10 - ``gnt-instance batch-create`` has been changed to use the bulk create
11 opcode from Ganeti. This lead to incompatible changes in the format of
12 the JSON file. It's now not a custom dict anymore but a dict
13 compatible with the ``OpInstanceCreate`` opcode.
14 - Parent directories for file storage need to be listed in
15 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
16 write the file automatically based on old configuration values, but it
17 can not distribute it across all nodes and the file contents should be
18 verified. Use ``gnt-cluster copyfile
19 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
20 upgraded. The reason for requiring this list of paths now is that
21 before it would have been possible to inject new paths via RPC,
22 allowing files to be created in arbitrary locations. The RPC protocol
23 is protected using SSL/X.509 certificates, but as a design principle
24 Ganeti does not permit arbitrary paths to be passed.
25 - The parsing of the variants file for OSes (see
26 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
27 lines and comment lines (starting with ``#``) are ignored for better
29 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
30 longer available. ``gnt-node add`` now invokes a new tool on the
31 destination node, named ``prepare-node-join``, to configure the SSH
32 daemon. Paramiko is no longer necessary to configure nodes' SSH
33 daemons via ``gnt-node add``.
34 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
35 for RAPI users. It allows granting permissions to query for
36 information to a specific user without giving
37 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
38 - A new tool named ``node-cleanup`` has been added. It cleans remains of
39 a cluster from a machine by stopping all daemons, removing
40 certificates and ssconf files. Unless the ``--no-backup`` option is
41 given, copies of the certificates are made.
42 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
43 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
44 the flag is not reset after a master failover.
45 - Instance creations now support the use of opportunistic locking,
46 potentially speeding up the (parallel) creation of multiple instances.
47 This feature is currently only available via the :doc:`RAPI
48 <rapi>` interface and when an instance allocator is used. If the
49 ``opportunistic_locking`` parameter is set the opcode will try to
50 acquire as many locks as possible, but will not wait for any locks
51 held by other opcodes. If the not enough resources can be found to
52 allocate the instance, the temporary error code
53 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
54 retried thereafter, with or without opportunistic locking.
55 - Man pages can now be included when the documentation is built, in
56 which case the output is in ``doc/man-html``. The configure-time
57 option is ``--enable-manpages-in-doc``. Sphinx 1.0 or higher is
64 *(Released Fri, 21 Dec 2012)*
66 Important behaviour change: hbal won't rebalance anymore instances which
67 have the ``auto_balance`` attribute set to false. This was the intention
68 all along, but until now it only skipped those from the N+1 memory
69 reservation (DRBD-specific).
71 A significant number of bug fixes in this release:
73 - Fixed disk adoption interaction with ipolicy checks.
74 - Fixed networking issues when instances are started, stopped or
75 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
76 - Fixed the warning in cluster verify for shared storage instances not
78 - Fixed removal of storage directory on shared file storage (issue 262).
79 - Fixed validation of LVM volume group name in OpClusterSetParams
80 (``gnt-cluster modify``) (issue 285).
81 - Fixed runtime memory increases (``gnt-instance modify -m``).
82 - Fixed live migration under Xen's ``xl`` mode.
83 - Fixed ``gnt-instance console`` with ``xl``.
84 - Fixed building with newer Haskell compiler/libraries.
85 - Fixed PID file writing in Haskell daemons (confd); this prevents
86 restart issues if confd was launched manually (outside of
87 ``daemon-util``) while another copy of it was running
88 - Fixed a type error when doing live migrations with KVM (issue 297) and
89 the error messages for failing migrations have been improved.
90 - Fixed opcode validation for the out-of-band commands (``gnt-node
92 - Fixed a type error when unsetting OS hypervisor parameters (issue
93 311); now it's possible to unset all OS-specific hypervisor
95 - Fixed the ``dry-run`` mode for many operations: verification of
96 results was over-zealous but didn't take into account the ``dry-run``
97 operation, resulting in "wrong" failures.
98 - Fixed bash completion in ``gnt-job list`` when the job queue has
99 hundreds of entries; especially with older ``bash`` versions, this
100 results in significant CPU usage.
102 And lastly, a few other improvements have been made:
104 - Added option to force master-failover without voting (issue 282).
105 - Clarified error message on lock conflict (issue 287).
106 - Logging of newly submitted jobs has been improved (issue 290).
107 - Hostname checks have been made uniform between instance rename and
109 - The ``--submit`` option is now supported by ``gnt-debug delay``.
110 - Shutting down the master daemon by sending SIGTERM now stops it from
111 processing jobs waiting for locks; instead, those jobs will be started
112 once again after the master daemon is started the next time (issue
114 - Support for Xen's ``xl`` program has been improved (besides the fixes
116 - Reduced logging noise in the Haskell confd daemon (only show one log
117 entry for each config reload, instead of two).
118 - Several man page updates and typo fixes.
124 *(Released Fri, 12 Oct 2012)*
126 A small bugfix release. Among the bugs fixed:
128 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
129 made the command unusable.
130 - Commands that issue many jobs don't fail anymore just because some jobs
131 take so long that other jobs are archived.
132 - Failures during ``gnt-instance reinstall`` are reflected by the exit
134 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
136 - When ``always_failover`` is set, ``--allow-failover`` is not required
137 in migrate commands anymore.
138 - ``bash_completion`` works even if extglob is disabled.
139 - Fixed bug with locks that made failover for RDB-based instances fail.
140 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
141 a random node instead of one based on the allocator metric.
142 - Support for newer versions of pylint and pep8.
143 - Hail doesn't fail anymore when trying to add an instance of type
144 ``file``, ``sharedfile`` or ``rbd``.
145 - Added new Makefile target to rebuild the whole distribution, so that
146 all files are included.
152 *(Released Fri, 27 Jul 2012)*
155 .. attention:: The ``LUXI`` protocol has been made more consistent
156 regarding its handling of command arguments. This, however, leads to
157 incompatibility issues with previous versions. Please ensure that you
158 restart Ganeti daemons soon after the upgrade, otherwise most
159 ``LUXI`` calls (job submission, setting/resetting the drain flag,
160 pausing/resuming the watcher, cancelling and archiving jobs, querying
161 the cluster configuration) will fail.
170 The current ``admin_up`` field, which used to denote whether an instance
171 should be running or not, has been removed. Instead, ``admin_state`` is
172 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
174 The rational behind this is that an instance being “down” can have
177 - it could be down during a reboot
178 - it could be temporarily be down for a reinstall
179 - or it could be down because it is deprecated and kept just for its
182 The previous Boolean state was making it difficult to do capacity
183 calculations: should Ganeti reserve memory for a down instance? Now, the
184 tri-state field makes it clear:
186 - in ``up`` and ``down`` state, all resources are reserved for the
187 instance, and it can be at any time brought up if it is down
188 - in ``offline`` state, only disk space is reserved for it, but not
191 The field can have an extra use: since the transition between ``up`` and
192 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
193 transition between ``offline`` and ``down`` is done via ``gnt-instance
194 modify``, it is possible to given different rights to users. For
195 example, owners of an instance could be allowed to start/stop it, but
196 not transition it out of the offline state.
198 Instance policies and specs
199 +++++++++++++++++++++++++++
201 In previous Ganeti versions, an instance creation request was not
202 limited on the minimum size and on the maximum size just by the cluster
203 resources. As such, any policy could be implemented only in third-party
204 clients (RAPI clients, or shell wrappers over ``gnt-*``
205 tools). Furthermore, calculating cluster capacity via ``hspace`` again
206 required external input with regards to instance sizes.
208 In order to improve these workflows and to allow for example better
209 per-node group differentiation, we introduced instance specs, which
212 - minimum instance disk size, disk count, memory size, cpu count
213 - maximum values for the above metrics
214 - and “standard” values (used in ``hspace`` to calculate the standard
217 The minimum/maximum values can be also customised at node-group level,
218 for example allowing more powerful hardware to support bigger instance
221 Beside the instance specs, there are a few other settings belonging to
222 the instance policy framework. It is possible now to customise, per
223 cluster and node-group:
225 - the list of allowed disk templates
226 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
227 - the maximum ratio of instance to spindles (see below for more
228 information) for local storage
230 All these together should allow all tools that talk to Ganeti to know
231 what are the ranges of allowed values for instances and the
232 over-subscription that is allowed.
234 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
235 instance configuration, and the physical CPU configuration from the
236 node. For the spindle ratios however, we didn't track before these
237 values, so new parameters have been added:
239 - a new node parameter ``spindle_count``, defaults to 1, customisable at
240 node group or node level
241 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
243 Note that spindles in this context doesn't need to mean actual
244 mechanical hard-drives; it's just a relative number for both the node
245 I/O capacity and instance I/O consumption.
247 Instance migration behaviour
248 ++++++++++++++++++++++++++++
250 While live-migration is in general desirable over failover, it is
251 possible that for some workloads it is actually worse, due to the
252 variable time of the “suspend” phase during live migration.
254 To allow the tools to work consistently over such instances (without
255 having to hard-code instance names), a new backend parameter
256 ``always_failover`` has been added to control the migration/failover
257 behaviour. When set to True, all migration requests for an instance will
258 instead fall-back to failover.
260 Instance memory ballooning
261 ++++++++++++++++++++++++++
263 Initial support for memory ballooning has been added. The memory for an
264 instance is no longer fixed (backend parameter ``memory``), but instead
265 can vary between minimum and maximum values (backend parameters
266 ``minmem`` and ``maxmem``). Currently we only change an instance's
269 - live migrating or failing over and instance and the target node
270 doesn't have enough memory
271 - user requests changing the memory via ``gnt-instance modify
277 In order to control the use of specific CPUs by instance, support for
278 controlling CPU pinning has been added for the Xen, HVM and LXC
279 hypervisors. This is controlled by a new hypervisor parameter
280 ``cpu_mask``; details about possible values for this are in the
281 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
282 VCPU-to-CPU mapping) form will work well only when all nodes in your
283 cluster have the same amount of CPUs.
288 Another area in which Ganeti was not customisable were the parameters
289 used for storage configuration, e.g. how many stripes to use for LVM,
290 DRBD resync configuration, etc.
292 To improve this area, we've added disks parameters, which are
293 customisable at cluster and node group level, and which allow to
294 specify various parameters for disks (DRBD has the most parameters
295 currently), for example:
297 - DRBD resync algorithm and parameters (e.g. speed)
298 - the default VG for meta-data volumes for DRBD
299 - number of stripes for LVM (plain disk template)
302 These parameters can be modified via ``gnt-cluster modify -D …`` and
303 ``gnt-group modify -D …``, and are used at either instance creation (in
304 case of LVM stripes, for example) or at disk “activation” time
307 Rados block device support
308 ++++++++++++++++++++++++++
310 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
311 denoted by the ``rbd`` disk template type. This is considered
312 experimental, feedback is welcome. For details on configuring it, see
313 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
318 The existing master IP functionality works well only in simple setups (a
319 single network shared by all nodes); however, if nodes belong to
320 different networks, then the ``/32`` setup and lack of routing
321 information is not enough.
323 To allow the master IP to function well in more complex cases, the
324 system was reworked as follows:
326 - a master IP netmask setting has been added
327 - the master IP activation/turn-down code was moved from the node daemon
329 - whether to run the Ganeti-supplied master IP script or a user-supplied
330 on is a ``gnt-cluster init`` setting
332 Details about the location of the standard and custom setup scripts are
333 in the man page :manpage:`gnt-cluster(8)`; for information about the
334 setup script protocol, look at the Ganeti-supplied script.
339 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
342 It is now possible to use TLS-protected connections, and when renewing
343 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
344 it is now possible to specify spice or spice CA certificates. Also, it
345 is possible to configure a password for SPICE sessions via the
346 hypervisor parameter ``spice_password_file``.
348 There are also new parameters to control the compression and streaming
349 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
350 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
351 for the spice parameters.
353 Lastly, it is now possible to see the SPICE connection information via
354 ``gnt-instance console``.
359 A new tool (``tools/ovfconverter``) has been added that supports
360 conversion between Ganeti and the `Open Virtualization Format
361 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
364 This relies on the ``qemu-img`` tool to convert the disk formats, so the
365 actual compatibility with other virtualization solutions depends on it.
370 The configuration query daemon (``ganeti-confd``) is now optional, and
371 has been rewritten in Haskell; whether to use the daemon at all, use the
372 Python (default) or the Haskell version is selectable at configure time
373 via the ``--enable-confd`` parameter, which can take one of the
374 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
375 daemon will result in a smaller footprint; for larger systems, we
376 welcome feedback on the Haskell version which might become the default
379 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
380 daemon running. The Python version doesn't implement the new call.
383 User interface changes
384 ~~~~~~~~~~~~~~~~~~~~~~
386 We have replaced the ``--disks`` option of ``gnt-instance
387 replace-disks`` with a more flexible ``--disk`` option, which allows
388 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
389 disk size and mode can be changed upon recreation (via ``gnt-instance
390 recreate-disks``, which accepts the same ``--disk`` option).
392 As many people are used to a ``show`` command, we have added that as an
393 alias to ``info`` on all ``gnt-*`` commands.
395 The ``gnt-instance grow-disk`` command has a new mode in which it can
396 accept the target size of the disk, instead of the delta; this can be
397 more safe since two runs in absolute mode will be idempotent, and
398 sometimes it's also easier to specify the desired size directly.
400 Also the handling of instances with regard to offline secondaries has
401 been improved. Instance operations should not fail because one of it's
402 secondary nodes is offline, even though it's safe to proceed.
404 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
405 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
406 minors to instance name.
411 RAPI coverage has improved, with (for example) new resources for
412 recreate-disks, node power-cycle, etc.
417 There is partial support for ``xl`` in the Xen hypervisor; feedback is
420 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
421 whether to still support Python 2.4 or move to Python 2.6 as minimum
424 Support for Fedora has been slightly improved; the provided example
425 init.d script should work better on it and the INSTALL file should
426 document the needed dependencies.
431 The deprecated ``QueryLocks`` LUXI request has been removed. Use
432 ``Query(what=QR_LOCK, ...)`` instead.
434 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
435 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
436 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
437 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
438 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
440 RAPI client: ``CertificateError`` now derives from
441 ``GanetiApiError``. This should make it more easy to handle Ganeti
444 Deprecation warnings due to PyCrypto/paramiko import in
445 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
446 make sure to run an up-to-date paramiko version, if you use this tool.
448 The QA scripts now depend on Python 2.5 or above (the main code base
449 still works with Python 2.4).
451 The configuration file (``config.data``) is now written without
452 indentation for performance reasons; if you want to edit it, it can be
453 re-formatted via ``tools/fmtjson``.
455 A number of bugs has been fixed in the cluster merge tool.
457 ``x509`` certification verification (used in import-export) has been
458 changed to allow the same clock skew as permitted by the cluster
459 verification. This will remove some rare but hard to diagnose errors in
466 *(Released Thu, 19 Jul 2012)*
468 Very few changes from rc4 to the final release, only bugfixes:
470 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
471 instance, fix CDROM booting for KVM instances)
472 - fixed node group modification of node parameters
473 - fixed issue in LUClusterVerifyGroup with multi-group clusters
474 - fixed generation of bash completion to ensure a stable ordering
481 *(Released Fri, 13 Jul 2012)*
483 Third release candidate for 2.6. The following changes were done from
486 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
487 - Fixed an inconsistency in the LUXI protocol with the provided
488 arguments (NOT backwards compatible)
489 - Fixed a bug with node groups ipolicy where ``min`` was greater than
490 the cluster ``std`` value
491 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
492 easier instance debugging on nodes (requires ``hconfd`` to work)
498 *(Released Tue, 03 Jul 2012)*
500 Second release candidate for 2.6. The following changes were done from
503 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
505 - Fixed a RAPI regression on missing beparams/memory
506 - Fixed redistribution of files on offline nodes
507 - Added possibility to run activate-disks even though secondaries are
508 offline. With this change it relaxes also the strictness on some other
509 commands which use activate disks internally:
510 * ``gnt-instance start|reboot|rename|backup|export``
511 - Made it possible to remove safely an instance if its secondaries are
513 - Made it possible to reinstall even though secondaries are offline
519 *(Released Mon, 25 Jun 2012)*
521 First release candidate for 2.6. The following changes were done from
524 - Fixed bugs with disk parameters and ``rbd`` templates as well as
526 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
527 behaviour. It supports now the modify operation
528 - ``hcheck`` implemented to analyze cluster health and possibility of
529 improving health by rebalance
530 - ``hbal`` has been improved in dealing with split instances
536 *(Released Mon, 11 Jun 2012)*
538 Second beta release of 2.6. The following changes were done from beta2
541 - Fixed ``daemon-util`` with non-root user models
542 - Fixed creation of plain instances with ``--no-wait-for-sync``
543 - Fix wrong iv_names when running ``cfgupgrade``
544 - Export more information in RAPI group queries
545 - Fixed bug when changing instance network interfaces
546 - Extended burnin to do NIC changes
547 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
548 - Changed default for DRBD barriers
549 - Fixed DRBD error reporting for syncer rate
550 - Verify the options on disk parameters
552 And of course various fixes to documentation and improved unittests and
559 *(Released Wed, 23 May 2012)*
561 First beta release of 2.6. The following changes were done from beta1 to
564 - integrated patch for distributions without ``start-stop-daemon``
565 - adapted example init.d script to work on Fedora
566 - fixed log handling in Haskell daemons
567 - adapted checks in the watcher for pycurl linked against libnss
568 - add partial support for ``xl`` instead of ``xm`` for Xen
569 - fixed a type issue in cluster verification
570 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
573 Plus integrated fixes from the 2.5 branch:
575 - fixed ``kvm-ifup`` to use ``/bin/bash``
576 - fixed parallel build failures
577 - KVM live migration when using a custom keymap
583 *(Released Tue, 24 Jul 2012)*
585 A small bugfix release, with no new features:
587 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
588 different default shell (e.g. Debian, Ubuntu)
589 - fixed KVM startup and live migration with a custom keymap (fixes Issue
590 243 and Debian bug #650664)
591 - fixed compatibility with KVM versions that don't support multiple boot
592 devices (fixes Issue 230 and Debian bug #624256)
594 Additionally, a few fixes were done to the build system (fixed parallel
595 build failures) and to the unittests (fixed race condition in test for
596 FileID functions, and the default enable/disable mode for QA test is now
603 *(Released Fri, 11 May 2012)*
605 A small bugfix release.
607 The main issues solved are on the topic of compatibility with newer LVM
610 - fixed parsing of ``lv_attr`` field
611 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
612 the ``--force`` flag is needed
614 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
615 to accept kernel 3.x too (was hardcoded to 2.6.*).
617 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
618 of hook results and that also made display of other errors suboptimal
619 was fixed; the code behaves now like 2.4 and earlier.
621 Another change in 2.5, the cleanup of the OS scripts environment, is too
622 aggressive: it removed even the ``PATH`` variable, which requires the OS
623 scripts to *always* need to export it. Since this is a bit too strict,
624 we now export a minimal PATH, the same that we export for hooks.
626 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
627 integrated into this release.
629 Finally, a few other miscellaneous changes were done (no new features,
630 just small improvements):
632 - Fix ``gnt-group --help`` display
633 - Fix hardcoded Xen kernel path
634 - Fix grow-disk handling of invalid units
635 - Update synopsis for ``gnt-cluster repair-disk-sizes``
636 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
642 *(Released Thu, 12 Apr 2012)*
644 Incompatible/important changes and bugfixes
645 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
647 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
648 resource's ``ip_check`` parameter changed from ``True`` to ``False``
649 to match the underlying LUXI interface.
650 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
651 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
652 not maintain backwards-compatibility as the underlying operation
653 changed in an incompatible way. The RAPI client can talk to old
654 servers, but it needs to be told so as the return value changed.
655 - When creating file-based instances via RAPI, the ``file_driver``
656 parameter no longer defaults to ``loop`` and must be specified.
657 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
659 - Support for the undocumented and deprecated RAPI instance creation
660 request format version 0 has been dropped. Use version 1, supported
661 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
662 - Pyparsing 1.4.6 or above is required, see :doc:`installation
663 documentation <install>`.
664 - The "cluster-verify" hooks are now executed per group by the
665 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
666 if you just run ``gnt-cluster verify``, which generates one opcode per
668 - The environment as passed to the OS scripts is cleared, and thus no
669 environment variables defined in the node daemon's environment will be
670 inherited by the scripts.
671 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
673 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
674 support operations involving multiple node groups.
675 - Offline nodes are ignored when failing over an instance.
676 - Support for KVM version 1.0, which changed the version reporting format
678 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
680 - ``Makefile`` is now compatible with Automake 1.11.2
681 - Includes all bugfixes made in the 2.4 series
686 - The ganeti-htools project has been merged into the ganeti-core source
687 tree and will be built as part of Ganeti (see :doc:`install-quick`).
688 - Implemented support for :doc:`shared storage <design-shared-storage>`.
689 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
690 GPT-style partition tables (requires `parted
691 <http://www.gnu.org/s/parted/>`_).
692 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
693 - Allowed adding tags on instance creation.
694 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
695 - Allow instances to be started in a paused state, enabling the user to
696 see the complete console output on boot using the console.
697 - Added new hypervisor flag to control default reboot behaviour
698 (``reboot_behavior``).
699 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
700 - Improved out-of-band management support:
702 - Added ``gnt-node health`` command reporting the health status of
704 - Added ``gnt-node power`` command to manage power status of nodes.
705 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
707 - Instance migration can fall back to failover if instance is not
709 - Filters can be used when listing nodes, instances, groups and locks;
710 see :manpage:`ganeti(7)` manpage.
711 - Added post-execution status as variables to :doc:`hooks <hooks>`
713 - Instance tags are exported/imported together with the instance.
714 - When given an explicit job ID, ``gnt-job info`` will work for archived
716 - Jobs can define dependencies on other jobs (not yet supported via
717 RAPI or command line, but used by internal commands and usable via
720 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
723 - Instance failover is now available as a RAPI resource
724 (``/2/instances/[instance_name]/failover``).
725 - ``gnt-instance info`` defaults to static information if primary node
727 - Opcodes have a new ``comment`` attribute.
728 - Added basic SPICE support to KVM hypervisor.
729 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
731 Node group improvements
732 ~~~~~~~~~~~~~~~~~~~~~~~
734 - ``gnt-cluster verify`` has been modified to check groups separately,
735 thereby improving performance.
736 - Node group support has been added to ``gnt-cluster verify-disks``,
737 which now operates per node group.
738 - Watcher has been changed to work better with node groups.
740 - One process and state file per node group.
741 - Slow watcher in one group doesn't block other group's watcher.
743 - Added new command, ``gnt-group evacuate``, to move all instances in a
744 node group to other groups.
745 - Added ``gnt-instance change-group`` to move an instance to another
747 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
748 per-group operations.
749 - Node groups can be tagged.
750 - Some operations switch from an exclusive to a shared lock as soon as
752 - Instance's primary and secondary nodes' groups are now available as
753 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
754 and ``snodes.group.uuid``).
759 - Numerous updates to documentation and manpages.
761 - :doc:`RAPI <rapi>` documentation now has detailed parameter
763 - Some opcode/job results are now also documented, see :doc:`RAPI
766 - A lockset's internal lock is now also visible in lock monitor.
767 - Log messages from job queue workers now contain information about the
768 opcode they're processing.
769 - ``gnt-instance console`` no longer requires the instance lock.
770 - A short delay when waiting for job changes reduces the number of LUXI
771 requests significantly.
772 - DRBD metadata volumes are overwritten with zeros during disk creation.
773 - Out-of-band commands no longer acquire the cluster lock in exclusive
775 - ``devel/upload`` now uses correct permissions for directories.
781 *(Released Fri, 23 Mar 2012)*
783 This was the sixth release candidate of the 2.5 series.
789 *(Released Mon, 9 Jan 2012)*
791 This was the fifth release candidate of the 2.5 series.
797 *(Released Thu, 27 Oct 2011)*
799 This was the fourth release candidate of the 2.5 series.
805 *(Released Wed, 26 Oct 2011)*
807 This was the third release candidate of the 2.5 series.
813 *(Released Tue, 18 Oct 2011)*
815 This was the second release candidate of the 2.5 series.
821 *(Released Tue, 4 Oct 2011)*
823 This was the first release candidate of the 2.5 series.
829 *(Released Wed, 31 Aug 2011)*
831 This was the third beta release of the 2.5 series.
837 *(Released Mon, 22 Aug 2011)*
839 This was the second beta release of the 2.5 series.
845 *(Released Fri, 12 Aug 2011)*
847 This was the first beta release of the 2.5 series.
853 *(Released Thu, 27 Oct 2011)*
855 - Fixed bug when parsing command line parameter values ending in
857 - Fixed assertion error after unclean master shutdown
858 - Disable HTTP client pool for RPC, significantly reducing memory usage
860 - Fixed queue archive creation with wrong permissions
866 *(Released Tue, 23 Aug 2011)*
870 - Fixed documentation for importing with ``--src-dir`` option
871 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
872 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
878 *(Released Fri, 5 Aug 2011)*
880 Many bug-fixes and a few small features:
882 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
883 issues when you tried to add an instance with two MAC addresses in one
885 - KVM: fixed per-instance stored UID value
886 - KVM: configure bridged NICs at migration start
887 - KVM: Fix a bug where instance will not start with never KVM versions
889 - Added OS search path to ``gnt-cluster info``
890 - Fixed an issue with ``file_storage_dir`` where you were forced to
891 provide an absolute path, but the documentation states it is a
892 relative path, the documentation was right
893 - Added a new parameter to instance stop/start called ``--no-remember``
894 that will make the state change to not be remembered
895 - Implemented ``no_remember`` at RAPI level
896 - Improved the documentation
897 - Node evacuation: don't call IAllocator if node is already empty
898 - Fixed bug in DRBD8 replace disks on current nodes
899 - Fixed bug in recreate-disks for DRBD instances
900 - Moved assertion checking locks in ``gnt-instance replace-disks``
901 causing it to abort with not owning the right locks for some situation
902 - Job queue: Fixed potential race condition when cancelling queued jobs
903 - Fixed off-by-one bug in job serial generation
904 - ``gnt-node volumes``: Fix instance names
905 - Fixed aliases in bash completion
906 - Fixed a bug in reopening log files after being sent a SIGHUP
907 - Added a flag to burnin to allow specifying VCPU count
908 - Bugfixes to non-root Ganeti configuration
914 *(Released Thu, 12 May 2011)*
916 Many bug-fixes and a few new small features:
918 - Fixed a bug related to log opening failures
919 - Fixed a bug in instance listing with orphan instances
920 - Fixed a bug which prevented resetting the cluster-level node parameter
921 ``oob_program`` to the default
922 - Many fixes related to the ``cluster-merge`` tool
923 - Fixed a race condition in the lock monitor, which caused failures
924 during (at least) creation of many instances in parallel
925 - Improved output for gnt-job info
926 - Removed the quiet flag on some ssh calls which prevented debugging
928 - Improved the N+1 failure messages in cluster verify by actually
929 showing the memory values (needed and available)
930 - Increased lock attempt timeouts so that when executing long operations
931 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
932 too early and thus prevent the use of the 'fair' mechanism
933 - Changed instance query data (``gnt-instance info``) to not acquire
934 locks unless needed, thus allowing its use on locked instance if only
935 static information is asked for
936 - Improved behaviour with filesystems that do not support rename on an
938 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
939 kept locks on all nodes during the wipe, which is unneeded
940 - Fixed ``gnt-watcher`` handling of errors during hooks execution
941 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
942 10GiB) which caused the wipe to fail right at the end in some cases
943 - Fixed master IP activation when doing master failover with no-voting
944 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
945 the master node itself
946 - Fixed potential data-loss in under disk full conditions, where Ganeti
947 wouldn't check correctly the return code and would consider
948 partially-written files 'correct'
949 - Fixed bug related to multiple VGs and DRBD disk replacing
950 - Added new disk parameter ``metavg`` that allows placement of the meta
951 device for DRBD in a different volume group
952 - Fixed error handling in the node daemon when the system libc doesn't
953 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
954 - Fixed lock release during replace-disks, which kept cluster-wide locks
955 when doing disk replaces with an iallocator script
956 - Added check for missing bridges in cluster verify
957 - Handle EPIPE errors while writing to the terminal better, so that
958 piping the output to e.g. ``less`` doesn't cause a backtrace
959 - Fixed rare case where a ^C during Luxi calls could have been
960 interpreted as server errors, instead of simply terminating
961 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
963 - Added a few more parameters to the KVM hypervisor, allowing a second
964 CDROM, custom disk type for CDROMs and a floppy image
965 - Removed redundant message in instance rename when the name is given
967 - Added option to ``gnt-instance recreate-disks`` to allow creating the
968 disks on new nodes, allowing recreation when the original instance
969 nodes are completely gone
970 - Added option when converting disk templates to DRBD to skip waiting
971 for the resync, in order to make the instance available sooner
972 - Added two new variables to the OS scripts environment (containing the
974 - Made the root_path and optional parameter for the xen-pvm hypervisor,
975 to allow use of ``pvgrub`` as bootloader
976 - Changed the instance memory modifications to only check out-of-memory
977 conditions on memory increases, and turned the secondary node warnings
978 into errors (they can still be overridden via ``--force``)
979 - Fixed the handling of a corner case when the Python installation gets
980 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
981 to execute a command that doesn't exist
982 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
983 node of the instance returned failures during instance shutdown; this
984 adds the option ``--ignore-consistency`` to gnt-instance move
986 And as usual, various improvements to the error messages, documentation
993 *(Released Wed, 09 Mar 2011)*
995 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
996 the RAPI users file if run twice (even with ``--dry-run``).
998 The release fixes that bug (nothing else changed).
1004 *(Released Mon, 07 Mar 2011)*
1006 Final 2.4.0 release. Just a few small fixes:
1008 - Fixed RAPI node evacuate
1009 - Fixed the kvm-ifup script
1010 - Fixed internal error handling for special job cases
1011 - Updated man page to specify the escaping feature for options
1017 *(Released Mon, 28 Feb 2011)*
1019 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1020 that this feature wiped the disks of the wrong instance, leading to loss
1025 - Fixed title of query field containing instance name
1026 - Expanded the glossary in the documentation
1027 - Fixed one unittest (internal issue)
1033 *(Released Mon, 21 Feb 2011)*
1035 A number of bug fixes plus just a couple functionality changes.
1037 On the user-visible side, the ``gnt-* list`` command output has changed
1038 with respect to "special" field states. The current rc1 style of display
1039 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1040 the default output mode special fields states are displayed as follows:
1042 - Offline resource: ``*``
1043 - Unavailable/not applicable: ``-``
1044 - Data missing (RPC failure): ``?``
1045 - Unknown field: ``??``
1047 Another user-visible change is the addition of ``--force-join`` to
1052 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1053 - Fixed regression in RAPI/instance reinstall where all parameters were
1054 required (instead of optional)
1055 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1056 - Fixed iallocator usage (offline nodes were not considered offline)
1057 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1058 - Fixed hypervisor and OS parameter validation with respect to
1059 non-vm_capable nodes
1060 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1062 - Fixed ``tools/listrunner`` with respect to agent-based usage
1068 *(Released Fri, 4 Feb 2011)*
1070 Many changes and fixes since the beta1 release. While there were some
1071 internal changes, the code has been mostly stabilised for the RC
1074 Note: the dumb allocator was removed in this release, as it was not kept
1075 up-to-date with the IAllocator protocol changes. It is recommended to
1076 use the ``hail`` command from the ganeti-htools package.
1078 Note: the 2.4 and up versions of Ganeti are not compatible with the
1079 0.2.x branch of ganeti-htools. You need to upgrade to
1080 ganeti-htools-0.3.0 (or later).
1082 Regressions fixed from 2.3
1083 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1085 - Fixed the ``gnt-cluster verify-disks`` command
1086 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1088 - Fixed disk adoption breakage
1089 - Fixed wrong headers in instance listing for field aliases
1094 - Fixed corner case in KVM handling of NICs
1095 - Fixed many cases of wrong handling of non-vm_capable nodes
1096 - Fixed a bug where a missing instance symlink was not possible to
1097 recreate with any ``gnt-*`` command (now ``gnt-instance
1098 activate-disks`` does it)
1099 - Fixed the volume group name as reported by ``gnt-cluster
1101 - Increased timeouts for the import-export code, hopefully leading to
1102 fewer aborts due network or instance timeouts
1103 - Fixed bug in ``gnt-node list-storage``
1104 - Fixed bug where not all daemons were started on cluster
1105 initialisation, but only at the first watcher run
1106 - Fixed many bugs in the OOB implementation
1107 - Fixed watcher behaviour in presence of instances with offline
1109 - Fixed instance list output for instances running on the wrong node
1110 - a few fixes to the cluster-merge tool, but it still cannot merge
1111 multi-node groups (currently it is not recommended to use this tool)
1117 - Improved network configuration for the KVM hypervisor
1118 - Added e1000 as a supported NIC for Xen-HVM
1119 - Improved the lvmstrap tool to also be able to use partitions, as
1120 opposed to full disks
1121 - Improved speed of disk wiping (the cluster parameter
1122 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1123 of instance creations
1124 - Added documentation for the OS parameters
1125 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1126 hypervisor is not responding
1127 - Added display of blacklisted and hidden OS information in
1128 ``gnt-cluster info``
1129 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1130 NIC and node parameters, which might create problems with currently
1131 invalid (but undetected) configuration files, but prevents validation
1132 failures when unrelated parameters are modified
1133 - Changed cluster initialisation to wait for the master daemon to become
1135 - Expanded the RAPI interface:
1137 - Added config redistribution resource
1138 - Added activation/deactivation of instance disks
1139 - Added export of console information
1141 - Implemented log file reopening on SIGHUP, which allows using
1142 logrotate(8) for the Ganeti log files
1143 - Added a basic OOB helper script as an example
1149 *(Released Fri, 14 Jan 2011)*
1154 - Fixed timezone issues when formatting timestamps
1155 - Added support for node groups, available via ``gnt-group`` and other
1157 - Added out-of-band framework and management, see :doc:`design
1158 document <design-oob>`
1159 - Removed support for roman numbers from ``gnt-node list`` and
1160 ``gnt-instance list``.
1161 - Allowed modification of master network interface via ``gnt-cluster
1162 modify --master-netdev``
1163 - Accept offline secondaries while shutting down instance disks
1164 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1165 - Added support for multiple LVM volume groups
1166 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1168 - Added commands to list available fields:
1170 - ``gnt-node list-fields``
1171 - ``gnt-group list-fields``
1172 - ``gnt-instance list-fields``
1174 - Updated documentation and man pages
1179 - Moved ``rapi_users`` file into separate directory, now named
1180 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1182 - Added new tool for running commands on many machines,
1183 ``tools/ganeti-listrunner``
1184 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1185 improving the ``gnt-instance console`` output
1186 - Allowed customisation of disk index separator at ``configure`` time
1187 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1188 - Added support for non-partitioned md disks in ``lvmstrap``
1189 - Added script to gracefully power off KVM instances
1190 - Split ``utils`` module into smaller parts
1191 - Changed query operations to return more detailed information, e.g.
1192 whether an information is unavailable due to an offline node. To use
1193 this new functionality, the LUXI call ``Query`` must be used. Field
1194 information is now stored by the master daemon and can be retrieved
1195 using ``QueryFields``. Instances, nodes and groups can also be queried
1196 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1197 exposed via RAPI). The following commands make use of this
1198 infrastructure change:
1200 - ``gnt-group list``
1201 - ``gnt-group list-fields``
1203 - ``gnt-node list-fields``
1204 - ``gnt-instance list``
1205 - ``gnt-instance list-fields``
1206 - ``gnt-debug locks``
1211 - New RAPI resources (see :doc:`rapi`):
1215 - ``/2/groups/[group_name]``
1216 - ``/2/groups/[group_name]/assign-nodes``
1217 - ``/2/groups/[group_name]/modify``
1218 - ``/2/groups/[group_name]/rename``
1219 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1223 - Implemented ``no_install`` for instance creation
1224 - Implemented OS parameters for instance reinstallation, allowing
1225 use of special settings on reinstallation (e.g. for preserving data)
1230 - Added IPv6 support in import/export
1231 - Pause DRBD synchronization while wiping disks on instance creation
1232 - Updated unittests and QA scripts
1233 - Improved network parameters passed to KVM
1234 - Converted man pages from docbook to reStructuredText
1240 *(Released Mon, 20 Dec 2010)*
1242 Released version 2.3.1~rc1 without any changes.
1248 *(Released Wed, 1 Dec 2010)*
1250 - impexpd: Disable OpenSSL compression in socat if possible (backport
1251 from master, commit e90739d625b, see :doc:`installation guide
1252 <install-quick>` for details)
1253 - Changed unittest coverage report to exclude test scripts
1254 - Added script to check version format
1260 *(Released Wed, 1 Dec 2010)*
1262 Released version 2.3.0~rc1 without any changes.
1268 *(Released Fri, 19 Nov 2010)*
1270 A number of bugfixes and documentation updates:
1272 - Update ganeti-os-interface documentation
1273 - Fixed a bug related to duplicate MACs or similar items which should be
1275 - Fix breakage in OS state modify
1276 - Reinstall instance: disallow offline secondaries (fixes bug related to
1277 OS changing but reinstall failing)
1278 - plus all the other fixes between 2.2.1 and 2.2.2
1284 *(Released Tue, 2 Nov 2010)*
1286 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1287 - Fixed master failover race with watcher
1288 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1290 - Accept previously stopped instance for export with instance removal
1291 - Simplify and extend the environment variables for instance OS scripts
1292 - Added new node flags, ``master_capable`` and ``vm_capable``
1293 - Added optional instance disk wiping prior during allocation. This is a
1294 cluster-wide option and can be set/modified using
1295 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1296 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1297 :doc:`install-quick`
1298 - Added a new watcher option (``--ignore-pause``)
1299 - Added option to ignore offline node on instance start/stop
1300 (``--ignore-offline``)
1301 - Allow overriding OS parameters with ``gnt-instance reinstall``
1302 - Added ability to change node's secondary IP address using ``gnt-node
1304 - Implemented privilege separation for all daemons except
1305 ``ganeti-noded``, see ``configure`` options
1306 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1308 - Implemented job priorities (see ``ganeti(7)`` manpage)
1309 - Ignore failures while shutting down instances during failover from
1311 - Exit daemon's bootstrap process only once daemon is ready
1312 - Export more information via ``LUInstanceQuery``/remote API
1313 - Improved documentation, QA and unittests
1314 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1315 restart if the file was created or changed
1316 - Added LUXI protocol version sent with each request and response,
1317 allowing detection of server/client mismatches
1318 - Moved the Python scripts among gnt-* and ganeti-* into modules
1319 - Moved all code related to setting up SSH to an external script,
1321 - Infrastructure changes for node group support in future versions
1327 *(Released Fri, 19 Nov 2010)*
1329 A few small bugs fixed, and some improvements to the build system:
1331 - Fix documentation regarding conversion to drbd
1332 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1334 - Fix error handling in node modify with multiple changes
1335 - Allow remote imports without checked names
1341 *(Released Tue, 19 Oct 2010)*
1343 - Disable SSL session ID cache in RPC client
1349 *(Released Thu, 14 Oct 2010)*
1351 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1352 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1354 - Fix problems with interaction between Curl and Python's HTTP server,
1355 resulting in increased speed in many RPC calls
1356 - Improve our release script to prevent breakage with older aclocal and
1363 *(Released Thu, 7 Oct 2010)*
1365 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1366 value retrieved from master
1367 - Added support for blacklisted or hidden OS definitions
1368 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1369 - Added support for -mem-path in KVM hypervisor abstraction layer
1370 - Allow overriding instance parameters in tool for inter-cluster
1371 instance moves (``tools/move-instance``)
1372 - Improved opcode summaries (e.g. in ``gnt-job list``)
1373 - Improve consistency of OS listing by sorting it
1374 - Documentation updates
1380 *(Released Fri, 8 Oct 2010)*
1382 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1388 *(Released Mon, 4 Oct 2010)*
1390 - Fixed regression in ``gnt-instance rename``
1396 *(Released Wed, 22 Sep 2010)*
1398 - Fixed OS_VARIANT variable for OS scripts
1399 - Fixed cluster tag operations via RAPI
1400 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1401 - Disabled RAPI CA checks in watcher
1407 *(Released Mon, 23 Aug 2010)*
1409 - Support DRBD versions of the format "a.b.c.d"
1411 - Re-introduce support for usage from multiple threads in RAPI client
1412 - Instance renames and modify via RAPI
1413 - Work around race condition between processing and archival in job
1415 - Mark opcodes following failed one as failed, too
1416 - Job field ``lock_status`` was removed due to difficulties making it
1417 work with the changed job queue in Ganeti 2.2; a better way to monitor
1418 locks is expected for a later 2.2.x release
1419 - Fixed dry-run behaviour with many commands
1420 - Support ``ssh-agent`` again when adding nodes
1421 - Many additional bugfixes
1427 *(Released Fri, 30 Jul 2010)*
1429 Important change: the internal RPC mechanism between Ganeti nodes has
1430 changed from using a home-grown http library (based on the Python base
1431 libraries) to use the PycURL library. This requires that PycURL is
1432 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1433 against GnuTLS by default. cURL's support for GnuTLS had known issues
1434 before cURL 7.21.0 and we recommend using the latest cURL release or
1435 linking against OpenSSL. Most other distributions already link PycURL
1436 and cURL against OpenSSL. The command::
1438 python -c 'import pycurl; print pycurl.version'
1440 can be used to determine the libraries PycURL and cURL are linked
1443 Other significant changes:
1445 - Rewrote much of the internals of the job queue, in order to achieve
1446 better parallelism; this decouples job query operations from the job
1447 processing, and it should allow much nicer behaviour of the master
1448 daemon under load, and it also has uncovered some long-standing bugs
1449 related to the job serialisation (now fixed)
1450 - Added a default iallocator setting to the cluster parameters,
1451 eliminating the need to always pass nodes or an iallocator for
1452 operations that require selection of new node(s)
1453 - Added experimental support for the LXC virtualization method
1454 - Added support for OS parameters, which allows the installation of
1455 instances to pass parameter to OS scripts in order to customise the
1457 - Added a hypervisor parameter controlling the migration type (live or
1458 non-live), since hypervisors have various levels of reliability; this
1459 has renamed the 'live' parameter to 'mode'
1460 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1461 logical volumes, meaning that cluster verify will ignore them and not
1462 flag their presence as errors
1463 - The watcher will now reset the error count for failed instances after
1464 8 hours, thus allowing self-healing if the problem that caused the
1465 instances to be down/fail to start has cleared in the meantime
1466 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1467 check for, and warn, if the drbd module parameter ``usermode_helper``
1468 is not consistent with the cluster-wide setting; this is needed to
1469 make diagnose easier of failed drbd creations
1470 - Started adding base IPv6 support, but this is not yet
1471 enabled/available for use
1472 - Rename operations (cluster, instance) will now return the new name,
1473 which is especially useful if a short name was passed in
1474 - Added support for instance migration in RAPI
1475 - Added a tool to pre-configure nodes for the SSH setup, before joining
1476 them to the cluster; this will allow in the future a simplified model
1477 for node joining (but not yet fully enabled in 2.2); this needs the
1478 paramiko python library
1479 - Fixed handling of name-resolving errors
1480 - Fixed consistency of job results on the error path
1481 - Fixed master-failover race condition when executed multiple times in
1483 - Fixed many bugs related to the job queue (mostly introduced during the
1484 2.2 development cycle, so not all are impacting 2.1)
1485 - Fixed instance migration with missing disk symlinks
1486 - Fixed handling of unknown jobs in ``gnt-job archive``
1487 - And many other small fixes/improvements
1491 - Enhanced both the unittest and the QA coverage
1492 - Switched the opcode validation to a generic model, and extended the
1493 validation to all opcode parameters
1494 - Changed more parts of the code that write shell scripts to use the
1496 - Switched the master daemon to use the asyncore library for the Luxi
1503 *(Released Thu, 17 Jun 2010)*
1505 - Added tool (``move-instance``) and infrastructure to move instances
1506 between separate clusters (see :doc:`separate documentation
1507 <move-instance>` and :doc:`design document <design-2.2>`)
1508 - Added per-request RPC timeout
1509 - RAPI now requires a Content-Type header for requests with a body (e.g.
1510 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1511 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1512 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1514 - Implemented initial support for running Ganeti daemons as separate
1515 users, see configure-time flags ``--with-user-prefix`` and
1516 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1518 - Instances can be removed after export (``gnt-backup export
1519 --remove-instance``)
1520 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1521 key (instead of 1024 bit)
1522 - Added new cluster configuration file for cluster domain secret
1523 - Import/export now use SSL instead of SSH
1524 - Added support for showing estimated time when exporting an instance,
1525 see the ``ganeti-os-interface(7)`` manpage and look for
1532 *(Released Tue, 16 Nov 2010)*
1534 Some more bugfixes. Unless critical bugs occur, this will be the last
1537 - Fix case of MAC special-values
1538 - Fix mac checker regex
1539 - backend: Fix typo causing "out of range" error
1540 - Add missing --units in gnt-instance list man page
1546 *(Released Tue, 24 Aug 2010)*
1549 - Don't ignore secondary node silently on non-mirrored disk templates
1551 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1552 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1553 - Properly document the usb_mouse parameter
1554 - Fix path in ganeti-rapi(8) (issue 116)
1555 - Adjust error message when the ganeti user's .ssh directory is
1557 - Add same-node-check when changing the disk template to drbd
1563 *(Released Fri, 16 Jul 2010)*
1566 - Add an option to only select some reboot types during qa/burnin.
1567 (on some hypervisors consequent reboots are not supported)
1568 - Fix infrequent race condition in master failover. Sometimes the old
1569 master ip address would be still detected as up for a short time
1570 after it was removed, causing failover to fail.
1571 - Decrease mlockall warnings when the ctypes module is missing. On
1572 Python 2.4 we support running even if no ctypes module is installed,
1573 but we were too verbose about this issue.
1574 - Fix building on old distributions, on which man doesn't have a
1576 - Fix RAPI not to ignore the MAC address on instance creation
1577 - Implement the old instance creation format in the RAPI client.
1583 *(Released Thu, 01 Jul 2010)*
1585 A small bugfix release:
1586 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1587 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1588 a non-existing option
1589 - Fix instance create: the --force-variant option was ignored
1590 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1591 - Fix modify node storage with non-FQDN arguments
1592 - Fix RAPI client to authenticate under Python 2.6 when used
1593 for more than 5 requests needing authentication
1594 - Fix gnt-instance modify -t (storage) giving a wrong error message
1595 when converting a non-shutdown drbd instance to plain
1601 *(Released Fri, 18 Jun 2010)*
1603 A small bugfix release:
1605 - Fix live migration of KVM instances started with older Ganeti
1606 versions which had fewer hypervisor parameters
1607 - Fix gnt-instance grow-disk on down instances
1608 - Fix an error-reporting bug during instance migration
1609 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1610 silently ignoring broken ones
1611 - Fix an RPC error reporting bug affecting, for example, RAPI client
1613 - Fix bug triggered by different API version os-es on different nodes
1614 - Fix a bug in instance startup with custom hvparams: OS level
1615 parameters would fail to be applied.
1616 - Fix the RAPI client under Python 2.6 (but more work is needed to
1617 make it work completely well with OpenSSL)
1618 - Fix handling of errors when resolving names from DNS
1624 *(Released Thu, 3 Jun 2010)*
1626 A medium sized development cycle. Some new features, and some
1627 fixes/small improvements/cleanups.
1629 Significant features
1630 ~~~~~~~~~~~~~~~~~~~~
1632 The node deamon now tries to mlock itself into memory, unless the
1633 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1634 its logs, and falls back to console logging. This allows emergency
1635 features such as ``gnt-node powercycle`` to work even in the event of a
1636 broken node disk (tested offlining the disk hosting the node's
1637 filesystem and dropping its memory caches; don't try this at home)
1639 KVM: add vhost-net acceleration support. It can be tested with a new
1640 enough version of the kernel and of qemu-kvm.
1642 KVM: Add instance chrooting feature. If you use privilege dropping for
1643 your VMs you can also now force them to chroot to an empty directory,
1644 before starting the emulated guest.
1646 KVM: Add maximum migration bandwith and maximum downtime tweaking
1647 support (requires a new-enough version of qemu-kvm).
1649 Cluster verify will now warn if the master node doesn't have the master
1650 ip configured on it.
1652 Add a new (incompatible) instance creation request format to RAPI which
1653 supports all parameters (previously only a subset was supported, and it
1654 wasn't possible to extend the old format to accomodate all the new
1655 features. The old format is still supported, and a client can check for
1656 this feature, before using it, by checking for its presence in the
1657 ``features`` RAPI resource.
1659 Now with ancient latin support. Try it passing the ``--roman`` option to
1660 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1661 (requires the python-roman module to be installed, in order to work).
1666 As usual many internal code refactorings, documentation updates, and
1669 - Lots of improvements and cleanups to the experimental Remote API
1670 (RAPI) client library.
1671 - A new unit test suite for the core daemon libraries.
1672 - A fix to creating missing directories makes sure the umask is not
1673 applied anymore. This enforces the same directory permissions
1675 - Better handling terminating daemons with ctrl+c (used when running
1676 them in debugging mode).
1677 - Fix a race condition in live migrating a KVM instance, when stat()
1678 on the old proc status file returned EINVAL, which is an unexpected
1680 - Fixed manpage checking with newer man and utf-8 charachters. But now
1681 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1687 *(Released Fri, 7 May 2010)*
1689 Fix a bug which prevented untagged KVM instances from starting.
1695 *(Released Fri, 7 May 2010)*
1697 Another release with a long development cycle, during which many
1698 different features were added.
1700 Significant features
1701 ~~~~~~~~~~~~~~~~~~~~
1703 The KVM hypervisor now can run the individual instances as non-root, to
1704 reduce the impact of a VM being hijacked due to bugs in the
1705 hypervisor. It is possible to run all instances as a single (non-root)
1706 user, to manually specify a user for each instance, or to dynamically
1707 allocate a user out of a cluster-wide pool to each instance, with the
1708 guarantee that no two instances will run under the same user ID on any
1711 An experimental RAPI client library, that can be used standalone
1712 (without the other Ganeti libraries), is provided in the source tree as
1713 ``lib/rapi/client.py``. Note this client might change its interface in
1714 the future, as we iterate on its capabilities.
1716 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1717 replace the cluster's certificates and crypto keys. This might help in
1718 case they have been compromised, or have simply expired.
1720 A new disk option for instance creation has been added that allows one
1721 to "adopt" currently existing logical volumes, with data
1722 preservation. This should allow easier migration to Ganeti from
1723 unmanaged (or managed via other software) instances.
1725 Another disk improvement is the possibility to convert between redundant
1726 (DRBD) and plain (LVM) disk configuration for an instance. This should
1727 allow better scalability (starting with one node and growing the
1728 cluster, or shrinking a two-node cluster to one node).
1730 A new feature that could help with automated node failovers has been
1731 implemented: if a node sees itself as offline (by querying the master
1732 candidates), it will try to shutdown (hard) all instances and any active
1733 DRBD devices. This reduces the risk of duplicate instances if an
1734 external script automatically failovers the instances on such nodes. To
1735 enable this, the cluster parameter ``maintain_node_health`` should be
1736 enabled; in the future this option (per the name) will enable other
1737 automatic maintenance features.
1739 Instance export/import now will reuse the original instance
1740 specifications for all parameters; that means exporting an instance,
1741 deleting it and the importing it back should give an almost identical
1742 instance. Note that the default import behaviour has changed from
1743 before, where it created only one NIC; now it recreates the original
1746 Cluster verify has added a few new checks: SSL certificates validity,
1747 /etc/hosts consistency across the cluster, etc.
1752 As usual, many internal changes were done, documentation fixes,
1755 - Fixed cluster initialization with disabled cluster storage (regression
1756 introduced in 2.1.1)
1757 - File-based storage supports growing the disks
1758 - Fixed behaviour of node role changes
1759 - Fixed cluster verify for some corner cases, plus a general rewrite of
1760 cluster verify to allow future extension with more checks
1761 - Fixed log spamming by watcher and node daemon (regression introduced
1763 - Fixed possible validation issues when changing the list of enabled
1765 - Fixed cleanup of /etc/hosts during node removal
1766 - Fixed RAPI response for invalid methods
1767 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1768 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1769 from ide disks, etc.)
1770 - Allow OS changes without re-installation (to record a changed OS
1771 outside of Ganeti, or to allow OS renames)
1772 - Allow instance creation without OS installation (useful for example if
1773 the OS will be installed manually, or restored from a backup not in
1775 - Implemented option to make cluster ``copyfile`` use the replication
1777 - Added list of enabled hypervisors to ssconf (possibly useful for
1779 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1781 - A partial form of node re-IP is possible via node readd, which now
1782 allows changed node primary IP
1783 - Command line utilities now show an informational message if the job is
1785 - The logs of the master daemon now show the PID/UID/GID of the
1792 *(Released Fri, 12 Mar 2010)*
1794 During the 2.1.0 long release candidate cycle, a lot of improvements and
1795 changes have accumulated with were released later as 2.1.1.
1800 The node evacuate command (``gnt-node evacuate``) was significantly
1801 rewritten, and as such the IAllocator protocol was changed - a new
1802 request type has been added. This unfortunate change during a stable
1803 series is designed to improve performance of node evacuations; on
1804 clusters with more than about five nodes and which are well-balanced,
1805 evacuation should proceed in parallel for all instances of the node
1806 being evacuated. As such, any existing IAllocator scripts need to be
1807 updated, otherwise the above command will fail due to the unknown
1808 request. The provided "dumb" allocator has not been updated; but the
1809 ganeti-htools package supports the new protocol since version 0.2.4.
1811 Another important change is increased validation of node and instance
1812 names. This might create problems in special cases, if invalid host
1813 names are being used.
1815 Also, a new layer of hypervisor parameters has been added, that sits at
1816 OS level between the cluster defaults and the instance ones. This allows
1817 customisation of virtualization parameters depending on the installed
1818 OS. For example instances with OS 'X' may have a different KVM kernel
1819 (or any other parameter) than the cluster defaults. This is intended to
1820 help managing a multiple OSes on the same cluster, without manual
1821 modification of each instance's parameters.
1823 A tool for merging clusters, ``cluster-merge``, has been added in the
1824 tools sub-directory.
1829 - Improved the int/float conversions that should make the code more
1830 robust in face of errors from the node daemons
1831 - Fixed the remove node code in case of internal configuration errors
1832 - Fixed the node daemon behaviour in face of inconsistent queue
1833 directory (e.g. read-only file-system where we can't open the files
1835 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1836 now it either aborts cleanly or, if given the new "auto_promote"
1837 parameter, will automatically promote other nodes as needed
1838 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1839 completely prevent Ganeti from working
1840 - Fixed bug for instance export when not all disks were successfully
1842 - Fixed behaviour of node add when the new node is slow in starting up
1844 - Fixed handling of signals in the LUXI client, which should improve
1845 behaviour of command-line scripts
1846 - Added checks for invalid node/instance names in the configuration (now
1847 flagged during cluster verify)
1848 - Fixed watcher behaviour for disk activation errors
1849 - Fixed two potentially endless loops in http library, which led to the
1850 RAPI daemon hanging and consuming 100% CPU in some cases
1851 - Fixed bug in RAPI daemon related to hashed passwords
1852 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1854 - Enhanced compatibility with non-Debian OSes, but not using absolute
1855 path in some commands and allowing customisation of the ssh
1856 configuration directory
1857 - Fixed possible future issue with new Python versions by abiding to the
1858 proper use of ``__slots__`` attribute on classes
1859 - Added checks that should prevent directory traversal attacks
1860 - Many documentation fixes based on feedback from users
1865 - Added an "early_release" more for instance replace disks and node
1866 evacuate, where we release locks earlier and thus allow higher
1867 parallelism within the cluster
1868 - Added watcher hooks, intended to allow the watcher to restart other
1869 daemons (e.g. from the ganeti-nbma project), but they can be used of
1870 course for any other purpose
1871 - Added a compile-time disable for DRBD barriers, to increase
1872 performance if the administrator trusts the power supply or the
1873 storage system to not lose writes
1874 - Added the option of using syslog for logging instead of, or in
1875 addition to, Ganeti's own log files
1876 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1877 can indeed boot from a paravirtual NIC
1878 - Added a generic debug level for many operations; while this is not
1879 used widely yet, it allows one to pass the debug value all the way to
1881 - Enhanced the hooks environment for instance moves (failovers,
1882 migrations) where the primary/secondary nodes changed during the
1883 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1884 - Enhanced data validations for many user-supplied values; one important
1885 item is the restrictions imposed on instance and node names, which
1886 might reject some (invalid) host names
1887 - Add a configure-time option to disable file-based storage, if it's not
1888 needed; this allows greater security separation between the master
1889 node and the other nodes from the point of view of the inter-node RPC
1891 - Added user notification in interactive tools if job is waiting in the
1892 job queue or trying to acquire locks
1893 - Added log messages when a job is waiting for locks
1894 - Added filtering by node tags in instance operations which admit
1895 multiple instances (start, stop, reboot, reinstall)
1896 - Added a new tool for cluster mergers, ``cluster-merge``
1897 - Parameters from command line which are of the form ``a=b,c=d`` can now
1898 use backslash escapes to pass in values which contain commas,
1899 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1901 - For KVM, the instance name is the first parameter passed to KVM, so
1902 that it's more visible in the process list
1908 *(Released Tue, 2 Mar 2010)*
1910 Ganeti 2.1 brings many improvements with it. Major changes:
1912 - Added infrastructure to ease automated disk repairs
1913 - Added new daemon to export configuration data in a cheaper way than
1914 using the remote API
1915 - Instance NICs can now be routed instead of being associated with a
1917 - Improved job locking logic to reduce impact of jobs acquiring multiple
1918 locks waiting for other long-running jobs
1920 In-depth implementation details can be found in the Ganeti 2.1 design
1926 - Added chroot hypervisor
1927 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1929 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1930 ``bootloader_path`` and ``bootloader_args``)
1931 - Added the ``use_localtime`` option for the xen-hvm and kvm
1932 hypervisors, and the default value for this has changed to false (in
1933 2.0 xen-hvm always enabled it)
1934 - Added luxi call to submit multiple jobs in one go
1935 - Added cluster initialization option to not modify ``/etc/hosts``
1937 - Added network interface parameters
1938 - Added dry run mode to some LUs
1939 - Added RAPI resources:
1941 - ``/2/instances/[instance_name]/info``
1942 - ``/2/instances/[instance_name]/replace-disks``
1943 - ``/2/nodes/[node_name]/evacuate``
1944 - ``/2/nodes/[node_name]/migrate``
1945 - ``/2/nodes/[node_name]/role``
1946 - ``/2/nodes/[node_name]/storage``
1947 - ``/2/nodes/[node_name]/storage/modify``
1948 - ``/2/nodes/[node_name]/storage/repair``
1950 - Added OpCodes to evacuate or migrate all instances on a node
1951 - Added new command to list storage elements on nodes (``gnt-node
1952 list-storage``) and modify them (``gnt-node modify-storage``)
1953 - Added new ssconf files with master candidate IP address
1954 (``ssconf_master_candidates_ips``), node primary IP address
1955 (``ssconf_node_primary_ips``) and node secondary IP address
1956 (``ssconf_node_secondary_ips``)
1957 - Added ``ganeti-confd`` and a client library to query the Ganeti
1958 configuration via UDP
1959 - Added ability to run hooks after cluster initialization and before
1961 - Added automatic mode for disk replace (``gnt-instance replace-disks
1963 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1964 after catastrophic data-loss
1965 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1967 - Added ``gnt-instance move`` command to move instances
1968 - Added ``gnt-cluster watcher`` command to control watcher
1969 - Added ``gnt-node powercycle`` command to powercycle nodes
1970 - Added new job status field ``lock_status``
1971 - Added parseable error codes to cluster verification (``gnt-cluster
1972 verify --error-codes``) and made output less verbose (use
1973 ``--verbose`` to restore previous behaviour)
1974 - Added UUIDs to the main config entities (cluster, nodes, instances)
1975 - Added support for OS variants
1976 - Added support for hashed passwords in the Ganeti remote API users file
1978 - Added option to specify maximum timeout on instance shutdown
1979 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1980 - Added new helper script to start and stop Ganeti daemons
1981 (``daemon-util``), with the intent to reduce the work necessary to
1982 adjust Ganeti for non-Debian distributions and to start/stop daemons
1984 - Added more unittests
1985 - Fixed critical bug in ganeti-masterd startup
1986 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1987 now customisable at the cluster level for both the KVM and Xen
1988 hypervisors using the new ``migration_port`` parameter
1989 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1990 reinstalling an instance
1991 - Allowed ``@`` in tag names
1992 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1993 - Many documentation updates
1994 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1995 - ``gnt-instance reinstall`` can now reinstall multiple instances
1996 - Updated many command line parameters
1997 - Introduced new OS API version 15
1998 - No longer support a default hypervisor
1999 - Treat virtual LVs as inexistent
2000 - Improved job locking logic to reduce lock contention
2001 - Match instance and node names case insensitively
2002 - Reimplemented bash completion script to be more complete
2009 *(Released Thu, 4 Feb 2010)*
2011 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2012 - Fix a string formatting bug
2013 - Improve safety of the code in some error paths
2014 - Improve data validation in the master of values returned from nodes
2020 *(Released Thu, 17 Dec 2009)*
2022 - Fix security issue due to missing validation of iallocator names; this
2023 allows local and remote execution of arbitrary executables
2024 - Fix failure of gnt-node list during instance removal
2025 - Ship the RAPI documentation in the archive
2031 *(Released Wed, 30 Sep 2009)*
2033 - Fixed many wrong messages
2034 - Fixed a few bugs related to the locking library
2035 - Fixed MAC checking at instance creation time
2036 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2037 - Fixed a few issues related to signal handling in both daemons and
2039 - Fixed the example startup script provided
2040 - Fixed insserv dependencies in the example startup script (patch from
2042 - Fixed handling of drained nodes in the iallocator framework
2043 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2045 - Fixed error related to invalid job IDs in job polling
2046 - Fixed job/opcode persistence on unclean master shutdown
2047 - Fixed handling of partial job processing after unclean master
2049 - Fixed error reporting from LUs, previously all errors were converted
2050 into execution errors
2051 - Fixed error reporting from burnin
2052 - Decreased significantly the memory usage of the job queue
2053 - Optimised slightly multi-job submission
2054 - Optimised slightly opcode loading
2055 - Backported the multi-job submit framework from the development
2056 branch; multi-instance start and stop should be faster
2057 - Added script to clean archived jobs after 21 days; this will reduce
2058 the size of the queue directory
2059 - Added some extra checks in disk size tracking
2060 - Added an example ethers hook script
2061 - Added a cluster parameter that prevents Ganeti from modifying of
2063 - Added more node information to RAPI responses
2064 - Added a ``gnt-job watch`` command that allows following the ouput of a
2066 - Added a bind-address option to ganeti-rapi
2067 - Added more checks to the configuration verify
2068 - Enhanced the burnin script such that some operations can be retried
2070 - Converted instance reinstall to multi-instance model
2076 *(Released Fri, 7 Aug 2009)*
2078 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2079 to allow using the pre-2.0.2 behaviour in activation, if any existing
2080 instances have mismatched disk sizes in the configuration
2081 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2082 any configuration mismatches for disk sizes
2083 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2084 failover to work on two-node clusters
2085 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2087 - Fixed detection of OS script errors in ``gnt-backup export``
2088 - Fixed exit code of ``gnt-backup export``
2094 *(Released Fri, 17 Jul 2009)*
2096 - Added experimental support for stripped logical volumes; this should
2097 enhance performance but comes with a higher complexity in the block
2098 device handling; stripping is only enabled when passing
2099 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2100 affected even in the non-stripped mode
2101 - Improved resiliency against transient failures at the end of DRBD
2102 resyncs, and in general of DRBD resync checks
2103 - Fixed a couple of issues with exports and snapshot errors
2104 - Fixed a couple of issues in instance listing
2105 - Added display of the disk size in ``gnt-instance info``
2106 - Fixed checking for valid OSes in instance creation
2107 - Fixed handling of the "vcpus" parameter in instance listing and in
2108 general of invalid parameters
2109 - Fixed http server library, and thus RAPI, to handle invalid
2110 username/password combinations correctly; this means that now they
2111 report unauthorized for queries too, not only for modifications,
2112 allowing earlier detect of configuration problems
2113 - Added a new "role" node list field, equivalent to the master/master
2114 candidate/drained/offline flags combinations
2115 - Fixed cluster modify and changes of candidate pool size
2116 - Fixed cluster verify error messages for wrong files on regular nodes
2117 - Fixed a couple of issues with node demotion from master candidate role
2118 - Fixed node readd issues
2119 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2120 - Added a new ``--no-voting`` option for masterfailover to fix failover
2121 on two-nodes clusters when the former master node is unreachable
2122 - Added instance reinstall over RAPI
2128 *(Released Tue, 16 Jun 2009)*
2130 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2131 allow re-adding the start in single-user option (regression from 1.2)
2132 - the watcher writes the instance status to a file, to allow monitoring
2133 to report the instance status (from the master) based on cached
2134 results of the watcher's queries; while this can get stale if the
2135 watcher is being locked due to other work on the cluster, this is
2136 still an improvement
2137 - the watcher now also restarts the node daemon and the rapi daemon if
2139 - fixed the watcher to handle full and drained queue cases
2140 - hooks export more instance data in the environment, which helps if
2141 hook scripts need to take action based on the instance's properties
2142 (no longer need to query back into ganeti)
2143 - instance failovers when the instance is stopped do not check for free
2144 RAM, so that failing over a stopped instance is possible in low memory
2146 - rapi uses queries for tags instead of jobs (for less job traffic), and
2147 for cluster tags it won't talk to masterd at all but read them from
2149 - a couple of error handling fixes in RAPI
2150 - drbd handling: improved the error handling of inconsistent disks after
2151 resync to reduce the frequency of "there are some degraded disks for
2152 this instance" messages
2153 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2154 error handling path called a wrong function name)
2160 *(Released Wed, 27 May 2009)*
2162 - no changes from rc5
2168 *(Released Wed, 20 May 2009)*
2170 - fix a couple of bugs (validation, argument checks)
2171 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2172 - some small improvements to RAPI and IAllocator
2173 - make watcher automatically start the master daemon if down
2179 *(Released Mon, 27 Apr 2009)*
2181 - change the OS list to not require locks; this helps with big clusters
2182 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2183 volume group is broken
2184 - ``gnt-instance info``, without any arguments, doesn't run for all
2185 instances anymore; either pass ``--all`` or pass the desired
2186 instances; this helps against mistakes on big clusters where listing
2187 the information for all instances takes a long time
2188 - miscellaneous doc and man pages fixes
2194 *(Released Wed, 8 Apr 2009)*
2196 - Change the internal locking model of some ``gnt-node`` commands, in
2197 order to reduce contention (and blocking of master daemon) when
2198 batching many creation/reinstall jobs
2199 - Fixes to Xen soft reboot
2200 - No longer build documentation at build time, instead distribute it in
2201 the archive, in order to reduce the need for the whole docbook/rst
2208 *(Released Fri, 27 Mar 2009)*
2210 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2211 - Fix watcher startup sequence, improves the behaviour of busy clusters
2212 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2213 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2214 bind address and other places
2215 - Some documentation fixes and updates
2221 *(Released Mon, 2 Mar 2009)*
2223 - More documentation updates, now all docs should be more-or-less
2225 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2227 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2229 - Fix an issue related to $libdir/run/ganeti and cluster creation
2235 *(Released Thu, 19 Feb 2009)*
2237 - Xen PVM and KVM have switched the default value for the instance root
2238 disk to the first partition on the first drive, instead of the whole
2239 drive; this means that the OS installation scripts must be changed
2241 - Man pages have been updated
2242 - RAPI has been switched by default to HTTPS, and the exported functions
2243 should all work correctly
2244 - RAPI v1 has been removed
2245 - Many improvements to the KVM hypervisor
2246 - Block device errors are now better reported
2247 - Many other bugfixes and small improvements
2253 *(Released Mon, 26 Jan 2009)*
2255 - Version 2 is a general rewrite of the code and therefore the
2256 differences are too many to list, see the design document for 2.0 in
2257 the ``doc/`` subdirectory for more details
2258 - In this beta version there is not yet a migration path from 1.2 (there
2259 will be one in the final 2.0 release)
2260 - A few significant changes are:
2262 - all commands are executed by a daemon (``ganeti-masterd``) and the
2263 various ``gnt-*`` commands are just front-ends to it
2264 - all the commands are entered into, and executed from a job queue,
2265 see the ``gnt-job(8)`` manpage
2266 - the RAPI daemon supports read-write operations, secured by basic
2267 HTTP authentication on top of HTTPS
2268 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2269 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2270 to migrate to DRBD 8 first while still running Ganeti 1.2)
2271 - DRBD devices are using statically allocated minor numbers, which
2272 will be assigned to existing instances during the migration process
2273 - there is support for both Xen PVM and Xen HVM instances running on
2275 - KVM virtualization is supported too
2276 - file-based storage has been implemented, which means that it is
2277 possible to run the cluster without LVM and DRBD storage, for
2278 example using a shared filesystem exported from shared storage (and
2279 still have live migration)
2285 *(Released Tue, 13 Jan 2009)*
2287 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2288 - Reuse the old instance mac address by default on instance import, if
2289 the instance name is the same.
2290 - Handle situations in which the node info rpc returns incomplete
2292 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2293 - Improved version of batcher:
2295 - state file support
2296 - instance mac address support
2297 - support for HVM clusters/instances
2299 - Add an option to show the number of cpu sockets and nodes in
2301 - Support OSes that handle more than one version of the OS api (but do
2302 not change the current API in any other way)
2303 - Fix ``gnt-node migrate``
2304 - ``gnt-debug`` man page
2305 - Fixes various more typos and small issues
2306 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2312 *(Released Wed, 24 Sep 2008)*
2314 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2315 type of disk exported to fully virtualized instances.
2316 - provide access to the serial console of HVM instances
2317 - instance auto_balance flag, set by default. If turned off it will
2318 avoid warnings on cluster verify if there is not enough memory to fail
2319 over an instance. in the future it will prevent automatically failing
2320 it over when we will support that.
2321 - batcher tool for instance creation, see ``tools/README.batcher``
2322 - ``gnt-instance reinstall --select-os`` to interactively select a new
2323 operating system when reinstalling an instance.
2324 - when changing the memory amount on instance modify a check has been
2325 added that the instance will be able to start. also warnings are
2326 emitted if the instance will not be able to fail over, if auto_balance
2328 - documentation fixes
2329 - sync fields between ``gnt-instance list/modify/add/import``
2330 - fix a race condition in drbd when the sync speed was set after giving
2331 the device a remote peer.
2337 *(Released Tue, 22 Jul 2008)*
2339 - note: the allowed size and number of tags per object were reduced
2340 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2341 - fixed twisted 8.x compatibility
2342 - fixed ``gnt-instance replace-disks`` with iallocator
2343 - add TCP keepalives on twisted connections to detect restarted nodes
2344 - disk increase support, see ``gnt-instance grow-disk``
2345 - implement bulk node/instance query for RAPI
2346 - add tags in node/instance listing (optional)
2347 - experimental migration (and live migration) support, read the man page
2348 for ``gnt-instance migrate``
2349 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2350 has some small improvements in handling its state file
2356 *(Released Fri, 13 Jun 2008)*
2358 - Experimental readonly, REST-based remote API implementation;
2359 automatically started on master node, TCP port 5080, if enabled by
2360 ``--enable-rapi`` parameter to configure script.
2361 - Instance allocator support. Add and import instance accept a
2362 ``--iallocator`` parameter, and call that instance allocator to decide
2363 which node to use for the instance. The iallocator document describes
2364 what's expected from an allocator script.
2365 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2366 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2367 node is lost there is still enough memory to fail over the instances
2369 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2370 ``gnt-cluster verify``, to check for site-specific compliance. All the
2371 hooks will run, and their output, if any, will be displayed. Any
2372 failing hook will make the verification return an error value.
2373 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2374 primary and secondary interfaces
2375 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2376 that is still declared as part of the cluster and has failed.
2377 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2378 specifying output fields, that just adds the chosen fields to the
2380 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2381 export from the filesystem.
2382 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2383 have been added. Using them you can enable/disable acpi and pae
2384 support, and specify a path for a cd image to be exported to the
2385 instance. These parameters as the name suggest only work on HVM
2387 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2388 PAE support will be set to the previously hardcoded values, but the
2389 (previously hardcoded) path to the CDROM ISO image will be unset and
2390 if required, needs to be set manually with ``gnt-instance modify``
2392 - The address to which an instance's VNC console is bound is now
2393 selectable per-instance, rather than being cluster wide. Of course
2394 this only applies to instances controlled via VNC, so currently just
2395 applies to HVM clusters.
2401 *(Released Mon, 18 Feb 2008)*
2403 - more tweaks to the disk activation code (especially helpful for DRBD)
2404 - change the default ``gnt-instance list`` output format, now there is
2405 one combined status field (see the manpage for the exact values this
2407 - some more fixes for the mac export to hooks change
2408 - make Ganeti not break with DRBD 8.2.x (which changed the version
2409 format in ``/proc/drbd``) (issue 24)
2410 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2411 template, allowing migration from DRBD0.7+MD to DRBD8
2417 *(Released Wed, 30 Jan 2008)*
2419 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2421 - add command aliases infrastructure and a few aliases
2422 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2423 man pages and the ``--help`` option of ``gnt-node
2424 list``/``gnt-instance list``
2425 - fix ``gnt-backup list`` with down nodes (issue 21)
2426 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2427 - fix the dist archive and add a check for including svn/git files in
2429 - some developer-related changes: improve the burnin and the QA suite,
2430 add an upload script for testing during development
2436 *(Released Wed, 16 Jan 2008)*
2438 - experimental HVM support, read the install document, section
2439 "Initializing the cluster"
2440 - allow for the PVM hypervisor per-instance kernel and initrd paths
2441 - add a new command ``gnt-cluster verify-disks`` which uses a new
2442 algorithm to improve the reconnection of the DRBD pairs if the device
2443 on the secondary node has gone away
2444 - make logical volume code auto-activate LVs at disk activation time
2445 - slightly improve the speed of activating disks
2446 - allow specification of the MAC address at instance creation time, and
2447 changing it later via ``gnt-instance modify``
2448 - fix handling of external commands that generate lots of output on
2450 - update documentation with regard to minimum version of DRBD8 supported
2456 *(Released Tue, 4 Dec 2007)*
2458 - Log the ``xm create`` output to the node daemon log on failure (to
2459 help diagnosing the error)
2460 - In debug mode, log all external commands output if failed to the logs
2461 - Change parsing of lvm commands to ignore stderr
2467 *(Released Wed, 28 Nov 2007)*
2469 - Another round of updates to the DRBD 8 code to deal with more failures
2470 in the replace secondary node operation
2471 - Some more logging of failures in disk operations (lvm, drbd)
2472 - A few documentation updates
2479 *(Released Tue, 13 Nov 2007)*
2481 - Change configuration file format from Python's Pickle to JSON.
2482 Upgrading is possible using the cfgupgrade utility.
2483 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2484 faster replace disks and is more stable (DRBD 8 has many improvements
2485 compared to DRBD 0.7)
2486 - Added command line tags support (see man pages for ``gnt-instance``,
2487 ``gnt-node``, ``gnt-cluster``)
2488 - Added instance rename support
2489 - Added multi-instance startup/shutdown
2490 - Added cluster rename support
2491 - Added ``gnt-node evacuate`` to simplify some node operations
2492 - Added instance reboot operation that can speedup reboot as compared to
2494 - Soften the requirement that hostnames are in FQDN format
2495 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2497 - Removed dependency on debian's patched fping that uses the
2498 non-standard ``-S`` option
2499 - Now the OS definitions are searched for in multiple, configurable
2500 paths (easier for distros to package)
2501 - Some changes to the hooks infrastructure (especially the new
2502 post-configuration update hook)
2503 - Other small bugfixes
2505 .. vim: set textwidth=72 syntax=rst :