instances
2021 - Added ``gnt-cluster watcher`` command to control watcher
2022 - Added ``gnt-node powercycle`` command to powercycle nodes
2023 - Added new job status field ``lock_status``
2024 - Added parseable error codes to cluster verification (``gnt-cluster
2025 verify --error-codes``) and made output less verbose (use
2026 ``--verbose`` to restore previous behaviour)
2027 - Added UUIDs to the main config entities (cluster, nodes, instances)
2028 - Added support for OS variants
2029 - Added support for hashed passwords in the Ganeti remote API users file
2031 - Added option to specify maximum timeout on instance shutdown
2032 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2033 - Added new helper script to start and stop Ganeti daemons
2034 (``daemon-util``), with the intent to reduce the work necessary to
2035 adjust Ganeti for non-Debian distributions and to start/stop daemons
2037 - Added more unittests
2038 - Fixed critical bug in ganeti-masterd startup
2039 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2040 now customisable at the cluster level for both the KVM and Xen
2041 hypervisors using the new ``migration_port`` parameter
2042 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2043 reinstalling an instance
2044 - Allowed ``@`` in tag names
2045 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2046 - Many documentation updates
2047 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2048 - ``gnt-instance reinstall`` can now reinstall multiple instances
2049 - Updated many command line parameters
2050 - Introduced new OS API version 15
2051 - No longer support a default hypervisor
2052 - Treat virtual LVs as inexistent
2053 - Improved job locking logic to reduce lock contention
2054 - Match instance and node names case insensitively
2055 - Reimplemented bash completion script to be more complete
2062 *(Released Thu, 4 Feb 2010)*
2064 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2065 - Fix a string formatting bug
2066 - Improve safety of the code in some error paths
2067 - Improve data validation in the master of values returned from nodes
2073 *(Released Thu, 17 Dec 2009)*
2075 - Fix security issue due to missing validation of iallocator names; this
2076 allows local and remote execution of arbitrary executables
2077 - Fix failure of gnt-node list during instance removal
2078 - Ship the RAPI documentation in the archive
2084 *(Released Wed, 30 Sep 2009)*
2086 - Fixed many wrong messages
2087 - Fixed a few bugs related to the locking library
2088 - Fixed MAC checking at instance creation time
2089 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2090 - Fixed a few issues related to signal handling in both daemons and
2092 - Fixed the example startup script provided
2093 - Fixed insserv dependencies in the example startup script (patch from
2095 - Fixed handling of drained nodes in the iallocator framework
2096 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2098 - Fixed error related to invalid job IDs in job polling
2099 - Fixed job/opcode persistence on unclean master shutdown
2100 - Fixed handling of partial job processing after unclean master
2102 - Fixed error reporting from LUs, previously all errors were converted
2103 into execution errors
2104 - Fixed error reporting from burnin
2105 - Decreased significantly the memory usage of the job queue
2106 - Optimised slightly multi-job submission
2107 - Optimised slightly opcode loading
2108 - Backported the multi-job submit framework from the development
2109 branch; multi-instance start and stop should be faster
2110 - Added script to clean archived jobs after 21 days; this will reduce
2111 the size of the queue directory
2112 - Added some extra checks in disk size tracking
2113 - Added an example ethers hook script
2114 - Added a cluster parameter that prevents Ganeti from modifying of
2116 - Added more node information to RAPI responses
2117 - Added a ``gnt-job watch`` command that allows following the ouput of a
2119 - Added a bind-address option to ganeti-rapi
2120 - Added more checks to the configuration verify
2121 - Enhanced the burnin script such that some operations can be retried
2123 - Converted instance reinstall to multi-instance model
2129 *(Released Fri, 7 Aug 2009)*
2131 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2132 to allow using the pre-2.0.2 behaviour in activation, if any existing
2133 instances have mismatched disk sizes in the configuration
2134 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2135 any configuration mismatches for disk sizes
2136 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2137 failover to work on two-node clusters
2138 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2140 - Fixed detection of OS script errors in ``gnt-backup export``
2141 - Fixed exit code of ``gnt-backup export``
2147 *(Released Fri, 17 Jul 2009)*
2149 - Added experimental support for stripped logical volumes; this should
2150 enhance performance but comes with a higher complexity in the block
2151 device handling; stripping is only enabled when passing
2152 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2153 affected even in the non-stripped mode
2154 - Improved resiliency against transient failures at the end of DRBD
2155 resyncs, and in general of DRBD resync checks
2156 - Fixed a couple of issues with exports and snapshot errors
2157 - Fixed a couple of issues in instance listing
2158 - Added display of the disk size in ``gnt-instance info``
2159 - Fixed checking for valid OSes in instance creation
2160 - Fixed handling of the "vcpus" parameter in instance listing and in
2161 general of invalid parameters
2162 - Fixed http server library, and thus RAPI, to handle invalid
2163 username/password combinations correctly; this means that now they
2164 report unauthorized for queries too, not only for modifications,
2165 allowing earlier detect of configuration problems
2166 - Added a new "role" node list field, equivalent to the master/master
2167 candidate/drained/offline flags combinations
2168 - Fixed cluster modify and changes of candidate pool size
2169 - Fixed cluster verify error messages for wrong files on regular nodes
2170 - Fixed a couple of issues with node demotion from master candidate role
2171 - Fixed node readd issues
2172 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2173 - Added a new ``--no-voting`` option for masterfailover to fix failover
2174 on two-nodes clusters when the former master node is unreachable
2175 - Added instance reinstall over RAPI
2181 *(Released Tue, 16 Jun 2009)*
2183 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2184 allow re-adding the start in single-user option (regression from 1.2)
2185 - the watcher writes the instance status to a file, to allow monitoring
2186 to report the instance status (from the master) based on cached
2187 results of the watcher's queries; while this can get stale if the
2188 watcher is being locked due to other work on the cluster, this is
2189 still an improvement
2190 - the watcher now also restarts the node daemon and the rapi daemon if
2192 - fixed the watcher to handle full and drained queue cases
2193 - hooks export more instance data in the environment, which helps if
2194 hook scripts need to take action based on the instance's properties
2195 (no longer need to query back into ganeti)
2196 - instance failovers when the instance is stopped do not check for free
2197 RAM, so that failing over a stopped instance is possible in low memory
2199 - rapi uses queries for tags instead of jobs (for less job traffic), and
2200 for cluster tags it won't talk to masterd at all but read them from
2202 - a couple of error handling fixes in RAPI
2203 - drbd handling: improved the error handling of inconsistent disks after
2204 resync to reduce the frequency of "there are some degraded disks for
2205 this instance" messages
2206 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2207 error handling path called a wrong function name)
2213 *(Released Wed, 27 May 2009)*
2215 - no changes from rc5
2221 *(Released Wed, 20 May 2009)*
2223 - fix a couple of bugs (validation, argument checks)
2224 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2225 - some small improvements to RAPI and IAllocator
2226 - make watcher automatically start the master daemon if down
2232 *(Released Mon, 27 Apr 2009)*
2234 - change the OS list to not require locks; this helps with big clusters
2235 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2236 volume group is broken
2237 - ``gnt-instance info``, without any arguments, doesn't run for all
2238 instances anymore; either pass ``--all`` or pass the desired
2239 instances; this helps against mistakes on big clusters where listing
2240 the information for all instances takes a long time
2241 - miscellaneous doc and man pages fixes
2247 *(Released Wed, 8 Apr 2009)*
2249 - Change the internal locking model of some ``gnt-node`` commands, in
2250 order to reduce contention (and blocking of master daemon) when
2251 batching many creation/reinstall jobs
2252 - Fixes to Xen soft reboot
2253 - No longer build documentation at build time, instead distribute it in
2254 the archive, in order to reduce the need for the whole docbook/rst
2261 *(Released Fri, 27 Mar 2009)*
2263 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2264 - Fix watcher startup sequence, improves the behaviour of busy clusters
2265 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2266 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2267 bind address and other places
2268 - Some documentation fixes and updates
2274 *(Released Mon, 2 Mar 2009)*
2276 - More documentation updates, now all docs should be more-or-less
2278 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2280 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2282 - Fix an issue related to $libdir/run/ganeti and cluster creation
2288 *(Released Thu, 19 Feb 2009)*
2290 - Xen PVM and KVM have switched the default value for the instance root
2291 disk to the first partition on the first drive, instead of the whole
2292 drive; this means that the OS installation scripts must be changed
2294 - Man pages have been updated
2295 - RAPI has been switched by default to HTTPS, and the exported functions
2296 should all work correctly
2297 - RAPI v1 has been removed
2298 - Many improvements to the KVM hypervisor
2299 - Block device errors are now better reported
2300 - Many other bugfixes and small improvements
2306 *(Released Mon, 26 Jan 2009)*
2308 - Version 2 is a general rewrite of the code and therefore the
2309 differences are too many to list, see the design document for 2.0 in
2310 the ``doc/`` subdirectory for more details
2311 - In this beta version there is not yet a migration path from 1.2 (there
2312 will be one in the final 2.0 release)
2313 - A few significant changes are:
2315 - all commands are executed by a daemon (``ganeti-masterd``) and the
2316 various ``gnt-*`` commands are just front-ends to it
2317 - all the commands are entered into, and executed from a job queue,
2318 see the ``gnt-job(8)`` manpage
2319 - the RAPI daemon supports read-write operations, secured by basic
2320 HTTP authentication on top of HTTPS
2321 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2322 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2323 to migrate to DRBD 8 first while still running Ganeti 1.2)
2324 - DRBD devices are using statically allocated minor numbers, which
2325 will be assigned to existing instances during the migration process
2326 - there is support for both Xen PVM and Xen HVM instances running on
2328 - KVM virtualization is supported too
2329 - file-based storage has been implemented, which means that it is
2330 possible to run the cluster without LVM and DRBD storage, for
2331 example using a shared filesystem exported from shared storage (and
2332 still have live migration)
2338 *(Released Tue, 13 Jan 2009)*
2340 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2341 - Reuse the old instance mac address by default on instance import, if
2342 the instance name is the same.
2343 - Handle situations in which the node info rpc returns incomplete
2345 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2346 - Improved version of batcher:
2348 - state file support
2349 - instance mac address support
2350 - support for HVM clusters/instances
2352 - Add an option to show the number of cpu sockets and nodes in
2354 - Support OSes that handle more than one version of the OS api (but do
2355 not change the current API in any other way)
2356 - Fix ``gnt-node migrate``
2357 - ``gnt-debug`` man page
2358 - Fixes various more typos and small issues
2359 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2365 *(Released Wed, 24 Sep 2008)*
2367 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2368 type of disk exported to fully virtualized instances.
2369 - provide access to the serial console of HVM instances
2370 - instance auto_balance flag, set by default. If turned off it will
2371 avoid warnings on cluster verify if there is not enough memory to fail
2372 over an instance. in the future it will prevent automatically failing
2373 it over when we will support that.
2374 - batcher tool for instance creation, see ``tools/README.batcher``
2375 - ``gnt-instance reinstall --select-os`` to interactively select a new
2376 operating system when reinstalling an instance.
2377 - when changing the memory amount on instance modify a check has been
2378 added that the instance will be able to start. also warnings are
2379 emitted if the instance will not be able to fail over, if auto_balance
2381 - documentation fixes
2382 - sync fields between ``gnt-instance list/modify/add/import``
2383 - fix a race condition in drbd when the sync speed was set after giving
2384 the device a remote peer.
2390 *(Released Tue, 22 Jul 2008)*
2392 - note: the allowed size and number of tags per object were reduced
2393 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2394 - fixed twisted 8.x compatibility
2395 - fixed ``gnt-instance replace-disks`` with iallocator
2396 - add TCP keepalives on twisted connections to detect restarted nodes
2397 - disk increase support, see ``gnt-instance grow-disk``
2398 - implement bulk node/instance query for RAPI
2399 - add tags in node/instance listing (optional)
2400 - experimental migration (and live migration) support, read the man page
2401 for ``gnt-instance migrate``
2402 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2403 has some small improvements in handling its state file
2409 *(Released Fri, 13 Jun 2008)*
2411 - Experimental readonly, REST-based remote API implementation;
2412 automatically started on master node, TCP port 5080, if enabled by
2413 ``--enable-rapi`` parameter to configure script.
2414 - Instance allocator support. Add and import instance accept a
2415 ``--iallocator`` parameter, and call that instance allocator to decide
2416 which node to use for the instance. The iallocator document describes
2417 what's expected from an allocator script.
2418 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2419 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2420 node is lost there is still enough memory to fail over the instances
2422 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2423 ``gnt-cluster verify``, to check for site-specific compliance. All the
2424 hooks will run, and their output, if any, will be displayed. Any
2425 failing hook will make the verification return an error value.
2426 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2427 primary and secondary interfaces
2428 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2429 that is still declared as part of the cluster and has failed.
2430 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2431 specifying output fields, that just adds the chosen fields to the
2433 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2434 export from the filesystem.
2435 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2436 have been added. Using them you can enable/disable acpi and pae
2437 support, and specify a path for a cd image to be exported to the
2438 instance. These parameters as the name suggest only work on HVM
2440 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2441 PAE support will be set to the previously hardcoded values, but the
2442 (previously hardcoded) path to the CDROM ISO image will be unset and
2443 if required, needs to be set manually with ``gnt-instance modify``
2445 - The address to which an instance's VNC console is bound is now
2446 selectable per-instance, rather than being cluster wide. Of course
2447 this only applies to instances controlled via VNC, so currently just
2448 applies to HVM clusters.
2454 *(Released Mon, 18 Feb 2008)*
2456 - more tweaks to the disk activation code (especially helpful for DRBD)
2457 - change the default ``gnt-instance list`` output format, now there is
2458 one combined status field (see the manpage for the exact values this
2460 - some more fixes for the mac export to hooks change
2461 - make Ganeti not break with DRBD 8.2.x (which changed the version
2462 format in ``/proc/drbd``) (issue 24)
2463 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2464 template, allowing migration from DRBD0.7+MD to DRBD8
2470 *(Released Wed, 30 Jan 2008)*
2472 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2474 - add command aliases infrastructure and a few aliases
2475 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2476 man pages and the ``--help`` option of ``gnt-node
2477 list``/``gnt-instance list``
2478 - fix ``gnt-backup list`` with down nodes (issue 21)
2479 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2480 - fix the dist archive and add a check for including svn/git files in
2482 - some developer-related changes: improve the burnin and the QA suite,
2483 add an upload script for testing during development
2489 *(Released Wed, 16 Jan 2008)*
2491 - experimental HVM support, read the install document, section
2492 "Initializing the cluster"
2493 - allow for the PVM hypervisor per-instance kernel and initrd paths
2494 - add a new command ``gnt-cluster verify-disks`` which uses a new
2495 algorithm to improve the reconnection of the DRBD pairs if the device
2496 on the secondary node has gone away
2497 - make logical volume code auto-activate LVs at disk activation time
2498 - slightly improve the speed of activating disks
2499 - allow specification of the MAC address at instance creation time, and
2500 changing it later via ``gnt-instance modify``
2501 - fix handling of external commands that generate lots of output on
2503 - update documentation with regard to minimum version of DRBD8 supported
2509 *(Released Tue, 4 Dec 2007)*
2511 - Log the ``xm create`` output to the node daemon log on failure (to
2512 help diagnosing the error)
2513 - In debug mode, log all external commands output if failed to the logs
2514 - Change parsing of lvm commands to ignore stderr
2520 *(Released Wed, 28 Nov 2007)*
2522 - Another round of updates to the DRBD 8 code to deal with more failures
2523 in the replace secondary node operation
2524 - Some more logging of failures in disk operations (lvm, drbd)
2525 - A few documentation updates
2532 *(Released Tue, 13 Nov 2007)*
2534 - Change configuration file format from Python's Pickle to JSON.
2535 Upgrading is possible using the cfgupgrade utility.
2536 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2537 faster replace disks and is more stable (DRBD 8 has many improvements
2538 compared to DRBD 0.7)
2539 - Added command line tags support (see man pages for ``gnt-instance``,
2540 ``gnt-node``, ``gnt-cluster``)
2541 - Added instance rename support
2542 - Added multi-instance startup/shutdown
2543 - Added cluster rename support
2544 - Added ``gnt-node evacuate`` to simplify some node operations
2545 - Added instance reboot operation that can speedup reboot as compared to
2547 - Soften the requirement that hostnames are in FQDN format
2548 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2550 - Removed dependency on debian's patched fping that uses the
2551 non-standard ``-S`` option
2552 - Now the OS definitions are searched for in multiple, configurable
2553 paths (easier for distros to package)
2554 - Some changes to the hooks infrastructure (especially the new
2555 post-configuration update hook)
2556 - Other small bugfixes
2558 .. vim: set textwidth=72 syntax=rst :