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.
40 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
41 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
42 the flag is not reset after a master failover.
48 *(Released Fri, 12 Oct 2012)*
50 A small bugfix release. Among the bugs fixed:
52 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
53 made the command unusable.
54 - Commands that issue many jobs don't fail anymore just because some jobs
55 take so long that other jobs are archived.
56 - Failures during ``gnt-instance reinstall`` are reflected by the exit
58 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
60 - When ``always_failover`` is set, ``--allow-failover`` is not required
61 in migrate commands anymore.
62 - ``bash_completion`` works even if extglob is disabled.
63 - Fixed bug with locks that made failover for RDB-based instances fail.
64 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
65 a random node instead of one based on the allocator metric.
66 - Support for newer versions of pylint and pep8.
67 - Hail doesn't fail anymore when trying to add an instance of type
68 ``file``, ``sharedfile`` or ``rbd``.
69 - Added new Makefile target to rebuild the whole distribution, so that
70 all files are included.
76 *(Released Fri, 27 Jul 2012)*
79 .. attention:: The ``LUXI`` protocol has been made more consistent
80 regarding its handling of command arguments. This, however, leads to
81 incompatibility issues with previous versions. Please ensure that you
82 restart Ganeti daemons soon after the upgrade, otherwise most
83 ``LUXI`` calls (job submission, setting/resetting the drain flag,
84 pausing/resuming the watcher, cancelling and archiving jobs, querying
85 the cluster configuration) will fail.
94 The current ``admin_up`` field, which used to denote whether an instance
95 should be running or not, has been removed. Instead, ``admin_state`` is
96 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
98 The rational behind this is that an instance being “down” can have
101 - it could be down during a reboot
102 - it could be temporarily be down for a reinstall
103 - or it could be down because it is deprecated and kept just for its
106 The previous Boolean state was making it difficult to do capacity
107 calculations: should Ganeti reserve memory for a down instance? Now, the
108 tri-state field makes it clear:
110 - in ``up`` and ``down`` state, all resources are reserved for the
111 instance, and it can be at any time brought up if it is down
112 - in ``offline`` state, only disk space is reserved for it, but not
115 The field can have an extra use: since the transition between ``up`` and
116 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
117 transition between ``offline`` and ``down`` is done via ``gnt-instance
118 modify``, it is possible to given different rights to users. For
119 example, owners of an instance could be allowed to start/stop it, but
120 not transition it out of the offline state.
122 Instance policies and specs
123 +++++++++++++++++++++++++++
125 In previous Ganeti versions, an instance creation request was not
126 limited on the minimum size and on the maximum size just by the cluster
127 resources. As such, any policy could be implemented only in third-party
128 clients (RAPI clients, or shell wrappers over ``gnt-*``
129 tools). Furthermore, calculating cluster capacity via ``hspace`` again
130 required external input with regards to instance sizes.
132 In order to improve these workflows and to allow for example better
133 per-node group differentiation, we introduced instance specs, which
136 - minimum instance disk size, disk count, memory size, cpu count
137 - maximum values for the above metrics
138 - and “standard” values (used in ``hspace`` to calculate the standard
141 The minimum/maximum values can be also customised at node-group level,
142 for example allowing more powerful hardware to support bigger instance
145 Beside the instance specs, there are a few other settings belonging to
146 the instance policy framework. It is possible now to customise, per
147 cluster and node-group:
149 - the list of allowed disk templates
150 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
151 - the maximum ratio of instance to spindles (see below for more
152 information) for local storage
154 All these together should allow all tools that talk to Ganeti to know
155 what are the ranges of allowed values for instances and the
156 over-subscription that is allowed.
158 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
159 instance configuration, and the physical CPU configuration from the
160 node. For the spindle ratios however, we didn't track before these
161 values, so new parameters have been added:
163 - a new node parameter ``spindle_count``, defaults to 1, customisable at
164 node group or node level
165 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
167 Note that spindles in this context doesn't need to mean actual
168 mechanical hard-drives; it's just a relative number for both the node
169 I/O capacity and instance I/O consumption.
171 Instance migration behaviour
172 ++++++++++++++++++++++++++++
174 While live-migration is in general desirable over failover, it is
175 possible that for some workloads it is actually worse, due to the
176 variable time of the “suspend” phase during live migration.
178 To allow the tools to work consistently over such instances (without
179 having to hard-code instance names), a new backend parameter
180 ``always_failover`` has been added to control the migration/failover
181 behaviour. When set to True, all migration requests for an instance will
182 instead fall-back to failover.
184 Instance memory ballooning
185 ++++++++++++++++++++++++++
187 Initial support for memory ballooning has been added. The memory for an
188 instance is no longer fixed (backend parameter ``memory``), but instead
189 can vary between minimum and maximum values (backend parameters
190 ``minmem`` and ``maxmem``). Currently we only change an instance's
193 - live migrating or failing over and instance and the target node
194 doesn't have enough memory
195 - user requests changing the memory via ``gnt-instance modify
201 In order to control the use of specific CPUs by instance, support for
202 controlling CPU pinning has been added for the Xen, HVM and LXC
203 hypervisors. This is controlled by a new hypervisor parameter
204 ``cpu_mask``; details about possible values for this are in the
205 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
206 VCPU-to-CPU mapping) form will work well only when all nodes in your
207 cluster have the same amount of CPUs.
212 Another area in which Ganeti was not customisable were the parameters
213 used for storage configuration, e.g. how many stripes to use for LVM,
214 DRBD resync configuration, etc.
216 To improve this area, we've added disks parameters, which are
217 customisable at cluster and node group level, and which allow to
218 specify various parameters for disks (DRBD has the most parameters
219 currently), for example:
221 - DRBD resync algorithm and parameters (e.g. speed)
222 - the default VG for meta-data volumes for DRBD
223 - number of stripes for LVM (plain disk template)
226 These parameters can be modified via ``gnt-cluster modify -D …`` and
227 ``gnt-group modify -D …``, and are used at either instance creation (in
228 case of LVM stripes, for example) or at disk “activation” time
231 Rados block device support
232 ++++++++++++++++++++++++++
234 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
235 denoted by the ``rbd`` disk template type. This is considered
236 experimental, feedback is welcome. For details on configuring it, see
237 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
242 The existing master IP functionality works well only in simple setups (a
243 single network shared by all nodes); however, if nodes belong to
244 different networks, then the ``/32`` setup and lack of routing
245 information is not enough.
247 To allow the master IP to function well in more complex cases, the
248 system was reworked as follows:
250 - a master IP netmask setting has been added
251 - the master IP activation/turn-down code was moved from the node daemon
253 - whether to run the Ganeti-supplied master IP script or a user-supplied
254 on is a ``gnt-cluster init`` setting
256 Details about the location of the standard and custom setup scripts are
257 in the man page :manpage:`gnt-cluster(8)`; for information about the
258 setup script protocol, look at the Ganeti-supplied script.
263 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
266 It is now possible to use TLS-protected connections, and when renewing
267 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
268 it is now possible to specify spice or spice CA certificates. Also, it
269 is possible to configure a password for SPICE sessions via the
270 hypervisor parameter ``spice_password_file``.
272 There are also new parameters to control the compression and streaming
273 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
274 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
275 for the spice parameters.
277 Lastly, it is now possible to see the SPICE connection information via
278 ``gnt-instance console``.
283 A new tool (``tools/ovfconverter``) has been added that supports
284 conversion between Ganeti and the `Open Virtualization Format
285 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
288 This relies on the ``qemu-img`` tool to convert the disk formats, so the
289 actual compatibility with other virtualization solutions depends on it.
294 The configuration query daemon (``ganeti-confd``) is now optional, and
295 has been rewritten in Haskell; whether to use the daemon at all, use the
296 Python (default) or the Haskell version is selectable at configure time
297 via the ``--enable-confd`` parameter, which can take one of the
298 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
299 daemon will result in a smaller footprint; for larger systems, we
300 welcome feedback on the Haskell version which might become the default
303 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
304 daemon running. The Python version doesn't implement the new call.
307 User interface changes
308 ~~~~~~~~~~~~~~~~~~~~~~
310 We have replaced the ``--disks`` option of ``gnt-instance
311 replace-disks`` with a more flexible ``--disk`` option, which allows
312 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
313 disk size and mode can be changed upon recreation (via ``gnt-instance
314 recreate-disks``, which accepts the same ``--disk`` option).
316 As many people are used to a ``show`` command, we have added that as an
317 alias to ``info`` on all ``gnt-*`` commands.
319 The ``gnt-instance grow-disk`` command has a new mode in which it can
320 accept the target size of the disk, instead of the delta; this can be
321 more safe since two runs in absolute mode will be idempotent, and
322 sometimes it's also easier to specify the desired size directly.
324 Also the handling of instances with regard to offline secondaries has
325 been improved. Instance operations should not fail because one of it's
326 secondary nodes is offline, even though it's safe to proceed.
328 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
329 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
330 minors to instance name.
335 RAPI coverage has improved, with (for example) new resources for
336 recreate-disks, node power-cycle, etc.
341 There is partial support for ``xl`` in the Xen hypervisor; feedback is
344 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
345 whether to still support Python 2.4 or move to Python 2.6 as minimum
348 Support for Fedora has been slightly improved; the provided example
349 init.d script should work better on it and the INSTALL file should
350 document the needed dependencies.
355 The deprecated ``QueryLocks`` LUXI request has been removed. Use
356 ``Query(what=QR_LOCK, ...)`` instead.
358 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
359 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
360 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
361 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
362 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
364 RAPI client: ``CertificateError`` now derives from
365 ``GanetiApiError``. This should make it more easy to handle Ganeti
368 Deprecation warnings due to PyCrypto/paramiko import in
369 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
370 make sure to run an up-to-date paramiko version, if you use this tool.
372 The QA scripts now depend on Python 2.5 or above (the main code base
373 still works with Python 2.4).
375 The configuration file (``config.data``) is now written without
376 indentation for performance reasons; if you want to edit it, it can be
377 re-formatted via ``tools/fmtjson``.
379 A number of bugs has been fixed in the cluster merge tool.
381 ``x509`` certification verification (used in import-export) has been
382 changed to allow the same clock skew as permitted by the cluster
383 verification. This will remove some rare but hard to diagnose errors in
390 *(Released Thu, 19 Jul 2012)*
392 Very few changes from rc4 to the final release, only bugfixes:
394 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
395 instance, fix CDROM booting for KVM instances)
396 - fixed node group modification of node parameters
397 - fixed issue in LUClusterVerifyGroup with multi-group clusters
398 - fixed generation of bash completion to ensure a stable ordering
405 *(Released Fri, 13 Jul 2012)*
407 Third release candidate for 2.6. The following changes were done from
410 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
411 - Fixed an inconsistency in the LUXI protocol with the provided
412 arguments (NOT backwards compatible)
413 - Fixed a bug with node groups ipolicy where ``min`` was greater than
414 the cluster ``std`` value
415 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
416 easier instance debugging on nodes (requires ``hconfd`` to work)
422 *(Released Tue, 03 Jul 2012)*
424 Second release candidate for 2.6. The following changes were done from
427 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
429 - Fixed a RAPI regression on missing beparams/memory
430 - Fixed redistribution of files on offline nodes
431 - Added possibility to run activate-disks even though secondaries are
432 offline. With this change it relaxes also the strictness on some other
433 commands which use activate disks internally:
434 * ``gnt-instance start|reboot|rename|backup|export``
435 - Made it possible to remove safely an instance if its secondaries are
437 - Made it possible to reinstall even though secondaries are offline
443 *(Released Mon, 25 Jun 2012)*
445 First release candidate for 2.6. The following changes were done from
448 - Fixed bugs with disk parameters and ``rbd`` templates as well as
450 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
451 behaviour. It supports now the modify operation
452 - ``hcheck`` implemented to analyze cluster health and possibility of
453 improving health by rebalance
454 - ``hbal`` has been improved in dealing with split instances
460 *(Released Mon, 11 Jun 2012)*
462 Second beta release of 2.6. The following changes were done from beta2
465 - Fixed ``daemon-util`` with non-root user models
466 - Fixed creation of plain instances with ``--no-wait-for-sync``
467 - Fix wrong iv_names when running ``cfgupgrade``
468 - Export more information in RAPI group queries
469 - Fixed bug when changing instance network interfaces
470 - Extended burnin to do NIC changes
471 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
472 - Changed default for DRBD barriers
473 - Fixed DRBD error reporting for syncer rate
474 - Verify the options on disk parameters
476 And of course various fixes to documentation and improved unittests and
483 *(Released Wed, 23 May 2012)*
485 First beta release of 2.6. The following changes were done from beta1 to
488 - integrated patch for distributions without ``start-stop-daemon``
489 - adapted example init.d script to work on Fedora
490 - fixed log handling in Haskell daemons
491 - adapted checks in the watcher for pycurl linked against libnss
492 - add partial support for ``xl`` instead of ``xm`` for Xen
493 - fixed a type issue in cluster verification
494 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
497 Plus integrated fixes from the 2.5 branch:
499 - fixed ``kvm-ifup`` to use ``/bin/bash``
500 - fixed parallel build failures
501 - KVM live migration when using a custom keymap
507 *(Released Tue, 24 Jul 2012)*
509 A small bugfix release, with no new features:
511 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
512 different default shell (e.g. Debian, Ubuntu)
513 - fixed KVM startup and live migration with a custom keymap (fixes Issue
514 243 and Debian bug #650664)
515 - fixed compatibility with KVM versions that don't support multiple boot
516 devices (fixes Issue 230 and Debian bug #624256)
518 Additionally, a few fixes were done to the build system (fixed parallel
519 build failures) and to the unittests (fixed race condition in test for
520 FileID functions, and the default enable/disable mode for QA test is now
527 *(Released Fri, 11 May 2012)*
529 A small bugfix release.
531 The main issues solved are on the topic of compatibility with newer LVM
534 - fixed parsing of ``lv_attr`` field
535 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
536 the ``--force`` flag is needed
538 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
539 to accept kernel 3.x too (was hardcoded to 2.6.*).
541 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
542 of hook results and that also made display of other errors suboptimal
543 was fixed; the code behaves now like 2.4 and earlier.
545 Another change in 2.5, the cleanup of the OS scripts environment, is too
546 aggressive: it removed even the ``PATH`` variable, which requires the OS
547 scripts to *always* need to export it. Since this is a bit too strict,
548 we now export a minimal PATH, the same that we export for hooks.
550 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
551 integrated into this release.
553 Finally, a few other miscellaneous changes were done (no new features,
554 just small improvements):
556 - Fix ``gnt-group --help`` display
557 - Fix hardcoded Xen kernel path
558 - Fix grow-disk handling of invalid units
559 - Update synopsis for ``gnt-cluster repair-disk-sizes``
560 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
566 *(Released Thu, 12 Apr 2012)*
568 Incompatible/important changes and bugfixes
569 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
571 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
572 resource's ``ip_check`` parameter changed from ``True`` to ``False``
573 to match the underlying LUXI interface.
574 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
575 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
576 not maintain backwards-compatibility as the underlying operation
577 changed in an incompatible way. The RAPI client can talk to old
578 servers, but it needs to be told so as the return value changed.
579 - When creating file-based instances via RAPI, the ``file_driver``
580 parameter no longer defaults to ``loop`` and must be specified.
581 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
583 - Support for the undocumented and deprecated RAPI instance creation
584 request format version 0 has been dropped. Use version 1, supported
585 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
586 - Pyparsing 1.4.6 or above is required, see :doc:`installation
587 documentation <install>`.
588 - The "cluster-verify" hooks are now executed per group by the
589 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
590 if you just run ``gnt-cluster verify``, which generates one opcode per
592 - The environment as passed to the OS scripts is cleared, and thus no
593 environment variables defined in the node daemon's environment will be
594 inherited by the scripts.
595 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
597 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
598 support operations involving multiple node groups.
599 - Offline nodes are ignored when failing over an instance.
600 - Support for KVM version 1.0, which changed the version reporting format
602 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
604 - ``Makefile`` is now compatible with Automake 1.11.2
605 - Includes all bugfixes made in the 2.4 series
610 - The ganeti-htools project has been merged into the ganeti-core source
611 tree and will be built as part of Ganeti (see :doc:`install-quick`).
612 - Implemented support for :doc:`shared storage <design-shared-storage>`.
613 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
614 GPT-style partition tables (requires `parted
615 <http://www.gnu.org/s/parted/>`_).
616 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
617 - Allowed adding tags on instance creation.
618 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
619 - Allow instances to be started in a paused state, enabling the user to
620 see the complete console output on boot using the console.
621 - Added new hypervisor flag to control default reboot behaviour
622 (``reboot_behavior``).
623 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
624 - Improved out-of-band management support:
626 - Added ``gnt-node health`` command reporting the health status of
628 - Added ``gnt-node power`` command to manage power status of nodes.
629 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
631 - Instance migration can fall back to failover if instance is not
633 - Filters can be used when listing nodes, instances, groups and locks;
634 see *ganeti(7)* manpage.
635 - Added post-execution status as variables to :doc:`hooks <hooks>`
637 - Instance tags are exported/imported together with the instance.
638 - When given an explicit job ID, ``gnt-job info`` will work for archived
640 - Jobs can define dependencies on other jobs (not yet supported via
641 RAPI or command line, but used by internal commands and usable via
644 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
647 - Instance failover is now available as a RAPI resource
648 (``/2/instances/[instance_name]/failover``).
649 - ``gnt-instance info`` defaults to static information if primary node
651 - Opcodes have a new ``comment`` attribute.
652 - Added basic SPICE support to KVM hypervisor.
653 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
655 Node group improvements
656 ~~~~~~~~~~~~~~~~~~~~~~~
658 - ``gnt-cluster verify`` has been modified to check groups separately,
659 thereby improving performance.
660 - Node group support has been added to ``gnt-cluster verify-disks``,
661 which now operates per node group.
662 - Watcher has been changed to work better with node groups.
664 - One process and state file per node group.
665 - Slow watcher in one group doesn't block other group's watcher.
667 - Added new command, ``gnt-group evacuate``, to move all instances in a
668 node group to other groups.
669 - Added ``gnt-instance change-group`` to move an instance to another
671 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
672 per-group operations.
673 - Node groups can be tagged.
674 - Some operations switch from an exclusive to a shared lock as soon as
676 - Instance's primary and secondary nodes' groups are now available as
677 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
678 and ``snodes.group.uuid``).
683 - Numerous updates to documentation and manpages.
685 - :doc:`RAPI <rapi>` documentation now has detailed parameter
687 - Some opcode/job results are now also documented, see :doc:`RAPI
690 - A lockset's internal lock is now also visible in lock monitor.
691 - Log messages from job queue workers now contain information about the
692 opcode they're processing.
693 - ``gnt-instance console`` no longer requires the instance lock.
694 - A short delay when waiting for job changes reduces the number of LUXI
695 requests significantly.
696 - DRBD metadata volumes are overwritten with zeros during disk creation.
697 - Out-of-band commands no longer acquire the cluster lock in exclusive
699 - ``devel/upload`` now uses correct permissions for directories.
705 *(Released Fri, 23 Mar 2012)*
707 This was the sixth release candidate of the 2.5 series.
713 *(Released Mon, 9 Jan 2012)*
715 This was the fifth release candidate of the 2.5 series.
721 *(Released Thu, 27 Oct 2011)*
723 This was the fourth release candidate of the 2.5 series.
729 *(Released Wed, 26 Oct 2011)*
731 This was the third release candidate of the 2.5 series.
737 *(Released Tue, 18 Oct 2011)*
739 This was the second release candidate of the 2.5 series.
745 *(Released Tue, 4 Oct 2011)*
747 This was the first release candidate of the 2.5 series.
753 *(Released Wed, 31 Aug 2011)*
755 This was the third beta release of the 2.5 series.
761 *(Released Mon, 22 Aug 2011)*
763 This was the second beta release of the 2.5 series.
769 *(Released Fri, 12 Aug 2011)*
771 This was the first beta release of the 2.5 series.
777 *(Released Thu, 27 Oct 2011)*
779 - Fixed bug when parsing command line parameter values ending in
781 - Fixed assertion error after unclean master shutdown
782 - Disable HTTP client pool for RPC, significantly reducing memory usage
784 - Fixed queue archive creation with wrong permissions
790 *(Released Tue, 23 Aug 2011)*
794 - Fixed documentation for importing with ``--src-dir`` option
795 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
796 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
802 *(Released Fri, 5 Aug 2011)*
804 Many bug-fixes and a few small features:
806 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
807 issues when you tried to add an instance with two MAC addresses in one
809 - KVM: fixed per-instance stored UID value
810 - KVM: configure bridged NICs at migration start
811 - KVM: Fix a bug where instance will not start with never KVM versions
813 - Added OS search path to ``gnt-cluster info``
814 - Fixed an issue with ``file_storage_dir`` where you were forced to
815 provide an absolute path, but the documentation states it is a
816 relative path, the documentation was right
817 - Added a new parameter to instance stop/start called ``--no-remember``
818 that will make the state change to not be remembered
819 - Implemented ``no_remember`` at RAPI level
820 - Improved the documentation
821 - Node evacuation: don't call IAllocator if node is already empty
822 - Fixed bug in DRBD8 replace disks on current nodes
823 - Fixed bug in recreate-disks for DRBD instances
824 - Moved assertion checking locks in ``gnt-instance replace-disks``
825 causing it to abort with not owning the right locks for some situation
826 - Job queue: Fixed potential race condition when cancelling queued jobs
827 - Fixed off-by-one bug in job serial generation
828 - ``gnt-node volumes``: Fix instance names
829 - Fixed aliases in bash completion
830 - Fixed a bug in reopening log files after being sent a SIGHUP
831 - Added a flag to burnin to allow specifying VCPU count
832 - Bugfixes to non-root Ganeti configuration
838 *(Released Thu, 12 May 2011)*
840 Many bug-fixes and a few new small features:
842 - Fixed a bug related to log opening failures
843 - Fixed a bug in instance listing with orphan instances
844 - Fixed a bug which prevented resetting the cluster-level node parameter
845 ``oob_program`` to the default
846 - Many fixes related to the ``cluster-merge`` tool
847 - Fixed a race condition in the lock monitor, which caused failures
848 during (at least) creation of many instances in parallel
849 - Improved output for gnt-job info
850 - Removed the quiet flag on some ssh calls which prevented debugging
852 - Improved the N+1 failure messages in cluster verify by actually
853 showing the memory values (needed and available)
854 - Increased lock attempt timeouts so that when executing long operations
855 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
856 too early and thus prevent the use of the 'fair' mechanism
857 - Changed instance query data (``gnt-instance info``) to not acquire
858 locks unless needed, thus allowing its use on locked instance if only
859 static information is asked for
860 - Improved behaviour with filesystems that do not support rename on an
862 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
863 kept locks on all nodes during the wipe, which is unneeded
864 - Fixed ``gnt-watcher`` handling of errors during hooks execution
865 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
866 10GiB) which caused the wipe to fail right at the end in some cases
867 - Fixed master IP activation when doing master failover with no-voting
868 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
869 the master node itself
870 - Fixed potential data-loss in under disk full conditions, where Ganeti
871 wouldn't check correctly the return code and would consider
872 partially-written files 'correct'
873 - Fixed bug related to multiple VGs and DRBD disk replacing
874 - Added new disk parameter ``metavg`` that allows placement of the meta
875 device for DRBD in a different volume group
876 - Fixed error handling in the node daemon when the system libc doesn't
877 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
878 - Fixed lock release during replace-disks, which kept cluster-wide locks
879 when doing disk replaces with an iallocator script
880 - Added check for missing bridges in cluster verify
881 - Handle EPIPE errors while writing to the terminal better, so that
882 piping the output to e.g. ``less`` doesn't cause a backtrace
883 - Fixed rare case where a ^C during Luxi calls could have been
884 interpreted as server errors, instead of simply terminating
885 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
887 - Added a few more parameters to the KVM hypervisor, allowing a second
888 CDROM, custom disk type for CDROMs and a floppy image
889 - Removed redundant message in instance rename when the name is given
891 - Added option to ``gnt-instance recreate-disks`` to allow creating the
892 disks on new nodes, allowing recreation when the original instance
893 nodes are completely gone
894 - Added option when converting disk templates to DRBD to skip waiting
895 for the resync, in order to make the instance available sooner
896 - Added two new variables to the OS scripts environment (containing the
898 - Made the root_path and optional parameter for the xen-pvm hypervisor,
899 to allow use of ``pvgrub`` as bootloader
900 - Changed the instance memory modifications to only check out-of-memory
901 conditions on memory increases, and turned the secondary node warnings
902 into errors (they can still be overridden via ``--force``)
903 - Fixed the handling of a corner case when the Python installation gets
904 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
905 to execute a command that doesn't exist
906 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
907 node of the instance returned failures during instance shutdown; this
908 adds the option ``--ignore-consistency`` to gnt-instance move
910 And as usual, various improvements to the error messages, documentation
917 *(Released Wed, 09 Mar 2011)*
919 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
920 the RAPI users file if run twice (even with ``--dry-run``).
922 The release fixes that bug (nothing else changed).
928 *(Released Mon, 07 Mar 2011)*
930 Final 2.4.0 release. Just a few small fixes:
932 - Fixed RAPI node evacuate
933 - Fixed the kvm-ifup script
934 - Fixed internal error handling for special job cases
935 - Updated man page to specify the escaping feature for options
941 *(Released Mon, 28 Feb 2011)*
943 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
944 that this feature wiped the disks of the wrong instance, leading to loss
949 - Fixed title of query field containing instance name
950 - Expanded the glossary in the documentation
951 - Fixed one unittest (internal issue)
957 *(Released Mon, 21 Feb 2011)*
959 A number of bug fixes plus just a couple functionality changes.
961 On the user-visible side, the ``gnt-* list`` command output has changed
962 with respect to "special" field states. The current rc1 style of display
963 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
964 the default output mode special fields states are displayed as follows:
966 - Offline resource: ``*``
967 - Unavailable/not applicable: ``-``
968 - Data missing (RPC failure): ``?``
969 - Unknown field: ``??``
971 Another user-visible change is the addition of ``--force-join`` to
976 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
977 - Fixed regression in RAPI/instance reinstall where all parameters were
978 required (instead of optional)
979 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
980 - Fixed iallocator usage (offline nodes were not considered offline)
981 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
982 - Fixed hypervisor and OS parameter validation with respect to
984 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
986 - Fixed ``tools/listrunner`` with respect to agent-based usage
992 *(Released Fri, 4 Feb 2011)*
994 Many changes and fixes since the beta1 release. While there were some
995 internal changes, the code has been mostly stabilised for the RC
998 Note: the dumb allocator was removed in this release, as it was not kept
999 up-to-date with the IAllocator protocol changes. It is recommended to
1000 use the ``hail`` command from the ganeti-htools package.
1002 Note: the 2.4 and up versions of Ganeti are not compatible with the
1003 0.2.x branch of ganeti-htools. You need to upgrade to
1004 ganeti-htools-0.3.0 (or later).
1006 Regressions fixed from 2.3
1007 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1009 - Fixed the ``gnt-cluster verify-disks`` command
1010 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1012 - Fixed disk adoption breakage
1013 - Fixed wrong headers in instance listing for field aliases
1018 - Fixed corner case in KVM handling of NICs
1019 - Fixed many cases of wrong handling of non-vm_capable nodes
1020 - Fixed a bug where a missing instance symlink was not possible to
1021 recreate with any ``gnt-*`` command (now ``gnt-instance
1022 activate-disks`` does it)
1023 - Fixed the volume group name as reported by ``gnt-cluster
1025 - Increased timeouts for the import-export code, hopefully leading to
1026 fewer aborts due network or instance timeouts
1027 - Fixed bug in ``gnt-node list-storage``
1028 - Fixed bug where not all daemons were started on cluster
1029 initialisation, but only at the first watcher run
1030 - Fixed many bugs in the OOB implementation
1031 - Fixed watcher behaviour in presence of instances with offline
1033 - Fixed instance list output for instances running on the wrong node
1034 - a few fixes to the cluster-merge tool, but it still cannot merge
1035 multi-node groups (currently it is not recommended to use this tool)
1041 - Improved network configuration for the KVM hypervisor
1042 - Added e1000 as a supported NIC for Xen-HVM
1043 - Improved the lvmstrap tool to also be able to use partitions, as
1044 opposed to full disks
1045 - Improved speed of disk wiping (the cluster parameter
1046 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1047 of instance creations
1048 - Added documentation for the OS parameters
1049 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1050 hypervisor is not responding
1051 - Added display of blacklisted and hidden OS information in
1052 ``gnt-cluster info``
1053 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1054 NIC and node parameters, which might create problems with currently
1055 invalid (but undetected) configuration files, but prevents validation
1056 failures when unrelated parameters are modified
1057 - Changed cluster initialisation to wait for the master daemon to become
1059 - Expanded the RAPI interface:
1061 - Added config redistribution resource
1062 - Added activation/deactivation of instance disks
1063 - Added export of console information
1065 - Implemented log file reopening on SIGHUP, which allows using
1066 logrotate(8) for the Ganeti log files
1067 - Added a basic OOB helper script as an example
1073 *(Released Fri, 14 Jan 2011)*
1078 - Fixed timezone issues when formatting timestamps
1079 - Added support for node groups, available via ``gnt-group`` and other
1081 - Added out-of-band framework and management, see :doc:`design
1082 document <design-oob>`
1083 - Removed support for roman numbers from ``gnt-node list`` and
1084 ``gnt-instance list``.
1085 - Allowed modification of master network interface via ``gnt-cluster
1086 modify --master-netdev``
1087 - Accept offline secondaries while shutting down instance disks
1088 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1089 - Added support for multiple LVM volume groups
1090 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1092 - Added commands to list available fields:
1094 - ``gnt-node list-fields``
1095 - ``gnt-group list-fields``
1096 - ``gnt-instance list-fields``
1098 - Updated documentation and man pages
1103 - Moved ``rapi_users`` file into separate directory, now named
1104 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1106 - Added new tool for running commands on many machines,
1107 ``tools/ganeti-listrunner``
1108 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1109 improving the ``gnt-instance console`` output
1110 - Allowed customisation of disk index separator at ``configure`` time
1111 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1112 - Added support for non-partitioned md disks in ``lvmstrap``
1113 - Added script to gracefully power off KVM instances
1114 - Split ``utils`` module into smaller parts
1115 - Changed query operations to return more detailed information, e.g.
1116 whether an information is unavailable due to an offline node. To use
1117 this new functionality, the LUXI call ``Query`` must be used. Field
1118 information is now stored by the master daemon and can be retrieved
1119 using ``QueryFields``. Instances, nodes and groups can also be queried
1120 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1121 exposed via RAPI). The following commands make use of this
1122 infrastructure change:
1124 - ``gnt-group list``
1125 - ``gnt-group list-fields``
1127 - ``gnt-node list-fields``
1128 - ``gnt-instance list``
1129 - ``gnt-instance list-fields``
1130 - ``gnt-debug locks``
1135 - New RAPI resources (see :doc:`rapi`):
1139 - ``/2/groups/[group_name]``
1140 - ``/2/groups/[group_name]/assign-nodes``
1141 - ``/2/groups/[group_name]/modify``
1142 - ``/2/groups/[group_name]/rename``
1143 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1147 - Implemented ``no_install`` for instance creation
1148 - Implemented OS parameters for instance reinstallation, allowing
1149 use of special settings on reinstallation (e.g. for preserving data)
1154 - Added IPv6 support in import/export
1155 - Pause DRBD synchronization while wiping disks on instance creation
1156 - Updated unittests and QA scripts
1157 - Improved network parameters passed to KVM
1158 - Converted man pages from docbook to reStructuredText
1164 *(Released Mon, 20 Dec 2010)*
1166 Released version 2.3.1~rc1 without any changes.
1172 *(Released Wed, 1 Dec 2010)*
1174 - impexpd: Disable OpenSSL compression in socat if possible (backport
1175 from master, commit e90739d625b, see :doc:`installation guide
1176 <install-quick>` for details)
1177 - Changed unittest coverage report to exclude test scripts
1178 - Added script to check version format
1184 *(Released Wed, 1 Dec 2010)*
1186 Released version 2.3.0~rc1 without any changes.
1192 *(Released Fri, 19 Nov 2010)*
1194 A number of bugfixes and documentation updates:
1196 - Update ganeti-os-interface documentation
1197 - Fixed a bug related to duplicate MACs or similar items which should be
1199 - Fix breakage in OS state modify
1200 - Reinstall instance: disallow offline secondaries (fixes bug related to
1201 OS changing but reinstall failing)
1202 - plus all the other fixes between 2.2.1 and 2.2.2
1208 *(Released Tue, 2 Nov 2010)*
1210 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1211 - Fixed master failover race with watcher
1212 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1214 - Accept previously stopped instance for export with instance removal
1215 - Simplify and extend the environment variables for instance OS scripts
1216 - Added new node flags, ``master_capable`` and ``vm_capable``
1217 - Added optional instance disk wiping prior during allocation. This is a
1218 cluster-wide option and can be set/modified using
1219 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1220 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1221 :doc:`install-quick`
1222 - Added a new watcher option (``--ignore-pause``)
1223 - Added option to ignore offline node on instance start/stop
1224 (``--ignore-offline``)
1225 - Allow overriding OS parameters with ``gnt-instance reinstall``
1226 - Added ability to change node's secondary IP address using ``gnt-node
1228 - Implemented privilege separation for all daemons except
1229 ``ganeti-noded``, see ``configure`` options
1230 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1232 - Implemented job priorities (see ``ganeti(7)`` manpage)
1233 - Ignore failures while shutting down instances during failover from
1235 - Exit daemon's bootstrap process only once daemon is ready
1236 - Export more information via ``LUInstanceQuery``/remote API
1237 - Improved documentation, QA and unittests
1238 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1239 restart if the file was created or changed
1240 - Added LUXI protocol version sent with each request and response,
1241 allowing detection of server/client mismatches
1242 - Moved the Python scripts among gnt-* and ganeti-* into modules
1243 - Moved all code related to setting up SSH to an external script,
1245 - Infrastructure changes for node group support in future versions
1251 *(Released Fri, 19 Nov 2010)*
1253 A few small bugs fixed, and some improvements to the build system:
1255 - Fix documentation regarding conversion to drbd
1256 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1258 - Fix error handling in node modify with multiple changes
1259 - Allow remote imports without checked names
1265 *(Released Tue, 19 Oct 2010)*
1267 - Disable SSL session ID cache in RPC client
1273 *(Released Thu, 14 Oct 2010)*
1275 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1276 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1278 - Fix problems with interaction between Curl and Python's HTTP server,
1279 resulting in increased speed in many RPC calls
1280 - Improve our release script to prevent breakage with older aclocal and
1287 *(Released Thu, 7 Oct 2010)*
1289 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1290 value retrieved from master
1291 - Added support for blacklisted or hidden OS definitions
1292 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1293 - Added support for -mem-path in KVM hypervisor abstraction layer
1294 - Allow overriding instance parameters in tool for inter-cluster
1295 instance moves (``tools/move-instance``)
1296 - Improved opcode summaries (e.g. in ``gnt-job list``)
1297 - Improve consistency of OS listing by sorting it
1298 - Documentation updates
1304 *(Released Fri, 8 Oct 2010)*
1306 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1312 *(Released Mon, 4 Oct 2010)*
1314 - Fixed regression in ``gnt-instance rename``
1320 *(Released Wed, 22 Sep 2010)*
1322 - Fixed OS_VARIANT variable for OS scripts
1323 - Fixed cluster tag operations via RAPI
1324 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1325 - Disabled RAPI CA checks in watcher
1331 *(Released Mon, 23 Aug 2010)*
1333 - Support DRBD versions of the format "a.b.c.d"
1335 - Re-introduce support for usage from multiple threads in RAPI client
1336 - Instance renames and modify via RAPI
1337 - Work around race condition between processing and archival in job
1339 - Mark opcodes following failed one as failed, too
1340 - Job field ``lock_status`` was removed due to difficulties making it
1341 work with the changed job queue in Ganeti 2.2; a better way to monitor
1342 locks is expected for a later 2.2.x release
1343 - Fixed dry-run behaviour with many commands
1344 - Support ``ssh-agent`` again when adding nodes
1345 - Many additional bugfixes
1351 *(Released Fri, 30 Jul 2010)*
1353 Important change: the internal RPC mechanism between Ganeti nodes has
1354 changed from using a home-grown http library (based on the Python base
1355 libraries) to use the PycURL library. This requires that PycURL is
1356 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1357 against GnuTLS by default. cURL's support for GnuTLS had known issues
1358 before cURL 7.21.0 and we recommend using the latest cURL release or
1359 linking against OpenSSL. Most other distributions already link PycURL
1360 and cURL against OpenSSL. The command::
1362 python -c 'import pycurl; print pycurl.version'
1364 can be used to determine the libraries PycURL and cURL are linked
1367 Other significant changes:
1369 - Rewrote much of the internals of the job queue, in order to achieve
1370 better parallelism; this decouples job query operations from the job
1371 processing, and it should allow much nicer behaviour of the master
1372 daemon under load, and it also has uncovered some long-standing bugs
1373 related to the job serialisation (now fixed)
1374 - Added a default iallocator setting to the cluster parameters,
1375 eliminating the need to always pass nodes or an iallocator for
1376 operations that require selection of new node(s)
1377 - Added experimental support for the LXC virtualization method
1378 - Added support for OS parameters, which allows the installation of
1379 instances to pass parameter to OS scripts in order to customise the
1381 - Added a hypervisor parameter controlling the migration type (live or
1382 non-live), since hypervisors have various levels of reliability; this
1383 has renamed the 'live' parameter to 'mode'
1384 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1385 logical volumes, meaning that cluster verify will ignore them and not
1386 flag their presence as errors
1387 - The watcher will now reset the error count for failed instances after
1388 8 hours, thus allowing self-healing if the problem that caused the
1389 instances to be down/fail to start has cleared in the meantime
1390 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1391 check for, and warn, if the drbd module parameter ``usermode_helper``
1392 is not consistent with the cluster-wide setting; this is needed to
1393 make diagnose easier of failed drbd creations
1394 - Started adding base IPv6 support, but this is not yet
1395 enabled/available for use
1396 - Rename operations (cluster, instance) will now return the new name,
1397 which is especially useful if a short name was passed in
1398 - Added support for instance migration in RAPI
1399 - Added a tool to pre-configure nodes for the SSH setup, before joining
1400 them to the cluster; this will allow in the future a simplified model
1401 for node joining (but not yet fully enabled in 2.2); this needs the
1402 paramiko python library
1403 - Fixed handling of name-resolving errors
1404 - Fixed consistency of job results on the error path
1405 - Fixed master-failover race condition when executed multiple times in
1407 - Fixed many bugs related to the job queue (mostly introduced during the
1408 2.2 development cycle, so not all are impacting 2.1)
1409 - Fixed instance migration with missing disk symlinks
1410 - Fixed handling of unknown jobs in ``gnt-job archive``
1411 - And many other small fixes/improvements
1415 - Enhanced both the unittest and the QA coverage
1416 - Switched the opcode validation to a generic model, and extended the
1417 validation to all opcode parameters
1418 - Changed more parts of the code that write shell scripts to use the
1420 - Switched the master daemon to use the asyncore library for the Luxi
1427 *(Released Thu, 17 Jun 2010)*
1429 - Added tool (``move-instance``) and infrastructure to move instances
1430 between separate clusters (see :doc:`separate documentation
1431 <move-instance>` and :doc:`design document <design-2.2>`)
1432 - Added per-request RPC timeout
1433 - RAPI now requires a Content-Type header for requests with a body (e.g.
1434 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1435 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1436 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1438 - Implemented initial support for running Ganeti daemons as separate
1439 users, see configure-time flags ``--with-user-prefix`` and
1440 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1442 - Instances can be removed after export (``gnt-backup export
1443 --remove-instance``)
1444 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1445 key (instead of 1024 bit)
1446 - Added new cluster configuration file for cluster domain secret
1447 - Import/export now use SSL instead of SSH
1448 - Added support for showing estimated time when exporting an instance,
1449 see the ``ganeti-os-interface(7)`` manpage and look for
1456 *(Released Tue, 16 Nov 2010)*
1458 Some more bugfixes. Unless critical bugs occur, this will be the last
1461 - Fix case of MAC special-values
1462 - Fix mac checker regex
1463 - backend: Fix typo causing "out of range" error
1464 - Add missing --units in gnt-instance list man page
1470 *(Released Tue, 24 Aug 2010)*
1473 - Don't ignore secondary node silently on non-mirrored disk templates
1475 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1476 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1477 - Properly document the usb_mouse parameter
1478 - Fix path in ganeti-rapi(8) (issue 116)
1479 - Adjust error message when the ganeti user's .ssh directory is
1481 - Add same-node-check when changing the disk template to drbd
1487 *(Released Fri, 16 Jul 2010)*
1490 - Add an option to only select some reboot types during qa/burnin.
1491 (on some hypervisors consequent reboots are not supported)
1492 - Fix infrequent race condition in master failover. Sometimes the old
1493 master ip address would be still detected as up for a short time
1494 after it was removed, causing failover to fail.
1495 - Decrease mlockall warnings when the ctypes module is missing. On
1496 Python 2.4 we support running even if no ctypes module is installed,
1497 but we were too verbose about this issue.
1498 - Fix building on old distributions, on which man doesn't have a
1500 - Fix RAPI not to ignore the MAC address on instance creation
1501 - Implement the old instance creation format in the RAPI client.
1507 *(Released Thu, 01 Jul 2010)*
1509 A small bugfix release:
1510 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1511 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1512 a non-existing option
1513 - Fix instance create: the --force-variant option was ignored
1514 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1515 - Fix modify node storage with non-FQDN arguments
1516 - Fix RAPI client to authenticate under Python 2.6 when used
1517 for more than 5 requests needing authentication
1518 - Fix gnt-instance modify -t (storage) giving a wrong error message
1519 when converting a non-shutdown drbd instance to plain
1525 *(Released Fri, 18 Jun 2010)*
1527 A small bugfix release:
1529 - Fix live migration of KVM instances started with older Ganeti
1530 versions which had fewer hypervisor parameters
1531 - Fix gnt-instance grow-disk on down instances
1532 - Fix an error-reporting bug during instance migration
1533 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1534 silently ignoring broken ones
1535 - Fix an RPC error reporting bug affecting, for example, RAPI client
1537 - Fix bug triggered by different API version os-es on different nodes
1538 - Fix a bug in instance startup with custom hvparams: OS level
1539 parameters would fail to be applied.
1540 - Fix the RAPI client under Python 2.6 (but more work is needed to
1541 make it work completely well with OpenSSL)
1542 - Fix handling of errors when resolving names from DNS
1548 *(Released Thu, 3 Jun 2010)*
1550 A medium sized development cycle. Some new features, and some
1551 fixes/small improvements/cleanups.
1553 Significant features
1554 ~~~~~~~~~~~~~~~~~~~~
1556 The node deamon now tries to mlock itself into memory, unless the
1557 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1558 its logs, and falls back to console logging. This allows emergency
1559 features such as ``gnt-node powercycle`` to work even in the event of a
1560 broken node disk (tested offlining the disk hosting the node's
1561 filesystem and dropping its memory caches; don't try this at home)
1563 KVM: add vhost-net acceleration support. It can be tested with a new
1564 enough version of the kernel and of qemu-kvm.
1566 KVM: Add instance chrooting feature. If you use privilege dropping for
1567 your VMs you can also now force them to chroot to an empty directory,
1568 before starting the emulated guest.
1570 KVM: Add maximum migration bandwith and maximum downtime tweaking
1571 support (requires a new-enough version of qemu-kvm).
1573 Cluster verify will now warn if the master node doesn't have the master
1574 ip configured on it.
1576 Add a new (incompatible) instance creation request format to RAPI which
1577 supports all parameters (previously only a subset was supported, and it
1578 wasn't possible to extend the old format to accomodate all the new
1579 features. The old format is still supported, and a client can check for
1580 this feature, before using it, by checking for its presence in the
1581 ``features`` RAPI resource.
1583 Now with ancient latin support. Try it passing the ``--roman`` option to
1584 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1585 (requires the python-roman module to be installed, in order to work).
1590 As usual many internal code refactorings, documentation updates, and
1593 - Lots of improvements and cleanups to the experimental Remote API
1594 (RAPI) client library.
1595 - A new unit test suite for the core daemon libraries.
1596 - A fix to creating missing directories makes sure the umask is not
1597 applied anymore. This enforces the same directory permissions
1599 - Better handling terminating daemons with ctrl+c (used when running
1600 them in debugging mode).
1601 - Fix a race condition in live migrating a KVM instance, when stat()
1602 on the old proc status file returned EINVAL, which is an unexpected
1604 - Fixed manpage checking with newer man and utf-8 charachters. But now
1605 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1611 *(Released Fri, 7 May 2010)*
1613 Fix a bug which prevented untagged KVM instances from starting.
1619 *(Released Fri, 7 May 2010)*
1621 Another release with a long development cycle, during which many
1622 different features were added.
1624 Significant features
1625 ~~~~~~~~~~~~~~~~~~~~
1627 The KVM hypervisor now can run the individual instances as non-root, to
1628 reduce the impact of a VM being hijacked due to bugs in the
1629 hypervisor. It is possible to run all instances as a single (non-root)
1630 user, to manually specify a user for each instance, or to dynamically
1631 allocate a user out of a cluster-wide pool to each instance, with the
1632 guarantee that no two instances will run under the same user ID on any
1635 An experimental RAPI client library, that can be used standalone
1636 (without the other Ganeti libraries), is provided in the source tree as
1637 ``lib/rapi/client.py``. Note this client might change its interface in
1638 the future, as we iterate on its capabilities.
1640 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1641 replace the cluster's certificates and crypto keys. This might help in
1642 case they have been compromised, or have simply expired.
1644 A new disk option for instance creation has been added that allows one
1645 to "adopt" currently existing logical volumes, with data
1646 preservation. This should allow easier migration to Ganeti from
1647 unmanaged (or managed via other software) instances.
1649 Another disk improvement is the possibility to convert between redundant
1650 (DRBD) and plain (LVM) disk configuration for an instance. This should
1651 allow better scalability (starting with one node and growing the
1652 cluster, or shrinking a two-node cluster to one node).
1654 A new feature that could help with automated node failovers has been
1655 implemented: if a node sees itself as offline (by querying the master
1656 candidates), it will try to shutdown (hard) all instances and any active
1657 DRBD devices. This reduces the risk of duplicate instances if an
1658 external script automatically failovers the instances on such nodes. To
1659 enable this, the cluster parameter ``maintain_node_health`` should be
1660 enabled; in the future this option (per the name) will enable other
1661 automatic maintenance features.
1663 Instance export/import now will reuse the original instance
1664 specifications for all parameters; that means exporting an instance,
1665 deleting it and the importing it back should give an almost identical
1666 instance. Note that the default import behaviour has changed from
1667 before, where it created only one NIC; now it recreates the original
1670 Cluster verify has added a few new checks: SSL certificates validity,
1671 /etc/hosts consistency across the cluster, etc.
1676 As usual, many internal changes were done, documentation fixes,
1679 - Fixed cluster initialization with disabled cluster storage (regression
1680 introduced in 2.1.1)
1681 - File-based storage supports growing the disks
1682 - Fixed behaviour of node role changes
1683 - Fixed cluster verify for some corner cases, plus a general rewrite of
1684 cluster verify to allow future extension with more checks
1685 - Fixed log spamming by watcher and node daemon (regression introduced
1687 - Fixed possible validation issues when changing the list of enabled
1689 - Fixed cleanup of /etc/hosts during node removal
1690 - Fixed RAPI response for invalid methods
1691 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1692 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1693 from ide disks, etc.)
1694 - Allow OS changes without re-installation (to record a changed OS
1695 outside of Ganeti, or to allow OS renames)
1696 - Allow instance creation without OS installation (useful for example if
1697 the OS will be installed manually, or restored from a backup not in
1699 - Implemented option to make cluster ``copyfile`` use the replication
1701 - Added list of enabled hypervisors to ssconf (possibly useful for
1703 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1705 - A partial form of node re-IP is possible via node readd, which now
1706 allows changed node primary IP
1707 - Command line utilities now show an informational message if the job is
1709 - The logs of the master daemon now show the PID/UID/GID of the
1716 *(Released Fri, 12 Mar 2010)*
1718 During the 2.1.0 long release candidate cycle, a lot of improvements and
1719 changes have accumulated with were released later as 2.1.1.
1724 The node evacuate command (``gnt-node evacuate``) was significantly
1725 rewritten, and as such the IAllocator protocol was changed - a new
1726 request type has been added. This unfortunate change during a stable
1727 series is designed to improve performance of node evacuations; on
1728 clusters with more than about five nodes and which are well-balanced,
1729 evacuation should proceed in parallel for all instances of the node
1730 being evacuated. As such, any existing IAllocator scripts need to be
1731 updated, otherwise the above command will fail due to the unknown
1732 request. The provided "dumb" allocator has not been updated; but the
1733 ganeti-htools package supports the new protocol since version 0.2.4.
1735 Another important change is increased validation of node and instance
1736 names. This might create problems in special cases, if invalid host
1737 names are being used.
1739 Also, a new layer of hypervisor parameters has been added, that sits at
1740 OS level between the cluster defaults and the instance ones. This allows
1741 customisation of virtualization parameters depending on the installed
1742 OS. For example instances with OS 'X' may have a different KVM kernel
1743 (or any other parameter) than the cluster defaults. This is intended to
1744 help managing a multiple OSes on the same cluster, without manual
1745 modification of each instance's parameters.
1747 A tool for merging clusters, ``cluster-merge``, has been added in the
1748 tools sub-directory.
1753 - Improved the int/float conversions that should make the code more
1754 robust in face of errors from the node daemons
1755 - Fixed the remove node code in case of internal configuration errors
1756 - Fixed the node daemon behaviour in face of inconsistent queue
1757 directory (e.g. read-only file-system where we can't open the files
1759 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1760 now it either aborts cleanly or, if given the new "auto_promote"
1761 parameter, will automatically promote other nodes as needed
1762 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1763 completely prevent Ganeti from working
1764 - Fixed bug for instance export when not all disks were successfully
1766 - Fixed behaviour of node add when the new node is slow in starting up
1768 - Fixed handling of signals in the LUXI client, which should improve
1769 behaviour of command-line scripts
1770 - Added checks for invalid node/instance names in the configuration (now
1771 flagged during cluster verify)
1772 - Fixed watcher behaviour for disk activation errors
1773 - Fixed two potentially endless loops in http library, which led to the
1774 RAPI daemon hanging and consuming 100% CPU in some cases
1775 - Fixed bug in RAPI daemon related to hashed passwords
1776 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1778 - Enhanced compatibility with non-Debian OSes, but not using absolute
1779 path in some commands and allowing customisation of the ssh
1780 configuration directory
1781 - Fixed possible future issue with new Python versions by abiding to the
1782 proper use of ``__slots__`` attribute on classes
1783 - Added checks that should prevent directory traversal attacks
1784 - Many documentation fixes based on feedback from users
1789 - Added an "early_release" more for instance replace disks and node
1790 evacuate, where we release locks earlier and thus allow higher
1791 parallelism within the cluster
1792 - Added watcher hooks, intended to allow the watcher to restart other
1793 daemons (e.g. from the ganeti-nbma project), but they can be used of
1794 course for any other purpose
1795 - Added a compile-time disable for DRBD barriers, to increase
1796 performance if the administrator trusts the power supply or the
1797 storage system to not lose writes
1798 - Added the option of using syslog for logging instead of, or in
1799 addition to, Ganeti's own log files
1800 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1801 can indeed boot from a paravirtual NIC
1802 - Added a generic debug level for many operations; while this is not
1803 used widely yet, it allows one to pass the debug value all the way to
1805 - Enhanced the hooks environment for instance moves (failovers,
1806 migrations) where the primary/secondary nodes changed during the
1807 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1808 - Enhanced data validations for many user-supplied values; one important
1809 item is the restrictions imposed on instance and node names, which
1810 might reject some (invalid) host names
1811 - Add a configure-time option to disable file-based storage, if it's not
1812 needed; this allows greater security separation between the master
1813 node and the other nodes from the point of view of the inter-node RPC
1815 - Added user notification in interactive tools if job is waiting in the
1816 job queue or trying to acquire locks
1817 - Added log messages when a job is waiting for locks
1818 - Added filtering by node tags in instance operations which admit
1819 multiple instances (start, stop, reboot, reinstall)
1820 - Added a new tool for cluster mergers, ``cluster-merge``
1821 - Parameters from command line which are of the form ``a=b,c=d`` can now
1822 use backslash escapes to pass in values which contain commas,
1823 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1825 - For KVM, the instance name is the first parameter passed to KVM, so
1826 that it's more visible in the process list
1832 *(Released Tue, 2 Mar 2010)*
1834 Ganeti 2.1 brings many improvements with it. Major changes:
1836 - Added infrastructure to ease automated disk repairs
1837 - Added new daemon to export configuration data in a cheaper way than
1838 using the remote API
1839 - Instance NICs can now be routed instead of being associated with a
1841 - Improved job locking logic to reduce impact of jobs acquiring multiple
1842 locks waiting for other long-running jobs
1844 In-depth implementation details can be found in the Ganeti 2.1 design
1850 - Added chroot hypervisor
1851 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1853 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1854 ``bootloader_path`` and ``bootloader_args``)
1855 - Added the ``use_localtime`` option for the xen-hvm and kvm
1856 hypervisors, and the default value for this has changed to false (in
1857 2.0 xen-hvm always enabled it)
1858 - Added luxi call to submit multiple jobs in one go
1859 - Added cluster initialization option to not modify ``/etc/hosts``
1861 - Added network interface parameters
1862 - Added dry run mode to some LUs
1863 - Added RAPI resources:
1865 - ``/2/instances/[instance_name]/info``
1866 - ``/2/instances/[instance_name]/replace-disks``
1867 - ``/2/nodes/[node_name]/evacuate``
1868 - ``/2/nodes/[node_name]/migrate``
1869 - ``/2/nodes/[node_name]/role``
1870 - ``/2/nodes/[node_name]/storage``
1871 - ``/2/nodes/[node_name]/storage/modify``
1872 - ``/2/nodes/[node_name]/storage/repair``
1874 - Added OpCodes to evacuate or migrate all instances on a node
1875 - Added new command to list storage elements on nodes (``gnt-node
1876 list-storage``) and modify them (``gnt-node modify-storage``)
1877 - Added new ssconf files with master candidate IP address
1878 (``ssconf_master_candidates_ips``), node primary IP address
1879 (``ssconf_node_primary_ips``) and node secondary IP address
1880 (``ssconf_node_secondary_ips``)
1881 - Added ``ganeti-confd`` and a client library to query the Ganeti
1882 configuration via UDP
1883 - Added ability to run hooks after cluster initialization and before
1885 - Added automatic mode for disk replace (``gnt-instance replace-disks
1887 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1888 after catastrophic data-loss
1889 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1891 - Added ``gnt-instance move`` command to move instances
1892 - Added ``gnt-cluster watcher`` command to control watcher
1893 - Added ``gnt-node powercycle`` command to powercycle nodes
1894 - Added new job status field ``lock_status``
1895 - Added parseable error codes to cluster verification (``gnt-cluster
1896 verify --error-codes``) and made output less verbose (use
1897 ``--verbose`` to restore previous behaviour)
1898 - Added UUIDs to the main config entities (cluster, nodes, instances)
1899 - Added support for OS variants
1900 - Added support for hashed passwords in the Ganeti remote API users file
1902 - Added option to specify maximum timeout on instance shutdown
1903 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1904 - Added new helper script to start and stop Ganeti daemons
1905 (``daemon-util``), with the intent to reduce the work necessary to
1906 adjust Ganeti for non-Debian distributions and to start/stop daemons
1908 - Added more unittests
1909 - Fixed critical bug in ganeti-masterd startup
1910 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1911 now customisable at the cluster level for both the KVM and Xen
1912 hypervisors using the new ``migration_port`` parameter
1913 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1914 reinstalling an instance
1915 - Allowed ``@`` in tag names
1916 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1917 - Many documentation updates
1918 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1919 - ``gnt-instance reinstall`` can now reinstall multiple instances
1920 - Updated many command line parameters
1921 - Introduced new OS API version 15
1922 - No longer support a default hypervisor
1923 - Treat virtual LVs as inexistent
1924 - Improved job locking logic to reduce lock contention
1925 - Match instance and node names case insensitively
1926 - Reimplemented bash completion script to be more complete
1933 *(Released Thu, 4 Feb 2010)*
1935 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1936 - Fix a string formatting bug
1937 - Improve safety of the code in some error paths
1938 - Improve data validation in the master of values returned from nodes
1944 *(Released Thu, 17 Dec 2009)*
1946 - Fix security issue due to missing validation of iallocator names; this
1947 allows local and remote execution of arbitrary executables
1948 - Fix failure of gnt-node list during instance removal
1949 - Ship the RAPI documentation in the archive
1955 *(Released Wed, 30 Sep 2009)*
1957 - Fixed many wrong messages
1958 - Fixed a few bugs related to the locking library
1959 - Fixed MAC checking at instance creation time
1960 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1961 - Fixed a few issues related to signal handling in both daemons and
1963 - Fixed the example startup script provided
1964 - Fixed insserv dependencies in the example startup script (patch from
1966 - Fixed handling of drained nodes in the iallocator framework
1967 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1969 - Fixed error related to invalid job IDs in job polling
1970 - Fixed job/opcode persistence on unclean master shutdown
1971 - Fixed handling of partial job processing after unclean master
1973 - Fixed error reporting from LUs, previously all errors were converted
1974 into execution errors
1975 - Fixed error reporting from burnin
1976 - Decreased significantly the memory usage of the job queue
1977 - Optimised slightly multi-job submission
1978 - Optimised slightly opcode loading
1979 - Backported the multi-job submit framework from the development
1980 branch; multi-instance start and stop should be faster
1981 - Added script to clean archived jobs after 21 days; this will reduce
1982 the size of the queue directory
1983 - Added some extra checks in disk size tracking
1984 - Added an example ethers hook script
1985 - Added a cluster parameter that prevents Ganeti from modifying of
1987 - Added more node information to RAPI responses
1988 - Added a ``gnt-job watch`` command that allows following the ouput of a
1990 - Added a bind-address option to ganeti-rapi
1991 - Added more checks to the configuration verify
1992 - Enhanced the burnin script such that some operations can be retried
1994 - Converted instance reinstall to multi-instance model
2000 *(Released Fri, 7 Aug 2009)*
2002 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2003 to allow using the pre-2.0.2 behaviour in activation, if any existing
2004 instances have mismatched disk sizes in the configuration
2005 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2006 any configuration mismatches for disk sizes
2007 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2008 failover to work on two-node clusters
2009 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2011 - Fixed detection of OS script errors in ``gnt-backup export``
2012 - Fixed exit code of ``gnt-backup export``
2018 *(Released Fri, 17 Jul 2009)*
2020 - Added experimental support for stripped logical volumes; this should
2021 enhance performance but comes with a higher complexity in the block
2022 device handling; stripping is only enabled when passing
2023 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2024 affected even in the non-stripped mode
2025 - Improved resiliency against transient failures at the end of DRBD
2026 resyncs, and in general of DRBD resync checks
2027 - Fixed a couple of issues with exports and snapshot errors
2028 - Fixed a couple of issues in instance listing
2029 - Added display of the disk size in ``gnt-instance info``
2030 - Fixed checking for valid OSes in instance creation
2031 - Fixed handling of the "vcpus" parameter in instance listing and in
2032 general of invalid parameters
2033 - Fixed http server library, and thus RAPI, to handle invalid
2034 username/password combinations correctly; this means that now they
2035 report unauthorized for queries too, not only for modifications,
2036 allowing earlier detect of configuration problems
2037 - Added a new "role" node list field, equivalent to the master/master
2038 candidate/drained/offline flags combinations
2039 - Fixed cluster modify and changes of candidate pool size
2040 - Fixed cluster verify error messages for wrong files on regular nodes
2041 - Fixed a couple of issues with node demotion from master candidate role
2042 - Fixed node readd issues
2043 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2044 - Added a new ``--no-voting`` option for masterfailover to fix failover
2045 on two-nodes clusters when the former master node is unreachable
2046 - Added instance reinstall over RAPI
2052 *(Released Tue, 16 Jun 2009)*
2054 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2055 allow re-adding the start in single-user option (regression from 1.2)
2056 - the watcher writes the instance status to a file, to allow monitoring
2057 to report the instance status (from the master) based on cached
2058 results of the watcher's queries; while this can get stale if the
2059 watcher is being locked due to other work on the cluster, this is
2060 still an improvement
2061 - the watcher now also restarts the node daemon and the rapi daemon if
2063 - fixed the watcher to handle full and drained queue cases
2064 - hooks export more instance data in the environment, which helps if
2065 hook scripts need to take action based on the instance's properties
2066 (no longer need to query back into ganeti)
2067 - instance failovers when the instance is stopped do not check for free
2068 RAM, so that failing over a stopped instance is possible in low memory
2070 - rapi uses queries for tags instead of jobs (for less job traffic), and
2071 for cluster tags it won't talk to masterd at all but read them from
2073 - a couple of error handling fixes in RAPI
2074 - drbd handling: improved the error handling of inconsistent disks after
2075 resync to reduce the frequency of "there are some degraded disks for
2076 this instance" messages
2077 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2078 error handling path called a wrong function name)
2084 *(Released Wed, 27 May 2009)*
2086 - no changes from rc5
2092 *(Released Wed, 20 May 2009)*
2094 - fix a couple of bugs (validation, argument checks)
2095 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2096 - some small improvements to RAPI and IAllocator
2097 - make watcher automatically start the master daemon if down
2103 *(Released Mon, 27 Apr 2009)*
2105 - change the OS list to not require locks; this helps with big clusters
2106 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2107 volume group is broken
2108 - ``gnt-instance info``, without any arguments, doesn't run for all
2109 instances anymore; either pass ``--all`` or pass the desired
2110 instances; this helps against mistakes on big clusters where listing
2111 the information for all instances takes a long time
2112 - miscellaneous doc and man pages fixes
2118 *(Released Wed, 8 Apr 2009)*
2120 - Change the internal locking model of some ``gnt-node`` commands, in
2121 order to reduce contention (and blocking of master daemon) when
2122 batching many creation/reinstall jobs
2123 - Fixes to Xen soft reboot
2124 - No longer build documentation at build time, instead distribute it in
2125 the archive, in order to reduce the need for the whole docbook/rst
2132 *(Released Fri, 27 Mar 2009)*
2134 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2135 - Fix watcher startup sequence, improves the behaviour of busy clusters
2136 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2137 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2138 bind address and other places
2139 - Some documentation fixes and updates
2145 *(Released Mon, 2 Mar 2009)*
2147 - More documentation updates, now all docs should be more-or-less
2149 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2151 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2153 - Fix an issue related to $libdir/run/ganeti and cluster creation
2159 *(Released Thu, 19 Feb 2009)*
2161 - Xen PVM and KVM have switched the default value for the instance root
2162 disk to the first partition on the first drive, instead of the whole
2163 drive; this means that the OS installation scripts must be changed
2165 - Man pages have been updated
2166 - RAPI has been switched by default to HTTPS, and the exported functions
2167 should all work correctly
2168 - RAPI v1 has been removed
2169 - Many improvements to the KVM hypervisor
2170 - Block device errors are now better reported
2171 - Many other bugfixes and small improvements
2177 *(Released Mon, 26 Jan 2009)*
2179 - Version 2 is a general rewrite of the code and therefore the
2180 differences are too many to list, see the design document for 2.0 in
2181 the ``doc/`` subdirectory for more details
2182 - In this beta version there is not yet a migration path from 1.2 (there
2183 will be one in the final 2.0 release)
2184 - A few significant changes are:
2186 - all commands are executed by a daemon (``ganeti-masterd``) and the
2187 various ``gnt-*`` commands are just front-ends to it
2188 - all the commands are entered into, and executed from a job queue,
2189 see the ``gnt-job(8)`` manpage
2190 - the RAPI daemon supports read-write operations, secured by basic
2191 HTTP authentication on top of HTTPS
2192 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2193 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2194 to migrate to DRBD 8 first while still running Ganeti 1.2)
2195 - DRBD devices are using statically allocated minor numbers, which
2196 will be assigned to existing instances during the migration process
2197 - there is support for both Xen PVM and Xen HVM instances running on
2199 - KVM virtualization is supported too
2200 - file-based storage has been implemented, which means that it is
2201 possible to run the cluster without LVM and DRBD storage, for
2202 example using a shared filesystem exported from shared storage (and
2203 still have live migration)
2209 *(Released Tue, 13 Jan 2009)*
2211 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2212 - Reuse the old instance mac address by default on instance import, if
2213 the instance name is the same.
2214 - Handle situations in which the node info rpc returns incomplete
2216 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2217 - Improved version of batcher:
2219 - state file support
2220 - instance mac address support
2221 - support for HVM clusters/instances
2223 - Add an option to show the number of cpu sockets and nodes in
2225 - Support OSes that handle more than one version of the OS api (but do
2226 not change the current API in any other way)
2227 - Fix ``gnt-node migrate``
2228 - ``gnt-debug`` man page
2229 - Fixes various more typos and small issues
2230 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2236 *(Released Wed, 24 Sep 2008)*
2238 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2239 type of disk exported to fully virtualized instances.
2240 - provide access to the serial console of HVM instances
2241 - instance auto_balance flag, set by default. If turned off it will
2242 avoid warnings on cluster verify if there is not enough memory to fail
2243 over an instance. in the future it will prevent automatically failing
2244 it over when we will support that.
2245 - batcher tool for instance creation, see ``tools/README.batcher``
2246 - ``gnt-instance reinstall --select-os`` to interactively select a new
2247 operating system when reinstalling an instance.
2248 - when changing the memory amount on instance modify a check has been
2249 added that the instance will be able to start. also warnings are
2250 emitted if the instance will not be able to fail over, if auto_balance
2252 - documentation fixes
2253 - sync fields between ``gnt-instance list/modify/add/import``
2254 - fix a race condition in drbd when the sync speed was set after giving
2255 the device a remote peer.
2261 *(Released Tue, 22 Jul 2008)*
2263 - note: the allowed size and number of tags per object were reduced
2264 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2265 - fixed twisted 8.x compatibility
2266 - fixed ``gnt-instance replace-disks`` with iallocator
2267 - add TCP keepalives on twisted connections to detect restarted nodes
2268 - disk increase support, see ``gnt-instance grow-disk``
2269 - implement bulk node/instance query for RAPI
2270 - add tags in node/instance listing (optional)
2271 - experimental migration (and live migration) support, read the man page
2272 for ``gnt-instance migrate``
2273 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2274 has some small improvements in handling its state file
2280 *(Released Fri, 13 Jun 2008)*
2282 - Experimental readonly, REST-based remote API implementation;
2283 automatically started on master node, TCP port 5080, if enabled by
2284 ``--enable-rapi`` parameter to configure script.
2285 - Instance allocator support. Add and import instance accept a
2286 ``--iallocator`` parameter, and call that instance allocator to decide
2287 which node to use for the instance. The iallocator document describes
2288 what's expected from an allocator script.
2289 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2290 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2291 node is lost there is still enough memory to fail over the instances
2293 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2294 ``gnt-cluster verify``, to check for site-specific compliance. All the
2295 hooks will run, and their output, if any, will be displayed. Any
2296 failing hook will make the verification return an error value.
2297 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2298 primary and secondary interfaces
2299 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2300 that is still declared as part of the cluster and has failed.
2301 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2302 specifying output fields, that just adds the chosen fields to the
2304 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2305 export from the filesystem.
2306 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2307 have been added. Using them you can enable/disable acpi and pae
2308 support, and specify a path for a cd image to be exported to the
2309 instance. These parameters as the name suggest only work on HVM
2311 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2312 PAE support will be set to the previously hardcoded values, but the
2313 (previously hardcoded) path to the CDROM ISO image will be unset and
2314 if required, needs to be set manually with ``gnt-instance modify``
2316 - The address to which an instance's VNC console is bound is now
2317 selectable per-instance, rather than being cluster wide. Of course
2318 this only applies to instances controlled via VNC, so currently just
2319 applies to HVM clusters.
2325 *(Released Mon, 18 Feb 2008)*
2327 - more tweaks to the disk activation code (especially helpful for DRBD)
2328 - change the default ``gnt-instance list`` output format, now there is
2329 one combined status field (see the manpage for the exact values this
2331 - some more fixes for the mac export to hooks change
2332 - make Ganeti not break with DRBD 8.2.x (which changed the version
2333 format in ``/proc/drbd``) (issue 24)
2334 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2335 template, allowing migration from DRBD0.7+MD to DRBD8
2341 *(Released Wed, 30 Jan 2008)*
2343 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2345 - add command aliases infrastructure and a few aliases
2346 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2347 man pages and the ``--help`` option of ``gnt-node
2348 list``/``gnt-instance list``
2349 - fix ``gnt-backup list`` with down nodes (issue 21)
2350 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2351 - fix the dist archive and add a check for including svn/git files in
2353 - some developer-related changes: improve the burnin and the QA suite,
2354 add an upload script for testing during development
2360 *(Released Wed, 16 Jan 2008)*
2362 - experimental HVM support, read the install document, section
2363 "Initializing the cluster"
2364 - allow for the PVM hypervisor per-instance kernel and initrd paths
2365 - add a new command ``gnt-cluster verify-disks`` which uses a new
2366 algorithm to improve the reconnection of the DRBD pairs if the device
2367 on the secondary node has gone away
2368 - make logical volume code auto-activate LVs at disk activation time
2369 - slightly improve the speed of activating disks
2370 - allow specification of the MAC address at instance creation time, and
2371 changing it later via ``gnt-instance modify``
2372 - fix handling of external commands that generate lots of output on
2374 - update documentation with regard to minimum version of DRBD8 supported
2380 *(Released Tue, 4 Dec 2007)*
2382 - Log the ``xm create`` output to the node daemon log on failure (to
2383 help diagnosing the error)
2384 - In debug mode, log all external commands output if failed to the logs
2385 - Change parsing of lvm commands to ignore stderr
2391 *(Released Wed, 28 Nov 2007)*
2393 - Another round of updates to the DRBD 8 code to deal with more failures
2394 in the replace secondary node operation
2395 - Some more logging of failures in disk operations (lvm, drbd)
2396 - A few documentation updates
2403 *(Released Tue, 13 Nov 2007)*
2405 - Change configuration file format from Python's Pickle to JSON.
2406 Upgrading is possible using the cfgupgrade utility.
2407 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2408 faster replace disks and is more stable (DRBD 8 has many improvements
2409 compared to DRBD 0.7)
2410 - Added command line tags support (see man pages for ``gnt-instance``,
2411 ``gnt-node``, ``gnt-cluster``)
2412 - Added instance rename support
2413 - Added multi-instance startup/shutdown
2414 - Added cluster rename support
2415 - Added ``gnt-node evacuate`` to simplify some node operations
2416 - Added instance reboot operation that can speedup reboot as compared to
2418 - Soften the requirement that hostnames are in FQDN format
2419 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2421 - Removed dependency on debian's patched fping that uses the
2422 non-standard ``-S`` option
2423 - Now the OS definitions are searched for in multiple, configurable
2424 paths (easier for distros to package)
2425 - Some changes to the hooks infrastructure (especially the new
2426 post-configuration update hook)
2427 - Other small bugfixes
2429 .. vim: set textwidth=72 syntax=rst :