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``.
32 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
33 for RAPI users. It allows granting permissions to query for
34 information to a specific user without giving
35 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
36 - A new tool named ``node-cleanup`` has been added. It cleans remains of
37 a cluster from a machine by stopping all daemons, removing
38 certificates and ssconf files. Unless the ``--no-backup`` option is
39 given, copies of the certificates are made.
45 *(Released Fri, 12 Oct 2012)*
47 A small bugfix release. Among the bugs fixed:
49 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
50 made the command unusable.
51 - Commands that issue many jobs don't fail anymore just because some jobs
52 take so long that other jobs are archived.
53 - Failures during ``gnt-instance reinstall`` are reflected by the exit
55 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
57 - When ``always_failover`` is set, ``--allow-failover`` is not required
58 in migrate commands anymore.
59 - ``bash_completion`` works even if extglob is disabled.
60 - Fixed bug with locks that made failover for RDB-based instances fail.
61 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
62 a random node instead of one based on the allocator metric.
63 - 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``.
66 - Added new Makefile target to rebuild the whole distribution, so that
67 all files are included.
73 *(Released Fri, 27 Jul 2012)*
76 .. attention:: The ``LUXI`` protocol has been made more consistent
77 regarding its handling of command arguments. This, however, leads to
78 incompatibility issues with previous versions. Please ensure that you
79 restart Ganeti daemons soon after the upgrade, otherwise most
80 ``LUXI`` calls (job submission, setting/resetting the drain flag,
81 pausing/resuming the watcher, cancelling and archiving jobs, querying
82 the cluster configuration) will fail.
91 The current ``admin_up`` field, which used to denote whether an instance
92 should be running or not, has been removed. Instead, ``admin_state`` is
93 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
95 The rational behind this is that an instance being “down” can have
98 - it could be down during a reboot
99 - it could be temporarily be down for a reinstall
100 - or it could be down because it is deprecated and kept just for its
103 The previous Boolean state was making it difficult to do capacity
104 calculations: should Ganeti reserve memory for a down instance? Now, the
105 tri-state field makes it clear:
107 - in ``up`` and ``down`` state, all resources are reserved for the
108 instance, and it can be at any time brought up if it is down
109 - in ``offline`` state, only disk space is reserved for it, but not
112 The field can have an extra use: since the transition between ``up`` and
113 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
114 transition between ``offline`` and ``down`` is done via ``gnt-instance
115 modify``, it is possible to given different rights to users. For
116 example, owners of an instance could be allowed to start/stop it, but
117 not transition it out of the offline state.
119 Instance policies and specs
120 +++++++++++++++++++++++++++
122 In previous Ganeti versions, an instance creation request was not
123 limited on the minimum size and on the maximum size just by the cluster
124 resources. As such, any policy could be implemented only in third-party
125 clients (RAPI clients, or shell wrappers over ``gnt-*``
126 tools). Furthermore, calculating cluster capacity via ``hspace`` again
127 required external input with regards to instance sizes.
129 In order to improve these workflows and to allow for example better
130 per-node group differentiation, we introduced instance specs, which
133 - minimum instance disk size, disk count, memory size, cpu count
134 - maximum values for the above metrics
135 - and “standard” values (used in ``hspace`` to calculate the standard
138 The minimum/maximum values can be also customised at node-group level,
139 for example allowing more powerful hardware to support bigger instance
142 Beside the instance specs, there are a few other settings belonging to
143 the instance policy framework. It is possible now to customise, per
144 cluster and node-group:
146 - the list of allowed disk templates
147 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
148 - the maximum ratio of instance to spindles (see below for more
149 information) for local storage
151 All these together should allow all tools that talk to Ganeti to know
152 what are the ranges of allowed values for instances and the
153 over-subscription that is allowed.
155 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
156 instance configuration, and the physical CPU configuration from the
157 node. For the spindle ratios however, we didn't track before these
158 values, so new parameters have been added:
160 - a new node parameter ``spindle_count``, defaults to 1, customisable at
161 node group or node level
162 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
164 Note that spindles in this context doesn't need to mean actual
165 mechanical hard-drives; it's just a relative number for both the node
166 I/O capacity and instance I/O consumption.
168 Instance migration behaviour
169 ++++++++++++++++++++++++++++
171 While live-migration is in general desirable over failover, it is
172 possible that for some workloads it is actually worse, due to the
173 variable time of the “suspend” phase during live migration.
175 To allow the tools to work consistently over such instances (without
176 having to hard-code instance names), a new backend parameter
177 ``always_failover`` has been added to control the migration/failover
178 behaviour. When set to True, all migration requests for an instance will
179 instead fall-back to failover.
181 Instance memory ballooning
182 ++++++++++++++++++++++++++
184 Initial support for memory ballooning has been added. The memory for an
185 instance is no longer fixed (backend parameter ``memory``), but instead
186 can vary between minimum and maximum values (backend parameters
187 ``minmem`` and ``maxmem``). Currently we only change an instance's
190 - live migrating or failing over and instance and the target node
191 doesn't have enough memory
192 - user requests changing the memory via ``gnt-instance modify
198 In order to control the use of specific CPUs by instance, support for
199 controlling CPU pinning has been added for the Xen, HVM and LXC
200 hypervisors. This is controlled by a new hypervisor parameter
201 ``cpu_mask``; details about possible values for this are in the
202 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
203 VCPU-to-CPU mapping) form will work well only when all nodes in your
204 cluster have the same amount of CPUs.
209 Another area in which Ganeti was not customisable were the parameters
210 used for storage configuration, e.g. how many stripes to use for LVM,
211 DRBD resync configuration, etc.
213 To improve this area, we've added disks parameters, which are
214 customisable at cluster and node group level, and which allow to
215 specify various parameters for disks (DRBD has the most parameters
216 currently), for example:
218 - DRBD resync algorithm and parameters (e.g. speed)
219 - the default VG for meta-data volumes for DRBD
220 - number of stripes for LVM (plain disk template)
223 These parameters can be modified via ``gnt-cluster modify -D …`` and
224 ``gnt-group modify -D …``, and are used at either instance creation (in
225 case of LVM stripes, for example) or at disk “activation” time
228 Rados block device support
229 ++++++++++++++++++++++++++
231 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
232 denoted by the ``rbd`` disk template type. This is considered
233 experimental, feedback is welcome. For details on configuring it, see
234 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
239 The existing master IP functionality works well only in simple setups (a
240 single network shared by all nodes); however, if nodes belong to
241 different networks, then the ``/32`` setup and lack of routing
242 information is not enough.
244 To allow the master IP to function well in more complex cases, the
245 system was reworked as follows:
247 - a master IP netmask setting has been added
248 - the master IP activation/turn-down code was moved from the node daemon
250 - whether to run the Ganeti-supplied master IP script or a user-supplied
251 on is a ``gnt-cluster init`` setting
253 Details about the location of the standard and custom setup scripts are
254 in the man page :manpage:`gnt-cluster(8)`; for information about the
255 setup script protocol, look at the Ganeti-supplied script.
260 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
263 It is now possible to use TLS-protected connections, and when renewing
264 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
265 it is now possible to specify spice or spice CA certificates. Also, it
266 is possible to configure a password for SPICE sessions via the
267 hypervisor parameter ``spice_password_file``.
269 There are also new parameters to control the compression and streaming
270 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
271 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
272 for the spice parameters.
274 Lastly, it is now possible to see the SPICE connection information via
275 ``gnt-instance console``.
280 A new tool (``tools/ovfconverter``) has been added that supports
281 conversion between Ganeti and the `Open Virtualization Format
282 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
285 This relies on the ``qemu-img`` tool to convert the disk formats, so the
286 actual compatibility with other virtualization solutions depends on it.
291 The configuration query daemon (``ganeti-confd``) is now optional, and
292 has been rewritten in Haskell; whether to use the daemon at all, use the
293 Python (default) or the Haskell version is selectable at configure time
294 via the ``--enable-confd`` parameter, which can take one of the
295 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
296 daemon will result in a smaller footprint; for larger systems, we
297 welcome feedback on the Haskell version which might become the default
300 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
301 daemon running. The Python version doesn't implement the new call.
304 User interface changes
305 ~~~~~~~~~~~~~~~~~~~~~~
307 We have replaced the ``--disks`` option of ``gnt-instance
308 replace-disks`` with a more flexible ``--disk`` option, which allows
309 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
310 disk size and mode can be changed upon recreation (via ``gnt-instance
311 recreate-disks``, which accepts the same ``--disk`` option).
313 As many people are used to a ``show`` command, we have added that as an
314 alias to ``info`` on all ``gnt-*`` commands.
316 The ``gnt-instance grow-disk`` command has a new mode in which it can
317 accept the target size of the disk, instead of the delta; this can be
318 more safe since two runs in absolute mode will be idempotent, and
319 sometimes it's also easier to specify the desired size directly.
321 Also the handling of instances with regard to offline secondaries has
322 been improved. Instance operations should not fail because one of it's
323 secondary nodes is offline, even though it's safe to proceed.
325 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
326 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
327 minors to instance name.
332 RAPI coverage has improved, with (for example) new resources for
333 recreate-disks, node power-cycle, etc.
338 There is partial support for ``xl`` in the Xen hypervisor; feedback is
341 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
342 whether to still support Python 2.4 or move to Python 2.6 as minimum
345 Support for Fedora has been slightly improved; the provided example
346 init.d script should work better on it and the INSTALL file should
347 document the needed dependencies.
352 The deprecated ``QueryLocks`` LUXI request has been removed. Use
353 ``Query(what=QR_LOCK, ...)`` instead.
355 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
356 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
357 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
358 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
359 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
361 RAPI client: ``CertificateError`` now derives from
362 ``GanetiApiError``. This should make it more easy to handle Ganeti
365 Deprecation warnings due to PyCrypto/paramiko import in
366 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
367 make sure to run an up-to-date paramiko version, if you use this tool.
369 The QA scripts now depend on Python 2.5 or above (the main code base
370 still works with Python 2.4).
372 The configuration file (``config.data``) is now written without
373 indentation for performance reasons; if you want to edit it, it can be
374 re-formatted via ``tools/fmtjson``.
376 A number of bugs has been fixed in the cluster merge tool.
378 ``x509`` certification verification (used in import-export) has been
379 changed to allow the same clock skew as permitted by the cluster
380 verification. This will remove some rare but hard to diagnose errors in
387 *(Released Thu, 19 Jul 2012)*
389 Very few changes from rc4 to the final release, only bugfixes:
391 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
392 instance, fix CDROM booting for KVM instances)
393 - fixed node group modification of node parameters
394 - fixed issue in LUClusterVerifyGroup with multi-group clusters
395 - fixed generation of bash completion to ensure a stable ordering
402 *(Released Fri, 13 Jul 2012)*
404 Third release candidate for 2.6. The following changes were done from
407 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
408 - Fixed an inconsistency in the LUXI protocol with the provided
409 arguments (NOT backwards compatible)
410 - Fixed a bug with node groups ipolicy where ``min`` was greater than
411 the cluster ``std`` value
412 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
413 easier instance debugging on nodes (requires ``hconfd`` to work)
419 *(Released Tue, 03 Jul 2012)*
421 Second release candidate for 2.6. The following changes were done from
424 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
426 - Fixed a RAPI regression on missing beparams/memory
427 - Fixed redistribution of files on offline nodes
428 - Added possibility to run activate-disks even though secondaries are
429 offline. With this change it relaxes also the strictness on some other
430 commands which use activate disks internally:
431 * ``gnt-instance start|reboot|rename|backup|export``
432 - Made it possible to remove safely an instance if its secondaries are
434 - Made it possible to reinstall even though secondaries are offline
440 *(Released Mon, 25 Jun 2012)*
442 First release candidate for 2.6. The following changes were done from
445 - Fixed bugs with disk parameters and ``rbd`` templates as well as
447 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
448 behaviour. It supports now the modify operation
449 - ``hcheck`` implemented to analyze cluster health and possibility of
450 improving health by rebalance
451 - ``hbal`` has been improved in dealing with split instances
457 *(Released Mon, 11 Jun 2012)*
459 Second beta release of 2.6. The following changes were done from beta2
462 - Fixed ``daemon-util`` with non-root user models
463 - Fixed creation of plain instances with ``--no-wait-for-sync``
464 - Fix wrong iv_names when running ``cfgupgrade``
465 - Export more information in RAPI group queries
466 - Fixed bug when changing instance network interfaces
467 - Extended burnin to do NIC changes
468 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
469 - Changed default for DRBD barriers
470 - Fixed DRBD error reporting for syncer rate
471 - Verify the options on disk parameters
473 And of course various fixes to documentation and improved unittests and
480 *(Released Wed, 23 May 2012)*
482 First beta release of 2.6. The following changes were done from beta1 to
485 - integrated patch for distributions without ``start-stop-daemon``
486 - adapted example init.d script to work on Fedora
487 - fixed log handling in Haskell daemons
488 - adapted checks in the watcher for pycurl linked against libnss
489 - add partial support for ``xl`` instead of ``xm`` for Xen
490 - fixed a type issue in cluster verification
491 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
494 Plus integrated fixes from the 2.5 branch:
496 - fixed ``kvm-ifup`` to use ``/bin/bash``
497 - fixed parallel build failures
498 - KVM live migration when using a custom keymap
504 *(Released Tue, 24 Jul 2012)*
506 A small bugfix release, with no new features:
508 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
509 different default shell (e.g. Debian, Ubuntu)
510 - fixed KVM startup and live migration with a custom keymap (fixes Issue
511 243 and Debian bug #650664)
512 - fixed compatibility with KVM versions that don't support multiple boot
513 devices (fixes Issue 230 and Debian bug #624256)
515 Additionally, a few fixes were done to the build system (fixed parallel
516 build failures) and to the unittests (fixed race condition in test for
517 FileID functions, and the default enable/disable mode for QA test is now
524 *(Released Fri, 11 May 2012)*
526 A small bugfix release.
528 The main issues solved are on the topic of compatibility with newer LVM
531 - fixed parsing of ``lv_attr`` field
532 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
533 the ``--force`` flag is needed
535 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
536 to accept kernel 3.x too (was hardcoded to 2.6.*).
538 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
539 of hook results and that also made display of other errors suboptimal
540 was fixed; the code behaves now like 2.4 and earlier.
542 Another change in 2.5, the cleanup of the OS scripts environment, is too
543 aggressive: it removed even the ``PATH`` variable, which requires the OS
544 scripts to *always* need to export it. Since this is a bit too strict,
545 we now export a minimal PATH, the same that we export for hooks.
547 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
548 integrated into this release.
550 Finally, a few other miscellaneous changes were done (no new features,
551 just small improvements):
553 - Fix ``gnt-group --help`` display
554 - Fix hardcoded Xen kernel path
555 - Fix grow-disk handling of invalid units
556 - Update synopsis for ``gnt-cluster repair-disk-sizes``
557 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
563 *(Released Thu, 12 Apr 2012)*
565 Incompatible/important changes and bugfixes
566 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
568 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
569 resource's ``ip_check`` parameter changed from ``True`` to ``False``
570 to match the underlying LUXI interface.
571 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
572 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
573 not maintain backwards-compatibility as the underlying operation
574 changed in an incompatible way. The RAPI client can talk to old
575 servers, but it needs to be told so as the return value changed.
576 - When creating file-based instances via RAPI, the ``file_driver``
577 parameter no longer defaults to ``loop`` and must be specified.
578 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
580 - Support for the undocumented and deprecated RAPI instance creation
581 request format version 0 has been dropped. Use version 1, supported
582 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
583 - Pyparsing 1.4.6 or above is required, see :doc:`installation
584 documentation <install>`.
585 - The "cluster-verify" hooks are now executed per group by the
586 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
587 if you just run ``gnt-cluster verify``, which generates one opcode per
589 - The environment as passed to the OS scripts is cleared, and thus no
590 environment variables defined in the node daemon's environment will be
591 inherited by the scripts.
592 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
594 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
595 support operations involving multiple node groups.
596 - Offline nodes are ignored when failing over an instance.
597 - Support for KVM version 1.0, which changed the version reporting format
599 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
601 - ``Makefile`` is now compatible with Automake 1.11.2
602 - Includes all bugfixes made in the 2.4 series
607 - The ganeti-htools project has been merged into the ganeti-core source
608 tree and will be built as part of Ganeti (see :doc:`install-quick`).
609 - Implemented support for :doc:`shared storage <design-shared-storage>`.
610 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
611 GPT-style partition tables (requires `parted
612 <http://www.gnu.org/s/parted/>`_).
613 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
614 - Allowed adding tags on instance creation.
615 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
616 - Allow instances to be started in a paused state, enabling the user to
617 see the complete console output on boot using the console.
618 - Added new hypervisor flag to control default reboot behaviour
619 (``reboot_behavior``).
620 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
621 - Improved out-of-band management support:
623 - Added ``gnt-node health`` command reporting the health status of
625 - Added ``gnt-node power`` command to manage power status of nodes.
626 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
628 - Instance migration can fall back to failover if instance is not
630 - Filters can be used when listing nodes, instances, groups and locks;
631 see *ganeti(7)* manpage.
632 - Added post-execution status as variables to :doc:`hooks <hooks>`
634 - Instance tags are exported/imported together with the instance.
635 - When given an explicit job ID, ``gnt-job info`` will work for archived
637 - Jobs can define dependencies on other jobs (not yet supported via
638 RAPI or command line, but used by internal commands and usable via
641 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
644 - Instance failover is now available as a RAPI resource
645 (``/2/instances/[instance_name]/failover``).
646 - ``gnt-instance info`` defaults to static information if primary node
648 - Opcodes have a new ``comment`` attribute.
649 - Added basic SPICE support to KVM hypervisor.
650 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
652 Node group improvements
653 ~~~~~~~~~~~~~~~~~~~~~~~
655 - ``gnt-cluster verify`` has been modified to check groups separately,
656 thereby improving performance.
657 - Node group support has been added to ``gnt-cluster verify-disks``,
658 which now operates per node group.
659 - Watcher has been changed to work better with node groups.
661 - One process and state file per node group.
662 - Slow watcher in one group doesn't block other group's watcher.
664 - Added new command, ``gnt-group evacuate``, to move all instances in a
665 node group to other groups.
666 - Added ``gnt-instance change-group`` to move an instance to another
668 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
669 per-group operations.
670 - Node groups can be tagged.
671 - Some operations switch from an exclusive to a shared lock as soon as
673 - Instance's primary and secondary nodes' groups are now available as
674 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
675 and ``snodes.group.uuid``).
680 - Numerous updates to documentation and manpages.
682 - :doc:`RAPI <rapi>` documentation now has detailed parameter
684 - Some opcode/job results are now also documented, see :doc:`RAPI
687 - A lockset's internal lock is now also visible in lock monitor.
688 - Log messages from job queue workers now contain information about the
689 opcode they're processing.
690 - ``gnt-instance console`` no longer requires the instance lock.
691 - A short delay when waiting for job changes reduces the number of LUXI
692 requests significantly.
693 - DRBD metadata volumes are overwritten with zeros during disk creation.
694 - Out-of-band commands no longer acquire the cluster lock in exclusive
696 - ``devel/upload`` now uses correct permissions for directories.
702 *(Released Fri, 23 Mar 2012)*
704 This was the sixth release candidate of the 2.5 series.
710 *(Released Mon, 9 Jan 2012)*
712 This was the fifth release candidate of the 2.5 series.
718 *(Released Thu, 27 Oct 2011)*
720 This was the fourth release candidate of the 2.5 series.
726 *(Released Wed, 26 Oct 2011)*
728 This was the third release candidate of the 2.5 series.
734 *(Released Tue, 18 Oct 2011)*
736 This was the second release candidate of the 2.5 series.
742 *(Released Tue, 4 Oct 2011)*
744 This was the first release candidate of the 2.5 series.
750 *(Released Wed, 31 Aug 2011)*
752 This was the third beta release of the 2.5 series.
758 *(Released Mon, 22 Aug 2011)*
760 This was the second beta release of the 2.5 series.
766 *(Released Fri, 12 Aug 2011)*
768 This was the first beta release of the 2.5 series.
774 *(Released Thu, 27 Oct 2011)*
776 - Fixed bug when parsing command line parameter values ending in
778 - Fixed assertion error after unclean master shutdown
779 - Disable HTTP client pool for RPC, significantly reducing memory usage
781 - Fixed queue archive creation with wrong permissions
787 *(Released Tue, 23 Aug 2011)*
791 - Fixed documentation for importing with ``--src-dir`` option
792 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
793 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
799 *(Released Fri, 5 Aug 2011)*
801 Many bug-fixes and a few small features:
803 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
804 issues when you tried to add an instance with two MAC addresses in one
806 - KVM: fixed per-instance stored UID value
807 - KVM: configure bridged NICs at migration start
808 - KVM: Fix a bug where instance will not start with never KVM versions
810 - Added OS search path to ``gnt-cluster info``
811 - Fixed an issue with ``file_storage_dir`` where you were forced to
812 provide an absolute path, but the documentation states it is a
813 relative path, the documentation was right
814 - Added a new parameter to instance stop/start called ``--no-remember``
815 that will make the state change to not be remembered
816 - Implemented ``no_remember`` at RAPI level
817 - Improved the documentation
818 - Node evacuation: don't call IAllocator if node is already empty
819 - Fixed bug in DRBD8 replace disks on current nodes
820 - Fixed bug in recreate-disks for DRBD instances
821 - Moved assertion checking locks in ``gnt-instance replace-disks``
822 causing it to abort with not owning the right locks for some situation
823 - Job queue: Fixed potential race condition when cancelling queued jobs
824 - Fixed off-by-one bug in job serial generation
825 - ``gnt-node volumes``: Fix instance names
826 - Fixed aliases in bash completion
827 - Fixed a bug in reopening log files after being sent a SIGHUP
828 - Added a flag to burnin to allow specifying VCPU count
829 - Bugfixes to non-root Ganeti configuration
835 *(Released Thu, 12 May 2011)*
837 Many bug-fixes and a few new small features:
839 - Fixed a bug related to log opening failures
840 - Fixed a bug in instance listing with orphan instances
841 - Fixed a bug which prevented resetting the cluster-level node parameter
842 ``oob_program`` to the default
843 - Many fixes related to the ``cluster-merge`` tool
844 - Fixed a race condition in the lock monitor, which caused failures
845 during (at least) creation of many instances in parallel
846 - Improved output for gnt-job info
847 - Removed the quiet flag on some ssh calls which prevented debugging
849 - Improved the N+1 failure messages in cluster verify by actually
850 showing the memory values (needed and available)
851 - Increased lock attempt timeouts so that when executing long operations
852 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
853 too early and thus prevent the use of the 'fair' mechanism
854 - Changed instance query data (``gnt-instance info``) to not acquire
855 locks unless needed, thus allowing its use on locked instance if only
856 static information is asked for
857 - Improved behaviour with filesystems that do not support rename on an
859 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
860 kept locks on all nodes during the wipe, which is unneeded
861 - Fixed ``gnt-watcher`` handling of errors during hooks execution
862 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
863 10GiB) which caused the wipe to fail right at the end in some cases
864 - Fixed master IP activation when doing master failover with no-voting
865 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
866 the master node itself
867 - Fixed potential data-loss in under disk full conditions, where Ganeti
868 wouldn't check correctly the return code and would consider
869 partially-written files 'correct'
870 - Fixed bug related to multiple VGs and DRBD disk replacing
871 - Added new disk parameter ``metavg`` that allows placement of the meta
872 device for DRBD in a different volume group
873 - Fixed error handling in the node daemon when the system libc doesn't
874 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
875 - Fixed lock release during replace-disks, which kept cluster-wide locks
876 when doing disk replaces with an iallocator script
877 - Added check for missing bridges in cluster verify
878 - Handle EPIPE errors while writing to the terminal better, so that
879 piping the output to e.g. ``less`` doesn't cause a backtrace
880 - Fixed rare case where a ^C during Luxi calls could have been
881 interpreted as server errors, instead of simply terminating
882 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
884 - Added a few more parameters to the KVM hypervisor, allowing a second
885 CDROM, custom disk type for CDROMs and a floppy image
886 - Removed redundant message in instance rename when the name is given
888 - Added option to ``gnt-instance recreate-disks`` to allow creating the
889 disks on new nodes, allowing recreation when the original instance
890 nodes are completely gone
891 - Added option when converting disk templates to DRBD to skip waiting
892 for the resync, in order to make the instance available sooner
893 - Added two new variables to the OS scripts environment (containing the
895 - Made the root_path and optional parameter for the xen-pvm hypervisor,
896 to allow use of ``pvgrub`` as bootloader
897 - Changed the instance memory modifications to only check out-of-memory
898 conditions on memory increases, and turned the secondary node warnings
899 into errors (they can still be overridden via ``--force``)
900 - Fixed the handling of a corner case when the Python installation gets
901 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
902 to execute a command that doesn't exist
903 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
904 node of the instance returned failures during instance shutdown; this
905 adds the option ``--ignore-consistency`` to gnt-instance move
907 And as usual, various improvements to the error messages, documentation
914 *(Released Wed, 09 Mar 2011)*
916 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
917 the RAPI users file if run twice (even with ``--dry-run``).
919 The release fixes that bug (nothing else changed).
925 *(Released Mon, 07 Mar 2011)*
927 Final 2.4.0 release. Just a few small fixes:
929 - Fixed RAPI node evacuate
930 - Fixed the kvm-ifup script
931 - Fixed internal error handling for special job cases
932 - Updated man page to specify the escaping feature for options
938 *(Released Mon, 28 Feb 2011)*
940 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
941 that this feature wiped the disks of the wrong instance, leading to loss
946 - Fixed title of query field containing instance name
947 - Expanded the glossary in the documentation
948 - Fixed one unittest (internal issue)
954 *(Released Mon, 21 Feb 2011)*
956 A number of bug fixes plus just a couple functionality changes.
958 On the user-visible side, the ``gnt-* list`` command output has changed
959 with respect to "special" field states. The current rc1 style of display
960 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
961 the default output mode special fields states are displayed as follows:
963 - Offline resource: ``*``
964 - Unavailable/not applicable: ``-``
965 - Data missing (RPC failure): ``?``
966 - Unknown field: ``??``
968 Another user-visible change is the addition of ``--force-join`` to
973 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
974 - Fixed regression in RAPI/instance reinstall where all parameters were
975 required (instead of optional)
976 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
977 - Fixed iallocator usage (offline nodes were not considered offline)
978 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
979 - Fixed hypervisor and OS parameter validation with respect to
981 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
983 - Fixed ``tools/listrunner`` with respect to agent-based usage
989 *(Released Fri, 4 Feb 2011)*
991 Many changes and fixes since the beta1 release. While there were some
992 internal changes, the code has been mostly stabilised for the RC
995 Note: the dumb allocator was removed in this release, as it was not kept
996 up-to-date with the IAllocator protocol changes. It is recommended to
997 use the ``hail`` command from the ganeti-htools package.
999 Note: the 2.4 and up versions of Ganeti are not compatible with the
1000 0.2.x branch of ganeti-htools. You need to upgrade to
1001 ganeti-htools-0.3.0 (or later).
1003 Regressions fixed from 2.3
1004 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1006 - Fixed the ``gnt-cluster verify-disks`` command
1007 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1009 - Fixed disk adoption breakage
1010 - Fixed wrong headers in instance listing for field aliases
1015 - Fixed corner case in KVM handling of NICs
1016 - Fixed many cases of wrong handling of non-vm_capable nodes
1017 - Fixed a bug where a missing instance symlink was not possible to
1018 recreate with any ``gnt-*`` command (now ``gnt-instance
1019 activate-disks`` does it)
1020 - Fixed the volume group name as reported by ``gnt-cluster
1022 - Increased timeouts for the import-export code, hopefully leading to
1023 fewer aborts due network or instance timeouts
1024 - Fixed bug in ``gnt-node list-storage``
1025 - Fixed bug where not all daemons were started on cluster
1026 initialisation, but only at the first watcher run
1027 - Fixed many bugs in the OOB implementation
1028 - Fixed watcher behaviour in presence of instances with offline
1030 - Fixed instance list output for instances running on the wrong node
1031 - a few fixes to the cluster-merge tool, but it still cannot merge
1032 multi-node groups (currently it is not recommended to use this tool)
1038 - Improved network configuration for the KVM hypervisor
1039 - Added e1000 as a supported NIC for Xen-HVM
1040 - Improved the lvmstrap tool to also be able to use partitions, as
1041 opposed to full disks
1042 - Improved speed of disk wiping (the cluster parameter
1043 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1044 of instance creations
1045 - Added documentation for the OS parameters
1046 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1047 hypervisor is not responding
1048 - Added display of blacklisted and hidden OS information in
1049 ``gnt-cluster info``
1050 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1051 NIC and node parameters, which might create problems with currently
1052 invalid (but undetected) configuration files, but prevents validation
1053 failures when unrelated parameters are modified
1054 - Changed cluster initialisation to wait for the master daemon to become
1056 - Expanded the RAPI interface:
1058 - Added config redistribution resource
1059 - Added activation/deactivation of instance disks
1060 - Added export of console information
1062 - Implemented log file reopening on SIGHUP, which allows using
1063 logrotate(8) for the Ganeti log files
1064 - Added a basic OOB helper script as an example
1070 *(Released Fri, 14 Jan 2011)*
1075 - Fixed timezone issues when formatting timestamps
1076 - Added support for node groups, available via ``gnt-group`` and other
1078 - Added out-of-band framework and management, see :doc:`design
1079 document <design-oob>`
1080 - Removed support for roman numbers from ``gnt-node list`` and
1081 ``gnt-instance list``.
1082 - Allowed modification of master network interface via ``gnt-cluster
1083 modify --master-netdev``
1084 - Accept offline secondaries while shutting down instance disks
1085 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1086 - Added support for multiple LVM volume groups
1087 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1089 - Added commands to list available fields:
1091 - ``gnt-node list-fields``
1092 - ``gnt-group list-fields``
1093 - ``gnt-instance list-fields``
1095 - Updated documentation and man pages
1100 - Moved ``rapi_users`` file into separate directory, now named
1101 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1103 - Added new tool for running commands on many machines,
1104 ``tools/ganeti-listrunner``
1105 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1106 improving the ``gnt-instance console`` output
1107 - Allowed customisation of disk index separator at ``configure`` time
1108 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1109 - Added support for non-partitioned md disks in ``lvmstrap``
1110 - Added script to gracefully power off KVM instances
1111 - Split ``utils`` module into smaller parts
1112 - Changed query operations to return more detailed information, e.g.
1113 whether an information is unavailable due to an offline node. To use
1114 this new functionality, the LUXI call ``Query`` must be used. Field
1115 information is now stored by the master daemon and can be retrieved
1116 using ``QueryFields``. Instances, nodes and groups can also be queried
1117 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1118 exposed via RAPI). The following commands make use of this
1119 infrastructure change:
1121 - ``gnt-group list``
1122 - ``gnt-group list-fields``
1124 - ``gnt-node list-fields``
1125 - ``gnt-instance list``
1126 - ``gnt-instance list-fields``
1127 - ``gnt-debug locks``
1132 - New RAPI resources (see :doc:`rapi`):
1136 - ``/2/groups/[group_name]``
1137 - ``/2/groups/[group_name]/assign-nodes``
1138 - ``/2/groups/[group_name]/modify``
1139 - ``/2/groups/[group_name]/rename``
1140 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1144 - Implemented ``no_install`` for instance creation
1145 - Implemented OS parameters for instance reinstallation, allowing
1146 use of special settings on reinstallation (e.g. for preserving data)
1151 - Added IPv6 support in import/export
1152 - Pause DRBD synchronization while wiping disks on instance creation
1153 - Updated unittests and QA scripts
1154 - Improved network parameters passed to KVM
1155 - Converted man pages from docbook to reStructuredText
1161 *(Released Mon, 20 Dec 2010)*
1163 Released version 2.3.1~rc1 without any changes.
1169 *(Released Wed, 1 Dec 2010)*
1171 - impexpd: Disable OpenSSL compression in socat if possible (backport
1172 from master, commit e90739d625b, see :doc:`installation guide
1173 <install-quick>` for details)
1174 - Changed unittest coverage report to exclude test scripts
1175 - Added script to check version format
1181 *(Released Wed, 1 Dec 2010)*
1183 Released version 2.3.0~rc1 without any changes.
1189 *(Released Fri, 19 Nov 2010)*
1191 A number of bugfixes and documentation updates:
1193 - Update ganeti-os-interface documentation
1194 - Fixed a bug related to duplicate MACs or similar items which should be
1196 - Fix breakage in OS state modify
1197 - Reinstall instance: disallow offline secondaries (fixes bug related to
1198 OS changing but reinstall failing)
1199 - plus all the other fixes between 2.2.1 and 2.2.2
1205 *(Released Tue, 2 Nov 2010)*
1207 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1208 - Fixed master failover race with watcher
1209 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1211 - Accept previously stopped instance for export with instance removal
1212 - Simplify and extend the environment variables for instance OS scripts
1213 - Added new node flags, ``master_capable`` and ``vm_capable``
1214 - Added optional instance disk wiping prior during allocation. This is a
1215 cluster-wide option and can be set/modified using
1216 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1217 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1218 :doc:`install-quick`
1219 - Added a new watcher option (``--ignore-pause``)
1220 - Added option to ignore offline node on instance start/stop
1221 (``--ignore-offline``)
1222 - Allow overriding OS parameters with ``gnt-instance reinstall``
1223 - Added ability to change node's secondary IP address using ``gnt-node
1225 - Implemented privilege separation for all daemons except
1226 ``ganeti-noded``, see ``configure`` options
1227 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1229 - Implemented job priorities (see ``ganeti(7)`` manpage)
1230 - Ignore failures while shutting down instances during failover from
1232 - Exit daemon's bootstrap process only once daemon is ready
1233 - Export more information via ``LUInstanceQuery``/remote API
1234 - Improved documentation, QA and unittests
1235 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1236 restart if the file was created or changed
1237 - Added LUXI protocol version sent with each request and response,
1238 allowing detection of server/client mismatches
1239 - Moved the Python scripts among gnt-* and ganeti-* into modules
1240 - Moved all code related to setting up SSH to an external script,
1242 - Infrastructure changes for node group support in future versions
1248 *(Released Fri, 19 Nov 2010)*
1250 A few small bugs fixed, and some improvements to the build system:
1252 - Fix documentation regarding conversion to drbd
1253 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1255 - Fix error handling in node modify with multiple changes
1256 - Allow remote imports without checked names
1262 *(Released Tue, 19 Oct 2010)*
1264 - Disable SSL session ID cache in RPC client
1270 *(Released Thu, 14 Oct 2010)*
1272 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1273 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1275 - Fix problems with interaction between Curl and Python's HTTP server,
1276 resulting in increased speed in many RPC calls
1277 - Improve our release script to prevent breakage with older aclocal and
1284 *(Released Thu, 7 Oct 2010)*
1286 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1287 value retrieved from master
1288 - Added support for blacklisted or hidden OS definitions
1289 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1290 - Added support for -mem-path in KVM hypervisor abstraction layer
1291 - Allow overriding instance parameters in tool for inter-cluster
1292 instance moves (``tools/move-instance``)
1293 - Improved opcode summaries (e.g. in ``gnt-job list``)
1294 - Improve consistency of OS listing by sorting it
1295 - Documentation updates
1301 *(Released Fri, 8 Oct 2010)*
1303 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1309 *(Released Mon, 4 Oct 2010)*
1311 - Fixed regression in ``gnt-instance rename``
1317 *(Released Wed, 22 Sep 2010)*
1319 - Fixed OS_VARIANT variable for OS scripts
1320 - Fixed cluster tag operations via RAPI
1321 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1322 - Disabled RAPI CA checks in watcher
1328 *(Released Mon, 23 Aug 2010)*
1330 - Support DRBD versions of the format "a.b.c.d"
1332 - Re-introduce support for usage from multiple threads in RAPI client
1333 - Instance renames and modify via RAPI
1334 - Work around race condition between processing and archival in job
1336 - Mark opcodes following failed one as failed, too
1337 - Job field ``lock_status`` was removed due to difficulties making it
1338 work with the changed job queue in Ganeti 2.2; a better way to monitor
1339 locks is expected for a later 2.2.x release
1340 - Fixed dry-run behaviour with many commands
1341 - Support ``ssh-agent`` again when adding nodes
1342 - Many additional bugfixes
1348 *(Released Fri, 30 Jul 2010)*
1350 Important change: the internal RPC mechanism between Ganeti nodes has
1351 changed from using a home-grown http library (based on the Python base
1352 libraries) to use the PycURL library. This requires that PycURL is
1353 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1354 against GnuTLS by default. cURL's support for GnuTLS had known issues
1355 before cURL 7.21.0 and we recommend using the latest cURL release or
1356 linking against OpenSSL. Most other distributions already link PycURL
1357 and cURL against OpenSSL. The command::
1359 python -c 'import pycurl; print pycurl.version'
1361 can be used to determine the libraries PycURL and cURL are linked
1364 Other significant changes:
1366 - Rewrote much of the internals of the job queue, in order to achieve
1367 better parallelism; this decouples job query operations from the job
1368 processing, and it should allow much nicer behaviour of the master
1369 daemon under load, and it also has uncovered some long-standing bugs
1370 related to the job serialisation (now fixed)
1371 - Added a default iallocator setting to the cluster parameters,
1372 eliminating the need to always pass nodes or an iallocator for
1373 operations that require selection of new node(s)
1374 - Added experimental support for the LXC virtualization method
1375 - Added support for OS parameters, which allows the installation of
1376 instances to pass parameter to OS scripts in order to customise the
1378 - Added a hypervisor parameter controlling the migration type (live or
1379 non-live), since hypervisors have various levels of reliability; this
1380 has renamed the 'live' parameter to 'mode'
1381 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1382 logical volumes, meaning that cluster verify will ignore them and not
1383 flag their presence as errors
1384 - The watcher will now reset the error count for failed instances after
1385 8 hours, thus allowing self-healing if the problem that caused the
1386 instances to be down/fail to start has cleared in the meantime
1387 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1388 check for, and warn, if the drbd module parameter ``usermode_helper``
1389 is not consistent with the cluster-wide setting; this is needed to
1390 make diagnose easier of failed drbd creations
1391 - Started adding base IPv6 support, but this is not yet
1392 enabled/available for use
1393 - Rename operations (cluster, instance) will now return the new name,
1394 which is especially useful if a short name was passed in
1395 - Added support for instance migration in RAPI
1396 - Added a tool to pre-configure nodes for the SSH setup, before joining
1397 them to the cluster; this will allow in the future a simplified model
1398 for node joining (but not yet fully enabled in 2.2); this needs the
1399 paramiko python library
1400 - Fixed handling of name-resolving errors
1401 - Fixed consistency of job results on the error path
1402 - Fixed master-failover race condition when executed multiple times in
1404 - Fixed many bugs related to the job queue (mostly introduced during the
1405 2.2 development cycle, so not all are impacting 2.1)
1406 - Fixed instance migration with missing disk symlinks
1407 - Fixed handling of unknown jobs in ``gnt-job archive``
1408 - And many other small fixes/improvements
1412 - Enhanced both the unittest and the QA coverage
1413 - Switched the opcode validation to a generic model, and extended the
1414 validation to all opcode parameters
1415 - Changed more parts of the code that write shell scripts to use the
1417 - Switched the master daemon to use the asyncore library for the Luxi
1424 *(Released Thu, 17 Jun 2010)*
1426 - Added tool (``move-instance``) and infrastructure to move instances
1427 between separate clusters (see :doc:`separate documentation
1428 <move-instance>` and :doc:`design document <design-2.2>`)
1429 - Added per-request RPC timeout
1430 - RAPI now requires a Content-Type header for requests with a body (e.g.
1431 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1432 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1433 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1435 - Implemented initial support for running Ganeti daemons as separate
1436 users, see configure-time flags ``--with-user-prefix`` and
1437 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1439 - Instances can be removed after export (``gnt-backup export
1440 --remove-instance``)
1441 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1442 key (instead of 1024 bit)
1443 - Added new cluster configuration file for cluster domain secret
1444 - Import/export now use SSL instead of SSH
1445 - Added support for showing estimated time when exporting an instance,
1446 see the ``ganeti-os-interface(7)`` manpage and look for
1453 *(Released Tue, 16 Nov 2010)*
1455 Some more bugfixes. Unless critical bugs occur, this will be the last
1458 - Fix case of MAC special-values
1459 - Fix mac checker regex
1460 - backend: Fix typo causing "out of range" error
1461 - Add missing --units in gnt-instance list man page
1467 *(Released Tue, 24 Aug 2010)*
1470 - Don't ignore secondary node silently on non-mirrored disk templates
1472 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1473 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1474 - Properly document the usb_mouse parameter
1475 - Fix path in ganeti-rapi(8) (issue 116)
1476 - Adjust error message when the ganeti user's .ssh directory is
1478 - Add same-node-check when changing the disk template to drbd
1484 *(Released Fri, 16 Jul 2010)*
1487 - Add an option to only select some reboot types during qa/burnin.
1488 (on some hypervisors consequent reboots are not supported)
1489 - Fix infrequent race condition in master failover. Sometimes the old
1490 master ip address would be still detected as up for a short time
1491 after it was removed, causing failover to fail.
1492 - Decrease mlockall warnings when the ctypes module is missing. On
1493 Python 2.4 we support running even if no ctypes module is installed,
1494 but we were too verbose about this issue.
1495 - Fix building on old distributions, on which man doesn't have a
1497 - Fix RAPI not to ignore the MAC address on instance creation
1498 - Implement the old instance creation format in the RAPI client.
1504 *(Released Thu, 01 Jul 2010)*
1506 A small bugfix release:
1507 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1508 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1509 a non-existing option
1510 - Fix instance create: the --force-variant option was ignored
1511 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1512 - Fix modify node storage with non-FQDN arguments
1513 - Fix RAPI client to authenticate under Python 2.6 when used
1514 for more than 5 requests needing authentication
1515 - Fix gnt-instance modify -t (storage) giving a wrong error message
1516 when converting a non-shutdown drbd instance to plain
1522 *(Released Fri, 18 Jun 2010)*
1524 A small bugfix release:
1526 - Fix live migration of KVM instances started with older Ganeti
1527 versions which had fewer hypervisor parameters
1528 - Fix gnt-instance grow-disk on down instances
1529 - Fix an error-reporting bug during instance migration
1530 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1531 silently ignoring broken ones
1532 - Fix an RPC error reporting bug affecting, for example, RAPI client
1534 - Fix bug triggered by different API version os-es on different nodes
1535 - Fix a bug in instance startup with custom hvparams: OS level
1536 parameters would fail to be applied.
1537 - Fix the RAPI client under Python 2.6 (but more work is needed to
1538 make it work completely well with OpenSSL)
1539 - Fix handling of errors when resolving names from DNS
1545 *(Released Thu, 3 Jun 2010)*
1547 A medium sized development cycle. Some new features, and some
1548 fixes/small improvements/cleanups.
1550 Significant features
1551 ~~~~~~~~~~~~~~~~~~~~
1553 The node deamon now tries to mlock itself into memory, unless the
1554 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1555 its logs, and falls back to console logging. This allows emergency
1556 features such as ``gnt-node powercycle`` to work even in the event of a
1557 broken node disk (tested offlining the disk hosting the node's
1558 filesystem and dropping its memory caches; don't try this at home)
1560 KVM: add vhost-net acceleration support. It can be tested with a new
1561 enough version of the kernel and of qemu-kvm.
1563 KVM: Add instance chrooting feature. If you use privilege dropping for
1564 your VMs you can also now force them to chroot to an empty directory,
1565 before starting the emulated guest.
1567 KVM: Add maximum migration bandwith and maximum downtime tweaking
1568 support (requires a new-enough version of qemu-kvm).
1570 Cluster verify will now warn if the master node doesn't have the master
1571 ip configured on it.
1573 Add a new (incompatible) instance creation request format to RAPI which
1574 supports all parameters (previously only a subset was supported, and it
1575 wasn't possible to extend the old format to accomodate all the new
1576 features. The old format is still supported, and a client can check for
1577 this feature, before using it, by checking for its presence in the
1578 ``features`` RAPI resource.
1580 Now with ancient latin support. Try it passing the ``--roman`` option to
1581 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1582 (requires the python-roman module to be installed, in order to work).
1587 As usual many internal code refactorings, documentation updates, and
1590 - Lots of improvements and cleanups to the experimental Remote API
1591 (RAPI) client library.
1592 - A new unit test suite for the core daemon libraries.
1593 - A fix to creating missing directories makes sure the umask is not
1594 applied anymore. This enforces the same directory permissions
1596 - Better handling terminating daemons with ctrl+c (used when running
1597 them in debugging mode).
1598 - Fix a race condition in live migrating a KVM instance, when stat()
1599 on the old proc status file returned EINVAL, which is an unexpected
1601 - Fixed manpage checking with newer man and utf-8 charachters. But now
1602 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1608 *(Released Fri, 7 May 2010)*
1610 Fix a bug which prevented untagged KVM instances from starting.
1616 *(Released Fri, 7 May 2010)*
1618 Another release with a long development cycle, during which many
1619 different features were added.
1621 Significant features
1622 ~~~~~~~~~~~~~~~~~~~~
1624 The KVM hypervisor now can run the individual instances as non-root, to
1625 reduce the impact of a VM being hijacked due to bugs in the
1626 hypervisor. It is possible to run all instances as a single (non-root)
1627 user, to manually specify a user for each instance, or to dynamically
1628 allocate a user out of a cluster-wide pool to each instance, with the
1629 guarantee that no two instances will run under the same user ID on any
1632 An experimental RAPI client library, that can be used standalone
1633 (without the other Ganeti libraries), is provided in the source tree as
1634 ``lib/rapi/client.py``. Note this client might change its interface in
1635 the future, as we iterate on its capabilities.
1637 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1638 replace the cluster's certificates and crypto keys. This might help in
1639 case they have been compromised, or have simply expired.
1641 A new disk option for instance creation has been added that allows one
1642 to "adopt" currently existing logical volumes, with data
1643 preservation. This should allow easier migration to Ganeti from
1644 unmanaged (or managed via other software) instances.
1646 Another disk improvement is the possibility to convert between redundant
1647 (DRBD) and plain (LVM) disk configuration for an instance. This should
1648 allow better scalability (starting with one node and growing the
1649 cluster, or shrinking a two-node cluster to one node).
1651 A new feature that could help with automated node failovers has been
1652 implemented: if a node sees itself as offline (by querying the master
1653 candidates), it will try to shutdown (hard) all instances and any active
1654 DRBD devices. This reduces the risk of duplicate instances if an
1655 external script automatically failovers the instances on such nodes. To
1656 enable this, the cluster parameter ``maintain_node_health`` should be
1657 enabled; in the future this option (per the name) will enable other
1658 automatic maintenance features.
1660 Instance export/import now will reuse the original instance
1661 specifications for all parameters; that means exporting an instance,
1662 deleting it and the importing it back should give an almost identical
1663 instance. Note that the default import behaviour has changed from
1664 before, where it created only one NIC; now it recreates the original
1667 Cluster verify has added a few new checks: SSL certificates validity,
1668 /etc/hosts consistency across the cluster, etc.
1673 As usual, many internal changes were done, documentation fixes,
1676 - Fixed cluster initialization with disabled cluster storage (regression
1677 introduced in 2.1.1)
1678 - File-based storage supports growing the disks
1679 - Fixed behaviour of node role changes
1680 - Fixed cluster verify for some corner cases, plus a general rewrite of
1681 cluster verify to allow future extension with more checks
1682 - Fixed log spamming by watcher and node daemon (regression introduced
1684 - Fixed possible validation issues when changing the list of enabled
1686 - Fixed cleanup of /etc/hosts during node removal
1687 - Fixed RAPI response for invalid methods
1688 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1689 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1690 from ide disks, etc.)
1691 - Allow OS changes without re-installation (to record a changed OS
1692 outside of Ganeti, or to allow OS renames)
1693 - Allow instance creation without OS installation (useful for example if
1694 the OS will be installed manually, or restored from a backup not in
1696 - Implemented option to make cluster ``copyfile`` use the replication
1698 - Added list of enabled hypervisors to ssconf (possibly useful for
1700 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1702 - A partial form of node re-IP is possible via node readd, which now
1703 allows changed node primary IP
1704 - Command line utilities now show an informational message if the job is
1706 - The logs of the master daemon now show the PID/UID/GID of the
1713 *(Released Fri, 12 Mar 2010)*
1715 During the 2.1.0 long release candidate cycle, a lot of improvements and
1716 changes have accumulated with were released later as 2.1.1.
1721 The node evacuate command (``gnt-node evacuate``) was significantly
1722 rewritten, and as such the IAllocator protocol was changed - a new
1723 request type has been added. This unfortunate change during a stable
1724 series is designed to improve performance of node evacuations; on
1725 clusters with more than about five nodes and which are well-balanced,
1726 evacuation should proceed in parallel for all instances of the node
1727 being evacuated. As such, any existing IAllocator scripts need to be
1728 updated, otherwise the above command will fail due to the unknown
1729 request. The provided "dumb" allocator has not been updated; but the
1730 ganeti-htools package supports the new protocol since version 0.2.4.
1732 Another important change is increased validation of node and instance
1733 names. This might create problems in special cases, if invalid host
1734 names are being used.
1736 Also, a new layer of hypervisor parameters has been added, that sits at
1737 OS level between the cluster defaults and the instance ones. This allows
1738 customisation of virtualization parameters depending on the installed
1739 OS. For example instances with OS 'X' may have a different KVM kernel
1740 (or any other parameter) than the cluster defaults. This is intended to
1741 help managing a multiple OSes on the same cluster, without manual
1742 modification of each instance's parameters.
1744 A tool for merging clusters, ``cluster-merge``, has been added in the
1745 tools sub-directory.
1750 - Improved the int/float conversions that should make the code more
1751 robust in face of errors from the node daemons
1752 - Fixed the remove node code in case of internal configuration errors
1753 - Fixed the node daemon behaviour in face of inconsistent queue
1754 directory (e.g. read-only file-system where we can't open the files
1756 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1757 now it either aborts cleanly or, if given the new "auto_promote"
1758 parameter, will automatically promote other nodes as needed
1759 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1760 completely prevent Ganeti from working
1761 - Fixed bug for instance export when not all disks were successfully
1763 - Fixed behaviour of node add when the new node is slow in starting up
1765 - Fixed handling of signals in the LUXI client, which should improve
1766 behaviour of command-line scripts
1767 - Added checks for invalid node/instance names in the configuration (now
1768 flagged during cluster verify)
1769 - Fixed watcher behaviour for disk activation errors
1770 - Fixed two potentially endless loops in http library, which led to the
1771 RAPI daemon hanging and consuming 100% CPU in some cases
1772 - Fixed bug in RAPI daemon related to hashed passwords
1773 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1775 - Enhanced compatibility with non-Debian OSes, but not using absolute
1776 path in some commands and allowing customisation of the ssh
1777 configuration directory
1778 - Fixed possible future issue with new Python versions by abiding to the
1779 proper use of ``__slots__`` attribute on classes
1780 - Added checks that should prevent directory traversal attacks
1781 - Many documentation fixes based on feedback from users
1786 - Added an "early_release" more for instance replace disks and node
1787 evacuate, where we release locks earlier and thus allow higher
1788 parallelism within the cluster
1789 - Added watcher hooks, intended to allow the watcher to restart other
1790 daemons (e.g. from the ganeti-nbma project), but they can be used of
1791 course for any other purpose
1792 - Added a compile-time disable for DRBD barriers, to increase
1793 performance if the administrator trusts the power supply or the
1794 storage system to not lose writes
1795 - Added the option of using syslog for logging instead of, or in
1796 addition to, Ganeti's own log files
1797 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1798 can indeed boot from a paravirtual NIC
1799 - Added a generic debug level for many operations; while this is not
1800 used widely yet, it allows one to pass the debug value all the way to
1802 - Enhanced the hooks environment for instance moves (failovers,
1803 migrations) where the primary/secondary nodes changed during the
1804 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1805 - Enhanced data validations for many user-supplied values; one important
1806 item is the restrictions imposed on instance and node names, which
1807 might reject some (invalid) host names
1808 - Add a configure-time option to disable file-based storage, if it's not
1809 needed; this allows greater security separation between the master
1810 node and the other nodes from the point of view of the inter-node RPC
1812 - Added user notification in interactive tools if job is waiting in the
1813 job queue or trying to acquire locks
1814 - Added log messages when a job is waiting for locks
1815 - Added filtering by node tags in instance operations which admit
1816 multiple instances (start, stop, reboot, reinstall)
1817 - Added a new tool for cluster mergers, ``cluster-merge``
1818 - Parameters from command line which are of the form ``a=b,c=d`` can now
1819 use backslash escapes to pass in values which contain commas,
1820 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1822 - For KVM, the instance name is the first parameter passed to KVM, so
1823 that it's more visible in the process list
1829 *(Released Tue, 2 Mar 2010)*
1831 Ganeti 2.1 brings many improvements with it. Major changes:
1833 - Added infrastructure to ease automated disk repairs
1834 - Added new daemon to export configuration data in a cheaper way than
1835 using the remote API
1836 - Instance NICs can now be routed instead of being associated with a
1838 - Improved job locking logic to reduce impact of jobs acquiring multiple
1839 locks waiting for other long-running jobs
1841 In-depth implementation details can be found in the Ganeti 2.1 design
1847 - Added chroot hypervisor
1848 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1850 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1851 ``bootloader_path`` and ``bootloader_args``)
1852 - Added the ``use_localtime`` option for the xen-hvm and kvm
1853 hypervisors, and the default value for this has changed to false (in
1854 2.0 xen-hvm always enabled it)
1855 - Added luxi call to submit multiple jobs in one go
1856 - Added cluster initialization option to not modify ``/etc/hosts``
1858 - Added network interface parameters
1859 - Added dry run mode to some LUs
1860 - Added RAPI resources:
1862 - ``/2/instances/[instance_name]/info``
1863 - ``/2/instances/[instance_name]/replace-disks``
1864 - ``/2/nodes/[node_name]/evacuate``
1865 - ``/2/nodes/[node_name]/migrate``
1866 - ``/2/nodes/[node_name]/role``
1867 - ``/2/nodes/[node_name]/storage``
1868 - ``/2/nodes/[node_name]/storage/modify``
1869 - ``/2/nodes/[node_name]/storage/repair``
1871 - Added OpCodes to evacuate or migrate all instances on a node
1872 - Added new command to list storage elements on nodes (``gnt-node
1873 list-storage``) and modify them (``gnt-node modify-storage``)
1874 - Added new ssconf files with master candidate IP address
1875 (``ssconf_master_candidates_ips``), node primary IP address
1876 (``ssconf_node_primary_ips``) and node secondary IP address
1877 (``ssconf_node_secondary_ips``)
1878 - Added ``ganeti-confd`` and a client library to query the Ganeti
1879 configuration via UDP
1880 - Added ability to run hooks after cluster initialization and before
1882 - Added automatic mode for disk replace (``gnt-instance replace-disks
1884 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1885 after catastrophic data-loss
1886 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1888 - Added ``gnt-instance move`` command to move instances
1889 - Added ``gnt-cluster watcher`` command to control watcher
1890 - Added ``gnt-node powercycle`` command to powercycle nodes
1891 - Added new job status field ``lock_status``
1892 - Added parseable error codes to cluster verification (``gnt-cluster
1893 verify --error-codes``) and made output less verbose (use
1894 ``--verbose`` to restore previous behaviour)
1895 - Added UUIDs to the main config entities (cluster, nodes, instances)
1896 - Added support for OS variants
1897 - Added support for hashed passwords in the Ganeti remote API users file
1899 - Added option to specify maximum timeout on instance shutdown
1900 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1901 - Added new helper script to start and stop Ganeti daemons
1902 (``daemon-util``), with the intent to reduce the work necessary to
1903 adjust Ganeti for non-Debian distributions and to start/stop daemons
1905 - Added more unittests
1906 - Fixed critical bug in ganeti-masterd startup
1907 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1908 now customisable at the cluster level for both the KVM and Xen
1909 hypervisors using the new ``migration_port`` parameter
1910 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1911 reinstalling an instance
1912 - Allowed ``@`` in tag names
1913 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1914 - Many documentation updates
1915 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1916 - ``gnt-instance reinstall`` can now reinstall multiple instances
1917 - Updated many command line parameters
1918 - Introduced new OS API version 15
1919 - No longer support a default hypervisor
1920 - Treat virtual LVs as inexistent
1921 - Improved job locking logic to reduce lock contention
1922 - Match instance and node names case insensitively
1923 - Reimplemented bash completion script to be more complete
1930 *(Released Thu, 4 Feb 2010)*
1932 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1933 - Fix a string formatting bug
1934 - Improve safety of the code in some error paths
1935 - Improve data validation in the master of values returned from nodes
1941 *(Released Thu, 17 Dec 2009)*
1943 - Fix security issue due to missing validation of iallocator names; this
1944 allows local and remote execution of arbitrary executables
1945 - Fix failure of gnt-node list during instance removal
1946 - Ship the RAPI documentation in the archive
1952 *(Released Wed, 30 Sep 2009)*
1954 - Fixed many wrong messages
1955 - Fixed a few bugs related to the locking library
1956 - Fixed MAC checking at instance creation time
1957 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1958 - Fixed a few issues related to signal handling in both daemons and
1960 - Fixed the example startup script provided
1961 - Fixed insserv dependencies in the example startup script (patch from
1963 - Fixed handling of drained nodes in the iallocator framework
1964 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1966 - Fixed error related to invalid job IDs in job polling
1967 - Fixed job/opcode persistence on unclean master shutdown
1968 - Fixed handling of partial job processing after unclean master
1970 - Fixed error reporting from LUs, previously all errors were converted
1971 into execution errors
1972 - Fixed error reporting from burnin
1973 - Decreased significantly the memory usage of the job queue
1974 - Optimised slightly multi-job submission
1975 - Optimised slightly opcode loading
1976 - Backported the multi-job submit framework from the development
1977 branch; multi-instance start and stop should be faster
1978 - Added script to clean archived jobs after 21 days; this will reduce
1979 the size of the queue directory
1980 - Added some extra checks in disk size tracking
1981 - Added an example ethers hook script
1982 - Added a cluster parameter that prevents Ganeti from modifying of
1984 - Added more node information to RAPI responses
1985 - Added a ``gnt-job watch`` command that allows following the ouput of a
1987 - Added a bind-address option to ganeti-rapi
1988 - Added more checks to the configuration verify
1989 - Enhanced the burnin script such that some operations can be retried
1991 - Converted instance reinstall to multi-instance model
1997 *(Released Fri, 7 Aug 2009)*
1999 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2000 to allow using the pre-2.0.2 behaviour in activation, if any existing
2001 instances have mismatched disk sizes in the configuration
2002 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2003 any configuration mismatches for disk sizes
2004 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2005 failover to work on two-node clusters
2006 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2008 - Fixed detection of OS script errors in ``gnt-backup export``
2009 - Fixed exit code of ``gnt-backup export``
2015 *(Released Fri, 17 Jul 2009)*
2017 - Added experimental support for stripped logical volumes; this should
2018 enhance performance but comes with a higher complexity in the block
2019 device handling; stripping is only enabled when passing
2020 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2021 affected even in the non-stripped mode
2022 - Improved resiliency against transient failures at the end of DRBD
2023 resyncs, and in general of DRBD resync checks
2024 - Fixed a couple of issues with exports and snapshot errors
2025 - Fixed a couple of issues in instance listing
2026 - Added display of the disk size in ``gnt-instance info``
2027 - Fixed checking for valid OSes in instance creation
2028 - Fixed handling of the "vcpus" parameter in instance listing and in
2029 general of invalid parameters
2030 - Fixed http server library, and thus RAPI, to handle invalid
2031 username/password combinations correctly; this means that now they
2032 report unauthorized for queries too, not only for modifications,
2033 allowing earlier detect of configuration problems
2034 - Added a new "role" node list field, equivalent to the master/master
2035 candidate/drained/offline flags combinations
2036 - Fixed cluster modify and changes of candidate pool size
2037 - Fixed cluster verify error messages for wrong files on regular nodes
2038 - Fixed a couple of issues with node demotion from master candidate role
2039 - Fixed node readd issues
2040 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2041 - Added a new ``--no-voting`` option for masterfailover to fix failover
2042 on two-nodes clusters when the former master node is unreachable
2043 - Added instance reinstall over RAPI
2049 *(Released Tue, 16 Jun 2009)*
2051 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2052 allow re-adding the start in single-user option (regression from 1.2)
2053 - the watcher writes the instance status to a file, to allow monitoring
2054 to report the instance status (from the master) based on cached
2055 results of the watcher's queries; while this can get stale if the
2056 watcher is being locked due to other work on the cluster, this is
2057 still an improvement
2058 - the watcher now also restarts the node daemon and the rapi daemon if
2060 - fixed the watcher to handle full and drained queue cases
2061 - hooks export more instance data in the environment, which helps if
2062 hook scripts need to take action based on the instance's properties
2063 (no longer need to query back into ganeti)
2064 - instance failovers when the instance is stopped do not check for free
2065 RAM, so that failing over a stopped instance is possible in low memory
2067 - rapi uses queries for tags instead of jobs (for less job traffic), and
2068 for cluster tags it won't talk to masterd at all but read them from
2070 - a couple of error handling fixes in RAPI
2071 - drbd handling: improved the error handling of inconsistent disks after
2072 resync to reduce the frequency of "there are some degraded disks for
2073 this instance" messages
2074 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2075 error handling path called a wrong function name)
2081 *(Released Wed, 27 May 2009)*
2083 - no changes from rc5
2089 *(Released Wed, 20 May 2009)*
2091 - fix a couple of bugs (validation, argument checks)
2092 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2093 - some small improvements to RAPI and IAllocator
2094 - make watcher automatically start the master daemon if down
2100 *(Released Mon, 27 Apr 2009)*
2102 - change the OS list to not require locks; this helps with big clusters
2103 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2104 volume group is broken
2105 - ``gnt-instance info``, without any arguments, doesn't run for all
2106 instances anymore; either pass ``--all`` or pass the desired
2107 instances; this helps against mistakes on big clusters where listing
2108 the information for all instances takes a long time
2109 - miscellaneous doc and man pages fixes
2115 *(Released Wed, 8 Apr 2009)*
2117 - Change the internal locking model of some ``gnt-node`` commands, in
2118 order to reduce contention (and blocking of master daemon) when
2119 batching many creation/reinstall jobs
2120 - Fixes to Xen soft reboot
2121 - No longer build documentation at build time, instead distribute it in
2122 the archive, in order to reduce the need for the whole docbook/rst
2129 *(Released Fri, 27 Mar 2009)*
2131 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2132 - Fix watcher startup sequence, improves the behaviour of busy clusters
2133 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2134 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2135 bind address and other places
2136 - Some documentation fixes and updates
2142 *(Released Mon, 2 Mar 2009)*
2144 - More documentation updates, now all docs should be more-or-less
2146 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2148 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2150 - Fix an issue related to $libdir/run/ganeti and cluster creation
2156 *(Released Thu, 19 Feb 2009)*
2158 - Xen PVM and KVM have switched the default value for the instance root
2159 disk to the first partition on the first drive, instead of the whole
2160 drive; this means that the OS installation scripts must be changed
2162 - Man pages have been updated
2163 - RAPI has been switched by default to HTTPS, and the exported functions
2164 should all work correctly
2165 - RAPI v1 has been removed
2166 - Many improvements to the KVM hypervisor
2167 - Block device errors are now better reported
2168 - Many other bugfixes and small improvements
2174 *(Released Mon, 26 Jan 2009)*
2176 - Version 2 is a general rewrite of the code and therefore the
2177 differences are too many to list, see the design document for 2.0 in
2178 the ``doc/`` subdirectory for more details
2179 - In this beta version there is not yet a migration path from 1.2 (there
2180 will be one in the final 2.0 release)
2181 - A few significant changes are:
2183 - all commands are executed by a daemon (``ganeti-masterd``) and the
2184 various ``gnt-*`` commands are just front-ends to it
2185 - all the commands are entered into, and executed from a job queue,
2186 see the ``gnt-job(8)`` manpage
2187 - the RAPI daemon supports read-write operations, secured by basic
2188 HTTP authentication on top of HTTPS
2189 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2190 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2191 to migrate to DRBD 8 first while still running Ganeti 1.2)
2192 - DRBD devices are using statically allocated minor numbers, which
2193 will be assigned to existing instances during the migration process
2194 - there is support for both Xen PVM and Xen HVM instances running on
2196 - KVM virtualization is supported too
2197 - file-based storage has been implemented, which means that it is
2198 possible to run the cluster without LVM and DRBD storage, for
2199 example using a shared filesystem exported from shared storage (and
2200 still have live migration)
2206 *(Released Tue, 13 Jan 2009)*
2208 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2209 - Reuse the old instance mac address by default on instance import, if
2210 the instance name is the same.
2211 - Handle situations in which the node info rpc returns incomplete
2213 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2214 - Improved version of batcher:
2216 - state file support
2217 - instance mac address support
2218 - support for HVM clusters/instances
2220 - Add an option to show the number of cpu sockets and nodes in
2222 - Support OSes that handle more than one version of the OS api (but do
2223 not change the current API in any other way)
2224 - Fix ``gnt-node migrate``
2225 - ``gnt-debug`` man page
2226 - Fixes various more typos and small issues
2227 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2233 *(Released Wed, 24 Sep 2008)*
2235 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2236 type of disk exported to fully virtualized instances.
2237 - provide access to the serial console of HVM instances
2238 - instance auto_balance flag, set by default. If turned off it will
2239 avoid warnings on cluster verify if there is not enough memory to fail
2240 over an instance. in the future it will prevent automatically failing
2241 it over when we will support that.
2242 - batcher tool for instance creation, see ``tools/README.batcher``
2243 - ``gnt-instance reinstall --select-os`` to interactively select a new
2244 operating system when reinstalling an instance.
2245 - when changing the memory amount on instance modify a check has been
2246 added that the instance will be able to start. also warnings are
2247 emitted if the instance will not be able to fail over, if auto_balance
2249 - documentation fixes
2250 - sync fields between ``gnt-instance list/modify/add/import``
2251 - fix a race condition in drbd when the sync speed was set after giving
2252 the device a remote peer.
2258 *(Released Tue, 22 Jul 2008)*
2260 - note: the allowed size and number of tags per object were reduced
2261 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2262 - fixed twisted 8.x compatibility
2263 - fixed ``gnt-instance replace-disks`` with iallocator
2264 - add TCP keepalives on twisted connections to detect restarted nodes
2265 - disk increase support, see ``gnt-instance grow-disk``
2266 - implement bulk node/instance query for RAPI
2267 - add tags in node/instance listing (optional)
2268 - experimental migration (and live migration) support, read the man page
2269 for ``gnt-instance migrate``
2270 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2271 has some small improvements in handling its state file
2277 *(Released Fri, 13 Jun 2008)*
2279 - Experimental readonly, REST-based remote API implementation;
2280 automatically started on master node, TCP port 5080, if enabled by
2281 ``--enable-rapi`` parameter to configure script.
2282 - Instance allocator support. Add and import instance accept a
2283 ``--iallocator`` parameter, and call that instance allocator to decide
2284 which node to use for the instance. The iallocator document describes
2285 what's expected from an allocator script.
2286 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2287 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2288 node is lost there is still enough memory to fail over the instances
2290 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2291 ``gnt-cluster verify``, to check for site-specific compliance. All the
2292 hooks will run, and their output, if any, will be displayed. Any
2293 failing hook will make the verification return an error value.
2294 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2295 primary and secondary interfaces
2296 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2297 that is still declared as part of the cluster and has failed.
2298 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2299 specifying output fields, that just adds the chosen fields to the
2301 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2302 export from the filesystem.
2303 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2304 have been added. Using them you can enable/disable acpi and pae
2305 support, and specify a path for a cd image to be exported to the
2306 instance. These parameters as the name suggest only work on HVM
2308 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2309 PAE support will be set to the previously hardcoded values, but the
2310 (previously hardcoded) path to the CDROM ISO image will be unset and
2311 if required, needs to be set manually with ``gnt-instance modify``
2313 - The address to which an instance's VNC console is bound is now
2314 selectable per-instance, rather than being cluster wide. Of course
2315 this only applies to instances controlled via VNC, so currently just
2316 applies to HVM clusters.
2322 *(Released Mon, 18 Feb 2008)*
2324 - more tweaks to the disk activation code (especially helpful for DRBD)
2325 - change the default ``gnt-instance list`` output format, now there is
2326 one combined status field (see the manpage for the exact values this
2328 - some more fixes for the mac export to hooks change
2329 - make Ganeti not break with DRBD 8.2.x (which changed the version
2330 format in ``/proc/drbd``) (issue 24)
2331 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2332 template, allowing migration from DRBD0.7+MD to DRBD8
2338 *(Released Wed, 30 Jan 2008)*
2340 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2342 - add command aliases infrastructure and a few aliases
2343 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2344 man pages and the ``--help`` option of ``gnt-node
2345 list``/``gnt-instance list``
2346 - fix ``gnt-backup list`` with down nodes (issue 21)
2347 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2348 - fix the dist archive and add a check for including svn/git files in
2350 - some developer-related changes: improve the burnin and the QA suite,
2351 add an upload script for testing during development
2357 *(Released Wed, 16 Jan 2008)*
2359 - experimental HVM support, read the install document, section
2360 "Initializing the cluster"
2361 - allow for the PVM hypervisor per-instance kernel and initrd paths
2362 - add a new command ``gnt-cluster verify-disks`` which uses a new
2363 algorithm to improve the reconnection of the DRBD pairs if the device
2364 on the secondary node has gone away
2365 - make logical volume code auto-activate LVs at disk activation time
2366 - slightly improve the speed of activating disks
2367 - allow specification of the MAC address at instance creation time, and
2368 changing it later via ``gnt-instance modify``
2369 - fix handling of external commands that generate lots of output on
2371 - update documentation with regard to minimum version of DRBD8 supported
2377 *(Released Tue, 4 Dec 2007)*
2379 - Log the ``xm create`` output to the node daemon log on failure (to
2380 help diagnosing the error)
2381 - In debug mode, log all external commands output if failed to the logs
2382 - Change parsing of lvm commands to ignore stderr
2388 *(Released Wed, 28 Nov 2007)*
2390 - Another round of updates to the DRBD 8 code to deal with more failures
2391 in the replace secondary node operation
2392 - Some more logging of failures in disk operations (lvm, drbd)
2393 - A few documentation updates
2400 *(Released Tue, 13 Nov 2007)*
2402 - Change configuration file format from Python's Pickle to JSON.
2403 Upgrading is possible using the cfgupgrade utility.
2404 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2405 faster replace disks and is more stable (DRBD 8 has many improvements
2406 compared to DRBD 0.7)
2407 - Added command line tags support (see man pages for ``gnt-instance``,
2408 ``gnt-node``, ``gnt-cluster``)
2409 - Added instance rename support
2410 - Added multi-instance startup/shutdown
2411 - Added cluster rename support
2412 - Added ``gnt-node evacuate`` to simplify some node operations
2413 - Added instance reboot operation that can speedup reboot as compared to
2415 - Soften the requirement that hostnames are in FQDN format
2416 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2418 - Removed dependency on debian's patched fping that uses the
2419 non-standard ``-S`` option
2420 - Now the OS definitions are searched for in multiple, configurable
2421 paths (easier for distros to package)
2422 - Some changes to the hooks infrastructure (especially the new
2423 post-configuration update hook)
2424 - Other small bugfixes
2426 .. vim: set textwidth=72 syntax=rst :