10 - Disk adoption interaction with ipolicy checks has been fixed.
11 - Tap device's MAC prefix is now forcibly set to "fe" (issue 217).
12 - Option to force master-failover without voting added (issue 282).
13 - Removal of storage directory on shared file storage was fixed (issue
15 - Validation of LVM volume group name in OpClusterSetParams
16 (``gnt-cluster modify``) was corrected (issue 285).
17 - Clarified error message on lock conflict (issue 287).
18 - Logging of newly submitted jobs has been improved (issue 290).
19 - Hostname checks have been made uniform between instance rename and
21 - The ``--submit`` option is now supported by ``gnt-debug delay``.
22 - Shutting down the master daemon by sending SIGTERM now stops it from
23 processing jobs waiting for locks; instead, those jobs will be started
24 once again after the master daemon is started the next time (issue
26 - Support for Xen's ``xl`` program has been improved.
27 - A type error when doing live migrations with KVM has been fixed (issue
28 297) and error messages for failing migrations have been improved.
29 - Another type error when unsetting OS parameters has been fixed (issue
31 - Several man page updates and typo fixes.
37 *(Released Fri, 12 Oct 2012)*
39 A small bugfix release. Among the bugs fixed:
41 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
42 made the command unusable.
43 - Commands that issue many jobs don't fail anymore just because some jobs
44 take so long that other jobs are archived.
45 - Failures during ``gnt-instance reinstall`` are reflected by the exit
47 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
49 - When ``always_failover`` is set, ``--allow-failover`` is not required
50 in migrate commands anymore.
51 - ``bash_completion`` works even if extglob is disabled.
52 - Fixed bug with locks that made failover for RDB-based instances fail.
53 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
54 a random node instead of one based on the allocator metric.
55 - Support for newer versions of pylint and pep8.
56 - Hail doesn't fail anymore when trying to add an instance of type
57 ``file``, ``sharedfile`` or ``rbd``.
58 - Added new Makefile target to rebuild the whole distribution, so that
59 all files are included.
65 *(Released Fri, 27 Jul 2012)*
68 .. attention:: The ``LUXI`` protocol has been made more consistent
69 regarding its handling of command arguments. This, however, leads to
70 incompatibility issues with previous versions. Please ensure that you
71 restart Ganeti daemons soon after the upgrade, otherwise most
72 ``LUXI`` calls (job submission, setting/resetting the drain flag,
73 pausing/resuming the watcher, cancelling and archiving jobs, querying
74 the cluster configuration) will fail.
83 The current ``admin_up`` field, which used to denote whether an instance
84 should be running or not, has been removed. Instead, ``admin_state`` is
85 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
87 The rational behind this is that an instance being “down” can have
90 - it could be down during a reboot
91 - it could be temporarily be down for a reinstall
92 - or it could be down because it is deprecated and kept just for its
95 The previous Boolean state was making it difficult to do capacity
96 calculations: should Ganeti reserve memory for a down instance? Now, the
97 tri-state field makes it clear:
99 - in ``up`` and ``down`` state, all resources are reserved for the
100 instance, and it can be at any time brought up if it is down
101 - in ``offline`` state, only disk space is reserved for it, but not
104 The field can have an extra use: since the transition between ``up`` and
105 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
106 transition between ``offline`` and ``down`` is done via ``gnt-instance
107 modify``, it is possible to given different rights to users. For
108 example, owners of an instance could be allowed to start/stop it, but
109 not transition it out of the offline state.
111 Instance policies and specs
112 +++++++++++++++++++++++++++
114 In previous Ganeti versions, an instance creation request was not
115 limited on the minimum size and on the maximum size just by the cluster
116 resources. As such, any policy could be implemented only in third-party
117 clients (RAPI clients, or shell wrappers over ``gnt-*``
118 tools). Furthermore, calculating cluster capacity via ``hspace`` again
119 required external input with regards to instance sizes.
121 In order to improve these workflows and to allow for example better
122 per-node group differentiation, we introduced instance specs, which
125 - minimum instance disk size, disk count, memory size, cpu count
126 - maximum values for the above metrics
127 - and “standard” values (used in ``hspace`` to calculate the standard
130 The minimum/maximum values can be also customised at node-group level,
131 for example allowing more powerful hardware to support bigger instance
134 Beside the instance specs, there are a few other settings belonging to
135 the instance policy framework. It is possible now to customise, per
136 cluster and node-group:
138 - the list of allowed disk templates
139 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
140 - the maximum ratio of instance to spindles (see below for more
141 information) for local storage
143 All these together should allow all tools that talk to Ganeti to know
144 what are the ranges of allowed values for instances and the
145 over-subscription that is allowed.
147 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
148 instance configuration, and the physical CPU configuration from the
149 node. For the spindle ratios however, we didn't track before these
150 values, so new parameters have been added:
152 - a new node parameter ``spindle_count``, defaults to 1, customisable at
153 node group or node level
154 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
156 Note that spindles in this context doesn't need to mean actual
157 mechanical hard-drives; it's just a relative number for both the node
158 I/O capacity and instance I/O consumption.
160 Instance migration behaviour
161 ++++++++++++++++++++++++++++
163 While live-migration is in general desirable over failover, it is
164 possible that for some workloads it is actually worse, due to the
165 variable time of the “suspend” phase during live migration.
167 To allow the tools to work consistently over such instances (without
168 having to hard-code instance names), a new backend parameter
169 ``always_failover`` has been added to control the migration/failover
170 behaviour. When set to True, all migration requests for an instance will
171 instead fall-back to failover.
173 Instance memory ballooning
174 ++++++++++++++++++++++++++
176 Initial support for memory ballooning has been added. The memory for an
177 instance is no longer fixed (backend parameter ``memory``), but instead
178 can vary between minimum and maximum values (backend parameters
179 ``minmem`` and ``maxmem``). Currently we only change an instance's
182 - live migrating or failing over and instance and the target node
183 doesn't have enough memory
184 - user requests changing the memory via ``gnt-instance modify
190 In order to control the use of specific CPUs by instance, support for
191 controlling CPU pinning has been added for the Xen, HVM and LXC
192 hypervisors. This is controlled by a new hypervisor parameter
193 ``cpu_mask``; details about possible values for this are in the
194 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
195 VCPU-to-CPU mapping) form will work well only when all nodes in your
196 cluster have the same amount of CPUs.
201 Another area in which Ganeti was not customisable were the parameters
202 used for storage configuration, e.g. how many stripes to use for LVM,
203 DRBD resync configuration, etc.
205 To improve this area, we've added disks parameters, which are
206 customisable at cluster and node group level, and which allow to
207 specify various parameters for disks (DRBD has the most parameters
208 currently), for example:
210 - DRBD resync algorithm and parameters (e.g. speed)
211 - the default VG for meta-data volumes for DRBD
212 - number of stripes for LVM (plain disk template)
215 These parameters can be modified via ``gnt-cluster modify -D …`` and
216 ``gnt-group modify -D …``, and are used at either instance creation (in
217 case of LVM stripes, for example) or at disk “activation” time
220 Rados block device support
221 ++++++++++++++++++++++++++
223 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
224 denoted by the ``rbd`` disk template type. This is considered
225 experimental, feedback is welcome. For details on configuring it, see
226 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
231 The existing master IP functionality works well only in simple setups (a
232 single network shared by all nodes); however, if nodes belong to
233 different networks, then the ``/32`` setup and lack of routing
234 information is not enough.
236 To allow the master IP to function well in more complex cases, the
237 system was reworked as follows:
239 - a master IP netmask setting has been added
240 - the master IP activation/turn-down code was moved from the node daemon
242 - whether to run the Ganeti-supplied master IP script or a user-supplied
243 on is a ``gnt-cluster init`` setting
245 Details about the location of the standard and custom setup scripts are
246 in the man page :manpage:`gnt-cluster(8)`; for information about the
247 setup script protocol, look at the Ganeti-supplied script.
252 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
255 It is now possible to use TLS-protected connections, and when renewing
256 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
257 it is now possible to specify spice or spice CA certificates. Also, it
258 is possible to configure a password for SPICE sessions via the
259 hypervisor parameter ``spice_password_file``.
261 There are also new parameters to control the compression and streaming
262 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
263 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
264 for the spice parameters.
266 Lastly, it is now possible to see the SPICE connection information via
267 ``gnt-instance console``.
272 A new tool (``tools/ovfconverter``) has been added that supports
273 conversion between Ganeti and the `Open Virtualization Format
274 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
277 This relies on the ``qemu-img`` tool to convert the disk formats, so the
278 actual compatibility with other virtualization solutions depends on it.
283 The configuration query daemon (``ganeti-confd``) is now optional, and
284 has been rewritten in Haskell; whether to use the daemon at all, use the
285 Python (default) or the Haskell version is selectable at configure time
286 via the ``--enable-confd`` parameter, which can take one of the
287 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
288 daemon will result in a smaller footprint; for larger systems, we
289 welcome feedback on the Haskell version which might become the default
292 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
293 daemon running. The Python version doesn't implement the new call.
296 User interface changes
297 ~~~~~~~~~~~~~~~~~~~~~~
299 We have replaced the ``--disks`` option of ``gnt-instance
300 replace-disks`` with a more flexible ``--disk`` option, which allows
301 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
302 disk size and mode can be changed upon recreation (via ``gnt-instance
303 recreate-disks``, which accepts the same ``--disk`` option).
305 As many people are used to a ``show`` command, we have added that as an
306 alias to ``info`` on all ``gnt-*`` commands.
308 The ``gnt-instance grow-disk`` command has a new mode in which it can
309 accept the target size of the disk, instead of the delta; this can be
310 more safe since two runs in absolute mode will be idempotent, and
311 sometimes it's also easier to specify the desired size directly.
313 Also the handling of instances with regard to offline secondaries has
314 been improved. Instance operations should not fail because one of it's
315 secondary nodes is offline, even though it's safe to proceed.
317 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
318 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
319 minors to instance name.
324 RAPI coverage has improved, with (for example) new resources for
325 recreate-disks, node power-cycle, etc.
330 There is partial support for ``xl`` in the Xen hypervisor; feedback is
333 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
334 whether to still support Python 2.4 or move to Python 2.6 as minimum
337 Support for Fedora has been slightly improved; the provided example
338 init.d script should work better on it and the INSTALL file should
339 document the needed dependencies.
344 The deprecated ``QueryLocks`` LUXI request has been removed. Use
345 ``Query(what=QR_LOCK, ...)`` instead.
347 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
348 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
349 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
350 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
351 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
353 RAPI client: ``CertificateError`` now derives from
354 ``GanetiApiError``. This should make it more easy to handle Ganeti
357 Deprecation warnings due to PyCrypto/paramiko import in
358 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
359 make sure to run an up-to-date paramiko version, if you use this tool.
361 The QA scripts now depend on Python 2.5 or above (the main code base
362 still works with Python 2.4).
364 The configuration file (``config.data``) is now written without
365 indentation for performance reasons; if you want to edit it, it can be
366 re-formatted via ``tools/fmtjson``.
368 A number of bugs has been fixed in the cluster merge tool.
370 ``x509`` certification verification (used in import-export) has been
371 changed to allow the same clock skew as permitted by the cluster
372 verification. This will remove some rare but hard to diagnose errors in
379 *(Released Thu, 19 Jul 2012)*
381 Very few changes from rc4 to the final release, only bugfixes:
383 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
384 instance, fix CDROM booting for KVM instances)
385 - fixed node group modification of node parameters
386 - fixed issue in LUClusterVerifyGroup with multi-group clusters
387 - fixed generation of bash completion to ensure a stable ordering
394 *(Released Fri, 13 Jul 2012)*
396 Third release candidate for 2.6. The following changes were done from
399 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
400 - Fixed an inconsistency in the LUXI protocol with the provided
401 arguments (NOT backwards compatible)
402 - Fixed a bug with node groups ipolicy where ``min`` was greater than
403 the cluster ``std`` value
404 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
405 easier instance debugging on nodes (requires ``hconfd`` to work)
411 *(Released Tue, 03 Jul 2012)*
413 Second release candidate for 2.6. The following changes were done from
416 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
418 - Fixed a RAPI regression on missing beparams/memory
419 - Fixed redistribution of files on offline nodes
420 - Added possibility to run activate-disks even though secondaries are
421 offline. With this change it relaxes also the strictness on some other
422 commands which use activate disks internally:
423 * ``gnt-instance start|reboot|rename|backup|export``
424 - Made it possible to remove safely an instance if its secondaries are
426 - Made it possible to reinstall even though secondaries are offline
432 *(Released Mon, 25 Jun 2012)*
434 First release candidate for 2.6. The following changes were done from
437 - Fixed bugs with disk parameters and ``rbd`` templates as well as
439 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
440 behaviour. It supports now the modify operation
441 - ``hcheck`` implemented to analyze cluster health and possibility of
442 improving health by rebalance
443 - ``hbal`` has been improved in dealing with split instances
449 *(Released Mon, 11 Jun 2012)*
451 Second beta release of 2.6. The following changes were done from beta2
454 - Fixed ``daemon-util`` with non-root user models
455 - Fixed creation of plain instances with ``--no-wait-for-sync``
456 - Fix wrong iv_names when running ``cfgupgrade``
457 - Export more information in RAPI group queries
458 - Fixed bug when changing instance network interfaces
459 - Extended burnin to do NIC changes
460 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
461 - Changed default for DRBD barriers
462 - Fixed DRBD error reporting for syncer rate
463 - Verify the options on disk parameters
465 And of course various fixes to documentation and improved unittests and
472 *(Released Wed, 23 May 2012)*
474 First beta release of 2.6. The following changes were done from beta1 to
477 - integrated patch for distributions without ``start-stop-daemon``
478 - adapted example init.d script to work on Fedora
479 - fixed log handling in Haskell daemons
480 - adapted checks in the watcher for pycurl linked against libnss
481 - add partial support for ``xl`` instead of ``xm`` for Xen
482 - fixed a type issue in cluster verification
483 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
486 Plus integrated fixes from the 2.5 branch:
488 - fixed ``kvm-ifup`` to use ``/bin/bash``
489 - fixed parallel build failures
490 - KVM live migration when using a custom keymap
496 *(Released Tue, 24 Jul 2012)*
498 A small bugfix release, with no new features:
500 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
501 different default shell (e.g. Debian, Ubuntu)
502 - fixed KVM startup and live migration with a custom keymap (fixes Issue
503 243 and Debian bug #650664)
504 - fixed compatibility with KVM versions that don't support multiple boot
505 devices (fixes Issue 230 and Debian bug #624256)
507 Additionally, a few fixes were done to the build system (fixed parallel
508 build failures) and to the unittests (fixed race condition in test for
509 FileID functions, and the default enable/disable mode for QA test is now
516 *(Released Fri, 11 May 2012)*
518 A small bugfix release.
520 The main issues solved are on the topic of compatibility with newer LVM
523 - fixed parsing of ``lv_attr`` field
524 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
525 the ``--force`` flag is needed
527 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
528 to accept kernel 3.x too (was hardcoded to 2.6.*).
530 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
531 of hook results and that also made display of other errors suboptimal
532 was fixed; the code behaves now like 2.4 and earlier.
534 Another change in 2.5, the cleanup of the OS scripts environment, is too
535 aggressive: it removed even the ``PATH`` variable, which requires the OS
536 scripts to *always* need to export it. Since this is a bit too strict,
537 we now export a minimal PATH, the same that we export for hooks.
539 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
540 integrated into this release.
542 Finally, a few other miscellaneous changes were done (no new features,
543 just small improvements):
545 - Fix ``gnt-group --help`` display
546 - Fix hardcoded Xen kernel path
547 - Fix grow-disk handling of invalid units
548 - Update synopsis for ``gnt-cluster repair-disk-sizes``
549 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
555 *(Released Thu, 12 Apr 2012)*
557 Incompatible/important changes and bugfixes
558 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
560 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
561 resource's ``ip_check`` parameter changed from ``True`` to ``False``
562 to match the underlying LUXI interface.
563 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
564 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
565 not maintain backwards-compatibility as the underlying operation
566 changed in an incompatible way. The RAPI client can talk to old
567 servers, but it needs to be told so as the return value changed.
568 - When creating file-based instances via RAPI, the ``file_driver``
569 parameter no longer defaults to ``loop`` and must be specified.
570 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
572 - Support for the undocumented and deprecated RAPI instance creation
573 request format version 0 has been dropped. Use version 1, supported
574 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
575 - Pyparsing 1.4.6 or above is required, see :doc:`installation
576 documentation <install>`.
577 - The "cluster-verify" hooks are now executed per group by the
578 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
579 if you just run ``gnt-cluster verify``, which generates one opcode per
581 - The environment as passed to the OS scripts is cleared, and thus no
582 environment variables defined in the node daemon's environment will be
583 inherited by the scripts.
584 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
586 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
587 support operations involving multiple node groups.
588 - Offline nodes are ignored when failing over an instance.
589 - Support for KVM version 1.0, which changed the version reporting format
591 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
593 - ``Makefile`` is now compatible with Automake 1.11.2
594 - Includes all bugfixes made in the 2.4 series
599 - The ganeti-htools project has been merged into the ganeti-core source
600 tree and will be built as part of Ganeti (see :doc:`install-quick`).
601 - Implemented support for :doc:`shared storage <design-shared-storage>`.
602 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
603 GPT-style partition tables (requires `parted
604 <http://www.gnu.org/s/parted/>`_).
605 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
606 - Allowed adding tags on instance creation.
607 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
608 - Allow instances to be started in a paused state, enabling the user to
609 see the complete console output on boot using the console.
610 - Added new hypervisor flag to control default reboot behaviour
611 (``reboot_behavior``).
612 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
613 - Improved out-of-band management support:
615 - Added ``gnt-node health`` command reporting the health status of
617 - Added ``gnt-node power`` command to manage power status of nodes.
618 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
620 - Instance migration can fall back to failover if instance is not
622 - Filters can be used when listing nodes, instances, groups and locks;
623 see *ganeti(7)* manpage.
624 - Added post-execution status as variables to :doc:`hooks <hooks>`
626 - Instance tags are exported/imported together with the instance.
627 - When given an explicit job ID, ``gnt-job info`` will work for archived
629 - Jobs can define dependencies on other jobs (not yet supported via
630 RAPI or command line, but used by internal commands and usable via
633 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
636 - Instance failover is now available as a RAPI resource
637 (``/2/instances/[instance_name]/failover``).
638 - ``gnt-instance info`` defaults to static information if primary node
640 - Opcodes have a new ``comment`` attribute.
641 - Added basic SPICE support to KVM hypervisor.
642 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
644 Node group improvements
645 ~~~~~~~~~~~~~~~~~~~~~~~
647 - ``gnt-cluster verify`` has been modified to check groups separately,
648 thereby improving performance.
649 - Node group support has been added to ``gnt-cluster verify-disks``,
650 which now operates per node group.
651 - Watcher has been changed to work better with node groups.
653 - One process and state file per node group.
654 - Slow watcher in one group doesn't block other group's watcher.
656 - Added new command, ``gnt-group evacuate``, to move all instances in a
657 node group to other groups.
658 - Added ``gnt-instance change-group`` to move an instance to another
660 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
661 per-group operations.
662 - Node groups can be tagged.
663 - Some operations switch from an exclusive to a shared lock as soon as
665 - Instance's primary and secondary nodes' groups are now available as
666 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
667 and ``snodes.group.uuid``).
672 - Numerous updates to documentation and manpages.
674 - :doc:`RAPI <rapi>` documentation now has detailed parameter
676 - Some opcode/job results are now also documented, see :doc:`RAPI
679 - A lockset's internal lock is now also visible in lock monitor.
680 - Log messages from job queue workers now contain information about the
681 opcode they're processing.
682 - ``gnt-instance console`` no longer requires the instance lock.
683 - A short delay when waiting for job changes reduces the number of LUXI
684 requests significantly.
685 - DRBD metadata volumes are overwritten with zeros during disk creation.
686 - Out-of-band commands no longer acquire the cluster lock in exclusive
688 - ``devel/upload`` now uses correct permissions for directories.
694 *(Released Fri, 23 Mar 2012)*
696 This was the sixth release candidate of the 2.5 series.
702 *(Released Mon, 9 Jan 2012)*
704 This was the fifth release candidate of the 2.5 series.
710 *(Released Thu, 27 Oct 2011)*
712 This was the fourth release candidate of the 2.5 series.
718 *(Released Wed, 26 Oct 2011)*
720 This was the third release candidate of the 2.5 series.
726 *(Released Tue, 18 Oct 2011)*
728 This was the second release candidate of the 2.5 series.
734 *(Released Tue, 4 Oct 2011)*
736 This was the first release candidate of the 2.5 series.
742 *(Released Wed, 31 Aug 2011)*
744 This was the third beta release of the 2.5 series.
750 *(Released Mon, 22 Aug 2011)*
752 This was the second beta release of the 2.5 series.
758 *(Released Fri, 12 Aug 2011)*
760 This was the first beta release of the 2.5 series.
766 *(Released Thu, 27 Oct 2011)*
768 - Fixed bug when parsing command line parameter values ending in
770 - Fixed assertion error after unclean master shutdown
771 - Disable HTTP client pool for RPC, significantly reducing memory usage
773 - Fixed queue archive creation with wrong permissions
779 *(Released Tue, 23 Aug 2011)*
783 - Fixed documentation for importing with ``--src-dir`` option
784 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
785 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
791 *(Released Fri, 5 Aug 2011)*
793 Many bug-fixes and a few small features:
795 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
796 issues when you tried to add an instance with two MAC addresses in one
798 - KVM: fixed per-instance stored UID value
799 - KVM: configure bridged NICs at migration start
800 - KVM: Fix a bug where instance will not start with never KVM versions
802 - Added OS search path to ``gnt-cluster info``
803 - Fixed an issue with ``file_storage_dir`` where you were forced to
804 provide an absolute path, but the documentation states it is a
805 relative path, the documentation was right
806 - Added a new parameter to instance stop/start called ``--no-remember``
807 that will make the state change to not be remembered
808 - Implemented ``no_remember`` at RAPI level
809 - Improved the documentation
810 - Node evacuation: don't call IAllocator if node is already empty
811 - Fixed bug in DRBD8 replace disks on current nodes
812 - Fixed bug in recreate-disks for DRBD instances
813 - Moved assertion checking locks in ``gnt-instance replace-disks``
814 causing it to abort with not owning the right locks for some situation
815 - Job queue: Fixed potential race condition when cancelling queued jobs
816 - Fixed off-by-one bug in job serial generation
817 - ``gnt-node volumes``: Fix instance names
818 - Fixed aliases in bash completion
819 - Fixed a bug in reopening log files after being sent a SIGHUP
820 - Added a flag to burnin to allow specifying VCPU count
821 - Bugfixes to non-root Ganeti configuration
827 *(Released Thu, 12 May 2011)*
829 Many bug-fixes and a few new small features:
831 - Fixed a bug related to log opening failures
832 - Fixed a bug in instance listing with orphan instances
833 - Fixed a bug which prevented resetting the cluster-level node parameter
834 ``oob_program`` to the default
835 - Many fixes related to the ``cluster-merge`` tool
836 - Fixed a race condition in the lock monitor, which caused failures
837 during (at least) creation of many instances in parallel
838 - Improved output for gnt-job info
839 - Removed the quiet flag on some ssh calls which prevented debugging
841 - Improved the N+1 failure messages in cluster verify by actually
842 showing the memory values (needed and available)
843 - Increased lock attempt timeouts so that when executing long operations
844 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
845 too early and thus prevent the use of the 'fair' mechanism
846 - Changed instance query data (``gnt-instance info``) to not acquire
847 locks unless needed, thus allowing its use on locked instance if only
848 static information is asked for
849 - Improved behaviour with filesystems that do not support rename on an
851 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
852 kept locks on all nodes during the wipe, which is unneeded
853 - Fixed ``gnt-watcher`` handling of errors during hooks execution
854 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
855 10GiB) which caused the wipe to fail right at the end in some cases
856 - Fixed master IP activation when doing master failover with no-voting
857 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
858 the master node itself
859 - Fixed potential data-loss in under disk full conditions, where Ganeti
860 wouldn't check correctly the return code and would consider
861 partially-written files 'correct'
862 - Fixed bug related to multiple VGs and DRBD disk replacing
863 - Added new disk parameter ``metavg`` that allows placement of the meta
864 device for DRBD in a different volume group
865 - Fixed error handling in the node daemon when the system libc doesn't
866 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
867 - Fixed lock release during replace-disks, which kept cluster-wide locks
868 when doing disk replaces with an iallocator script
869 - Added check for missing bridges in cluster verify
870 - Handle EPIPE errors while writing to the terminal better, so that
871 piping the output to e.g. ``less`` doesn't cause a backtrace
872 - Fixed rare case where a ^C during Luxi calls could have been
873 interpreted as server errors, instead of simply terminating
874 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
876 - Added a few more parameters to the KVM hypervisor, allowing a second
877 CDROM, custom disk type for CDROMs and a floppy image
878 - Removed redundant message in instance rename when the name is given
880 - Added option to ``gnt-instance recreate-disks`` to allow creating the
881 disks on new nodes, allowing recreation when the original instance
882 nodes are completely gone
883 - Added option when converting disk templates to DRBD to skip waiting
884 for the resync, in order to make the instance available sooner
885 - Added two new variables to the OS scripts environment (containing the
887 - Made the root_path and optional parameter for the xen-pvm hypervisor,
888 to allow use of ``pvgrub`` as bootloader
889 - Changed the instance memory modifications to only check out-of-memory
890 conditions on memory increases, and turned the secondary node warnings
891 into errors (they can still be overridden via ``--force``)
892 - Fixed the handling of a corner case when the Python installation gets
893 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
894 to execute a command that doesn't exist
895 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
896 node of the instance returned failures during instance shutdown; this
897 adds the option ``--ignore-consistency`` to gnt-instance move
899 And as usual, various improvements to the error messages, documentation
906 *(Released Wed, 09 Mar 2011)*
908 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
909 the RAPI users file if run twice (even with ``--dry-run``).
911 The release fixes that bug (nothing else changed).
917 *(Released Mon, 07 Mar 2011)*
919 Final 2.4.0 release. Just a few small fixes:
921 - Fixed RAPI node evacuate
922 - Fixed the kvm-ifup script
923 - Fixed internal error handling for special job cases
924 - Updated man page to specify the escaping feature for options
930 *(Released Mon, 28 Feb 2011)*
932 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
933 that this feature wiped the disks of the wrong instance, leading to loss
938 - Fixed title of query field containing instance name
939 - Expanded the glossary in the documentation
940 - Fixed one unittest (internal issue)
946 *(Released Mon, 21 Feb 2011)*
948 A number of bug fixes plus just a couple functionality changes.
950 On the user-visible side, the ``gnt-* list`` command output has changed
951 with respect to "special" field states. The current rc1 style of display
952 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
953 the default output mode special fields states are displayed as follows:
955 - Offline resource: ``*``
956 - Unavailable/not applicable: ``-``
957 - Data missing (RPC failure): ``?``
958 - Unknown field: ``??``
960 Another user-visible change is the addition of ``--force-join`` to
965 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
966 - Fixed regression in RAPI/instance reinstall where all parameters were
967 required (instead of optional)
968 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
969 - Fixed iallocator usage (offline nodes were not considered offline)
970 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
971 - Fixed hypervisor and OS parameter validation with respect to
973 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
975 - Fixed ``tools/listrunner`` with respect to agent-based usage
981 *(Released Fri, 4 Feb 2011)*
983 Many changes and fixes since the beta1 release. While there were some
984 internal changes, the code has been mostly stabilised for the RC
987 Note: the dumb allocator was removed in this release, as it was not kept
988 up-to-date with the IAllocator protocol changes. It is recommended to
989 use the ``hail`` command from the ganeti-htools package.
991 Note: the 2.4 and up versions of Ganeti are not compatible with the
992 0.2.x branch of ganeti-htools. You need to upgrade to
993 ganeti-htools-0.3.0 (or later).
995 Regressions fixed from 2.3
996 ~~~~~~~~~~~~~~~~~~~~~~~~~~
998 - Fixed the ``gnt-cluster verify-disks`` command
999 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1001 - Fixed disk adoption breakage
1002 - Fixed wrong headers in instance listing for field aliases
1007 - Fixed corner case in KVM handling of NICs
1008 - Fixed many cases of wrong handling of non-vm_capable nodes
1009 - Fixed a bug where a missing instance symlink was not possible to
1010 recreate with any ``gnt-*`` command (now ``gnt-instance
1011 activate-disks`` does it)
1012 - Fixed the volume group name as reported by ``gnt-cluster
1014 - Increased timeouts for the import-export code, hopefully leading to
1015 fewer aborts due network or instance timeouts
1016 - Fixed bug in ``gnt-node list-storage``
1017 - Fixed bug where not all daemons were started on cluster
1018 initialisation, but only at the first watcher run
1019 - Fixed many bugs in the OOB implementation
1020 - Fixed watcher behaviour in presence of instances with offline
1022 - Fixed instance list output for instances running on the wrong node
1023 - a few fixes to the cluster-merge tool, but it still cannot merge
1024 multi-node groups (currently it is not recommended to use this tool)
1030 - Improved network configuration for the KVM hypervisor
1031 - Added e1000 as a supported NIC for Xen-HVM
1032 - Improved the lvmstrap tool to also be able to use partitions, as
1033 opposed to full disks
1034 - Improved speed of disk wiping (the cluster parameter
1035 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1036 of instance creations
1037 - Added documentation for the OS parameters
1038 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1039 hypervisor is not responding
1040 - Added display of blacklisted and hidden OS information in
1041 ``gnt-cluster info``
1042 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1043 NIC and node parameters, which might create problems with currently
1044 invalid (but undetected) configuration files, but prevents validation
1045 failures when unrelated parameters are modified
1046 - Changed cluster initialisation to wait for the master daemon to become
1048 - Expanded the RAPI interface:
1050 - Added config redistribution resource
1051 - Added activation/deactivation of instance disks
1052 - Added export of console information
1054 - Implemented log file reopening on SIGHUP, which allows using
1055 logrotate(8) for the Ganeti log files
1056 - Added a basic OOB helper script as an example
1062 *(Released Fri, 14 Jan 2011)*
1067 - Fixed timezone issues when formatting timestamps
1068 - Added support for node groups, available via ``gnt-group`` and other
1070 - Added out-of-band framework and management, see :doc:`design
1071 document <design-oob>`
1072 - Removed support for roman numbers from ``gnt-node list`` and
1073 ``gnt-instance list``.
1074 - Allowed modification of master network interface via ``gnt-cluster
1075 modify --master-netdev``
1076 - Accept offline secondaries while shutting down instance disks
1077 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1078 - Added support for multiple LVM volume groups
1079 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1081 - Added commands to list available fields:
1083 - ``gnt-node list-fields``
1084 - ``gnt-group list-fields``
1085 - ``gnt-instance list-fields``
1087 - Updated documentation and man pages
1092 - Moved ``rapi_users`` file into separate directory, now named
1093 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1095 - Added new tool for running commands on many machines,
1096 ``tools/ganeti-listrunner``
1097 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1098 improving the ``gnt-instance console`` output
1099 - Allowed customisation of disk index separator at ``configure`` time
1100 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1101 - Added support for non-partitioned md disks in ``lvmstrap``
1102 - Added script to gracefully power off KVM instances
1103 - Split ``utils`` module into smaller parts
1104 - Changed query operations to return more detailed information, e.g.
1105 whether an information is unavailable due to an offline node. To use
1106 this new functionality, the LUXI call ``Query`` must be used. Field
1107 information is now stored by the master daemon and can be retrieved
1108 using ``QueryFields``. Instances, nodes and groups can also be queried
1109 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1110 exposed via RAPI). The following commands make use of this
1111 infrastructure change:
1113 - ``gnt-group list``
1114 - ``gnt-group list-fields``
1116 - ``gnt-node list-fields``
1117 - ``gnt-instance list``
1118 - ``gnt-instance list-fields``
1119 - ``gnt-debug locks``
1124 - New RAPI resources (see :doc:`rapi`):
1128 - ``/2/groups/[group_name]``
1129 - ``/2/groups/[group_name]/assign-nodes``
1130 - ``/2/groups/[group_name]/modify``
1131 - ``/2/groups/[group_name]/rename``
1132 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1136 - Implemented ``no_install`` for instance creation
1137 - Implemented OS parameters for instance reinstallation, allowing
1138 use of special settings on reinstallation (e.g. for preserving data)
1143 - Added IPv6 support in import/export
1144 - Pause DRBD synchronization while wiping disks on instance creation
1145 - Updated unittests and QA scripts
1146 - Improved network parameters passed to KVM
1147 - Converted man pages from docbook to reStructuredText
1153 *(Released Mon, 20 Dec 2010)*
1155 Released version 2.3.1~rc1 without any changes.
1161 *(Released Wed, 1 Dec 2010)*
1163 - impexpd: Disable OpenSSL compression in socat if possible (backport
1164 from master, commit e90739d625b, see :doc:`installation guide
1165 <install-quick>` for details)
1166 - Changed unittest coverage report to exclude test scripts
1167 - Added script to check version format
1173 *(Released Wed, 1 Dec 2010)*
1175 Released version 2.3.0~rc1 without any changes.
1181 *(Released Fri, 19 Nov 2010)*
1183 A number of bugfixes and documentation updates:
1185 - Update ganeti-os-interface documentation
1186 - Fixed a bug related to duplicate MACs or similar items which should be
1188 - Fix breakage in OS state modify
1189 - Reinstall instance: disallow offline secondaries (fixes bug related to
1190 OS changing but reinstall failing)
1191 - plus all the other fixes between 2.2.1 and 2.2.2
1197 *(Released Tue, 2 Nov 2010)*
1199 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1200 - Fixed master failover race with watcher
1201 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1203 - Accept previously stopped instance for export with instance removal
1204 - Simplify and extend the environment variables for instance OS scripts
1205 - Added new node flags, ``master_capable`` and ``vm_capable``
1206 - Added optional instance disk wiping prior during allocation. This is a
1207 cluster-wide option and can be set/modified using
1208 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1209 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1210 :doc:`install-quick`
1211 - Added a new watcher option (``--ignore-pause``)
1212 - Added option to ignore offline node on instance start/stop
1213 (``--ignore-offline``)
1214 - Allow overriding OS parameters with ``gnt-instance reinstall``
1215 - Added ability to change node's secondary IP address using ``gnt-node
1217 - Implemented privilege separation for all daemons except
1218 ``ganeti-noded``, see ``configure`` options
1219 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1221 - Implemented job priorities (see ``ganeti(7)`` manpage)
1222 - Ignore failures while shutting down instances during failover from
1224 - Exit daemon's bootstrap process only once daemon is ready
1225 - Export more information via ``LUInstanceQuery``/remote API
1226 - Improved documentation, QA and unittests
1227 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1228 restart if the file was created or changed
1229 - Added LUXI protocol version sent with each request and response,
1230 allowing detection of server/client mismatches
1231 - Moved the Python scripts among gnt-* and ganeti-* into modules
1232 - Moved all code related to setting up SSH to an external script,
1234 - Infrastructure changes for node group support in future versions
1240 *(Released Fri, 19 Nov 2010)*
1242 A few small bugs fixed, and some improvements to the build system:
1244 - Fix documentation regarding conversion to drbd
1245 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1247 - Fix error handling in node modify with multiple changes
1248 - Allow remote imports without checked names
1254 *(Released Tue, 19 Oct 2010)*
1256 - Disable SSL session ID cache in RPC client
1262 *(Released Thu, 14 Oct 2010)*
1264 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1265 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1267 - Fix problems with interaction between Curl and Python's HTTP server,
1268 resulting in increased speed in many RPC calls
1269 - Improve our release script to prevent breakage with older aclocal and
1276 *(Released Thu, 7 Oct 2010)*
1278 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1279 value retrieved from master
1280 - Added support for blacklisted or hidden OS definitions
1281 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1282 - Added support for -mem-path in KVM hypervisor abstraction layer
1283 - Allow overriding instance parameters in tool for inter-cluster
1284 instance moves (``tools/move-instance``)
1285 - Improved opcode summaries (e.g. in ``gnt-job list``)
1286 - Improve consistency of OS listing by sorting it
1287 - Documentation updates
1293 *(Released Fri, 8 Oct 2010)*
1295 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1301 *(Released Mon, 4 Oct 2010)*
1303 - Fixed regression in ``gnt-instance rename``
1309 *(Released Wed, 22 Sep 2010)*
1311 - Fixed OS_VARIANT variable for OS scripts
1312 - Fixed cluster tag operations via RAPI
1313 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1314 - Disabled RAPI CA checks in watcher
1320 *(Released Mon, 23 Aug 2010)*
1322 - Support DRBD versions of the format "a.b.c.d"
1324 - Re-introduce support for usage from multiple threads in RAPI client
1325 - Instance renames and modify via RAPI
1326 - Work around race condition between processing and archival in job
1328 - Mark opcodes following failed one as failed, too
1329 - Job field ``lock_status`` was removed due to difficulties making it
1330 work with the changed job queue in Ganeti 2.2; a better way to monitor
1331 locks is expected for a later 2.2.x release
1332 - Fixed dry-run behaviour with many commands
1333 - Support ``ssh-agent`` again when adding nodes
1334 - Many additional bugfixes
1340 *(Released Fri, 30 Jul 2010)*
1342 Important change: the internal RPC mechanism between Ganeti nodes has
1343 changed from using a home-grown http library (based on the Python base
1344 libraries) to use the PycURL library. This requires that PycURL is
1345 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1346 against GnuTLS by default. cURL's support for GnuTLS had known issues
1347 before cURL 7.21.0 and we recommend using the latest cURL release or
1348 linking against OpenSSL. Most other distributions already link PycURL
1349 and cURL against OpenSSL. The command::
1351 python -c 'import pycurl; print pycurl.version'
1353 can be used to determine the libraries PycURL and cURL are linked
1356 Other significant changes:
1358 - Rewrote much of the internals of the job queue, in order to achieve
1359 better parallelism; this decouples job query operations from the job
1360 processing, and it should allow much nicer behaviour of the master
1361 daemon under load, and it also has uncovered some long-standing bugs
1362 related to the job serialisation (now fixed)
1363 - Added a default iallocator setting to the cluster parameters,
1364 eliminating the need to always pass nodes or an iallocator for
1365 operations that require selection of new node(s)
1366 - Added experimental support for the LXC virtualization method
1367 - Added support for OS parameters, which allows the installation of
1368 instances to pass parameter to OS scripts in order to customise the
1370 - Added a hypervisor parameter controlling the migration type (live or
1371 non-live), since hypervisors have various levels of reliability; this
1372 has renamed the 'live' parameter to 'mode'
1373 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1374 logical volumes, meaning that cluster verify will ignore them and not
1375 flag their presence as errors
1376 - The watcher will now reset the error count for failed instances after
1377 8 hours, thus allowing self-healing if the problem that caused the
1378 instances to be down/fail to start has cleared in the meantime
1379 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1380 check for, and warn, if the drbd module parameter ``usermode_helper``
1381 is not consistent with the cluster-wide setting; this is needed to
1382 make diagnose easier of failed drbd creations
1383 - Started adding base IPv6 support, but this is not yet
1384 enabled/available for use
1385 - Rename operations (cluster, instance) will now return the new name,
1386 which is especially useful if a short name was passed in
1387 - Added support for instance migration in RAPI
1388 - Added a tool to pre-configure nodes for the SSH setup, before joining
1389 them to the cluster; this will allow in the future a simplified model
1390 for node joining (but not yet fully enabled in 2.2); this needs the
1391 paramiko python library
1392 - Fixed handling of name-resolving errors
1393 - Fixed consistency of job results on the error path
1394 - Fixed master-failover race condition when executed multiple times in
1396 - Fixed many bugs related to the job queue (mostly introduced during the
1397 2.2 development cycle, so not all are impacting 2.1)
1398 - Fixed instance migration with missing disk symlinks
1399 - Fixed handling of unknown jobs in ``gnt-job archive``
1400 - And many other small fixes/improvements
1404 - Enhanced both the unittest and the QA coverage
1405 - Switched the opcode validation to a generic model, and extended the
1406 validation to all opcode parameters
1407 - Changed more parts of the code that write shell scripts to use the
1409 - Switched the master daemon to use the asyncore library for the Luxi
1416 *(Released Thu, 17 Jun 2010)*
1418 - Added tool (``move-instance``) and infrastructure to move instances
1419 between separate clusters (see :doc:`separate documentation
1420 <move-instance>` and :doc:`design document <design-2.2>`)
1421 - Added per-request RPC timeout
1422 - RAPI now requires a Content-Type header for requests with a body (e.g.
1423 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1424 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1425 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1427 - Implemented initial support for running Ganeti daemons as separate
1428 users, see configure-time flags ``--with-user-prefix`` and
1429 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1431 - Instances can be removed after export (``gnt-backup export
1432 --remove-instance``)
1433 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1434 key (instead of 1024 bit)
1435 - Added new cluster configuration file for cluster domain secret
1436 - Import/export now use SSL instead of SSH
1437 - Added support for showing estimated time when exporting an instance,
1438 see the ``ganeti-os-interface(7)`` manpage and look for
1445 *(Released Tue, 16 Nov 2010)*
1447 Some more bugfixes. Unless critical bugs occur, this will be the last
1450 - Fix case of MAC special-values
1451 - Fix mac checker regex
1452 - backend: Fix typo causing "out of range" error
1453 - Add missing --units in gnt-instance list man page
1459 *(Released Tue, 24 Aug 2010)*
1462 - Don't ignore secondary node silently on non-mirrored disk templates
1464 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1465 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1466 - Properly document the usb_mouse parameter
1467 - Fix path in ganeti-rapi(8) (issue 116)
1468 - Adjust error message when the ganeti user's .ssh directory is
1470 - Add same-node-check when changing the disk template to drbd
1476 *(Released Fri, 16 Jul 2010)*
1479 - Add an option to only select some reboot types during qa/burnin.
1480 (on some hypervisors consequent reboots are not supported)
1481 - Fix infrequent race condition in master failover. Sometimes the old
1482 master ip address would be still detected as up for a short time
1483 after it was removed, causing failover to fail.
1484 - Decrease mlockall warnings when the ctypes module is missing. On
1485 Python 2.4 we support running even if no ctypes module is installed,
1486 but we were too verbose about this issue.
1487 - Fix building on old distributions, on which man doesn't have a
1489 - Fix RAPI not to ignore the MAC address on instance creation
1490 - Implement the old instance creation format in the RAPI client.
1496 *(Released Thu, 01 Jul 2010)*
1498 A small bugfix release:
1499 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1500 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1501 a non-existing option
1502 - Fix instance create: the --force-variant option was ignored
1503 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1504 - Fix modify node storage with non-FQDN arguments
1505 - Fix RAPI client to authenticate under Python 2.6 when used
1506 for more than 5 requests needing authentication
1507 - Fix gnt-instance modify -t (storage) giving a wrong error message
1508 when converting a non-shutdown drbd instance to plain
1514 *(Released Fri, 18 Jun 2010)*
1516 A small bugfix release:
1518 - Fix live migration of KVM instances started with older Ganeti
1519 versions which had fewer hypervisor parameters
1520 - Fix gnt-instance grow-disk on down instances
1521 - Fix an error-reporting bug during instance migration
1522 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1523 silently ignoring broken ones
1524 - Fix an RPC error reporting bug affecting, for example, RAPI client
1526 - Fix bug triggered by different API version os-es on different nodes
1527 - Fix a bug in instance startup with custom hvparams: OS level
1528 parameters would fail to be applied.
1529 - Fix the RAPI client under Python 2.6 (but more work is needed to
1530 make it work completely well with OpenSSL)
1531 - Fix handling of errors when resolving names from DNS
1537 *(Released Thu, 3 Jun 2010)*
1539 A medium sized development cycle. Some new features, and some
1540 fixes/small improvements/cleanups.
1542 Significant features
1543 ~~~~~~~~~~~~~~~~~~~~
1545 The node deamon now tries to mlock itself into memory, unless the
1546 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1547 its logs, and falls back to console logging. This allows emergency
1548 features such as ``gnt-node powercycle`` to work even in the event of a
1549 broken node disk (tested offlining the disk hosting the node's
1550 filesystem and dropping its memory caches; don't try this at home)
1552 KVM: add vhost-net acceleration support. It can be tested with a new
1553 enough version of the kernel and of qemu-kvm.
1555 KVM: Add instance chrooting feature. If you use privilege dropping for
1556 your VMs you can also now force them to chroot to an empty directory,
1557 before starting the emulated guest.
1559 KVM: Add maximum migration bandwith and maximum downtime tweaking
1560 support (requires a new-enough version of qemu-kvm).
1562 Cluster verify will now warn if the master node doesn't have the master
1563 ip configured on it.
1565 Add a new (incompatible) instance creation request format to RAPI which
1566 supports all parameters (previously only a subset was supported, and it
1567 wasn't possible to extend the old format to accomodate all the new
1568 features. The old format is still supported, and a client can check for
1569 this feature, before using it, by checking for its presence in the
1570 ``features`` RAPI resource.
1572 Now with ancient latin support. Try it passing the ``--roman`` option to
1573 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1574 (requires the python-roman module to be installed, in order to work).
1579 As usual many internal code refactorings, documentation updates, and
1582 - Lots of improvements and cleanups to the experimental Remote API
1583 (RAPI) client library.
1584 - A new unit test suite for the core daemon libraries.
1585 - A fix to creating missing directories makes sure the umask is not
1586 applied anymore. This enforces the same directory permissions
1588 - Better handling terminating daemons with ctrl+c (used when running
1589 them in debugging mode).
1590 - Fix a race condition in live migrating a KVM instance, when stat()
1591 on the old proc status file returned EINVAL, which is an unexpected
1593 - Fixed manpage checking with newer man and utf-8 charachters. But now
1594 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1600 *(Released Fri, 7 May 2010)*
1602 Fix a bug which prevented untagged KVM instances from starting.
1608 *(Released Fri, 7 May 2010)*
1610 Another release with a long development cycle, during which many
1611 different features were added.
1613 Significant features
1614 ~~~~~~~~~~~~~~~~~~~~
1616 The KVM hypervisor now can run the individual instances as non-root, to
1617 reduce the impact of a VM being hijacked due to bugs in the
1618 hypervisor. It is possible to run all instances as a single (non-root)
1619 user, to manually specify a user for each instance, or to dynamically
1620 allocate a user out of a cluster-wide pool to each instance, with the
1621 guarantee that no two instances will run under the same user ID on any
1624 An experimental RAPI client library, that can be used standalone
1625 (without the other Ganeti libraries), is provided in the source tree as
1626 ``lib/rapi/client.py``. Note this client might change its interface in
1627 the future, as we iterate on its capabilities.
1629 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1630 replace the cluster's certificates and crypto keys. This might help in
1631 case they have been compromised, or have simply expired.
1633 A new disk option for instance creation has been added that allows one
1634 to "adopt" currently existing logical volumes, with data
1635 preservation. This should allow easier migration to Ganeti from
1636 unmanaged (or managed via other software) instances.
1638 Another disk improvement is the possibility to convert between redundant
1639 (DRBD) and plain (LVM) disk configuration for an instance. This should
1640 allow better scalability (starting with one node and growing the
1641 cluster, or shrinking a two-node cluster to one node).
1643 A new feature that could help with automated node failovers has been
1644 implemented: if a node sees itself as offline (by querying the master
1645 candidates), it will try to shutdown (hard) all instances and any active
1646 DRBD devices. This reduces the risk of duplicate instances if an
1647 external script automatically failovers the instances on such nodes. To
1648 enable this, the cluster parameter ``maintain_node_health`` should be
1649 enabled; in the future this option (per the name) will enable other
1650 automatic maintenance features.
1652 Instance export/import now will reuse the original instance
1653 specifications for all parameters; that means exporting an instance,
1654 deleting it and the importing it back should give an almost identical
1655 instance. Note that the default import behaviour has changed from
1656 before, where it created only one NIC; now it recreates the original
1659 Cluster verify has added a few new checks: SSL certificates validity,
1660 /etc/hosts consistency across the cluster, etc.
1665 As usual, many internal changes were done, documentation fixes,
1668 - Fixed cluster initialization with disabled cluster storage (regression
1669 introduced in 2.1.1)
1670 - File-based storage supports growing the disks
1671 - Fixed behaviour of node role changes
1672 - Fixed cluster verify for some corner cases, plus a general rewrite of
1673 cluster verify to allow future extension with more checks
1674 - Fixed log spamming by watcher and node daemon (regression introduced
1676 - Fixed possible validation issues when changing the list of enabled
1678 - Fixed cleanup of /etc/hosts during node removal
1679 - Fixed RAPI response for invalid methods
1680 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1681 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1682 from ide disks, etc.)
1683 - Allow OS changes without re-installation (to record a changed OS
1684 outside of Ganeti, or to allow OS renames)
1685 - Allow instance creation without OS installation (useful for example if
1686 the OS will be installed manually, or restored from a backup not in
1688 - Implemented option to make cluster ``copyfile`` use the replication
1690 - Added list of enabled hypervisors to ssconf (possibly useful for
1692 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1694 - A partial form of node re-IP is possible via node readd, which now
1695 allows changed node primary IP
1696 - Command line utilities now show an informational message if the job is
1698 - The logs of the master daemon now show the PID/UID/GID of the
1705 *(Released Fri, 12 Mar 2010)*
1707 During the 2.1.0 long release candidate cycle, a lot of improvements and
1708 changes have accumulated with were released later as 2.1.1.
1713 The node evacuate command (``gnt-node evacuate``) was significantly
1714 rewritten, and as such the IAllocator protocol was changed - a new
1715 request type has been added. This unfortunate change during a stable
1716 series is designed to improve performance of node evacuations; on
1717 clusters with more than about five nodes and which are well-balanced,
1718 evacuation should proceed in parallel for all instances of the node
1719 being evacuated. As such, any existing IAllocator scripts need to be
1720 updated, otherwise the above command will fail due to the unknown
1721 request. The provided "dumb" allocator has not been updated; but the
1722 ganeti-htools package supports the new protocol since version 0.2.4.
1724 Another important change is increased validation of node and instance
1725 names. This might create problems in special cases, if invalid host
1726 names are being used.
1728 Also, a new layer of hypervisor parameters has been added, that sits at
1729 OS level between the cluster defaults and the instance ones. This allows
1730 customisation of virtualization parameters depending on the installed
1731 OS. For example instances with OS 'X' may have a different KVM kernel
1732 (or any other parameter) than the cluster defaults. This is intended to
1733 help managing a multiple OSes on the same cluster, without manual
1734 modification of each instance's parameters.
1736 A tool for merging clusters, ``cluster-merge``, has been added in the
1737 tools sub-directory.
1742 - Improved the int/float conversions that should make the code more
1743 robust in face of errors from the node daemons
1744 - Fixed the remove node code in case of internal configuration errors
1745 - Fixed the node daemon behaviour in face of inconsistent queue
1746 directory (e.g. read-only file-system where we can't open the files
1748 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1749 now it either aborts cleanly or, if given the new "auto_promote"
1750 parameter, will automatically promote other nodes as needed
1751 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1752 completely prevent Ganeti from working
1753 - Fixed bug for instance export when not all disks were successfully
1755 - Fixed behaviour of node add when the new node is slow in starting up
1757 - Fixed handling of signals in the LUXI client, which should improve
1758 behaviour of command-line scripts
1759 - Added checks for invalid node/instance names in the configuration (now
1760 flagged during cluster verify)
1761 - Fixed watcher behaviour for disk activation errors
1762 - Fixed two potentially endless loops in http library, which led to the
1763 RAPI daemon hanging and consuming 100% CPU in some cases
1764 - Fixed bug in RAPI daemon related to hashed passwords
1765 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1767 - Enhanced compatibility with non-Debian OSes, but not using absolute
1768 path in some commands and allowing customisation of the ssh
1769 configuration directory
1770 - Fixed possible future issue with new Python versions by abiding to the
1771 proper use of ``__slots__`` attribute on classes
1772 - Added checks that should prevent directory traversal attacks
1773 - Many documentation fixes based on feedback from users
1778 - Added an "early_release" more for instance replace disks and node
1779 evacuate, where we release locks earlier and thus allow higher
1780 parallelism within the cluster
1781 - Added watcher hooks, intended to allow the watcher to restart other
1782 daemons (e.g. from the ganeti-nbma project), but they can be used of
1783 course for any other purpose
1784 - Added a compile-time disable for DRBD barriers, to increase
1785 performance if the administrator trusts the power supply or the
1786 storage system to not lose writes
1787 - Added the option of using syslog for logging instead of, or in
1788 addition to, Ganeti's own log files
1789 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1790 can indeed boot from a paravirtual NIC
1791 - Added a generic debug level for many operations; while this is not
1792 used widely yet, it allows one to pass the debug value all the way to
1794 - Enhanced the hooks environment for instance moves (failovers,
1795 migrations) where the primary/secondary nodes changed during the
1796 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1797 - Enhanced data validations for many user-supplied values; one important
1798 item is the restrictions imposed on instance and node names, which
1799 might reject some (invalid) host names
1800 - Add a configure-time option to disable file-based storage, if it's not
1801 needed; this allows greater security separation between the master
1802 node and the other nodes from the point of view of the inter-node RPC
1804 - Added user notification in interactive tools if job is waiting in the
1805 job queue or trying to acquire locks
1806 - Added log messages when a job is waiting for locks
1807 - Added filtering by node tags in instance operations which admit
1808 multiple instances (start, stop, reboot, reinstall)
1809 - Added a new tool for cluster mergers, ``cluster-merge``
1810 - Parameters from command line which are of the form ``a=b,c=d`` can now
1811 use backslash escapes to pass in values which contain commas,
1812 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1814 - For KVM, the instance name is the first parameter passed to KVM, so
1815 that it's more visible in the process list
1821 *(Released Tue, 2 Mar 2010)*
1823 Ganeti 2.1 brings many improvements with it. Major changes:
1825 - Added infrastructure to ease automated disk repairs
1826 - Added new daemon to export configuration data in a cheaper way than
1827 using the remote API
1828 - Instance NICs can now be routed instead of being associated with a
1830 - Improved job locking logic to reduce impact of jobs acquiring multiple
1831 locks waiting for other long-running jobs
1833 In-depth implementation details can be found in the Ganeti 2.1 design
1839 - Added chroot hypervisor
1840 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1842 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1843 ``bootloader_path`` and ``bootloader_args``)
1844 - Added the ``use_localtime`` option for the xen-hvm and kvm
1845 hypervisors, and the default value for this has changed to false (in
1846 2.0 xen-hvm always enabled it)
1847 - Added luxi call to submit multiple jobs in one go
1848 - Added cluster initialization option to not modify ``/etc/hosts``
1850 - Added network interface parameters
1851 - Added dry run mode to some LUs
1852 - Added RAPI resources:
1854 - ``/2/instances/[instance_name]/info``
1855 - ``/2/instances/[instance_name]/replace-disks``
1856 - ``/2/nodes/[node_name]/evacuate``
1857 - ``/2/nodes/[node_name]/migrate``
1858 - ``/2/nodes/[node_name]/role``
1859 - ``/2/nodes/[node_name]/storage``
1860 - ``/2/nodes/[node_name]/storage/modify``
1861 - ``/2/nodes/[node_name]/storage/repair``
1863 - Added OpCodes to evacuate or migrate all instances on a node
1864 - Added new command to list storage elements on nodes (``gnt-node
1865 list-storage``) and modify them (``gnt-node modify-storage``)
1866 - Added new ssconf files with master candidate IP address
1867 (``ssconf_master_candidates_ips``), node primary IP address
1868 (``ssconf_node_primary_ips``) and node secondary IP address
1869 (``ssconf_node_secondary_ips``)
1870 - Added ``ganeti-confd`` and a client library to query the Ganeti
1871 configuration via UDP
1872 - Added ability to run hooks after cluster initialization and before
1874 - Added automatic mode for disk replace (``gnt-instance replace-disks
1876 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1877 after catastrophic data-loss
1878 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1880 - Added ``gnt-instance move`` command to move instances
1881 - Added ``gnt-cluster watcher`` command to control watcher
1882 - Added ``gnt-node powercycle`` command to powercycle nodes
1883 - Added new job status field ``lock_status``
1884 - Added parseable error codes to cluster verification (``gnt-cluster
1885 verify --error-codes``) and made output less verbose (use
1886 ``--verbose`` to restore previous behaviour)
1887 - Added UUIDs to the main config entities (cluster, nodes, instances)
1888 - Added support for OS variants
1889 - Added support for hashed passwords in the Ganeti remote API users file
1891 - Added option to specify maximum timeout on instance shutdown
1892 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1893 - Added new helper script to start and stop Ganeti daemons
1894 (``daemon-util``), with the intent to reduce the work necessary to
1895 adjust Ganeti for non-Debian distributions and to start/stop daemons
1897 - Added more unittests
1898 - Fixed critical bug in ganeti-masterd startup
1899 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1900 now customisable at the cluster level for both the KVM and Xen
1901 hypervisors using the new ``migration_port`` parameter
1902 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1903 reinstalling an instance
1904 - Allowed ``@`` in tag names
1905 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1906 - Many documentation updates
1907 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1908 - ``gnt-instance reinstall`` can now reinstall multiple instances
1909 - Updated many command line parameters
1910 - Introduced new OS API version 15
1911 - No longer support a default hypervisor
1912 - Treat virtual LVs as inexistent
1913 - Improved job locking logic to reduce lock contention
1914 - Match instance and node names case insensitively
1915 - Reimplemented bash completion script to be more complete
1922 *(Released Thu, 4 Feb 2010)*
1924 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1925 - Fix a string formatting bug
1926 - Improve safety of the code in some error paths
1927 - Improve data validation in the master of values returned from nodes
1933 *(Released Thu, 17 Dec 2009)*
1935 - Fix security issue due to missing validation of iallocator names; this
1936 allows local and remote execution of arbitrary executables
1937 - Fix failure of gnt-node list during instance removal
1938 - Ship the RAPI documentation in the archive
1944 *(Released Wed, 30 Sep 2009)*
1946 - Fixed many wrong messages
1947 - Fixed a few bugs related to the locking library
1948 - Fixed MAC checking at instance creation time
1949 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1950 - Fixed a few issues related to signal handling in both daemons and
1952 - Fixed the example startup script provided
1953 - Fixed insserv dependencies in the example startup script (patch from
1955 - Fixed handling of drained nodes in the iallocator framework
1956 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1958 - Fixed error related to invalid job IDs in job polling
1959 - Fixed job/opcode persistence on unclean master shutdown
1960 - Fixed handling of partial job processing after unclean master
1962 - Fixed error reporting from LUs, previously all errors were converted
1963 into execution errors
1964 - Fixed error reporting from burnin
1965 - Decreased significantly the memory usage of the job queue
1966 - Optimised slightly multi-job submission
1967 - Optimised slightly opcode loading
1968 - Backported the multi-job submit framework from the development
1969 branch; multi-instance start and stop should be faster
1970 - Added script to clean archived jobs after 21 days; this will reduce
1971 the size of the queue directory
1972 - Added some extra checks in disk size tracking
1973 - Added an example ethers hook script
1974 - Added a cluster parameter that prevents Ganeti from modifying of
1976 - Added more node information to RAPI responses
1977 - Added a ``gnt-job watch`` command that allows following the ouput of a
1979 - Added a bind-address option to ganeti-rapi
1980 - Added more checks to the configuration verify
1981 - Enhanced the burnin script such that some operations can be retried
1983 - Converted instance reinstall to multi-instance model
1989 *(Released Fri, 7 Aug 2009)*
1991 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1992 to allow using the pre-2.0.2 behaviour in activation, if any existing
1993 instances have mismatched disk sizes in the configuration
1994 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
1995 any configuration mismatches for disk sizes
1996 - Added ``gnt-master cluste-failover --no-voting`` to allow master
1997 failover to work on two-node clusters
1998 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2000 - Fixed detection of OS script errors in ``gnt-backup export``
2001 - Fixed exit code of ``gnt-backup export``
2007 *(Released Fri, 17 Jul 2009)*
2009 - Added experimental support for stripped logical volumes; this should
2010 enhance performance but comes with a higher complexity in the block
2011 device handling; stripping is only enabled when passing
2012 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2013 affected even in the non-stripped mode
2014 - Improved resiliency against transient failures at the end of DRBD
2015 resyncs, and in general of DRBD resync checks
2016 - Fixed a couple of issues with exports and snapshot errors
2017 - Fixed a couple of issues in instance listing
2018 - Added display of the disk size in ``gnt-instance info``
2019 - Fixed checking for valid OSes in instance creation
2020 - Fixed handling of the "vcpus" parameter in instance listing and in
2021 general of invalid parameters
2022 - Fixed http server library, and thus RAPI, to handle invalid
2023 username/password combinations correctly; this means that now they
2024 report unauthorized for queries too, not only for modifications,
2025 allowing earlier detect of configuration problems
2026 - Added a new "role" node list field, equivalent to the master/master
2027 candidate/drained/offline flags combinations
2028 - Fixed cluster modify and changes of candidate pool size
2029 - Fixed cluster verify error messages for wrong files on regular nodes
2030 - Fixed a couple of issues with node demotion from master candidate role
2031 - Fixed node readd issues
2032 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2033 - Added a new ``--no-voting`` option for masterfailover to fix failover
2034 on two-nodes clusters when the former master node is unreachable
2035 - Added instance reinstall over RAPI
2041 *(Released Tue, 16 Jun 2009)*
2043 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2044 allow re-adding the start in single-user option (regression from 1.2)
2045 - the watcher writes the instance status to a file, to allow monitoring
2046 to report the instance status (from the master) based on cached
2047 results of the watcher's queries; while this can get stale if the
2048 watcher is being locked due to other work on the cluster, this is
2049 still an improvement
2050 - the watcher now also restarts the node daemon and the rapi daemon if
2052 - fixed the watcher to handle full and drained queue cases
2053 - hooks export more instance data in the environment, which helps if
2054 hook scripts need to take action based on the instance's properties
2055 (no longer need to query back into ganeti)
2056 - instance failovers when the instance is stopped do not check for free
2057 RAM, so that failing over a stopped instance is possible in low memory
2059 - rapi uses queries for tags instead of jobs (for less job traffic), and
2060 for cluster tags it won't talk to masterd at all but read them from
2062 - a couple of error handling fixes in RAPI
2063 - drbd handling: improved the error handling of inconsistent disks after
2064 resync to reduce the frequency of "there are some degraded disks for
2065 this instance" messages
2066 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2067 error handling path called a wrong function name)
2073 *(Released Wed, 27 May 2009)*
2075 - no changes from rc5
2081 *(Released Wed, 20 May 2009)*
2083 - fix a couple of bugs (validation, argument checks)
2084 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2085 - some small improvements to RAPI and IAllocator
2086 - make watcher automatically start the master daemon if down
2092 *(Released Mon, 27 Apr 2009)*
2094 - change the OS list to not require locks; this helps with big clusters
2095 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2096 volume group is broken
2097 - ``gnt-instance info``, without any arguments, doesn't run for all
2098 instances anymore; either pass ``--all`` or pass the desired
2099 instances; this helps against mistakes on big clusters where listing
2100 the information for all instances takes a long time
2101 - miscellaneous doc and man pages fixes
2107 *(Released Wed, 8 Apr 2009)*
2109 - Change the internal locking model of some ``gnt-node`` commands, in
2110 order to reduce contention (and blocking of master daemon) when
2111 batching many creation/reinstall jobs
2112 - Fixes to Xen soft reboot
2113 - No longer build documentation at build time, instead distribute it in
2114 the archive, in order to reduce the need for the whole docbook/rst
2121 *(Released Fri, 27 Mar 2009)*
2123 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2124 - Fix watcher startup sequence, improves the behaviour of busy clusters
2125 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2126 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2127 bind address and other places
2128 - Some documentation fixes and updates
2134 *(Released Mon, 2 Mar 2009)*
2136 - More documentation updates, now all docs should be more-or-less
2138 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2140 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2142 - Fix an issue related to $libdir/run/ganeti and cluster creation
2148 *(Released Thu, 19 Feb 2009)*
2150 - Xen PVM and KVM have switched the default value for the instance root
2151 disk to the first partition on the first drive, instead of the whole
2152 drive; this means that the OS installation scripts must be changed
2154 - Man pages have been updated
2155 - RAPI has been switched by default to HTTPS, and the exported functions
2156 should all work correctly
2157 - RAPI v1 has been removed
2158 - Many improvements to the KVM hypervisor
2159 - Block device errors are now better reported
2160 - Many other bugfixes and small improvements
2166 *(Released Mon, 26 Jan 2009)*
2168 - Version 2 is a general rewrite of the code and therefore the
2169 differences are too many to list, see the design document for 2.0 in
2170 the ``doc/`` subdirectory for more details
2171 - In this beta version there is not yet a migration path from 1.2 (there
2172 will be one in the final 2.0 release)
2173 - A few significant changes are:
2175 - all commands are executed by a daemon (``ganeti-masterd``) and the
2176 various ``gnt-*`` commands are just front-ends to it
2177 - all the commands are entered into, and executed from a job queue,
2178 see the ``gnt-job(8)`` manpage
2179 - the RAPI daemon supports read-write operations, secured by basic
2180 HTTP authentication on top of HTTPS
2181 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2182 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2183 to migrate to DRBD 8 first while still running Ganeti 1.2)
2184 - DRBD devices are using statically allocated minor numbers, which
2185 will be assigned to existing instances during the migration process
2186 - there is support for both Xen PVM and Xen HVM instances running on
2188 - KVM virtualization is supported too
2189 - file-based storage has been implemented, which means that it is
2190 possible to run the cluster without LVM and DRBD storage, for
2191 example using a shared filesystem exported from shared storage (and
2192 still have live migration)
2198 *(Released Tue, 13 Jan 2009)*
2200 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2201 - Reuse the old instance mac address by default on instance import, if
2202 the instance name is the same.
2203 - Handle situations in which the node info rpc returns incomplete
2205 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2206 - Improved version of batcher:
2208 - state file support
2209 - instance mac address support
2210 - support for HVM clusters/instances
2212 - Add an option to show the number of cpu sockets and nodes in
2214 - Support OSes that handle more than one version of the OS api (but do
2215 not change the current API in any other way)
2216 - Fix ``gnt-node migrate``
2217 - ``gnt-debug`` man page
2218 - Fixes various more typos and small issues
2219 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2225 *(Released Wed, 24 Sep 2008)*
2227 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2228 type of disk exported to fully virtualized instances.
2229 - provide access to the serial console of HVM instances
2230 - instance auto_balance flag, set by default. If turned off it will
2231 avoid warnings on cluster verify if there is not enough memory to fail
2232 over an instance. in the future it will prevent automatically failing
2233 it over when we will support that.
2234 - batcher tool for instance creation, see ``tools/README.batcher``
2235 - ``gnt-instance reinstall --select-os`` to interactively select a new
2236 operating system when reinstalling an instance.
2237 - when changing the memory amount on instance modify a check has been
2238 added that the instance will be able to start. also warnings are
2239 emitted if the instance will not be able to fail over, if auto_balance
2241 - documentation fixes
2242 - sync fields between ``gnt-instance list/modify/add/import``
2243 - fix a race condition in drbd when the sync speed was set after giving
2244 the device a remote peer.
2250 *(Released Tue, 22 Jul 2008)*
2252 - note: the allowed size and number of tags per object were reduced
2253 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2254 - fixed twisted 8.x compatibility
2255 - fixed ``gnt-instance replace-disks`` with iallocator
2256 - add TCP keepalives on twisted connections to detect restarted nodes
2257 - disk increase support, see ``gnt-instance grow-disk``
2258 - implement bulk node/instance query for RAPI
2259 - add tags in node/instance listing (optional)
2260 - experimental migration (and live migration) support, read the man page
2261 for ``gnt-instance migrate``
2262 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2263 has some small improvements in handling its state file
2269 *(Released Fri, 13 Jun 2008)*
2271 - Experimental readonly, REST-based remote API implementation;
2272 automatically started on master node, TCP port 5080, if enabled by
2273 ``--enable-rapi`` parameter to configure script.
2274 - Instance allocator support. Add and import instance accept a
2275 ``--iallocator`` parameter, and call that instance allocator to decide
2276 which node to use for the instance. The iallocator document describes
2277 what's expected from an allocator script.
2278 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2279 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2280 node is lost there is still enough memory to fail over the instances
2282 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2283 ``gnt-cluster verify``, to check for site-specific compliance. All the
2284 hooks will run, and their output, if any, will be displayed. Any
2285 failing hook will make the verification return an error value.
2286 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2287 primary and secondary interfaces
2288 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2289 that is still declared as part of the cluster and has failed.
2290 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2291 specifying output fields, that just adds the chosen fields to the
2293 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2294 export from the filesystem.
2295 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2296 have been added. Using them you can enable/disable acpi and pae
2297 support, and specify a path for a cd image to be exported to the
2298 instance. These parameters as the name suggest only work on HVM
2300 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2301 PAE support will be set to the previously hardcoded values, but the
2302 (previously hardcoded) path to the CDROM ISO image will be unset and
2303 if required, needs to be set manually with ``gnt-instance modify``
2305 - The address to which an instance's VNC console is bound is now
2306 selectable per-instance, rather than being cluster wide. Of course
2307 this only applies to instances controlled via VNC, so currently just
2308 applies to HVM clusters.
2314 *(Released Mon, 18 Feb 2008)*
2316 - more tweaks to the disk activation code (especially helpful for DRBD)
2317 - change the default ``gnt-instance list`` output format, now there is
2318 one combined status field (see the manpage for the exact values this
2320 - some more fixes for the mac export to hooks change
2321 - make Ganeti not break with DRBD 8.2.x (which changed the version
2322 format in ``/proc/drbd``) (issue 24)
2323 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2324 template, allowing migration from DRBD0.7+MD to DRBD8
2330 *(Released Wed, 30 Jan 2008)*
2332 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2334 - add command aliases infrastructure and a few aliases
2335 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2336 man pages and the ``--help`` option of ``gnt-node
2337 list``/``gnt-instance list``
2338 - fix ``gnt-backup list`` with down nodes (issue 21)
2339 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2340 - fix the dist archive and add a check for including svn/git files in
2342 - some developer-related changes: improve the burnin and the QA suite,
2343 add an upload script for testing during development
2349 *(Released Wed, 16 Jan 2008)*
2351 - experimental HVM support, read the install document, section
2352 "Initializing the cluster"
2353 - allow for the PVM hypervisor per-instance kernel and initrd paths
2354 - add a new command ``gnt-cluster verify-disks`` which uses a new
2355 algorithm to improve the reconnection of the DRBD pairs if the device
2356 on the secondary node has gone away
2357 - make logical volume code auto-activate LVs at disk activation time
2358 - slightly improve the speed of activating disks
2359 - allow specification of the MAC address at instance creation time, and
2360 changing it later via ``gnt-instance modify``
2361 - fix handling of external commands that generate lots of output on
2363 - update documentation with regard to minimum version of DRBD8 supported
2369 *(Released Tue, 4 Dec 2007)*
2371 - Log the ``xm create`` output to the node daemon log on failure (to
2372 help diagnosing the error)
2373 - In debug mode, log all external commands output if failed to the logs
2374 - Change parsing of lvm commands to ignore stderr
2380 *(Released Wed, 28 Nov 2007)*
2382 - Another round of updates to the DRBD 8 code to deal with more failures
2383 in the replace secondary node operation
2384 - Some more logging of failures in disk operations (lvm, drbd)
2385 - A few documentation updates
2392 *(Released Tue, 13 Nov 2007)*
2394 - Change configuration file format from Python's Pickle to JSON.
2395 Upgrading is possible using the cfgupgrade utility.
2396 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2397 faster replace disks and is more stable (DRBD 8 has many improvements
2398 compared to DRBD 0.7)
2399 - Added command line tags support (see man pages for ``gnt-instance``,
2400 ``gnt-node``, ``gnt-cluster``)
2401 - Added instance rename support
2402 - Added multi-instance startup/shutdown
2403 - Added cluster rename support
2404 - Added ``gnt-node evacuate`` to simplify some node operations
2405 - Added instance reboot operation that can speedup reboot as compared to
2407 - Soften the requirement that hostnames are in FQDN format
2408 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2410 - Removed dependency on debian's patched fping that uses the
2411 non-standard ``-S`` option
2412 - Now the OS definitions are searched for in multiple, configurable
2413 paths (easier for distros to package)
2414 - Some changes to the hooks infrastructure (especially the new
2415 post-configuration update hook)
2416 - Other small bugfixes
2418 .. vim: set textwidth=72 syntax=rst :