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 now to be listed in
15 ``$sysconfdir/ganeti/file-storage-paths``. ``cfgupgrade`` will write
16 the file automatically based on old configuration values, but it can
17 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(8)` has been slightly changed: now empty
27 lines and comment lines are ignored for better readability.
28 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced.
29 ``gnt-node add`` now invokes a new tool on the destination node, named
30 ``prepare-node-join``, to configure the SSH daemon. Paramiko is no
31 longer necessary to configure nodes' SSH daemons via ``gnt-node add``.
37 *(Released Fri, 12 Oct 2012)*
39 A small bugfix release.
41 Fix double use of PRIORITY_OPT in gnt-node migrate, that would make the
44 Commands that issue many jobs don't fail anymore just because some jobs
45 take so long that other jobs are archived.
47 Failures during gnt-instance reinstall are reflected by the exit status.
49 Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
52 When always_failover is set, --allow-failover is not required in migrate
55 bash_completion works even if extglob is disabled
57 Fix bug with locks that made failover for RDB-based instances fail.
59 Fix bug in non-mirrored instance allocation that would make Ganeti
60 choose a random node instead of one based on the allocator metric.
62 Support for newer versions of pylint and pep8.
64 Hail doesn't fail anymore when trying to add an instance of type
65 'file', 'sharedfile' or 'rbd'.
67 Add new Makefile target to rebuild the whole dist, so that all files are
74 *(Released Fri, 27 Jul 2012)*
77 .. attention:: The ``LUXI`` protocol has been made more consistent
78 regarding its handling of command arguments. This, however, leads to
79 incompatibility issues with previous versions. Please ensure that you
80 restart Ganeti daemons soon after the upgrade, otherwise most
81 ``LUXI`` calls (job submission, setting/resetting the drain flag,
82 pausing/resuming the watcher, cancelling and archiving jobs, querying
83 the cluster configuration) will fail.
92 The current ``admin_up`` field, which used to denote whether an instance
93 should be running or not, has been removed. Instead, ``admin_state`` is
94 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
96 The rational behind this is that an instance being “down” can have
99 - it could be down during a reboot
100 - it could be temporarily be down for a reinstall
101 - or it could be down because it is deprecated and kept just for its
104 The previous Boolean state was making it difficult to do capacity
105 calculations: should Ganeti reserve memory for a down instance? Now, the
106 tri-state field makes it clear:
108 - in ``up`` and ``down`` state, all resources are reserved for the
109 instance, and it can be at any time brought up if it is down
110 - in ``offline`` state, only disk space is reserved for it, but not
113 The field can have an extra use: since the transition between ``up`` and
114 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
115 transition between ``offline`` and ``down`` is done via ``gnt-instance
116 modify``, it is possible to given different rights to users. For
117 example, owners of an instance could be allowed to start/stop it, but
118 not transition it out of the offline state.
120 Instance policies and specs
121 +++++++++++++++++++++++++++
123 In previous Ganeti versions, an instance creation request was not
124 limited on the minimum size and on the maximum size just by the cluster
125 resources. As such, any policy could be implemented only in third-party
126 clients (RAPI clients, or shell wrappers over ``gnt-*``
127 tools). Furthermore, calculating cluster capacity via ``hspace`` again
128 required external input with regards to instance sizes.
130 In order to improve these workflows and to allow for example better
131 per-node group differentiation, we introduced instance specs, which
134 - minimum instance disk size, disk count, memory size, cpu count
135 - maximum values for the above metrics
136 - and “standard” values (used in ``hspace`` to calculate the standard
139 The minimum/maximum values can be also customised at node-group level,
140 for example allowing more powerful hardware to support bigger instance
143 Beside the instance specs, there are a few other settings belonging to
144 the instance policy framework. It is possible now to customise, per
145 cluster and node-group:
147 - the list of allowed disk templates
148 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
149 - the maximum ratio of instance to spindles (see below for more
150 information) for local storage
152 All these together should allow all tools that talk to Ganeti to know
153 what are the ranges of allowed values for instances and the
154 over-subscription that is allowed.
156 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
157 instance configuration, and the physical CPU configuration from the
158 node. For the spindle ratios however, we didn't track before these
159 values, so new parameters have been added:
161 - a new node parameter ``spindle_count``, defaults to 1, customisable at
162 node group or node level
163 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
165 Note that spindles in this context doesn't need to mean actual
166 mechanical hard-drives; it's just a relative number for both the node
167 I/O capacity and instance I/O consumption.
169 Instance migration behaviour
170 ++++++++++++++++++++++++++++
172 While live-migration is in general desirable over failover, it is
173 possible that for some workloads it is actually worse, due to the
174 variable time of the “suspend” phase during live migration.
176 To allow the tools to work consistently over such instances (without
177 having to hard-code instance names), a new backend parameter
178 ``always_failover`` has been added to control the migration/failover
179 behaviour. When set to True, all migration requests for an instance will
180 instead fall-back to failover.
182 Instance memory ballooning
183 ++++++++++++++++++++++++++
185 Initial support for memory ballooning has been added. The memory for an
186 instance is no longer fixed (backend parameter ``memory``), but instead
187 can vary between minimum and maximum values (backend parameters
188 ``minmem`` and ``maxmem``). Currently we only change an instance's
191 - live migrating or failing over and instance and the target node
192 doesn't have enough memory
193 - user requests changing the memory via ``gnt-instance modify
199 In order to control the use of specific CPUs by instance, support for
200 controlling CPU pinning has been added for the Xen, HVM and LXC
201 hypervisors. This is controlled by a new hypervisor parameter
202 ``cpu_mask``; details about possible values for this are in the
203 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
204 VCPU-to-CPU mapping) form will work well only when all nodes in your
205 cluster have the same amount of CPUs.
210 Another area in which Ganeti was not customisable were the parameters
211 used for storage configuration, e.g. how many stripes to use for LVM,
212 DRBD resync configuration, etc.
214 To improve this area, we've added disks parameters, which are
215 customisable at cluster and node group level, and which allow to
216 specify various parameters for disks (DRBD has the most parameters
217 currently), for example:
219 - DRBD resync algorithm and parameters (e.g. speed)
220 - the default VG for meta-data volumes for DRBD
221 - number of stripes for LVM (plain disk template)
224 These parameters can be modified via ``gnt-cluster modify -D …`` and
225 ``gnt-group modify -D …``, and are used at either instance creation (in
226 case of LVM stripes, for example) or at disk “activation” time
229 Rados block device support
230 ++++++++++++++++++++++++++
232 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
233 denoted by the ``rbd`` disk template type. This is considered
234 experimental, feedback is welcome. For details on configuring it, see
235 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
240 The existing master IP functionality works well only in simple setups (a
241 single network shared by all nodes); however, if nodes belong to
242 different networks, then the ``/32`` setup and lack of routing
243 information is not enough.
245 To allow the master IP to function well in more complex cases, the
246 system was reworked as follows:
248 - a master IP netmask setting has been added
249 - the master IP activation/turn-down code was moved from the node daemon
251 - whether to run the Ganeti-supplied master IP script or a user-supplied
252 on is a ``gnt-cluster init`` setting
254 Details about the location of the standard and custom setup scripts are
255 in the man page :manpage:`gnt-cluster(8)`; for information about the
256 setup script protocol, look at the Ganeti-supplied script.
261 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
264 It is now possible to use TLS-protected connections, and when renewing
265 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
266 it is now possible to specify spice or spice CA certificates. Also, it
267 is possible to configure a password for SPICE sessions via the
268 hypervisor parameter ``spice_password_file``.
270 There are also new parameters to control the compression and streaming
271 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
272 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
273 for the spice parameters.
275 Lastly, it is now possible to see the SPICE connection information via
276 ``gnt-instance console``.
281 A new tool (``tools/ovfconverter``) has been added that supports
282 conversion between Ganeti and the `Open Virtualization Format
283 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
286 This relies on the ``qemu-img`` tool to convert the disk formats, so the
287 actual compatibility with other virtualization solutions depends on it.
292 The configuration query daemon (``ganeti-confd``) is now optional, and
293 has been rewritten in Haskell; whether to use the daemon at all, use the
294 Python (default) or the Haskell version is selectable at configure time
295 via the ``--enable-confd`` parameter, which can take one of the
296 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
297 daemon will result in a smaller footprint; for larger systems, we
298 welcome feedback on the Haskell version which might become the default
301 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
302 daemon running. The Python version doesn't implement the new call.
305 User interface changes
306 ~~~~~~~~~~~~~~~~~~~~~~
308 We have replaced the ``--disks`` option of ``gnt-instance
309 replace-disks`` with a more flexible ``--disk`` option, which allows
310 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
311 disk size and mode can be changed upon recreation (via ``gnt-instance
312 recreate-disks``, which accepts the same ``--disk`` option).
314 As many people are used to a ``show`` command, we have added that as an
315 alias to ``info`` on all ``gnt-*`` commands.
317 The ``gnt-instance grow-disk`` command has a new mode in which it can
318 accept the target size of the disk, instead of the delta; this can be
319 more safe since two runs in absolute mode will be idempotent, and
320 sometimes it's also easier to specify the desired size directly.
322 Also the handling of instances with regard to offline secondaries has
323 been improved. Instance operations should not fail because one of it's
324 secondary nodes is offline, even though it's safe to proceed.
326 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
327 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
328 minors to instance name.
333 RAPI coverage has improved, with (for example) new resources for
334 recreate-disks, node power-cycle, etc.
339 There is partial support for ``xl`` in the Xen hypervisor; feedback is
342 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
343 whether to still support Python 2.4 or move to Python 2.6 as minimum
346 Support for Fedora has been slightly improved; the provided example
347 init.d script should work better on it and the INSTALL file should
348 document the needed dependencies.
353 The deprecated ``QueryLocks`` LUXI request has been removed. Use
354 ``Query(what=QR_LOCK, ...)`` instead.
356 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
357 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
358 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
359 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
360 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
362 RAPI client: ``CertificateError`` now derives from
363 ``GanetiApiError``. This should make it more easy to handle Ganeti
366 Deprecation warnings due to PyCrypto/paramiko import in
367 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
368 make sure to run an up-to-date paramiko version, if you use this tool.
370 The QA scripts now depend on Python 2.5 or above (the main code base
371 still works with Python 2.4).
373 The configuration file (``config.data``) is now written without
374 indentation for performance reasons; if you want to edit it, it can be
375 re-formatted via ``tools/fmtjson``.
377 A number of bugs has been fixed in the cluster merge tool.
379 ``x509`` certification verification (used in import-export) has been
380 changed to allow the same clock skew as permitted by the cluster
381 verification. This will remove some rare but hard to diagnose errors in
388 *(Released Thu, 19 Jul 2012)*
390 Very few changes from rc4 to the final release, only bugfixes:
392 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
393 instance, fix CDROM booting for KVM instances)
394 - fixed node group modification of node parameters
395 - fixed issue in LUClusterVerifyGroup with multi-group clusters
396 - fixed generation of bash completion to ensure a stable ordering
403 *(Released Fri, 13 Jul 2012)*
405 Third release candidate for 2.6. The following changes were done from
408 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
409 - Fixed an inconsistency in the LUXI protocol with the provided
410 arguments (NOT backwards compatible)
411 - Fixed a bug with node groups ipolicy where ``min`` was greater than
412 the cluster ``std`` value
413 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
414 easier instance debugging on nodes (requires ``hconfd`` to work)
420 *(Released Tue, 03 Jul 2012)*
422 Second release candidate for 2.6. The following changes were done from
425 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
427 - Fixed a RAPI regression on missing beparams/memory
428 - Fixed redistribution of files on offline nodes
429 - Added possibility to run activate-disks even though secondaries are
430 offline. With this change it relaxes also the strictness on some other
431 commands which use activate disks internally:
432 * ``gnt-instance start|reboot|rename|backup|export``
433 - Made it possible to remove safely an instance if its secondaries are
435 - Made it possible to reinstall even though secondaries are offline
441 *(Released Mon, 25 Jun 2012)*
443 First release candidate for 2.6. The following changes were done from
446 - Fixed bugs with disk parameters and ``rbd`` templates as well as
448 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
449 behaviour. It supports now the modify operation
450 - ``hcheck`` implemented to analyze cluster health and possibility of
451 improving health by rebalance
452 - ``hbal`` has been improved in dealing with split instances
458 *(Released Mon, 11 Jun 2012)*
460 Second beta release of 2.6. The following changes were done from beta2
463 - Fixed ``daemon-util`` with non-root user models
464 - Fixed creation of plain instances with ``--no-wait-for-sync``
465 - Fix wrong iv_names when running ``cfgupgrade``
466 - Export more information in RAPI group queries
467 - Fixed bug when changing instance network interfaces
468 - Extended burnin to do NIC changes
469 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
470 - Changed default for DRBD barriers
471 - Fixed DRBD error reporting for syncer rate
472 - Verify the options on disk parameters
474 And of course various fixes to documentation and improved unittests and
481 *(Released Wed, 23 May 2012)*
483 First beta release of 2.6. The following changes were done from beta1 to
486 - integrated patch for distributions without ``start-stop-daemon``
487 - adapted example init.d script to work on Fedora
488 - fixed log handling in Haskell daemons
489 - adapted checks in the watcher for pycurl linked against libnss
490 - add partial support for ``xl`` instead of ``xm`` for Xen
491 - fixed a type issue in cluster verification
492 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
495 Plus integrated fixes from the 2.5 branch:
497 - fixed ``kvm-ifup`` to use ``/bin/bash``
498 - fixed parallel build failures
499 - KVM live migration when using a custom keymap
505 *(Released Tue, 24 Jul 2012)*
507 A small bugfix release, with no new features:
509 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
510 different default shell (e.g. Debian, Ubuntu)
511 - fixed KVM startup and live migration with a custom keymap (fixes Issue
512 243 and Debian bug #650664)
513 - fixed compatibility with KVM versions that don't support multiple boot
514 devices (fixes Issue 230 and Debian bug #624256)
516 Additionally, a few fixes were done to the build system (fixed parallel
517 build failures) and to the unittests (fixed race condition in test for
518 FileID functions, and the default enable/disable mode for QA test is now
525 *(Released Fri, 11 May 2012)*
527 A small bugfix release.
529 The main issues solved are on the topic of compatibility with newer LVM
532 - fixed parsing of ``lv_attr`` field
533 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
534 the ``--force`` flag is needed
536 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
537 to accept kernel 3.x too (was hardcoded to 2.6.*).
539 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
540 of hook results and that also made display of other errors suboptimal
541 was fixed; the code behaves now like 2.4 and earlier.
543 Another change in 2.5, the cleanup of the OS scripts environment, is too
544 aggressive: it removed even the ``PATH`` variable, which requires the OS
545 scripts to *always* need to export it. Since this is a bit too strict,
546 we now export a minimal PATH, the same that we export for hooks.
548 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
549 integrated into this release.
551 Finally, a few other miscellaneous changes were done (no new features,
552 just small improvements):
554 - Fix ``gnt-group --help`` display
555 - Fix hardcoded Xen kernel path
556 - Fix grow-disk handling of invalid units
557 - Update synopsis for ``gnt-cluster repair-disk-sizes``
558 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
564 *(Released Thu, 12 Apr 2012)*
566 Incompatible/important changes and bugfixes
567 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
569 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
570 resource's ``ip_check`` parameter changed from ``True`` to ``False``
571 to match the underlying LUXI interface.
572 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
573 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
574 not maintain backwards-compatibility as the underlying operation
575 changed in an incompatible way. The RAPI client can talk to old
576 servers, but it needs to be told so as the return value changed.
577 - When creating file-based instances via RAPI, the ``file_driver``
578 parameter no longer defaults to ``loop`` and must be specified.
579 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
581 - Support for the undocumented and deprecated RAPI instance creation
582 request format version 0 has been dropped. Use version 1, supported
583 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
584 - Pyparsing 1.4.6 or above is required, see :doc:`installation
585 documentation <install>`.
586 - The "cluster-verify" hooks are now executed per group by the
587 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
588 if you just run ``gnt-cluster verify``, which generates one opcode per
590 - The environment as passed to the OS scripts is cleared, and thus no
591 environment variables defined in the node daemon's environment will be
592 inherited by the scripts.
593 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
595 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
596 support operations involving multiple node groups.
597 - Offline nodes are ignored when failing over an instance.
598 - Support for KVM version 1.0, which changed the version reporting format
600 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
602 - ``Makefile`` is now compatible with Automake 1.11.2
603 - Includes all bugfixes made in the 2.4 series
608 - The ganeti-htools project has been merged into the ganeti-core source
609 tree and will be built as part of Ganeti (see :doc:`install-quick`).
610 - Implemented support for :doc:`shared storage <design-shared-storage>`.
611 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
612 GPT-style partition tables (requires `parted
613 <http://www.gnu.org/s/parted/>`_).
614 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
615 - Allowed adding tags on instance creation.
616 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
617 - Allow instances to be started in a paused state, enabling the user to
618 see the complete console output on boot using the console.
619 - Added new hypervisor flag to control default reboot behaviour
620 (``reboot_behavior``).
621 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
622 - Improved out-of-band management support:
624 - Added ``gnt-node health`` command reporting the health status of
626 - Added ``gnt-node power`` command to manage power status of nodes.
627 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
629 - Instance migration can fall back to failover if instance is not
631 - Filters can be used when listing nodes, instances, groups and locks;
632 see *ganeti(7)* manpage.
633 - Added post-execution status as variables to :doc:`hooks <hooks>`
635 - Instance tags are exported/imported together with the instance.
636 - When given an explicit job ID, ``gnt-job info`` will work for archived
638 - Jobs can define dependencies on other jobs (not yet supported via
639 RAPI or command line, but used by internal commands and usable via
642 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
645 - Instance failover is now available as a RAPI resource
646 (``/2/instances/[instance_name]/failover``).
647 - ``gnt-instance info`` defaults to static information if primary node
649 - Opcodes have a new ``comment`` attribute.
650 - Added basic SPICE support to KVM hypervisor.
651 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
653 Node group improvements
654 ~~~~~~~~~~~~~~~~~~~~~~~
656 - ``gnt-cluster verify`` has been modified to check groups separately,
657 thereby improving performance.
658 - Node group support has been added to ``gnt-cluster verify-disks``,
659 which now operates per node group.
660 - Watcher has been changed to work better with node groups.
662 - One process and state file per node group.
663 - Slow watcher in one group doesn't block other group's watcher.
665 - Added new command, ``gnt-group evacuate``, to move all instances in a
666 node group to other groups.
667 - Added ``gnt-instance change-group`` to move an instance to another
669 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
670 per-group operations.
671 - Node groups can be tagged.
672 - Some operations switch from an exclusive to a shared lock as soon as
674 - Instance's primary and secondary nodes' groups are now available as
675 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
676 and ``snodes.group.uuid``).
681 - Numerous updates to documentation and manpages.
683 - :doc:`RAPI <rapi>` documentation now has detailed parameter
685 - Some opcode/job results are now also documented, see :doc:`RAPI
688 - A lockset's internal lock is now also visible in lock monitor.
689 - Log messages from job queue workers now contain information about the
690 opcode they're processing.
691 - ``gnt-instance console`` no longer requires the instance lock.
692 - A short delay when waiting for job changes reduces the number of LUXI
693 requests significantly.
694 - DRBD metadata volumes are overwritten with zeros during disk creation.
695 - Out-of-band commands no longer acquire the cluster lock in exclusive
697 - ``devel/upload`` now uses correct permissions for directories.
703 *(Released Fri, 23 Mar 2012)*
705 This was the sixth release candidate of the 2.5 series.
711 *(Released Mon, 9 Jan 2012)*
713 This was the fifth release candidate of the 2.5 series.
719 *(Released Thu, 27 Oct 2011)*
721 This was the fourth release candidate of the 2.5 series.
727 *(Released Wed, 26 Oct 2011)*
729 This was the third release candidate of the 2.5 series.
735 *(Released Tue, 18 Oct 2011)*
737 This was the second release candidate of the 2.5 series.
743 *(Released Tue, 4 Oct 2011)*
745 This was the first release candidate of the 2.5 series.
751 *(Released Wed, 31 Aug 2011)*
753 This was the third beta release of the 2.5 series.
759 *(Released Mon, 22 Aug 2011)*
761 This was the second beta release of the 2.5 series.
767 *(Released Fri, 12 Aug 2011)*
769 This was the first beta release of the 2.5 series.
775 *(Released Thu, 27 Oct 2011)*
777 - Fixed bug when parsing command line parameter values ending in
779 - Fixed assertion error after unclean master shutdown
780 - Disable HTTP client pool for RPC, significantly reducing memory usage
782 - Fixed queue archive creation with wrong permissions
788 *(Released Tue, 23 Aug 2011)*
792 - Fixed documentation for importing with ``--src-dir`` option
793 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
794 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
800 *(Released Fri, 5 Aug 2011)*
802 Many bug-fixes and a few small features:
804 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
805 issues when you tried to add an instance with two MAC addresses in one
807 - KVM: fixed per-instance stored UID value
808 - KVM: configure bridged NICs at migration start
809 - KVM: Fix a bug where instance will not start with never KVM versions
811 - Added OS search path to ``gnt-cluster info``
812 - Fixed an issue with ``file_storage_dir`` where you were forced to
813 provide an absolute path, but the documentation states it is a
814 relative path, the documentation was right
815 - Added a new parameter to instance stop/start called ``--no-remember``
816 that will make the state change to not be remembered
817 - Implemented ``no_remember`` at RAPI level
818 - Improved the documentation
819 - Node evacuation: don't call IAllocator if node is already empty
820 - Fixed bug in DRBD8 replace disks on current nodes
821 - Fixed bug in recreate-disks for DRBD instances
822 - Moved assertion checking locks in ``gnt-instance replace-disks``
823 causing it to abort with not owning the right locks for some situation
824 - Job queue: Fixed potential race condition when cancelling queued jobs
825 - Fixed off-by-one bug in job serial generation
826 - ``gnt-node volumes``: Fix instance names
827 - Fixed aliases in bash completion
828 - Fixed a bug in reopening log files after being sent a SIGHUP
829 - Added a flag to burnin to allow specifying VCPU count
830 - Bugfixes to non-root Ganeti configuration
836 *(Released Thu, 12 May 2011)*
838 Many bug-fixes and a few new small features:
840 - Fixed a bug related to log opening failures
841 - Fixed a bug in instance listing with orphan instances
842 - Fixed a bug which prevented resetting the cluster-level node parameter
843 ``oob_program`` to the default
844 - Many fixes related to the ``cluster-merge`` tool
845 - Fixed a race condition in the lock monitor, which caused failures
846 during (at least) creation of many instances in parallel
847 - Improved output for gnt-job info
848 - Removed the quiet flag on some ssh calls which prevented debugging
850 - Improved the N+1 failure messages in cluster verify by actually
851 showing the memory values (needed and available)
852 - Increased lock attempt timeouts so that when executing long operations
853 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
854 too early and thus prevent the use of the 'fair' mechanism
855 - Changed instance query data (``gnt-instance info``) to not acquire
856 locks unless needed, thus allowing its use on locked instance if only
857 static information is asked for
858 - Improved behaviour with filesystems that do not support rename on an
860 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
861 kept locks on all nodes during the wipe, which is unneeded
862 - Fixed ``gnt-watcher`` handling of errors during hooks execution
863 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
864 10GiB) which caused the wipe to fail right at the end in some cases
865 - Fixed master IP activation when doing master failover with no-voting
866 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
867 the master node itself
868 - Fixed potential data-loss in under disk full conditions, where Ganeti
869 wouldn't check correctly the return code and would consider
870 partially-written files 'correct'
871 - Fixed bug related to multiple VGs and DRBD disk replacing
872 - Added new disk parameter ``metavg`` that allows placement of the meta
873 device for DRBD in a different volume group
874 - Fixed error handling in the node daemon when the system libc doesn't
875 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
876 - Fixed lock release during replace-disks, which kept cluster-wide locks
877 when doing disk replaces with an iallocator script
878 - Added check for missing bridges in cluster verify
879 - Handle EPIPE errors while writing to the terminal better, so that
880 piping the output to e.g. ``less`` doesn't cause a backtrace
881 - Fixed rare case where a ^C during Luxi calls could have been
882 interpreted as server errors, instead of simply terminating
883 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
885 - Added a few more parameters to the KVM hypervisor, allowing a second
886 CDROM, custom disk type for CDROMs and a floppy image
887 - Removed redundant message in instance rename when the name is given
889 - Added option to ``gnt-instance recreate-disks`` to allow creating the
890 disks on new nodes, allowing recreation when the original instance
891 nodes are completely gone
892 - Added option when converting disk templates to DRBD to skip waiting
893 for the resync, in order to make the instance available sooner
894 - Added two new variables to the OS scripts environment (containing the
896 - Made the root_path and optional parameter for the xen-pvm hypervisor,
897 to allow use of ``pvgrub`` as bootloader
898 - Changed the instance memory modifications to only check out-of-memory
899 conditions on memory increases, and turned the secondary node warnings
900 into errors (they can still be overridden via ``--force``)
901 - Fixed the handling of a corner case when the Python installation gets
902 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
903 to execute a command that doesn't exist
904 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
905 node of the instance returned failures during instance shutdown; this
906 adds the option ``--ignore-consistency`` to gnt-instance move
908 And as usual, various improvements to the error messages, documentation
915 *(Released Wed, 09 Mar 2011)*
917 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
918 the RAPI users file if run twice (even with ``--dry-run``).
920 The release fixes that bug (nothing else changed).
926 *(Released Mon, 07 Mar 2011)*
928 Final 2.4.0 release. Just a few small fixes:
930 - Fixed RAPI node evacuate
931 - Fixed the kvm-ifup script
932 - Fixed internal error handling for special job cases
933 - Updated man page to specify the escaping feature for options
939 *(Released Mon, 28 Feb 2011)*
941 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
942 that this feature wiped the disks of the wrong instance, leading to loss
947 - Fixed title of query field containing instance name
948 - Expanded the glossary in the documentation
949 - Fixed one unittest (internal issue)
955 *(Released Mon, 21 Feb 2011)*
957 A number of bug fixes plus just a couple functionality changes.
959 On the user-visible side, the ``gnt-* list`` command output has changed
960 with respect to "special" field states. The current rc1 style of display
961 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
962 the default output mode special fields states are displayed as follows:
964 - Offline resource: ``*``
965 - Unavailable/not applicable: ``-``
966 - Data missing (RPC failure): ``?``
967 - Unknown field: ``??``
969 Another user-visible change is the addition of ``--force-join`` to
974 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
975 - Fixed regression in RAPI/instance reinstall where all parameters were
976 required (instead of optional)
977 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
978 - Fixed iallocator usage (offline nodes were not considered offline)
979 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
980 - Fixed hypervisor and OS parameter validation with respect to
982 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
984 - Fixed ``tools/listrunner`` with respect to agent-based usage
990 *(Released Fri, 4 Feb 2011)*
992 Many changes and fixes since the beta1 release. While there were some
993 internal changes, the code has been mostly stabilised for the RC
996 Note: the dumb allocator was removed in this release, as it was not kept
997 up-to-date with the IAllocator protocol changes. It is recommended to
998 use the ``hail`` command from the ganeti-htools package.
1000 Note: the 2.4 and up versions of Ganeti are not compatible with the
1001 0.2.x branch of ganeti-htools. You need to upgrade to
1002 ganeti-htools-0.3.0 (or later).
1004 Regressions fixed from 2.3
1005 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1007 - Fixed the ``gnt-cluster verify-disks`` command
1008 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1010 - Fixed disk adoption breakage
1011 - Fixed wrong headers in instance listing for field aliases
1016 - Fixed corner case in KVM handling of NICs
1017 - Fixed many cases of wrong handling of non-vm_capable nodes
1018 - Fixed a bug where a missing instance symlink was not possible to
1019 recreate with any ``gnt-*`` command (now ``gnt-instance
1020 activate-disks`` does it)
1021 - Fixed the volume group name as reported by ``gnt-cluster
1023 - Increased timeouts for the import-export code, hopefully leading to
1024 fewer aborts due network or instance timeouts
1025 - Fixed bug in ``gnt-node list-storage``
1026 - Fixed bug where not all daemons were started on cluster
1027 initialisation, but only at the first watcher run
1028 - Fixed many bugs in the OOB implementation
1029 - Fixed watcher behaviour in presence of instances with offline
1031 - Fixed instance list output for instances running on the wrong node
1032 - a few fixes to the cluster-merge tool, but it still cannot merge
1033 multi-node groups (currently it is not recommended to use this tool)
1039 - Improved network configuration for the KVM hypervisor
1040 - Added e1000 as a supported NIC for Xen-HVM
1041 - Improved the lvmstrap tool to also be able to use partitions, as
1042 opposed to full disks
1043 - Improved speed of disk wiping (the cluster parameter
1044 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1045 of instance creations
1046 - Added documentation for the OS parameters
1047 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1048 hypervisor is not responding
1049 - Added display of blacklisted and hidden OS information in
1050 ``gnt-cluster info``
1051 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1052 NIC and node parameters, which might create problems with currently
1053 invalid (but undetected) configuration files, but prevents validation
1054 failures when unrelated parameters are modified
1055 - Changed cluster initialisation to wait for the master daemon to become
1057 - Expanded the RAPI interface:
1059 - Added config redistribution resource
1060 - Added activation/deactivation of instance disks
1061 - Added export of console information
1063 - Implemented log file reopening on SIGHUP, which allows using
1064 logrotate(8) for the Ganeti log files
1065 - Added a basic OOB helper script as an example
1071 *(Released Fri, 14 Jan 2011)*
1076 - Fixed timezone issues when formatting timestamps
1077 - Added support for node groups, available via ``gnt-group`` and other
1079 - Added out-of-band framework and management, see :doc:`design
1080 document <design-oob>`
1081 - Removed support for roman numbers from ``gnt-node list`` and
1082 ``gnt-instance list``.
1083 - Allowed modification of master network interface via ``gnt-cluster
1084 modify --master-netdev``
1085 - Accept offline secondaries while shutting down instance disks
1086 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1087 - Added support for multiple LVM volume groups
1088 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1090 - Added commands to list available fields:
1092 - ``gnt-node list-fields``
1093 - ``gnt-group list-fields``
1094 - ``gnt-instance list-fields``
1096 - Updated documentation and man pages
1101 - Moved ``rapi_users`` file into separate directory, now named
1102 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1104 - Added new tool for running commands on many machines,
1105 ``tools/ganeti-listrunner``
1106 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1107 improving the ``gnt-instance console`` output
1108 - Allowed customisation of disk index separator at ``configure`` time
1109 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1110 - Added support for non-partitioned md disks in ``lvmstrap``
1111 - Added script to gracefully power off KVM instances
1112 - Split ``utils`` module into smaller parts
1113 - Changed query operations to return more detailed information, e.g.
1114 whether an information is unavailable due to an offline node. To use
1115 this new functionality, the LUXI call ``Query`` must be used. Field
1116 information is now stored by the master daemon and can be retrieved
1117 using ``QueryFields``. Instances, nodes and groups can also be queried
1118 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1119 exposed via RAPI). The following commands make use of this
1120 infrastructure change:
1122 - ``gnt-group list``
1123 - ``gnt-group list-fields``
1125 - ``gnt-node list-fields``
1126 - ``gnt-instance list``
1127 - ``gnt-instance list-fields``
1128 - ``gnt-debug locks``
1133 - New RAPI resources (see :doc:`rapi`):
1137 - ``/2/groups/[group_name]``
1138 - ``/2/groups/[group_name]/assign-nodes``
1139 - ``/2/groups/[group_name]/modify``
1140 - ``/2/groups/[group_name]/rename``
1141 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1145 - Implemented ``no_install`` for instance creation
1146 - Implemented OS parameters for instance reinstallation, allowing
1147 use of special settings on reinstallation (e.g. for preserving data)
1152 - Added IPv6 support in import/export
1153 - Pause DRBD synchronization while wiping disks on instance creation
1154 - Updated unittests and QA scripts
1155 - Improved network parameters passed to KVM
1156 - Converted man pages from docbook to reStructuredText
1162 *(Released Mon, 20 Dec 2010)*
1164 Released version 2.3.1~rc1 without any changes.
1170 *(Released Wed, 1 Dec 2010)*
1172 - impexpd: Disable OpenSSL compression in socat if possible (backport
1173 from master, commit e90739d625b, see :doc:`installation guide
1174 <install-quick>` for details)
1175 - Changed unittest coverage report to exclude test scripts
1176 - Added script to check version format
1182 *(Released Wed, 1 Dec 2010)*
1184 Released version 2.3.0~rc1 without any changes.
1190 *(Released Fri, 19 Nov 2010)*
1192 A number of bugfixes and documentation updates:
1194 - Update ganeti-os-interface documentation
1195 - Fixed a bug related to duplicate MACs or similar items which should be
1197 - Fix breakage in OS state modify
1198 - Reinstall instance: disallow offline secondaries (fixes bug related to
1199 OS changing but reinstall failing)
1200 - plus all the other fixes between 2.2.1 and 2.2.2
1206 *(Released Tue, 2 Nov 2010)*
1208 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1209 - Fixed master failover race with watcher
1210 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1212 - Accept previously stopped instance for export with instance removal
1213 - Simplify and extend the environment variables for instance OS scripts
1214 - Added new node flags, ``master_capable`` and ``vm_capable``
1215 - Added optional instance disk wiping prior during allocation. This is a
1216 cluster-wide option and can be set/modified using
1217 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1218 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1219 :doc:`install-quick`
1220 - Added a new watcher option (``--ignore-pause``)
1221 - Added option to ignore offline node on instance start/stop
1222 (``--ignore-offline``)
1223 - Allow overriding OS parameters with ``gnt-instance reinstall``
1224 - Added ability to change node's secondary IP address using ``gnt-node
1226 - Implemented privilege separation for all daemons except
1227 ``ganeti-noded``, see ``configure`` options
1228 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1230 - Implemented job priorities (see ``ganeti(7)`` manpage)
1231 - Ignore failures while shutting down instances during failover from
1233 - Exit daemon's bootstrap process only once daemon is ready
1234 - Export more information via ``LUInstanceQuery``/remote API
1235 - Improved documentation, QA and unittests
1236 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1237 restart if the file was created or changed
1238 - Added LUXI protocol version sent with each request and response,
1239 allowing detection of server/client mismatches
1240 - Moved the Python scripts among gnt-* and ganeti-* into modules
1241 - Moved all code related to setting up SSH to an external script,
1243 - Infrastructure changes for node group support in future versions
1249 *(Released Fri, 19 Nov 2010)*
1251 A few small bugs fixed, and some improvements to the build system:
1253 - Fix documentation regarding conversion to drbd
1254 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1256 - Fix error handling in node modify with multiple changes
1257 - Allow remote imports without checked names
1263 *(Released Tue, 19 Oct 2010)*
1265 - Disable SSL session ID cache in RPC client
1271 *(Released Thu, 14 Oct 2010)*
1273 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1274 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1276 - Fix problems with interaction between Curl and Python's HTTP server,
1277 resulting in increased speed in many RPC calls
1278 - Improve our release script to prevent breakage with older aclocal and
1285 *(Released Thu, 7 Oct 2010)*
1287 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1288 value retrieved from master
1289 - Added support for blacklisted or hidden OS definitions
1290 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1291 - Added support for -mem-path in KVM hypervisor abstraction layer
1292 - Allow overriding instance parameters in tool for inter-cluster
1293 instance moves (``tools/move-instance``)
1294 - Improved opcode summaries (e.g. in ``gnt-job list``)
1295 - Improve consistency of OS listing by sorting it
1296 - Documentation updates
1302 *(Released Fri, 8 Oct 2010)*
1304 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1310 *(Released Mon, 4 Oct 2010)*
1312 - Fixed regression in ``gnt-instance rename``
1318 *(Released Wed, 22 Sep 2010)*
1320 - Fixed OS_VARIANT variable for OS scripts
1321 - Fixed cluster tag operations via RAPI
1322 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1323 - Disabled RAPI CA checks in watcher
1329 *(Released Mon, 23 Aug 2010)*
1331 - Support DRBD versions of the format "a.b.c.d"
1333 - Re-introduce support for usage from multiple threads in RAPI client
1334 - Instance renames and modify via RAPI
1335 - Work around race condition between processing and archival in job
1337 - Mark opcodes following failed one as failed, too
1338 - Job field ``lock_status`` was removed due to difficulties making it
1339 work with the changed job queue in Ganeti 2.2; a better way to monitor
1340 locks is expected for a later 2.2.x release
1341 - Fixed dry-run behaviour with many commands
1342 - Support ``ssh-agent`` again when adding nodes
1343 - Many additional bugfixes
1349 *(Released Fri, 30 Jul 2010)*
1351 Important change: the internal RPC mechanism between Ganeti nodes has
1352 changed from using a home-grown http library (based on the Python base
1353 libraries) to use the PycURL library. This requires that PycURL is
1354 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1355 against GnuTLS by default. cURL's support for GnuTLS had known issues
1356 before cURL 7.21.0 and we recommend using the latest cURL release or
1357 linking against OpenSSL. Most other distributions already link PycURL
1358 and cURL against OpenSSL. The command::
1360 python -c 'import pycurl; print pycurl.version'
1362 can be used to determine the libraries PycURL and cURL are linked
1365 Other significant changes:
1367 - Rewrote much of the internals of the job queue, in order to achieve
1368 better parallelism; this decouples job query operations from the job
1369 processing, and it should allow much nicer behaviour of the master
1370 daemon under load, and it also has uncovered some long-standing bugs
1371 related to the job serialisation (now fixed)
1372 - Added a default iallocator setting to the cluster parameters,
1373 eliminating the need to always pass nodes or an iallocator for
1374 operations that require selection of new node(s)
1375 - Added experimental support for the LXC virtualization method
1376 - Added support for OS parameters, which allows the installation of
1377 instances to pass parameter to OS scripts in order to customise the
1379 - Added a hypervisor parameter controlling the migration type (live or
1380 non-live), since hypervisors have various levels of reliability; this
1381 has renamed the 'live' parameter to 'mode'
1382 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1383 logical volumes, meaning that cluster verify will ignore them and not
1384 flag their presence as errors
1385 - The watcher will now reset the error count for failed instances after
1386 8 hours, thus allowing self-healing if the problem that caused the
1387 instances to be down/fail to start has cleared in the meantime
1388 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1389 check for, and warn, if the drbd module parameter ``usermode_helper``
1390 is not consistent with the cluster-wide setting; this is needed to
1391 make diagnose easier of failed drbd creations
1392 - Started adding base IPv6 support, but this is not yet
1393 enabled/available for use
1394 - Rename operations (cluster, instance) will now return the new name,
1395 which is especially useful if a short name was passed in
1396 - Added support for instance migration in RAPI
1397 - Added a tool to pre-configure nodes for the SSH setup, before joining
1398 them to the cluster; this will allow in the future a simplified model
1399 for node joining (but not yet fully enabled in 2.2); this needs the
1400 paramiko python library
1401 - Fixed handling of name-resolving errors
1402 - Fixed consistency of job results on the error path
1403 - Fixed master-failover race condition when executed multiple times in
1405 - Fixed many bugs related to the job queue (mostly introduced during the
1406 2.2 development cycle, so not all are impacting 2.1)
1407 - Fixed instance migration with missing disk symlinks
1408 - Fixed handling of unknown jobs in ``gnt-job archive``
1409 - And many other small fixes/improvements
1413 - Enhanced both the unittest and the QA coverage
1414 - Switched the opcode validation to a generic model, and extended the
1415 validation to all opcode parameters
1416 - Changed more parts of the code that write shell scripts to use the
1418 - Switched the master daemon to use the asyncore library for the Luxi
1425 *(Released Thu, 17 Jun 2010)*
1427 - Added tool (``move-instance``) and infrastructure to move instances
1428 between separate clusters (see :doc:`separate documentation
1429 <move-instance>` and :doc:`design document <design-2.2>`)
1430 - Added per-request RPC timeout
1431 - RAPI now requires a Content-Type header for requests with a body (e.g.
1432 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1433 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1434 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1436 - Implemented initial support for running Ganeti daemons as separate
1437 users, see configure-time flags ``--with-user-prefix`` and
1438 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1440 - Instances can be removed after export (``gnt-backup export
1441 --remove-instance``)
1442 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1443 key (instead of 1024 bit)
1444 - Added new cluster configuration file for cluster domain secret
1445 - Import/export now use SSL instead of SSH
1446 - Added support for showing estimated time when exporting an instance,
1447 see the ``ganeti-os-interface(7)`` manpage and look for
1454 *(Released Tue, 16 Nov 2010)*
1456 Some more bugfixes. Unless critical bugs occur, this will be the last
1459 - Fix case of MAC special-values
1460 - Fix mac checker regex
1461 - backend: Fix typo causing "out of range" error
1462 - Add missing --units in gnt-instance list man page
1468 *(Released Tue, 24 Aug 2010)*
1471 - Don't ignore secondary node silently on non-mirrored disk templates
1473 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1474 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1475 - Properly document the usb_mouse parameter
1476 - Fix path in ganeti-rapi(8) (issue 116)
1477 - Adjust error message when the ganeti user's .ssh directory is
1479 - Add same-node-check when changing the disk template to drbd
1485 *(Released Fri, 16 Jul 2010)*
1488 - Add an option to only select some reboot types during qa/burnin.
1489 (on some hypervisors consequent reboots are not supported)
1490 - Fix infrequent race condition in master failover. Sometimes the old
1491 master ip address would be still detected as up for a short time
1492 after it was removed, causing failover to fail.
1493 - Decrease mlockall warnings when the ctypes module is missing. On
1494 Python 2.4 we support running even if no ctypes module is installed,
1495 but we were too verbose about this issue.
1496 - Fix building on old distributions, on which man doesn't have a
1498 - Fix RAPI not to ignore the MAC address on instance creation
1499 - Implement the old instance creation format in the RAPI client.
1505 *(Released Thu, 01 Jul 2010)*
1507 A small bugfix release:
1508 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1509 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1510 a non-existing option
1511 - Fix instance create: the --force-variant option was ignored
1512 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1513 - Fix modify node storage with non-FQDN arguments
1514 - Fix RAPI client to authenticate under Python 2.6 when used
1515 for more than 5 requests needing authentication
1516 - Fix gnt-instance modify -t (storage) giving a wrong error message
1517 when converting a non-shutdown drbd instance to plain
1523 *(Released Fri, 18 Jun 2010)*
1525 A small bugfix release:
1527 - Fix live migration of KVM instances started with older Ganeti
1528 versions which had fewer hypervisor parameters
1529 - Fix gnt-instance grow-disk on down instances
1530 - Fix an error-reporting bug during instance migration
1531 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1532 silently ignoring broken ones
1533 - Fix an RPC error reporting bug affecting, for example, RAPI client
1535 - Fix bug triggered by different API version os-es on different nodes
1536 - Fix a bug in instance startup with custom hvparams: OS level
1537 parameters would fail to be applied.
1538 - Fix the RAPI client under Python 2.6 (but more work is needed to
1539 make it work completely well with OpenSSL)
1540 - Fix handling of errors when resolving names from DNS
1546 *(Released Thu, 3 Jun 2010)*
1548 A medium sized development cycle. Some new features, and some
1549 fixes/small improvements/cleanups.
1551 Significant features
1552 ~~~~~~~~~~~~~~~~~~~~
1554 The node deamon now tries to mlock itself into memory, unless the
1555 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1556 its logs, and falls back to console logging. This allows emergency
1557 features such as ``gnt-node powercycle`` to work even in the event of a
1558 broken node disk (tested offlining the disk hosting the node's
1559 filesystem and dropping its memory caches; don't try this at home)
1561 KVM: add vhost-net acceleration support. It can be tested with a new
1562 enough version of the kernel and of qemu-kvm.
1564 KVM: Add instance chrooting feature. If you use privilege dropping for
1565 your VMs you can also now force them to chroot to an empty directory,
1566 before starting the emulated guest.
1568 KVM: Add maximum migration bandwith and maximum downtime tweaking
1569 support (requires a new-enough version of qemu-kvm).
1571 Cluster verify will now warn if the master node doesn't have the master
1572 ip configured on it.
1574 Add a new (incompatible) instance creation request format to RAPI which
1575 supports all parameters (previously only a subset was supported, and it
1576 wasn't possible to extend the old format to accomodate all the new
1577 features. The old format is still supported, and a client can check for
1578 this feature, before using it, by checking for its presence in the
1579 ``features`` RAPI resource.
1581 Now with ancient latin support. Try it passing the ``--roman`` option to
1582 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1583 (requires the python-roman module to be installed, in order to work).
1588 As usual many internal code refactorings, documentation updates, and
1591 - Lots of improvements and cleanups to the experimental Remote API
1592 (RAPI) client library.
1593 - A new unit test suite for the core daemon libraries.
1594 - A fix to creating missing directories makes sure the umask is not
1595 applied anymore. This enforces the same directory permissions
1597 - Better handling terminating daemons with ctrl+c (used when running
1598 them in debugging mode).
1599 - Fix a race condition in live migrating a KVM instance, when stat()
1600 on the old proc status file returned EINVAL, which is an unexpected
1602 - Fixed manpage checking with newer man and utf-8 charachters. But now
1603 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1609 *(Released Fri, 7 May 2010)*
1611 Fix a bug which prevented untagged KVM instances from starting.
1617 *(Released Fri, 7 May 2010)*
1619 Another release with a long development cycle, during which many
1620 different features were added.
1622 Significant features
1623 ~~~~~~~~~~~~~~~~~~~~
1625 The KVM hypervisor now can run the individual instances as non-root, to
1626 reduce the impact of a VM being hijacked due to bugs in the
1627 hypervisor. It is possible to run all instances as a single (non-root)
1628 user, to manually specify a user for each instance, or to dynamically
1629 allocate a user out of a cluster-wide pool to each instance, with the
1630 guarantee that no two instances will run under the same user ID on any
1633 An experimental RAPI client library, that can be used standalone
1634 (without the other Ganeti libraries), is provided in the source tree as
1635 ``lib/rapi/client.py``. Note this client might change its interface in
1636 the future, as we iterate on its capabilities.
1638 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1639 replace the cluster's certificates and crypto keys. This might help in
1640 case they have been compromised, or have simply expired.
1642 A new disk option for instance creation has been added that allows one
1643 to "adopt" currently existing logical volumes, with data
1644 preservation. This should allow easier migration to Ganeti from
1645 unmanaged (or managed via other software) instances.
1647 Another disk improvement is the possibility to convert between redundant
1648 (DRBD) and plain (LVM) disk configuration for an instance. This should
1649 allow better scalability (starting with one node and growing the
1650 cluster, or shrinking a two-node cluster to one node).
1652 A new feature that could help with automated node failovers has been
1653 implemented: if a node sees itself as offline (by querying the master
1654 candidates), it will try to shutdown (hard) all instances and any active
1655 DRBD devices. This reduces the risk of duplicate instances if an
1656 external script automatically failovers the instances on such nodes. To
1657 enable this, the cluster parameter ``maintain_node_health`` should be
1658 enabled; in the future this option (per the name) will enable other
1659 automatic maintenance features.
1661 Instance export/import now will reuse the original instance
1662 specifications for all parameters; that means exporting an instance,
1663 deleting it and the importing it back should give an almost identical
1664 instance. Note that the default import behaviour has changed from
1665 before, where it created only one NIC; now it recreates the original
1668 Cluster verify has added a few new checks: SSL certificates validity,
1669 /etc/hosts consistency across the cluster, etc.
1674 As usual, many internal changes were done, documentation fixes,
1677 - Fixed cluster initialization with disabled cluster storage (regression
1678 introduced in 2.1.1)
1679 - File-based storage supports growing the disks
1680 - Fixed behaviour of node role changes
1681 - Fixed cluster verify for some corner cases, plus a general rewrite of
1682 cluster verify to allow future extension with more checks
1683 - Fixed log spamming by watcher and node daemon (regression introduced
1685 - Fixed possible validation issues when changing the list of enabled
1687 - Fixed cleanup of /etc/hosts during node removal
1688 - Fixed RAPI response for invalid methods
1689 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1690 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1691 from ide disks, etc.)
1692 - Allow OS changes without re-installation (to record a changed OS
1693 outside of Ganeti, or to allow OS renames)
1694 - Allow instance creation without OS installation (useful for example if
1695 the OS will be installed manually, or restored from a backup not in
1697 - Implemented option to make cluster ``copyfile`` use the replication
1699 - Added list of enabled hypervisors to ssconf (possibly useful for
1701 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1703 - A partial form of node re-IP is possible via node readd, which now
1704 allows changed node primary IP
1705 - Command line utilities now show an informational message if the job is
1707 - The logs of the master daemon now show the PID/UID/GID of the
1714 *(Released Fri, 12 Mar 2010)*
1716 During the 2.1.0 long release candidate cycle, a lot of improvements and
1717 changes have accumulated with were released later as 2.1.1.
1722 The node evacuate command (``gnt-node evacuate``) was significantly
1723 rewritten, and as such the IAllocator protocol was changed - a new
1724 request type has been added. This unfortunate change during a stable
1725 series is designed to improve performance of node evacuations; on
1726 clusters with more than about five nodes and which are well-balanced,
1727 evacuation should proceed in parallel for all instances of the node
1728 being evacuated. As such, any existing IAllocator scripts need to be
1729 updated, otherwise the above command will fail due to the unknown
1730 request. The provided "dumb" allocator has not been updated; but the
1731 ganeti-htools package supports the new protocol since version 0.2.4.
1733 Another important change is increased validation of node and instance
1734 names. This might create problems in special cases, if invalid host
1735 names are being used.
1737 Also, a new layer of hypervisor parameters has been added, that sits at
1738 OS level between the cluster defaults and the instance ones. This allows
1739 customisation of virtualization parameters depending on the installed
1740 OS. For example instances with OS 'X' may have a different KVM kernel
1741 (or any other parameter) than the cluster defaults. This is intended to
1742 help managing a multiple OSes on the same cluster, without manual
1743 modification of each instance's parameters.
1745 A tool for merging clusters, ``cluster-merge``, has been added in the
1746 tools sub-directory.
1751 - Improved the int/float conversions that should make the code more
1752 robust in face of errors from the node daemons
1753 - Fixed the remove node code in case of internal configuration errors
1754 - Fixed the node daemon behaviour in face of inconsistent queue
1755 directory (e.g. read-only file-system where we can't open the files
1757 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1758 now it either aborts cleanly or, if given the new "auto_promote"
1759 parameter, will automatically promote other nodes as needed
1760 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1761 completely prevent Ganeti from working
1762 - Fixed bug for instance export when not all disks were successfully
1764 - Fixed behaviour of node add when the new node is slow in starting up
1766 - Fixed handling of signals in the LUXI client, which should improve
1767 behaviour of command-line scripts
1768 - Added checks for invalid node/instance names in the configuration (now
1769 flagged during cluster verify)
1770 - Fixed watcher behaviour for disk activation errors
1771 - Fixed two potentially endless loops in http library, which led to the
1772 RAPI daemon hanging and consuming 100% CPU in some cases
1773 - Fixed bug in RAPI daemon related to hashed passwords
1774 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1776 - Enhanced compatibility with non-Debian OSes, but not using absolute
1777 path in some commands and allowing customisation of the ssh
1778 configuration directory
1779 - Fixed possible future issue with new Python versions by abiding to the
1780 proper use of ``__slots__`` attribute on classes
1781 - Added checks that should prevent directory traversal attacks
1782 - Many documentation fixes based on feedback from users
1787 - Added an "early_release" more for instance replace disks and node
1788 evacuate, where we release locks earlier and thus allow higher
1789 parallelism within the cluster
1790 - Added watcher hooks, intended to allow the watcher to restart other
1791 daemons (e.g. from the ganeti-nbma project), but they can be used of
1792 course for any other purpose
1793 - Added a compile-time disable for DRBD barriers, to increase
1794 performance if the administrator trusts the power supply or the
1795 storage system to not lose writes
1796 - Added the option of using syslog for logging instead of, or in
1797 addition to, Ganeti's own log files
1798 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1799 can indeed boot from a paravirtual NIC
1800 - Added a generic debug level for many operations; while this is not
1801 used widely yet, it allows one to pass the debug value all the way to
1803 - Enhanced the hooks environment for instance moves (failovers,
1804 migrations) where the primary/secondary nodes changed during the
1805 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1806 - Enhanced data validations for many user-supplied values; one important
1807 item is the restrictions imposed on instance and node names, which
1808 might reject some (invalid) host names
1809 - Add a configure-time option to disable file-based storage, if it's not
1810 needed; this allows greater security separation between the master
1811 node and the other nodes from the point of view of the inter-node RPC
1813 - Added user notification in interactive tools if job is waiting in the
1814 job queue or trying to acquire locks
1815 - Added log messages when a job is waiting for locks
1816 - Added filtering by node tags in instance operations which admit
1817 multiple instances (start, stop, reboot, reinstall)
1818 - Added a new tool for cluster mergers, ``cluster-merge``
1819 - Parameters from command line which are of the form ``a=b,c=d`` can now
1820 use backslash escapes to pass in values which contain commas,
1821 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1823 - For KVM, the instance name is the first parameter passed to KVM, so
1824 that it's more visible in the process list
1830 *(Released Tue, 2 Mar 2010)*
1832 Ganeti 2.1 brings many improvements with it. Major changes:
1834 - Added infrastructure to ease automated disk repairs
1835 - Added new daemon to export configuration data in a cheaper way than
1836 using the remote API
1837 - Instance NICs can now be routed instead of being associated with a
1839 - Improved job locking logic to reduce impact of jobs acquiring multiple
1840 locks waiting for other long-running jobs
1842 In-depth implementation details can be found in the Ganeti 2.1 design
1848 - Added chroot hypervisor
1849 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1851 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1852 ``bootloader_path`` and ``bootloader_args``)
1853 - Added the ``use_localtime`` option for the xen-hvm and kvm
1854 hypervisors, and the default value for this has changed to false (in
1855 2.0 xen-hvm always enabled it)
1856 - Added luxi call to submit multiple jobs in one go
1857 - Added cluster initialization option to not modify ``/etc/hosts``
1859 - Added network interface parameters
1860 - Added dry run mode to some LUs
1861 - Added RAPI resources:
1863 - ``/2/instances/[instance_name]/info``
1864 - ``/2/instances/[instance_name]/replace-disks``
1865 - ``/2/nodes/[node_name]/evacuate``
1866 - ``/2/nodes/[node_name]/migrate``
1867 - ``/2/nodes/[node_name]/role``
1868 - ``/2/nodes/[node_name]/storage``
1869 - ``/2/nodes/[node_name]/storage/modify``
1870 - ``/2/nodes/[node_name]/storage/repair``
1872 - Added OpCodes to evacuate or migrate all instances on a node
1873 - Added new command to list storage elements on nodes (``gnt-node
1874 list-storage``) and modify them (``gnt-node modify-storage``)
1875 - Added new ssconf files with master candidate IP address
1876 (``ssconf_master_candidates_ips``), node primary IP address
1877 (``ssconf_node_primary_ips``) and node secondary IP address
1878 (``ssconf_node_secondary_ips``)
1879 - Added ``ganeti-confd`` and a client library to query the Ganeti
1880 configuration via UDP
1881 - Added ability to run hooks after cluster initialization and before
1883 - Added automatic mode for disk replace (``gnt-instance replace-disks
1885 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1886 after catastrophic data-loss
1887 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1889 - Added ``gnt-instance move`` command to move instances
1890 - Added ``gnt-cluster watcher`` command to control watcher
1891 - Added ``gnt-node powercycle`` command to powercycle nodes
1892 - Added new job status field ``lock_status``
1893 - Added parseable error codes to cluster verification (``gnt-cluster
1894 verify --error-codes``) and made output less verbose (use
1895 ``--verbose`` to restore previous behaviour)
1896 - Added UUIDs to the main config entities (cluster, nodes, instances)
1897 - Added support for OS variants
1898 - Added support for hashed passwords in the Ganeti remote API users file
1900 - Added option to specify maximum timeout on instance shutdown
1901 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1902 - Added new helper script to start and stop Ganeti daemons
1903 (``daemon-util``), with the intent to reduce the work necessary to
1904 adjust Ganeti for non-Debian distributions and to start/stop daemons
1906 - Added more unittests
1907 - Fixed critical bug in ganeti-masterd startup
1908 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1909 now customisable at the cluster level for both the KVM and Xen
1910 hypervisors using the new ``migration_port`` parameter
1911 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1912 reinstalling an instance
1913 - Allowed ``@`` in tag names
1914 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1915 - Many documentation updates
1916 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1917 - ``gnt-instance reinstall`` can now reinstall multiple instances
1918 - Updated many command line parameters
1919 - Introduced new OS API version 15
1920 - No longer support a default hypervisor
1921 - Treat virtual LVs as inexistent
1922 - Improved job locking logic to reduce lock contention
1923 - Match instance and node names case insensitively
1924 - Reimplemented bash completion script to be more complete
1931 *(Released Thu, 4 Feb 2010)*
1933 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1934 - Fix a string formatting bug
1935 - Improve safety of the code in some error paths
1936 - Improve data validation in the master of values returned from nodes
1942 *(Released Thu, 17 Dec 2009)*
1944 - Fix security issue due to missing validation of iallocator names; this
1945 allows local and remote execution of arbitrary executables
1946 - Fix failure of gnt-node list during instance removal
1947 - Ship the RAPI documentation in the archive
1953 *(Released Wed, 30 Sep 2009)*
1955 - Fixed many wrong messages
1956 - Fixed a few bugs related to the locking library
1957 - Fixed MAC checking at instance creation time
1958 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1959 - Fixed a few issues related to signal handling in both daemons and
1961 - Fixed the example startup script provided
1962 - Fixed insserv dependencies in the example startup script (patch from
1964 - Fixed handling of drained nodes in the iallocator framework
1965 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1967 - Fixed error related to invalid job IDs in job polling
1968 - Fixed job/opcode persistence on unclean master shutdown
1969 - Fixed handling of partial job processing after unclean master
1971 - Fixed error reporting from LUs, previously all errors were converted
1972 into execution errors
1973 - Fixed error reporting from burnin
1974 - Decreased significantly the memory usage of the job queue
1975 - Optimised slightly multi-job submission
1976 - Optimised slightly opcode loading
1977 - Backported the multi-job submit framework from the development
1978 branch; multi-instance start and stop should be faster
1979 - Added script to clean archived jobs after 21 days; this will reduce
1980 the size of the queue directory
1981 - Added some extra checks in disk size tracking
1982 - Added an example ethers hook script
1983 - Added a cluster parameter that prevents Ganeti from modifying of
1985 - Added more node information to RAPI responses
1986 - Added a ``gnt-job watch`` command that allows following the ouput of a
1988 - Added a bind-address option to ganeti-rapi
1989 - Added more checks to the configuration verify
1990 - Enhanced the burnin script such that some operations can be retried
1992 - Converted instance reinstall to multi-instance model
1998 *(Released Fri, 7 Aug 2009)*
2000 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2001 to allow using the pre-2.0.2 behaviour in activation, if any existing
2002 instances have mismatched disk sizes in the configuration
2003 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2004 any configuration mismatches for disk sizes
2005 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2006 failover to work on two-node clusters
2007 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2009 - Fixed detection of OS script errors in ``gnt-backup export``
2010 - Fixed exit code of ``gnt-backup export``
2016 *(Released Fri, 17 Jul 2009)*
2018 - Added experimental support for stripped logical volumes; this should
2019 enhance performance but comes with a higher complexity in the block
2020 device handling; stripping is only enabled when passing
2021 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2022 affected even in the non-stripped mode
2023 - Improved resiliency against transient failures at the end of DRBD
2024 resyncs, and in general of DRBD resync checks
2025 - Fixed a couple of issues with exports and snapshot errors
2026 - Fixed a couple of issues in instance listing
2027 - Added display of the disk size in ``gnt-instance info``
2028 - Fixed checking for valid OSes in instance creation
2029 - Fixed handling of the "vcpus" parameter in instance listing and in
2030 general of invalid parameters
2031 - Fixed http server library, and thus RAPI, to handle invalid
2032 username/password combinations correctly; this means that now they
2033 report unauthorized for queries too, not only for modifications,
2034 allowing earlier detect of configuration problems
2035 - Added a new "role" node list field, equivalent to the master/master
2036 candidate/drained/offline flags combinations
2037 - Fixed cluster modify and changes of candidate pool size
2038 - Fixed cluster verify error messages for wrong files on regular nodes
2039 - Fixed a couple of issues with node demotion from master candidate role
2040 - Fixed node readd issues
2041 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2042 - Added a new ``--no-voting`` option for masterfailover to fix failover
2043 on two-nodes clusters when the former master node is unreachable
2044 - Added instance reinstall over RAPI
2050 *(Released Tue, 16 Jun 2009)*
2052 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2053 allow re-adding the start in single-user option (regression from 1.2)
2054 - the watcher writes the instance status to a file, to allow monitoring
2055 to report the instance status (from the master) based on cached
2056 results of the watcher's queries; while this can get stale if the
2057 watcher is being locked due to other work on the cluster, this is
2058 still an improvement
2059 - the watcher now also restarts the node daemon and the rapi daemon if
2061 - fixed the watcher to handle full and drained queue cases
2062 - hooks export more instance data in the environment, which helps if
2063 hook scripts need to take action based on the instance's properties
2064 (no longer need to query back into ganeti)
2065 - instance failovers when the instance is stopped do not check for free
2066 RAM, so that failing over a stopped instance is possible in low memory
2068 - rapi uses queries for tags instead of jobs (for less job traffic), and
2069 for cluster tags it won't talk to masterd at all but read them from
2071 - a couple of error handling fixes in RAPI
2072 - drbd handling: improved the error handling of inconsistent disks after
2073 resync to reduce the frequency of "there are some degraded disks for
2074 this instance" messages
2075 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2076 error handling path called a wrong function name)
2082 *(Released Wed, 27 May 2009)*
2084 - no changes from rc5
2090 *(Released Wed, 20 May 2009)*
2092 - fix a couple of bugs (validation, argument checks)
2093 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2094 - some small improvements to RAPI and IAllocator
2095 - make watcher automatically start the master daemon if down
2101 *(Released Mon, 27 Apr 2009)*
2103 - change the OS list to not require locks; this helps with big clusters
2104 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2105 volume group is broken
2106 - ``gnt-instance info``, without any arguments, doesn't run for all
2107 instances anymore; either pass ``--all`` or pass the desired
2108 instances; this helps against mistakes on big clusters where listing
2109 the information for all instances takes a long time
2110 - miscellaneous doc and man pages fixes
2116 *(Released Wed, 8 Apr 2009)*
2118 - Change the internal locking model of some ``gnt-node`` commands, in
2119 order to reduce contention (and blocking of master daemon) when
2120 batching many creation/reinstall jobs
2121 - Fixes to Xen soft reboot
2122 - No longer build documentation at build time, instead distribute it in
2123 the archive, in order to reduce the need for the whole docbook/rst
2130 *(Released Fri, 27 Mar 2009)*
2132 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2133 - Fix watcher startup sequence, improves the behaviour of busy clusters
2134 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2135 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2136 bind address and other places
2137 - Some documentation fixes and updates
2143 *(Released Mon, 2 Mar 2009)*
2145 - More documentation updates, now all docs should be more-or-less
2147 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2149 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2151 - Fix an issue related to $libdir/run/ganeti and cluster creation
2157 *(Released Thu, 19 Feb 2009)*
2159 - Xen PVM and KVM have switched the default value for the instance root
2160 disk to the first partition on the first drive, instead of the whole
2161 drive; this means that the OS installation scripts must be changed
2163 - Man pages have been updated
2164 - RAPI has been switched by default to HTTPS, and the exported functions
2165 should all work correctly
2166 - RAPI v1 has been removed
2167 - Many improvements to the KVM hypervisor
2168 - Block device errors are now better reported
2169 - Many other bugfixes and small improvements
2175 *(Released Mon, 26 Jan 2009)*
2177 - Version 2 is a general rewrite of the code and therefore the
2178 differences are too many to list, see the design document for 2.0 in
2179 the ``doc/`` subdirectory for more details
2180 - In this beta version there is not yet a migration path from 1.2 (there
2181 will be one in the final 2.0 release)
2182 - A few significant changes are:
2184 - all commands are executed by a daemon (``ganeti-masterd``) and the
2185 various ``gnt-*`` commands are just front-ends to it
2186 - all the commands are entered into, and executed from a job queue,
2187 see the ``gnt-job(8)`` manpage
2188 - the RAPI daemon supports read-write operations, secured by basic
2189 HTTP authentication on top of HTTPS
2190 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2191 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2192 to migrate to DRBD 8 first while still running Ganeti 1.2)
2193 - DRBD devices are using statically allocated minor numbers, which
2194 will be assigned to existing instances during the migration process
2195 - there is support for both Xen PVM and Xen HVM instances running on
2197 - KVM virtualization is supported too
2198 - file-based storage has been implemented, which means that it is
2199 possible to run the cluster without LVM and DRBD storage, for
2200 example using a shared filesystem exported from shared storage (and
2201 still have live migration)
2207 *(Released Tue, 13 Jan 2009)*
2209 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2210 - Reuse the old instance mac address by default on instance import, if
2211 the instance name is the same.
2212 - Handle situations in which the node info rpc returns incomplete
2214 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2215 - Improved version of batcher:
2217 - state file support
2218 - instance mac address support
2219 - support for HVM clusters/instances
2221 - Add an option to show the number of cpu sockets and nodes in
2223 - Support OSes that handle more than one version of the OS api (but do
2224 not change the current API in any other way)
2225 - Fix ``gnt-node migrate``
2226 - ``gnt-debug`` man page
2227 - Fixes various more typos and small issues
2228 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2234 *(Released Wed, 24 Sep 2008)*
2236 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2237 type of disk exported to fully virtualized instances.
2238 - provide access to the serial console of HVM instances
2239 - instance auto_balance flag, set by default. If turned off it will
2240 avoid warnings on cluster verify if there is not enough memory to fail
2241 over an instance. in the future it will prevent automatically failing
2242 it over when we will support that.
2243 - batcher tool for instance creation, see ``tools/README.batcher``
2244 - ``gnt-instance reinstall --select-os`` to interactively select a new
2245 operating system when reinstalling an instance.
2246 - when changing the memory amount on instance modify a check has been
2247 added that the instance will be able to start. also warnings are
2248 emitted if the instance will not be able to fail over, if auto_balance
2250 - documentation fixes
2251 - sync fields between ``gnt-instance list/modify/add/import``
2252 - fix a race condition in drbd when the sync speed was set after giving
2253 the device a remote peer.
2259 *(Released Tue, 22 Jul 2008)*
2261 - note: the allowed size and number of tags per object were reduced
2262 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2263 - fixed twisted 8.x compatibility
2264 - fixed ``gnt-instance replace-disks`` with iallocator
2265 - add TCP keepalives on twisted connections to detect restarted nodes
2266 - disk increase support, see ``gnt-instance grow-disk``
2267 - implement bulk node/instance query for RAPI
2268 - add tags in node/instance listing (optional)
2269 - experimental migration (and live migration) support, read the man page
2270 for ``gnt-instance migrate``
2271 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2272 has some small improvements in handling its state file
2278 *(Released Fri, 13 Jun 2008)*
2280 - Experimental readonly, REST-based remote API implementation;
2281 automatically started on master node, TCP port 5080, if enabled by
2282 ``--enable-rapi`` parameter to configure script.
2283 - Instance allocator support. Add and import instance accept a
2284 ``--iallocator`` parameter, and call that instance allocator to decide
2285 which node to use for the instance. The iallocator document describes
2286 what's expected from an allocator script.
2287 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2288 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2289 node is lost there is still enough memory to fail over the instances
2291 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2292 ``gnt-cluster verify``, to check for site-specific compliance. All the
2293 hooks will run, and their output, if any, will be displayed. Any
2294 failing hook will make the verification return an error value.
2295 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2296 primary and secondary interfaces
2297 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2298 that is still declared as part of the cluster and has failed.
2299 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2300 specifying output fields, that just adds the chosen fields to the
2302 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2303 export from the filesystem.
2304 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2305 have been added. Using them you can enable/disable acpi and pae
2306 support, and specify a path for a cd image to be exported to the
2307 instance. These parameters as the name suggest only work on HVM
2309 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2310 PAE support will be set to the previously hardcoded values, but the
2311 (previously hardcoded) path to the CDROM ISO image will be unset and
2312 if required, needs to be set manually with ``gnt-instance modify``
2314 - The address to which an instance's VNC console is bound is now
2315 selectable per-instance, rather than being cluster wide. Of course
2316 this only applies to instances controlled via VNC, so currently just
2317 applies to HVM clusters.
2323 *(Released Mon, 18 Feb 2008)*
2325 - more tweaks to the disk activation code (especially helpful for DRBD)
2326 - change the default ``gnt-instance list`` output format, now there is
2327 one combined status field (see the manpage for the exact values this
2329 - some more fixes for the mac export to hooks change
2330 - make Ganeti not break with DRBD 8.2.x (which changed the version
2331 format in ``/proc/drbd``) (issue 24)
2332 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2333 template, allowing migration from DRBD0.7+MD to DRBD8
2339 *(Released Wed, 30 Jan 2008)*
2341 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2343 - add command aliases infrastructure and a few aliases
2344 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2345 man pages and the ``--help`` option of ``gnt-node
2346 list``/``gnt-instance list``
2347 - fix ``gnt-backup list`` with down nodes (issue 21)
2348 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2349 - fix the dist archive and add a check for including svn/git files in
2351 - some developer-related changes: improve the burnin and the QA suite,
2352 add an upload script for testing during development
2358 *(Released Wed, 16 Jan 2008)*
2360 - experimental HVM support, read the install document, section
2361 "Initializing the cluster"
2362 - allow for the PVM hypervisor per-instance kernel and initrd paths
2363 - add a new command ``gnt-cluster verify-disks`` which uses a new
2364 algorithm to improve the reconnection of the DRBD pairs if the device
2365 on the secondary node has gone away
2366 - make logical volume code auto-activate LVs at disk activation time
2367 - slightly improve the speed of activating disks
2368 - allow specification of the MAC address at instance creation time, and
2369 changing it later via ``gnt-instance modify``
2370 - fix handling of external commands that generate lots of output on
2372 - update documentation with regard to minimum version of DRBD8 supported
2378 *(Released Tue, 4 Dec 2007)*
2380 - Log the ``xm create`` output to the node daemon log on failure (to
2381 help diagnosing the error)
2382 - In debug mode, log all external commands output if failed to the logs
2383 - Change parsing of lvm commands to ignore stderr
2389 *(Released Wed, 28 Nov 2007)*
2391 - Another round of updates to the DRBD 8 code to deal with more failures
2392 in the replace secondary node operation
2393 - Some more logging of failures in disk operations (lvm, drbd)
2394 - A few documentation updates
2401 *(Released Tue, 13 Nov 2007)*
2403 - Change configuration file format from Python's Pickle to JSON.
2404 Upgrading is possible using the cfgupgrade utility.
2405 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2406 faster replace disks and is more stable (DRBD 8 has many improvements
2407 compared to DRBD 0.7)
2408 - Added command line tags support (see man pages for ``gnt-instance``,
2409 ``gnt-node``, ``gnt-cluster``)
2410 - Added instance rename support
2411 - Added multi-instance startup/shutdown
2412 - Added cluster rename support
2413 - Added ``gnt-node evacuate`` to simplify some node operations
2414 - Added instance reboot operation that can speedup reboot as compared to
2416 - Soften the requirement that hostnames are in FQDN format
2417 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2419 - Removed dependency on debian's patched fping that uses the
2420 non-standard ``-S`` option
2421 - Now the OS definitions are searched for in multiple, configurable
2422 paths (easier for distros to package)
2423 - Some changes to the hooks infrastructure (especially the new
2424 post-configuration update hook)
2425 - Other small bugfixes
2427 .. vim: set textwidth=72 syntax=rst :