10 - ``gnt-instance batch-create`` has been changed to use the bulk create
11 opcode from Ganeti. This lead to incompatible changes in the format of
12 the JSON file. It's now not a custom dict anymore but a dict
13 compatible with the ``OpInstanceCreate`` opcode.
14 - Parent directories for file storage need to be listed in
15 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
16 write the file automatically based on old configuration values, but it
17 can not distribute it across all nodes and the file contents should be
18 verified. Use ``gnt-cluster copyfile
19 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
20 upgraded. The reason for requiring this list of paths now is that
21 before it would have been possible to inject new paths via RPC,
22 allowing files to be created in arbitrary locations. The RPC protocol
23 is protected using SSL/X.509 certificates, but as a design principle
24 Ganeti does not permit arbitrary paths to be passed.
25 - The parsing of the variants file for OSes (see
26 :manpage:`ganeti-os-interface(8)` has been slightly changed: now empty
27 lines and comment lines (starting with ``#``) are ignored for better
29 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
30 longer available. ``gnt-node add`` now invokes a new tool on the
31 destination node, named ``prepare-node-join``, to configure the SSH
32 daemon. Paramiko is no longer necessary to configure nodes' SSH
33 daemons via ``gnt-node add``.
34 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
35 for RAPI users. It allows granting permissions to query for
36 information to a specific user without giving
37 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
38 - A new tool named ``node-cleanup`` has been added. It cleans remains of
39 a cluster from a machine by stopping all daemons, removing
40 certificates and ssconf files. Unless the ``--no-backup`` option is
41 given, copies of the certificates are made.
42 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
43 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
44 the flag is not reset after a master failover.
45 - Instance creations now support the use of opportunistic locking,
46 potentially speeding up the (parallel) creation of multiple instances.
47 This feature is currently only available via the :doc:`RAPI
48 <rapi>` interface and when an instance allocator is used. If the
49 ``opportunistic_locking`` parameter is set the opcode will try to
50 acquire as many locks as possible, but will not wait for any locks
51 held by other opcodes. If the not enough resources can be found to
52 allocate the instance, the temporary error code
53 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
54 retried thereafter, with or without opportunistic locking.
60 *(Released Fri, 21 Dec 2012)*
62 Important behaviour change: hbal won't rebalance anymore instances which
63 have the ``auto_balance`` attribute set to false. This was the intention
64 all along, but until now it only skipped those from the N+1 memory
65 reservation (DRBD-specific).
67 A significant number of bug fixes in this release:
69 - Fixed disk adoption interaction with ipolicy checks.
70 - Fixed networking issues when instances are started, stopped or
71 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
72 - Fixed the warning in cluster verify for shared storage instances not
74 - Fixed removal of storage directory on shared file storage (issue 262).
75 - Fixed validation of LVM volume group name in OpClusterSetParams
76 (``gnt-cluster modify``) (issue 285).
77 - Fixed runtime memory increases (``gnt-instance modify -m``).
78 - Fixed live migration under Xen's ``xl`` mode.
79 - Fixed ``gnt-instance console`` with ``xl``.
80 - Fixed building with newer Haskell compiler/libraries.
81 - Fixed PID file writing in Haskell daemons (confd); this prevents
82 restart issues if confd was launched manually (outside of
83 ``daemon-util``) while another copy of it was running
84 - Fixed a type error when doing live migrations with KVM (issue 297) and
85 the error messages for failing migrations have been improved.
86 - Fixed opcode validation for the out-of-band commands (``gnt-node
88 - Fixed a type error when unsetting OS hypervisor parameters (issue
89 311); now it's possible to unset all OS-specific hypervisor
91 - Fixed the ``dry-run`` mode for many operations: verification of
92 results was over-zealous but didn't take into account the ``dry-run``
93 operation, resulting in "wrong" failures.
94 - Fixed bash completion in ``gnt-job list`` when the job queue has
95 hundreds of entries; especially with older ``bash`` versions, this
96 results in significant CPU usage.
98 And lastly, a few other improvements have been made:
100 - Added option to force master-failover without voting (issue 282).
101 - Clarified error message on lock conflict (issue 287).
102 - Logging of newly submitted jobs has been improved (issue 290).
103 - Hostname checks have been made uniform between instance rename and
105 - The ``--submit`` option is now supported by ``gnt-debug delay``.
106 - Shutting down the master daemon by sending SIGTERM now stops it from
107 processing jobs waiting for locks; instead, those jobs will be started
108 once again after the master daemon is started the next time (issue
110 - Support for Xen's ``xl`` program has been improved (besides the fixes
112 - Reduced logging noise in the Haskell confd daemon (only show one log
113 entry for each config reload, instead of two).
114 - Several man page updates and typo fixes.
120 *(Released Fri, 12 Oct 2012)*
122 A small bugfix release. Among the bugs fixed:
124 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
125 made the command unusable.
126 - Commands that issue many jobs don't fail anymore just because some jobs
127 take so long that other jobs are archived.
128 - Failures during ``gnt-instance reinstall`` are reflected by the exit
130 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
132 - When ``always_failover`` is set, ``--allow-failover`` is not required
133 in migrate commands anymore.
134 - ``bash_completion`` works even if extglob is disabled.
135 - Fixed bug with locks that made failover for RDB-based instances fail.
136 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
137 a random node instead of one based on the allocator metric.
138 - Support for newer versions of pylint and pep8.
139 - Hail doesn't fail anymore when trying to add an instance of type
140 ``file``, ``sharedfile`` or ``rbd``.
141 - Added new Makefile target to rebuild the whole distribution, so that
142 all files are included.
148 *(Released Fri, 27 Jul 2012)*
151 .. attention:: The ``LUXI`` protocol has been made more consistent
152 regarding its handling of command arguments. This, however, leads to
153 incompatibility issues with previous versions. Please ensure that you
154 restart Ganeti daemons soon after the upgrade, otherwise most
155 ``LUXI`` calls (job submission, setting/resetting the drain flag,
156 pausing/resuming the watcher, cancelling and archiving jobs, querying
157 the cluster configuration) will fail.
166 The current ``admin_up`` field, which used to denote whether an instance
167 should be running or not, has been removed. Instead, ``admin_state`` is
168 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
170 The rational behind this is that an instance being “down” can have
173 - it could be down during a reboot
174 - it could be temporarily be down for a reinstall
175 - or it could be down because it is deprecated and kept just for its
178 The previous Boolean state was making it difficult to do capacity
179 calculations: should Ganeti reserve memory for a down instance? Now, the
180 tri-state field makes it clear:
182 - in ``up`` and ``down`` state, all resources are reserved for the
183 instance, and it can be at any time brought up if it is down
184 - in ``offline`` state, only disk space is reserved for it, but not
187 The field can have an extra use: since the transition between ``up`` and
188 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
189 transition between ``offline`` and ``down`` is done via ``gnt-instance
190 modify``, it is possible to given different rights to users. For
191 example, owners of an instance could be allowed to start/stop it, but
192 not transition it out of the offline state.
194 Instance policies and specs
195 +++++++++++++++++++++++++++
197 In previous Ganeti versions, an instance creation request was not
198 limited on the minimum size and on the maximum size just by the cluster
199 resources. As such, any policy could be implemented only in third-party
200 clients (RAPI clients, or shell wrappers over ``gnt-*``
201 tools). Furthermore, calculating cluster capacity via ``hspace`` again
202 required external input with regards to instance sizes.
204 In order to improve these workflows and to allow for example better
205 per-node group differentiation, we introduced instance specs, which
208 - minimum instance disk size, disk count, memory size, cpu count
209 - maximum values for the above metrics
210 - and “standard” values (used in ``hspace`` to calculate the standard
213 The minimum/maximum values can be also customised at node-group level,
214 for example allowing more powerful hardware to support bigger instance
217 Beside the instance specs, there are a few other settings belonging to
218 the instance policy framework. It is possible now to customise, per
219 cluster and node-group:
221 - the list of allowed disk templates
222 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
223 - the maximum ratio of instance to spindles (see below for more
224 information) for local storage
226 All these together should allow all tools that talk to Ganeti to know
227 what are the ranges of allowed values for instances and the
228 over-subscription that is allowed.
230 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
231 instance configuration, and the physical CPU configuration from the
232 node. For the spindle ratios however, we didn't track before these
233 values, so new parameters have been added:
235 - a new node parameter ``spindle_count``, defaults to 1, customisable at
236 node group or node level
237 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
239 Note that spindles in this context doesn't need to mean actual
240 mechanical hard-drives; it's just a relative number for both the node
241 I/O capacity and instance I/O consumption.
243 Instance migration behaviour
244 ++++++++++++++++++++++++++++
246 While live-migration is in general desirable over failover, it is
247 possible that for some workloads it is actually worse, due to the
248 variable time of the “suspend” phase during live migration.
250 To allow the tools to work consistently over such instances (without
251 having to hard-code instance names), a new backend parameter
252 ``always_failover`` has been added to control the migration/failover
253 behaviour. When set to True, all migration requests for an instance will
254 instead fall-back to failover.
256 Instance memory ballooning
257 ++++++++++++++++++++++++++
259 Initial support for memory ballooning has been added. The memory for an
260 instance is no longer fixed (backend parameter ``memory``), but instead
261 can vary between minimum and maximum values (backend parameters
262 ``minmem`` and ``maxmem``). Currently we only change an instance's
265 - live migrating or failing over and instance and the target node
266 doesn't have enough memory
267 - user requests changing the memory via ``gnt-instance modify
273 In order to control the use of specific CPUs by instance, support for
274 controlling CPU pinning has been added for the Xen, HVM and LXC
275 hypervisors. This is controlled by a new hypervisor parameter
276 ``cpu_mask``; details about possible values for this are in the
277 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
278 VCPU-to-CPU mapping) form will work well only when all nodes in your
279 cluster have the same amount of CPUs.
284 Another area in which Ganeti was not customisable were the parameters
285 used for storage configuration, e.g. how many stripes to use for LVM,
286 DRBD resync configuration, etc.
288 To improve this area, we've added disks parameters, which are
289 customisable at cluster and node group level, and which allow to
290 specify various parameters for disks (DRBD has the most parameters
291 currently), for example:
293 - DRBD resync algorithm and parameters (e.g. speed)
294 - the default VG for meta-data volumes for DRBD
295 - number of stripes for LVM (plain disk template)
298 These parameters can be modified via ``gnt-cluster modify -D …`` and
299 ``gnt-group modify -D …``, and are used at either instance creation (in
300 case of LVM stripes, for example) or at disk “activation” time
303 Rados block device support
304 ++++++++++++++++++++++++++
306 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
307 denoted by the ``rbd`` disk template type. This is considered
308 experimental, feedback is welcome. For details on configuring it, see
309 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
314 The existing master IP functionality works well only in simple setups (a
315 single network shared by all nodes); however, if nodes belong to
316 different networks, then the ``/32`` setup and lack of routing
317 information is not enough.
319 To allow the master IP to function well in more complex cases, the
320 system was reworked as follows:
322 - a master IP netmask setting has been added
323 - the master IP activation/turn-down code was moved from the node daemon
325 - whether to run the Ganeti-supplied master IP script or a user-supplied
326 on is a ``gnt-cluster init`` setting
328 Details about the location of the standard and custom setup scripts are
329 in the man page :manpage:`gnt-cluster(8)`; for information about the
330 setup script protocol, look at the Ganeti-supplied script.
335 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
338 It is now possible to use TLS-protected connections, and when renewing
339 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
340 it is now possible to specify spice or spice CA certificates. Also, it
341 is possible to configure a password for SPICE sessions via the
342 hypervisor parameter ``spice_password_file``.
344 There are also new parameters to control the compression and streaming
345 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
346 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
347 for the spice parameters.
349 Lastly, it is now possible to see the SPICE connection information via
350 ``gnt-instance console``.
355 A new tool (``tools/ovfconverter``) has been added that supports
356 conversion between Ganeti and the `Open Virtualization Format
357 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
360 This relies on the ``qemu-img`` tool to convert the disk formats, so the
361 actual compatibility with other virtualization solutions depends on it.
366 The configuration query daemon (``ganeti-confd``) is now optional, and
367 has been rewritten in Haskell; whether to use the daemon at all, use the
368 Python (default) or the Haskell version is selectable at configure time
369 via the ``--enable-confd`` parameter, which can take one of the
370 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
371 daemon will result in a smaller footprint; for larger systems, we
372 welcome feedback on the Haskell version which might become the default
375 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
376 daemon running. The Python version doesn't implement the new call.
379 User interface changes
380 ~~~~~~~~~~~~~~~~~~~~~~
382 We have replaced the ``--disks`` option of ``gnt-instance
383 replace-disks`` with a more flexible ``--disk`` option, which allows
384 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
385 disk size and mode can be changed upon recreation (via ``gnt-instance
386 recreate-disks``, which accepts the same ``--disk`` option).
388 As many people are used to a ``show`` command, we have added that as an
389 alias to ``info`` on all ``gnt-*`` commands.
391 The ``gnt-instance grow-disk`` command has a new mode in which it can
392 accept the target size of the disk, instead of the delta; this can be
393 more safe since two runs in absolute mode will be idempotent, and
394 sometimes it's also easier to specify the desired size directly.
396 Also the handling of instances with regard to offline secondaries has
397 been improved. Instance operations should not fail because one of it's
398 secondary nodes is offline, even though it's safe to proceed.
400 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
401 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
402 minors to instance name.
407 RAPI coverage has improved, with (for example) new resources for
408 recreate-disks, node power-cycle, etc.
413 There is partial support for ``xl`` in the Xen hypervisor; feedback is
416 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
417 whether to still support Python 2.4 or move to Python 2.6 as minimum
420 Support for Fedora has been slightly improved; the provided example
421 init.d script should work better on it and the INSTALL file should
422 document the needed dependencies.
427 The deprecated ``QueryLocks`` LUXI request has been removed. Use
428 ``Query(what=QR_LOCK, ...)`` instead.
430 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
431 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
432 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
433 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
434 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
436 RAPI client: ``CertificateError`` now derives from
437 ``GanetiApiError``. This should make it more easy to handle Ganeti
440 Deprecation warnings due to PyCrypto/paramiko import in
441 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
442 make sure to run an up-to-date paramiko version, if you use this tool.
444 The QA scripts now depend on Python 2.5 or above (the main code base
445 still works with Python 2.4).
447 The configuration file (``config.data``) is now written without
448 indentation for performance reasons; if you want to edit it, it can be
449 re-formatted via ``tools/fmtjson``.
451 A number of bugs has been fixed in the cluster merge tool.
453 ``x509`` certification verification (used in import-export) has been
454 changed to allow the same clock skew as permitted by the cluster
455 verification. This will remove some rare but hard to diagnose errors in
462 *(Released Thu, 19 Jul 2012)*
464 Very few changes from rc4 to the final release, only bugfixes:
466 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
467 instance, fix CDROM booting for KVM instances)
468 - fixed node group modification of node parameters
469 - fixed issue in LUClusterVerifyGroup with multi-group clusters
470 - fixed generation of bash completion to ensure a stable ordering
477 *(Released Fri, 13 Jul 2012)*
479 Third release candidate for 2.6. The following changes were done from
482 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
483 - Fixed an inconsistency in the LUXI protocol with the provided
484 arguments (NOT backwards compatible)
485 - Fixed a bug with node groups ipolicy where ``min`` was greater than
486 the cluster ``std`` value
487 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
488 easier instance debugging on nodes (requires ``hconfd`` to work)
494 *(Released Tue, 03 Jul 2012)*
496 Second release candidate for 2.6. The following changes were done from
499 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
501 - Fixed a RAPI regression on missing beparams/memory
502 - Fixed redistribution of files on offline nodes
503 - Added possibility to run activate-disks even though secondaries are
504 offline. With this change it relaxes also the strictness on some other
505 commands which use activate disks internally:
506 * ``gnt-instance start|reboot|rename|backup|export``
507 - Made it possible to remove safely an instance if its secondaries are
509 - Made it possible to reinstall even though secondaries are offline
515 *(Released Mon, 25 Jun 2012)*
517 First release candidate for 2.6. The following changes were done from
520 - Fixed bugs with disk parameters and ``rbd`` templates as well as
522 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
523 behaviour. It supports now the modify operation
524 - ``hcheck`` implemented to analyze cluster health and possibility of
525 improving health by rebalance
526 - ``hbal`` has been improved in dealing with split instances
532 *(Released Mon, 11 Jun 2012)*
534 Second beta release of 2.6. The following changes were done from beta2
537 - Fixed ``daemon-util`` with non-root user models
538 - Fixed creation of plain instances with ``--no-wait-for-sync``
539 - Fix wrong iv_names when running ``cfgupgrade``
540 - Export more information in RAPI group queries
541 - Fixed bug when changing instance network interfaces
542 - Extended burnin to do NIC changes
543 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
544 - Changed default for DRBD barriers
545 - Fixed DRBD error reporting for syncer rate
546 - Verify the options on disk parameters
548 And of course various fixes to documentation and improved unittests and
555 *(Released Wed, 23 May 2012)*
557 First beta release of 2.6. The following changes were done from beta1 to
560 - integrated patch for distributions without ``start-stop-daemon``
561 - adapted example init.d script to work on Fedora
562 - fixed log handling in Haskell daemons
563 - adapted checks in the watcher for pycurl linked against libnss
564 - add partial support for ``xl`` instead of ``xm`` for Xen
565 - fixed a type issue in cluster verification
566 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
569 Plus integrated fixes from the 2.5 branch:
571 - fixed ``kvm-ifup`` to use ``/bin/bash``
572 - fixed parallel build failures
573 - KVM live migration when using a custom keymap
579 *(Released Tue, 24 Jul 2012)*
581 A small bugfix release, with no new features:
583 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
584 different default shell (e.g. Debian, Ubuntu)
585 - fixed KVM startup and live migration with a custom keymap (fixes Issue
586 243 and Debian bug #650664)
587 - fixed compatibility with KVM versions that don't support multiple boot
588 devices (fixes Issue 230 and Debian bug #624256)
590 Additionally, a few fixes were done to the build system (fixed parallel
591 build failures) and to the unittests (fixed race condition in test for
592 FileID functions, and the default enable/disable mode for QA test is now
599 *(Released Fri, 11 May 2012)*
601 A small bugfix release.
603 The main issues solved are on the topic of compatibility with newer LVM
606 - fixed parsing of ``lv_attr`` field
607 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
608 the ``--force`` flag is needed
610 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
611 to accept kernel 3.x too (was hardcoded to 2.6.*).
613 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
614 of hook results and that also made display of other errors suboptimal
615 was fixed; the code behaves now like 2.4 and earlier.
617 Another change in 2.5, the cleanup of the OS scripts environment, is too
618 aggressive: it removed even the ``PATH`` variable, which requires the OS
619 scripts to *always* need to export it. Since this is a bit too strict,
620 we now export a minimal PATH, the same that we export for hooks.
622 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
623 integrated into this release.
625 Finally, a few other miscellaneous changes were done (no new features,
626 just small improvements):
628 - Fix ``gnt-group --help`` display
629 - Fix hardcoded Xen kernel path
630 - Fix grow-disk handling of invalid units
631 - Update synopsis for ``gnt-cluster repair-disk-sizes``
632 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
638 *(Released Thu, 12 Apr 2012)*
640 Incompatible/important changes and bugfixes
641 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
643 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
644 resource's ``ip_check`` parameter changed from ``True`` to ``False``
645 to match the underlying LUXI interface.
646 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
647 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
648 not maintain backwards-compatibility as the underlying operation
649 changed in an incompatible way. The RAPI client can talk to old
650 servers, but it needs to be told so as the return value changed.
651 - When creating file-based instances via RAPI, the ``file_driver``
652 parameter no longer defaults to ``loop`` and must be specified.
653 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
655 - Support for the undocumented and deprecated RAPI instance creation
656 request format version 0 has been dropped. Use version 1, supported
657 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
658 - Pyparsing 1.4.6 or above is required, see :doc:`installation
659 documentation <install>`.
660 - The "cluster-verify" hooks are now executed per group by the
661 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
662 if you just run ``gnt-cluster verify``, which generates one opcode per
664 - The environment as passed to the OS scripts is cleared, and thus no
665 environment variables defined in the node daemon's environment will be
666 inherited by the scripts.
667 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
669 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
670 support operations involving multiple node groups.
671 - Offline nodes are ignored when failing over an instance.
672 - Support for KVM version 1.0, which changed the version reporting format
674 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
676 - ``Makefile`` is now compatible with Automake 1.11.2
677 - Includes all bugfixes made in the 2.4 series
682 - The ganeti-htools project has been merged into the ganeti-core source
683 tree and will be built as part of Ganeti (see :doc:`install-quick`).
684 - Implemented support for :doc:`shared storage <design-shared-storage>`.
685 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
686 GPT-style partition tables (requires `parted
687 <http://www.gnu.org/s/parted/>`_).
688 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
689 - Allowed adding tags on instance creation.
690 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
691 - Allow instances to be started in a paused state, enabling the user to
692 see the complete console output on boot using the console.
693 - Added new hypervisor flag to control default reboot behaviour
694 (``reboot_behavior``).
695 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
696 - Improved out-of-band management support:
698 - Added ``gnt-node health`` command reporting the health status of
700 - Added ``gnt-node power`` command to manage power status of nodes.
701 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
703 - Instance migration can fall back to failover if instance is not
705 - Filters can be used when listing nodes, instances, groups and locks;
706 see *ganeti(7)* manpage.
707 - Added post-execution status as variables to :doc:`hooks <hooks>`
709 - Instance tags are exported/imported together with the instance.
710 - When given an explicit job ID, ``gnt-job info`` will work for archived
712 - Jobs can define dependencies on other jobs (not yet supported via
713 RAPI or command line, but used by internal commands and usable via
716 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
719 - Instance failover is now available as a RAPI resource
720 (``/2/instances/[instance_name]/failover``).
721 - ``gnt-instance info`` defaults to static information if primary node
723 - Opcodes have a new ``comment`` attribute.
724 - Added basic SPICE support to KVM hypervisor.
725 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
727 Node group improvements
728 ~~~~~~~~~~~~~~~~~~~~~~~
730 - ``gnt-cluster verify`` has been modified to check groups separately,
731 thereby improving performance.
732 - Node group support has been added to ``gnt-cluster verify-disks``,
733 which now operates per node group.
734 - Watcher has been changed to work better with node groups.
736 - One process and state file per node group.
737 - Slow watcher in one group doesn't block other group's watcher.
739 - Added new command, ``gnt-group evacuate``, to move all instances in a
740 node group to other groups.
741 - Added ``gnt-instance change-group`` to move an instance to another
743 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
744 per-group operations.
745 - Node groups can be tagged.
746 - Some operations switch from an exclusive to a shared lock as soon as
748 - Instance's primary and secondary nodes' groups are now available as
749 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
750 and ``snodes.group.uuid``).
755 - Numerous updates to documentation and manpages.
757 - :doc:`RAPI <rapi>` documentation now has detailed parameter
759 - Some opcode/job results are now also documented, see :doc:`RAPI
762 - A lockset's internal lock is now also visible in lock monitor.
763 - Log messages from job queue workers now contain information about the
764 opcode they're processing.
765 - ``gnt-instance console`` no longer requires the instance lock.
766 - A short delay when waiting for job changes reduces the number of LUXI
767 requests significantly.
768 - DRBD metadata volumes are overwritten with zeros during disk creation.
769 - Out-of-band commands no longer acquire the cluster lock in exclusive
771 - ``devel/upload`` now uses correct permissions for directories.
777 *(Released Fri, 23 Mar 2012)*
779 This was the sixth release candidate of the 2.5 series.
785 *(Released Mon, 9 Jan 2012)*
787 This was the fifth release candidate of the 2.5 series.
793 *(Released Thu, 27 Oct 2011)*
795 This was the fourth release candidate of the 2.5 series.
801 *(Released Wed, 26 Oct 2011)*
803 This was the third release candidate of the 2.5 series.
809 *(Released Tue, 18 Oct 2011)*
811 This was the second release candidate of the 2.5 series.
817 *(Released Tue, 4 Oct 2011)*
819 This was the first release candidate of the 2.5 series.
825 *(Released Wed, 31 Aug 2011)*
827 This was the third beta release of the 2.5 series.
833 *(Released Mon, 22 Aug 2011)*
835 This was the second beta release of the 2.5 series.
841 *(Released Fri, 12 Aug 2011)*
843 This was the first beta release of the 2.5 series.
849 *(Released Thu, 27 Oct 2011)*
851 - Fixed bug when parsing command line parameter values ending in
853 - Fixed assertion error after unclean master shutdown
854 - Disable HTTP client pool for RPC, significantly reducing memory usage
856 - Fixed queue archive creation with wrong permissions
862 *(Released Tue, 23 Aug 2011)*
866 - Fixed documentation for importing with ``--src-dir`` option
867 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
868 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
874 *(Released Fri, 5 Aug 2011)*
876 Many bug-fixes and a few small features:
878 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
879 issues when you tried to add an instance with two MAC addresses in one
881 - KVM: fixed per-instance stored UID value
882 - KVM: configure bridged NICs at migration start
883 - KVM: Fix a bug where instance will not start with never KVM versions
885 - Added OS search path to ``gnt-cluster info``
886 - Fixed an issue with ``file_storage_dir`` where you were forced to
887 provide an absolute path, but the documentation states it is a
888 relative path, the documentation was right
889 - Added a new parameter to instance stop/start called ``--no-remember``
890 that will make the state change to not be remembered
891 - Implemented ``no_remember`` at RAPI level
892 - Improved the documentation
893 - Node evacuation: don't call IAllocator if node is already empty
894 - Fixed bug in DRBD8 replace disks on current nodes
895 - Fixed bug in recreate-disks for DRBD instances
896 - Moved assertion checking locks in ``gnt-instance replace-disks``
897 causing it to abort with not owning the right locks for some situation
898 - Job queue: Fixed potential race condition when cancelling queued jobs
899 - Fixed off-by-one bug in job serial generation
900 - ``gnt-node volumes``: Fix instance names
901 - Fixed aliases in bash completion
902 - Fixed a bug in reopening log files after being sent a SIGHUP
903 - Added a flag to burnin to allow specifying VCPU count
904 - Bugfixes to non-root Ganeti configuration
910 *(Released Thu, 12 May 2011)*
912 Many bug-fixes and a few new small features:
914 - Fixed a bug related to log opening failures
915 - Fixed a bug in instance listing with orphan instances
916 - Fixed a bug which prevented resetting the cluster-level node parameter
917 ``oob_program`` to the default
918 - Many fixes related to the ``cluster-merge`` tool
919 - Fixed a race condition in the lock monitor, which caused failures
920 during (at least) creation of many instances in parallel
921 - Improved output for gnt-job info
922 - Removed the quiet flag on some ssh calls which prevented debugging
924 - Improved the N+1 failure messages in cluster verify by actually
925 showing the memory values (needed and available)
926 - Increased lock attempt timeouts so that when executing long operations
927 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
928 too early and thus prevent the use of the 'fair' mechanism
929 - Changed instance query data (``gnt-instance info``) to not acquire
930 locks unless needed, thus allowing its use on locked instance if only
931 static information is asked for
932 - Improved behaviour with filesystems that do not support rename on an
934 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
935 kept locks on all nodes during the wipe, which is unneeded
936 - Fixed ``gnt-watcher`` handling of errors during hooks execution
937 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
938 10GiB) which caused the wipe to fail right at the end in some cases
939 - Fixed master IP activation when doing master failover with no-voting
940 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
941 the master node itself
942 - Fixed potential data-loss in under disk full conditions, where Ganeti
943 wouldn't check correctly the return code and would consider
944 partially-written files 'correct'
945 - Fixed bug related to multiple VGs and DRBD disk replacing
946 - Added new disk parameter ``metavg`` that allows placement of the meta
947 device for DRBD in a different volume group
948 - Fixed error handling in the node daemon when the system libc doesn't
949 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
950 - Fixed lock release during replace-disks, which kept cluster-wide locks
951 when doing disk replaces with an iallocator script
952 - Added check for missing bridges in cluster verify
953 - Handle EPIPE errors while writing to the terminal better, so that
954 piping the output to e.g. ``less`` doesn't cause a backtrace
955 - Fixed rare case where a ^C during Luxi calls could have been
956 interpreted as server errors, instead of simply terminating
957 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
959 - Added a few more parameters to the KVM hypervisor, allowing a second
960 CDROM, custom disk type for CDROMs and a floppy image
961 - Removed redundant message in instance rename when the name is given
963 - Added option to ``gnt-instance recreate-disks`` to allow creating the
964 disks on new nodes, allowing recreation when the original instance
965 nodes are completely gone
966 - Added option when converting disk templates to DRBD to skip waiting
967 for the resync, in order to make the instance available sooner
968 - Added two new variables to the OS scripts environment (containing the
970 - Made the root_path and optional parameter for the xen-pvm hypervisor,
971 to allow use of ``pvgrub`` as bootloader
972 - Changed the instance memory modifications to only check out-of-memory
973 conditions on memory increases, and turned the secondary node warnings
974 into errors (they can still be overridden via ``--force``)
975 - Fixed the handling of a corner case when the Python installation gets
976 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
977 to execute a command that doesn't exist
978 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
979 node of the instance returned failures during instance shutdown; this
980 adds the option ``--ignore-consistency`` to gnt-instance move
982 And as usual, various improvements to the error messages, documentation
989 *(Released Wed, 09 Mar 2011)*
991 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
992 the RAPI users file if run twice (even with ``--dry-run``).
994 The release fixes that bug (nothing else changed).
1000 *(Released Mon, 07 Mar 2011)*
1002 Final 2.4.0 release. Just a few small fixes:
1004 - Fixed RAPI node evacuate
1005 - Fixed the kvm-ifup script
1006 - Fixed internal error handling for special job cases
1007 - Updated man page to specify the escaping feature for options
1013 *(Released Mon, 28 Feb 2011)*
1015 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1016 that this feature wiped the disks of the wrong instance, leading to loss
1021 - Fixed title of query field containing instance name
1022 - Expanded the glossary in the documentation
1023 - Fixed one unittest (internal issue)
1029 *(Released Mon, 21 Feb 2011)*
1031 A number of bug fixes plus just a couple functionality changes.
1033 On the user-visible side, the ``gnt-* list`` command output has changed
1034 with respect to "special" field states. The current rc1 style of display
1035 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1036 the default output mode special fields states are displayed as follows:
1038 - Offline resource: ``*``
1039 - Unavailable/not applicable: ``-``
1040 - Data missing (RPC failure): ``?``
1041 - Unknown field: ``??``
1043 Another user-visible change is the addition of ``--force-join`` to
1048 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1049 - Fixed regression in RAPI/instance reinstall where all parameters were
1050 required (instead of optional)
1051 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1052 - Fixed iallocator usage (offline nodes were not considered offline)
1053 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1054 - Fixed hypervisor and OS parameter validation with respect to
1055 non-vm_capable nodes
1056 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1058 - Fixed ``tools/listrunner`` with respect to agent-based usage
1064 *(Released Fri, 4 Feb 2011)*
1066 Many changes and fixes since the beta1 release. While there were some
1067 internal changes, the code has been mostly stabilised for the RC
1070 Note: the dumb allocator was removed in this release, as it was not kept
1071 up-to-date with the IAllocator protocol changes. It is recommended to
1072 use the ``hail`` command from the ganeti-htools package.
1074 Note: the 2.4 and up versions of Ganeti are not compatible with the
1075 0.2.x branch of ganeti-htools. You need to upgrade to
1076 ganeti-htools-0.3.0 (or later).
1078 Regressions fixed from 2.3
1079 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1081 - Fixed the ``gnt-cluster verify-disks`` command
1082 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1084 - Fixed disk adoption breakage
1085 - Fixed wrong headers in instance listing for field aliases
1090 - Fixed corner case in KVM handling of NICs
1091 - Fixed many cases of wrong handling of non-vm_capable nodes
1092 - Fixed a bug where a missing instance symlink was not possible to
1093 recreate with any ``gnt-*`` command (now ``gnt-instance
1094 activate-disks`` does it)
1095 - Fixed the volume group name as reported by ``gnt-cluster
1097 - Increased timeouts for the import-export code, hopefully leading to
1098 fewer aborts due network or instance timeouts
1099 - Fixed bug in ``gnt-node list-storage``
1100 - Fixed bug where not all daemons were started on cluster
1101 initialisation, but only at the first watcher run
1102 - Fixed many bugs in the OOB implementation
1103 - Fixed watcher behaviour in presence of instances with offline
1105 - Fixed instance list output for instances running on the wrong node
1106 - a few fixes to the cluster-merge tool, but it still cannot merge
1107 multi-node groups (currently it is not recommended to use this tool)
1113 - Improved network configuration for the KVM hypervisor
1114 - Added e1000 as a supported NIC for Xen-HVM
1115 - Improved the lvmstrap tool to also be able to use partitions, as
1116 opposed to full disks
1117 - Improved speed of disk wiping (the cluster parameter
1118 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1119 of instance creations
1120 - Added documentation for the OS parameters
1121 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1122 hypervisor is not responding
1123 - Added display of blacklisted and hidden OS information in
1124 ``gnt-cluster info``
1125 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1126 NIC and node parameters, which might create problems with currently
1127 invalid (but undetected) configuration files, but prevents validation
1128 failures when unrelated parameters are modified
1129 - Changed cluster initialisation to wait for the master daemon to become
1131 - Expanded the RAPI interface:
1133 - Added config redistribution resource
1134 - Added activation/deactivation of instance disks
1135 - Added export of console information
1137 - Implemented log file reopening on SIGHUP, which allows using
1138 logrotate(8) for the Ganeti log files
1139 - Added a basic OOB helper script as an example
1145 *(Released Fri, 14 Jan 2011)*
1150 - Fixed timezone issues when formatting timestamps
1151 - Added support for node groups, available via ``gnt-group`` and other
1153 - Added out-of-band framework and management, see :doc:`design
1154 document <design-oob>`
1155 - Removed support for roman numbers from ``gnt-node list`` and
1156 ``gnt-instance list``.
1157 - Allowed modification of master network interface via ``gnt-cluster
1158 modify --master-netdev``
1159 - Accept offline secondaries while shutting down instance disks
1160 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1161 - Added support for multiple LVM volume groups
1162 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1164 - Added commands to list available fields:
1166 - ``gnt-node list-fields``
1167 - ``gnt-group list-fields``
1168 - ``gnt-instance list-fields``
1170 - Updated documentation and man pages
1175 - Moved ``rapi_users`` file into separate directory, now named
1176 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1178 - Added new tool for running commands on many machines,
1179 ``tools/ganeti-listrunner``
1180 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1181 improving the ``gnt-instance console`` output
1182 - Allowed customisation of disk index separator at ``configure`` time
1183 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1184 - Added support for non-partitioned md disks in ``lvmstrap``
1185 - Added script to gracefully power off KVM instances
1186 - Split ``utils`` module into smaller parts
1187 - Changed query operations to return more detailed information, e.g.
1188 whether an information is unavailable due to an offline node. To use
1189 this new functionality, the LUXI call ``Query`` must be used. Field
1190 information is now stored by the master daemon and can be retrieved
1191 using ``QueryFields``. Instances, nodes and groups can also be queried
1192 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1193 exposed via RAPI). The following commands make use of this
1194 infrastructure change:
1196 - ``gnt-group list``
1197 - ``gnt-group list-fields``
1199 - ``gnt-node list-fields``
1200 - ``gnt-instance list``
1201 - ``gnt-instance list-fields``
1202 - ``gnt-debug locks``
1207 - New RAPI resources (see :doc:`rapi`):
1211 - ``/2/groups/[group_name]``
1212 - ``/2/groups/[group_name]/assign-nodes``
1213 - ``/2/groups/[group_name]/modify``
1214 - ``/2/groups/[group_name]/rename``
1215 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1219 - Implemented ``no_install`` for instance creation
1220 - Implemented OS parameters for instance reinstallation, allowing
1221 use of special settings on reinstallation (e.g. for preserving data)
1226 - Added IPv6 support in import/export
1227 - Pause DRBD synchronization while wiping disks on instance creation
1228 - Updated unittests and QA scripts
1229 - Improved network parameters passed to KVM
1230 - Converted man pages from docbook to reStructuredText
1236 *(Released Mon, 20 Dec 2010)*
1238 Released version 2.3.1~rc1 without any changes.
1244 *(Released Wed, 1 Dec 2010)*
1246 - impexpd: Disable OpenSSL compression in socat if possible (backport
1247 from master, commit e90739d625b, see :doc:`installation guide
1248 <install-quick>` for details)
1249 - Changed unittest coverage report to exclude test scripts
1250 - Added script to check version format
1256 *(Released Wed, 1 Dec 2010)*
1258 Released version 2.3.0~rc1 without any changes.
1264 *(Released Fri, 19 Nov 2010)*
1266 A number of bugfixes and documentation updates:
1268 - Update ganeti-os-interface documentation
1269 - Fixed a bug related to duplicate MACs or similar items which should be
1271 - Fix breakage in OS state modify
1272 - Reinstall instance: disallow offline secondaries (fixes bug related to
1273 OS changing but reinstall failing)
1274 - plus all the other fixes between 2.2.1 and 2.2.2
1280 *(Released Tue, 2 Nov 2010)*
1282 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1283 - Fixed master failover race with watcher
1284 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1286 - Accept previously stopped instance for export with instance removal
1287 - Simplify and extend the environment variables for instance OS scripts
1288 - Added new node flags, ``master_capable`` and ``vm_capable``
1289 - Added optional instance disk wiping prior during allocation. This is a
1290 cluster-wide option and can be set/modified using
1291 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1292 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1293 :doc:`install-quick`
1294 - Added a new watcher option (``--ignore-pause``)
1295 - Added option to ignore offline node on instance start/stop
1296 (``--ignore-offline``)
1297 - Allow overriding OS parameters with ``gnt-instance reinstall``
1298 - Added ability to change node's secondary IP address using ``gnt-node
1300 - Implemented privilege separation for all daemons except
1301 ``ganeti-noded``, see ``configure`` options
1302 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1304 - Implemented job priorities (see ``ganeti(7)`` manpage)
1305 - Ignore failures while shutting down instances during failover from
1307 - Exit daemon's bootstrap process only once daemon is ready
1308 - Export more information via ``LUInstanceQuery``/remote API
1309 - Improved documentation, QA and unittests
1310 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1311 restart if the file was created or changed
1312 - Added LUXI protocol version sent with each request and response,
1313 allowing detection of server/client mismatches
1314 - Moved the Python scripts among gnt-* and ganeti-* into modules
1315 - Moved all code related to setting up SSH to an external script,
1317 - Infrastructure changes for node group support in future versions
1323 *(Released Fri, 19 Nov 2010)*
1325 A few small bugs fixed, and some improvements to the build system:
1327 - Fix documentation regarding conversion to drbd
1328 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1330 - Fix error handling in node modify with multiple changes
1331 - Allow remote imports without checked names
1337 *(Released Tue, 19 Oct 2010)*
1339 - Disable SSL session ID cache in RPC client
1345 *(Released Thu, 14 Oct 2010)*
1347 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1348 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1350 - Fix problems with interaction between Curl and Python's HTTP server,
1351 resulting in increased speed in many RPC calls
1352 - Improve our release script to prevent breakage with older aclocal and
1359 *(Released Thu, 7 Oct 2010)*
1361 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1362 value retrieved from master
1363 - Added support for blacklisted or hidden OS definitions
1364 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1365 - Added support for -mem-path in KVM hypervisor abstraction layer
1366 - Allow overriding instance parameters in tool for inter-cluster
1367 instance moves (``tools/move-instance``)
1368 - Improved opcode summaries (e.g. in ``gnt-job list``)
1369 - Improve consistency of OS listing by sorting it
1370 - Documentation updates
1376 *(Released Fri, 8 Oct 2010)*
1378 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1384 *(Released Mon, 4 Oct 2010)*
1386 - Fixed regression in ``gnt-instance rename``
1392 *(Released Wed, 22 Sep 2010)*
1394 - Fixed OS_VARIANT variable for OS scripts
1395 - Fixed cluster tag operations via RAPI
1396 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1397 - Disabled RAPI CA checks in watcher
1403 *(Released Mon, 23 Aug 2010)*
1405 - Support DRBD versions of the format "a.b.c.d"
1407 - Re-introduce support for usage from multiple threads in RAPI client
1408 - Instance renames and modify via RAPI
1409 - Work around race condition between processing and archival in job
1411 - Mark opcodes following failed one as failed, too
1412 - Job field ``lock_status`` was removed due to difficulties making it
1413 work with the changed job queue in Ganeti 2.2; a better way to monitor
1414 locks is expected for a later 2.2.x release
1415 - Fixed dry-run behaviour with many commands
1416 - Support ``ssh-agent`` again when adding nodes
1417 - Many additional bugfixes
1423 *(Released Fri, 30 Jul 2010)*
1425 Important change: the internal RPC mechanism between Ganeti nodes has
1426 changed from using a home-grown http library (based on the Python base
1427 libraries) to use the PycURL library. This requires that PycURL is
1428 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1429 against GnuTLS by default. cURL's support for GnuTLS had known issues
1430 before cURL 7.21.0 and we recommend using the latest cURL release or
1431 linking against OpenSSL. Most other distributions already link PycURL
1432 and cURL against OpenSSL. The command::
1434 python -c 'import pycurl; print pycurl.version'
1436 can be used to determine the libraries PycURL and cURL are linked
1439 Other significant changes:
1441 - Rewrote much of the internals of the job queue, in order to achieve
1442 better parallelism; this decouples job query operations from the job
1443 processing, and it should allow much nicer behaviour of the master
1444 daemon under load, and it also has uncovered some long-standing bugs
1445 related to the job serialisation (now fixed)
1446 - Added a default iallocator setting to the cluster parameters,
1447 eliminating the need to always pass nodes or an iallocator for
1448 operations that require selection of new node(s)
1449 - Added experimental support for the LXC virtualization method
1450 - Added support for OS parameters, which allows the installation of
1451 instances to pass parameter to OS scripts in order to customise the
1453 - Added a hypervisor parameter controlling the migration type (live or
1454 non-live), since hypervisors have various levels of reliability; this
1455 has renamed the 'live' parameter to 'mode'
1456 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1457 logical volumes, meaning that cluster verify will ignore them and not
1458 flag their presence as errors
1459 - The watcher will now reset the error count for failed instances after
1460 8 hours, thus allowing self-healing if the problem that caused the
1461 instances to be down/fail to start has cleared in the meantime
1462 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1463 check for, and warn, if the drbd module parameter ``usermode_helper``
1464 is not consistent with the cluster-wide setting; this is needed to
1465 make diagnose easier of failed drbd creations
1466 - Started adding base IPv6 support, but this is not yet
1467 enabled/available for use
1468 - Rename operations (cluster, instance) will now return the new name,
1469 which is especially useful if a short name was passed in
1470 - Added support for instance migration in RAPI
1471 - Added a tool to pre-configure nodes for the SSH setup, before joining
1472 them to the cluster; this will allow in the future a simplified model
1473 for node joining (but not yet fully enabled in 2.2); this needs the
1474 paramiko python library
1475 - Fixed handling of name-resolving errors
1476 - Fixed consistency of job results on the error path
1477 - Fixed master-failover race condition when executed multiple times in
1479 - Fixed many bugs related to the job queue (mostly introduced during the
1480 2.2 development cycle, so not all are impacting 2.1)
1481 - Fixed instance migration with missing disk symlinks
1482 - Fixed handling of unknown jobs in ``gnt-job archive``
1483 - And many other small fixes/improvements
1487 - Enhanced both the unittest and the QA coverage
1488 - Switched the opcode validation to a generic model, and extended the
1489 validation to all opcode parameters
1490 - Changed more parts of the code that write shell scripts to use the
1492 - Switched the master daemon to use the asyncore library for the Luxi
1499 *(Released Thu, 17 Jun 2010)*
1501 - Added tool (``move-instance``) and infrastructure to move instances
1502 between separate clusters (see :doc:`separate documentation
1503 <move-instance>` and :doc:`design document <design-2.2>`)
1504 - Added per-request RPC timeout
1505 - RAPI now requires a Content-Type header for requests with a body (e.g.
1506 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1507 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1508 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1510 - Implemented initial support for running Ganeti daemons as separate
1511 users, see configure-time flags ``--with-user-prefix`` and
1512 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1514 - Instances can be removed after export (``gnt-backup export
1515 --remove-instance``)
1516 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1517 key (instead of 1024 bit)
1518 - Added new cluster configuration file for cluster domain secret
1519 - Import/export now use SSL instead of SSH
1520 - Added support for showing estimated time when exporting an instance,
1521 see the ``ganeti-os-interface(7)`` manpage and look for
1528 *(Released Tue, 16 Nov 2010)*
1530 Some more bugfixes. Unless critical bugs occur, this will be the last
1533 - Fix case of MAC special-values
1534 - Fix mac checker regex
1535 - backend: Fix typo causing "out of range" error
1536 - Add missing --units in gnt-instance list man page
1542 *(Released Tue, 24 Aug 2010)*
1545 - Don't ignore secondary node silently on non-mirrored disk templates
1547 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1548 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1549 - Properly document the usb_mouse parameter
1550 - Fix path in ganeti-rapi(8) (issue 116)
1551 - Adjust error message when the ganeti user's .ssh directory is
1553 - Add same-node-check when changing the disk template to drbd
1559 *(Released Fri, 16 Jul 2010)*
1562 - Add an option to only select some reboot types during qa/burnin.
1563 (on some hypervisors consequent reboots are not supported)
1564 - Fix infrequent race condition in master failover. Sometimes the old
1565 master ip address would be still detected as up for a short time
1566 after it was removed, causing failover to fail.
1567 - Decrease mlockall warnings when the ctypes module is missing. On
1568 Python 2.4 we support running even if no ctypes module is installed,
1569 but we were too verbose about this issue.
1570 - Fix building on old distributions, on which man doesn't have a
1572 - Fix RAPI not to ignore the MAC address on instance creation
1573 - Implement the old instance creation format in the RAPI client.
1579 *(Released Thu, 01 Jul 2010)*
1581 A small bugfix release:
1582 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1583 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1584 a non-existing option
1585 - Fix instance create: the --force-variant option was ignored
1586 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1587 - Fix modify node storage with non-FQDN arguments
1588 - Fix RAPI client to authenticate under Python 2.6 when used
1589 for more than 5 requests needing authentication
1590 - Fix gnt-instance modify -t (storage) giving a wrong error message
1591 when converting a non-shutdown drbd instance to plain
1597 *(Released Fri, 18 Jun 2010)*
1599 A small bugfix release:
1601 - Fix live migration of KVM instances started with older Ganeti
1602 versions which had fewer hypervisor parameters
1603 - Fix gnt-instance grow-disk on down instances
1604 - Fix an error-reporting bug during instance migration
1605 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1606 silently ignoring broken ones
1607 - Fix an RPC error reporting bug affecting, for example, RAPI client
1609 - Fix bug triggered by different API version os-es on different nodes
1610 - Fix a bug in instance startup with custom hvparams: OS level
1611 parameters would fail to be applied.
1612 - Fix the RAPI client under Python 2.6 (but more work is needed to
1613 make it work completely well with OpenSSL)
1614 - Fix handling of errors when resolving names from DNS
1620 *(Released Thu, 3 Jun 2010)*
1622 A medium sized development cycle. Some new features, and some
1623 fixes/small improvements/cleanups.
1625 Significant features
1626 ~~~~~~~~~~~~~~~~~~~~
1628 The node deamon now tries to mlock itself into memory, unless the
1629 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1630 its logs, and falls back to console logging. This allows emergency
1631 features such as ``gnt-node powercycle`` to work even in the event of a
1632 broken node disk (tested offlining the disk hosting the node's
1633 filesystem and dropping its memory caches; don't try this at home)
1635 KVM: add vhost-net acceleration support. It can be tested with a new
1636 enough version of the kernel and of qemu-kvm.
1638 KVM: Add instance chrooting feature. If you use privilege dropping for
1639 your VMs you can also now force them to chroot to an empty directory,
1640 before starting the emulated guest.
1642 KVM: Add maximum migration bandwith and maximum downtime tweaking
1643 support (requires a new-enough version of qemu-kvm).
1645 Cluster verify will now warn if the master node doesn't have the master
1646 ip configured on it.
1648 Add a new (incompatible) instance creation request format to RAPI which
1649 supports all parameters (previously only a subset was supported, and it
1650 wasn't possible to extend the old format to accomodate all the new
1651 features. The old format is still supported, and a client can check for
1652 this feature, before using it, by checking for its presence in the
1653 ``features`` RAPI resource.
1655 Now with ancient latin support. Try it passing the ``--roman`` option to
1656 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1657 (requires the python-roman module to be installed, in order to work).
1662 As usual many internal code refactorings, documentation updates, and
1665 - Lots of improvements and cleanups to the experimental Remote API
1666 (RAPI) client library.
1667 - A new unit test suite for the core daemon libraries.
1668 - A fix to creating missing directories makes sure the umask is not
1669 applied anymore. This enforces the same directory permissions
1671 - Better handling terminating daemons with ctrl+c (used when running
1672 them in debugging mode).
1673 - Fix a race condition in live migrating a KVM instance, when stat()
1674 on the old proc status file returned EINVAL, which is an unexpected
1676 - Fixed manpage checking with newer man and utf-8 charachters. But now
1677 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1683 *(Released Fri, 7 May 2010)*
1685 Fix a bug which prevented untagged KVM instances from starting.
1691 *(Released Fri, 7 May 2010)*
1693 Another release with a long development cycle, during which many
1694 different features were added.
1696 Significant features
1697 ~~~~~~~~~~~~~~~~~~~~
1699 The KVM hypervisor now can run the individual instances as non-root, to
1700 reduce the impact of a VM being hijacked due to bugs in the
1701 hypervisor. It is possible to run all instances as a single (non-root)
1702 user, to manually specify a user for each instance, or to dynamically
1703 allocate a user out of a cluster-wide pool to each instance, with the
1704 guarantee that no two instances will run under the same user ID on any
1707 An experimental RAPI client library, that can be used standalone
1708 (without the other Ganeti libraries), is provided in the source tree as
1709 ``lib/rapi/client.py``. Note this client might change its interface in
1710 the future, as we iterate on its capabilities.
1712 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1713 replace the cluster's certificates and crypto keys. This might help in
1714 case they have been compromised, or have simply expired.
1716 A new disk option for instance creation has been added that allows one
1717 to "adopt" currently existing logical volumes, with data
1718 preservation. This should allow easier migration to Ganeti from
1719 unmanaged (or managed via other software) instances.
1721 Another disk improvement is the possibility to convert between redundant
1722 (DRBD) and plain (LVM) disk configuration for an instance. This should
1723 allow better scalability (starting with one node and growing the
1724 cluster, or shrinking a two-node cluster to one node).
1726 A new feature that could help with automated node failovers has been
1727 implemented: if a node sees itself as offline (by querying the master
1728 candidates), it will try to shutdown (hard) all instances and any active
1729 DRBD devices. This reduces the risk of duplicate instances if an
1730 external script automatically failovers the instances on such nodes. To
1731 enable this, the cluster parameter ``maintain_node_health`` should be
1732 enabled; in the future this option (per the name) will enable other
1733 automatic maintenance features.
1735 Instance export/import now will reuse the original instance
1736 specifications for all parameters; that means exporting an instance,
1737 deleting it and the importing it back should give an almost identical
1738 instance. Note that the default import behaviour has changed from
1739 before, where it created only one NIC; now it recreates the original
1742 Cluster verify has added a few new checks: SSL certificates validity,
1743 /etc/hosts consistency across the cluster, etc.
1748 As usual, many internal changes were done, documentation fixes,
1751 - Fixed cluster initialization with disabled cluster storage (regression
1752 introduced in 2.1.1)
1753 - File-based storage supports growing the disks
1754 - Fixed behaviour of node role changes
1755 - Fixed cluster verify for some corner cases, plus a general rewrite of
1756 cluster verify to allow future extension with more checks
1757 - Fixed log spamming by watcher and node daemon (regression introduced
1759 - Fixed possible validation issues when changing the list of enabled
1761 - Fixed cleanup of /etc/hosts during node removal
1762 - Fixed RAPI response for invalid methods
1763 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1764 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1765 from ide disks, etc.)
1766 - Allow OS changes without re-installation (to record a changed OS
1767 outside of Ganeti, or to allow OS renames)
1768 - Allow instance creation without OS installation (useful for example if
1769 the OS will be installed manually, or restored from a backup not in
1771 - Implemented option to make cluster ``copyfile`` use the replication
1773 - Added list of enabled hypervisors to ssconf (possibly useful for
1775 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1777 - A partial form of node re-IP is possible via node readd, which now
1778 allows changed node primary IP
1779 - Command line utilities now show an informational message if the job is
1781 - The logs of the master daemon now show the PID/UID/GID of the
1788 *(Released Fri, 12 Mar 2010)*
1790 During the 2.1.0 long release candidate cycle, a lot of improvements and
1791 changes have accumulated with were released later as 2.1.1.
1796 The node evacuate command (``gnt-node evacuate``) was significantly
1797 rewritten, and as such the IAllocator protocol was changed - a new
1798 request type has been added. This unfortunate change during a stable
1799 series is designed to improve performance of node evacuations; on
1800 clusters with more than about five nodes and which are well-balanced,
1801 evacuation should proceed in parallel for all instances of the node
1802 being evacuated. As such, any existing IAllocator scripts need to be
1803 updated, otherwise the above command will fail due to the unknown
1804 request. The provided "dumb" allocator has not been updated; but the
1805 ganeti-htools package supports the new protocol since version 0.2.4.
1807 Another important change is increased validation of node and instance
1808 names. This might create problems in special cases, if invalid host
1809 names are being used.
1811 Also, a new layer of hypervisor parameters has been added, that sits at
1812 OS level between the cluster defaults and the instance ones. This allows
1813 customisation of virtualization parameters depending on the installed
1814 OS. For example instances with OS 'X' may have a different KVM kernel
1815 (or any other parameter) than the cluster defaults. This is intended to
1816 help managing a multiple OSes on the same cluster, without manual
1817 modification of each instance's parameters.
1819 A tool for merging clusters, ``cluster-merge``, has been added in the
1820 tools sub-directory.
1825 - Improved the int/float conversions that should make the code more
1826 robust in face of errors from the node daemons
1827 - Fixed the remove node code in case of internal configuration errors
1828 - Fixed the node daemon behaviour in face of inconsistent queue
1829 directory (e.g. read-only file-system where we can't open the files
1831 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1832 now it either aborts cleanly or, if given the new "auto_promote"
1833 parameter, will automatically promote other nodes as needed
1834 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1835 completely prevent Ganeti from working
1836 - Fixed bug for instance export when not all disks were successfully
1838 - Fixed behaviour of node add when the new node is slow in starting up
1840 - Fixed handling of signals in the LUXI client, which should improve
1841 behaviour of command-line scripts
1842 - Added checks for invalid node/instance names in the configuration (now
1843 flagged during cluster verify)
1844 - Fixed watcher behaviour for disk activation errors
1845 - Fixed two potentially endless loops in http library, which led to the
1846 RAPI daemon hanging and consuming 100% CPU in some cases
1847 - Fixed bug in RAPI daemon related to hashed passwords
1848 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1850 - Enhanced compatibility with non-Debian OSes, but not using absolute
1851 path in some commands and allowing customisation of the ssh
1852 configuration directory
1853 - Fixed possible future issue with new Python versions by abiding to the
1854 proper use of ``__slots__`` attribute on classes
1855 - Added checks that should prevent directory traversal attacks
1856 - Many documentation fixes based on feedback from users
1861 - Added an "early_release" more for instance replace disks and node
1862 evacuate, where we release locks earlier and thus allow higher
1863 parallelism within the cluster
1864 - Added watcher hooks, intended to allow the watcher to restart other
1865 daemons (e.g. from the ganeti-nbma project), but they can be used of
1866 course for any other purpose
1867 - Added a compile-time disable for DRBD barriers, to increase
1868 performance if the administrator trusts the power supply or the
1869 storage system to not lose writes
1870 - Added the option of using syslog for logging instead of, or in
1871 addition to, Ganeti's own log files
1872 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1873 can indeed boot from a paravirtual NIC
1874 - Added a generic debug level for many operations; while this is not
1875 used widely yet, it allows one to pass the debug value all the way to
1877 - Enhanced the hooks environment for instance moves (failovers,
1878 migrations) where the primary/secondary nodes changed during the
1879 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1880 - Enhanced data validations for many user-supplied values; one important
1881 item is the restrictions imposed on instance and node names, which
1882 might reject some (invalid) host names
1883 - Add a configure-time option to disable file-based storage, if it's not
1884 needed; this allows greater security separation between the master
1885 node and the other nodes from the point of view of the inter-node RPC
1887 - Added user notification in interactive tools if job is waiting in the
1888 job queue or trying to acquire locks
1889 - Added log messages when a job is waiting for locks
1890 - Added filtering by node tags in instance operations which admit
1891 multiple instances (start, stop, reboot, reinstall)
1892 - Added a new tool for cluster mergers, ``cluster-merge``
1893 - Parameters from command line which are of the form ``a=b,c=d`` can now
1894 use backslash escapes to pass in values which contain commas,
1895 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1897 - For KVM, the instance name is the first parameter passed to KVM, so
1898 that it's more visible in the process list
1904 *(Released Tue, 2 Mar 2010)*
1906 Ganeti 2.1 brings many improvements with it. Major changes:
1908 - Added infrastructure to ease automated disk repairs
1909 - Added new daemon to export configuration data in a cheaper way than
1910 using the remote API
1911 - Instance NICs can now be routed instead of being associated with a
1913 - Improved job locking logic to reduce impact of jobs acquiring multiple
1914 locks waiting for other long-running jobs
1916 In-depth implementation details can be found in the Ganeti 2.1 design
1922 - Added chroot hypervisor
1923 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1925 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1926 ``bootloader_path`` and ``bootloader_args``)
1927 - Added the ``use_localtime`` option for the xen-hvm and kvm
1928 hypervisors, and the default value for this has changed to false (in
1929 2.0 xen-hvm always enabled it)
1930 - Added luxi call to submit multiple jobs in one go
1931 - Added cluster initialization option to not modify ``/etc/hosts``
1933 - Added network interface parameters
1934 - Added dry run mode to some LUs
1935 - Added RAPI resources:
1937 - ``/2/instances/[instance_name]/info``
1938 - ``/2/instances/[instance_name]/replace-disks``
1939 - ``/2/nodes/[node_name]/evacuate``
1940 - ``/2/nodes/[node_name]/migrate``
1941 - ``/2/nodes/[node_name]/role``
1942 - ``/2/nodes/[node_name]/storage``
1943 - ``/2/nodes/[node_name]/storage/modify``
1944 - ``/2/nodes/[node_name]/storage/repair``
1946 - Added OpCodes to evacuate or migrate all instances on a node
1947 - Added new command to list storage elements on nodes (``gnt-node
1948 list-storage``) and modify them (``gnt-node modify-storage``)
1949 - Added new ssconf files with master candidate IP address
1950 (``ssconf_master_candidates_ips``), node primary IP address
1951 (``ssconf_node_primary_ips``) and node secondary IP address
1952 (``ssconf_node_secondary_ips``)
1953 - Added ``ganeti-confd`` and a client library to query the Ganeti
1954 configuration via UDP
1955 - Added ability to run hooks after cluster initialization and before
1957 - Added automatic mode for disk replace (``gnt-instance replace-disks
1959 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1960 after catastrophic data-loss
1961 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1963 - Added ``gnt-instance move`` command to move instances
1964 - Added ``gnt-cluster watcher`` command to control watcher
1965 - Added ``gnt-node powercycle`` command to powercycle nodes
1966 - Added new job status field ``lock_status``
1967 - Added parseable error codes to cluster verification (``gnt-cluster
1968 verify --error-codes``) and made output less verbose (use
1969 ``--verbose`` to restore previous behaviour)
1970 - Added UUIDs to the main config entities (cluster, nodes, instances)
1971 - Added support for OS variants
1972 - Added support for hashed passwords in the Ganeti remote API users file
1974 - Added option to specify maximum timeout on instance shutdown
1975 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1976 - Added new helper script to start and stop Ganeti daemons
1977 (``daemon-util``), with the intent to reduce the work necessary to
1978 adjust Ganeti for non-Debian distributions and to start/stop daemons
1980 - Added more unittests
1981 - Fixed critical bug in ganeti-masterd startup
1982 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1983 now customisable at the cluster level for both the KVM and Xen
1984 hypervisors using the new ``migration_port`` parameter
1985 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1986 reinstalling an instance
1987 - Allowed ``@`` in tag names
1988 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1989 - Many documentation updates
1990 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1991 - ``gnt-instance reinstall`` can now reinstall multiple instances
1992 - Updated many command line parameters
1993 - Introduced new OS API version 15
1994 - No longer support a default hypervisor
1995 - Treat virtual LVs as inexistent
1996 - Improved job locking logic to reduce lock contention
1997 - Match instance and node names case insensitively
1998 - Reimplemented bash completion script to be more complete
2005 *(Released Thu, 4 Feb 2010)*
2007 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2008 - Fix a string formatting bug
2009 - Improve safety of the code in some error paths
2010 - Improve data validation in the master of values returned from nodes
2016 *(Released Thu, 17 Dec 2009)*
2018 - Fix security issue due to missing validation of iallocator names; this
2019 allows local and remote execution of arbitrary executables
2020 - Fix failure of gnt-node list during instance removal
2021 - Ship the RAPI documentation in the archive
2027 *(Released Wed, 30 Sep 2009)*
2029 - Fixed many wrong messages
2030 - Fixed a few bugs related to the locking library
2031 - Fixed MAC checking at instance creation time
2032 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2033 - Fixed a few issues related to signal handling in both daemons and
2035 - Fixed the example startup script provided
2036 - Fixed insserv dependencies in the example startup script (patch from
2038 - Fixed handling of drained nodes in the iallocator framework
2039 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2041 - Fixed error related to invalid job IDs in job polling
2042 - Fixed job/opcode persistence on unclean master shutdown
2043 - Fixed handling of partial job processing after unclean master
2045 - Fixed error reporting from LUs, previously all errors were converted
2046 into execution errors
2047 - Fixed error reporting from burnin
2048 - Decreased significantly the memory usage of the job queue
2049 - Optimised slightly multi-job submission
2050 - Optimised slightly opcode loading
2051 - Backported the multi-job submit framework from the development
2052 branch; multi-instance start and stop should be faster
2053 - Added script to clean archived jobs after 21 days; this will reduce
2054 the size of the queue directory
2055 - Added some extra checks in disk size tracking
2056 - Added an example ethers hook script
2057 - Added a cluster parameter that prevents Ganeti from modifying of
2059 - Added more node information to RAPI responses
2060 - Added a ``gnt-job watch`` command that allows following the ouput of a
2062 - Added a bind-address option to ganeti-rapi
2063 - Added more checks to the configuration verify
2064 - Enhanced the burnin script such that some operations can be retried
2066 - Converted instance reinstall to multi-instance model
2072 *(Released Fri, 7 Aug 2009)*
2074 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2075 to allow using the pre-2.0.2 behaviour in activation, if any existing
2076 instances have mismatched disk sizes in the configuration
2077 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2078 any configuration mismatches for disk sizes
2079 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2080 failover to work on two-node clusters
2081 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2083 - Fixed detection of OS script errors in ``gnt-backup export``
2084 - Fixed exit code of ``gnt-backup export``
2090 *(Released Fri, 17 Jul 2009)*
2092 - Added experimental support for stripped logical volumes; this should
2093 enhance performance but comes with a higher complexity in the block
2094 device handling; stripping is only enabled when passing
2095 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2096 affected even in the non-stripped mode
2097 - Improved resiliency against transient failures at the end of DRBD
2098 resyncs, and in general of DRBD resync checks
2099 - Fixed a couple of issues with exports and snapshot errors
2100 - Fixed a couple of issues in instance listing
2101 - Added display of the disk size in ``gnt-instance info``
2102 - Fixed checking for valid OSes in instance creation
2103 - Fixed handling of the "vcpus" parameter in instance listing and in
2104 general of invalid parameters
2105 - Fixed http server library, and thus RAPI, to handle invalid
2106 username/password combinations correctly; this means that now they
2107 report unauthorized for queries too, not only for modifications,
2108 allowing earlier detect of configuration problems
2109 - Added a new "role" node list field, equivalent to the master/master
2110 candidate/drained/offline flags combinations
2111 - Fixed cluster modify and changes of candidate pool size
2112 - Fixed cluster verify error messages for wrong files on regular nodes
2113 - Fixed a couple of issues with node demotion from master candidate role
2114 - Fixed node readd issues
2115 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2116 - Added a new ``--no-voting`` option for masterfailover to fix failover
2117 on two-nodes clusters when the former master node is unreachable
2118 - Added instance reinstall over RAPI
2124 *(Released Tue, 16 Jun 2009)*
2126 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2127 allow re-adding the start in single-user option (regression from 1.2)
2128 - the watcher writes the instance status to a file, to allow monitoring
2129 to report the instance status (from the master) based on cached
2130 results of the watcher's queries; while this can get stale if the
2131 watcher is being locked due to other work on the cluster, this is
2132 still an improvement
2133 - the watcher now also restarts the node daemon and the rapi daemon if
2135 - fixed the watcher to handle full and drained queue cases
2136 - hooks export more instance data in the environment, which helps if
2137 hook scripts need to take action based on the instance's properties
2138 (no longer need to query back into ganeti)
2139 - instance failovers when the instance is stopped do not check for free
2140 RAM, so that failing over a stopped instance is possible in low memory
2142 - rapi uses queries for tags instead of jobs (for less job traffic), and
2143 for cluster tags it won't talk to masterd at all but read them from
2145 - a couple of error handling fixes in RAPI
2146 - drbd handling: improved the error handling of inconsistent disks after
2147 resync to reduce the frequency of "there are some degraded disks for
2148 this instance" messages
2149 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2150 error handling path called a wrong function name)
2156 *(Released Wed, 27 May 2009)*
2158 - no changes from rc5
2164 *(Released Wed, 20 May 2009)*
2166 - fix a couple of bugs (validation, argument checks)
2167 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2168 - some small improvements to RAPI and IAllocator
2169 - make watcher automatically start the master daemon if down
2175 *(Released Mon, 27 Apr 2009)*
2177 - change the OS list to not require locks; this helps with big clusters
2178 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2179 volume group is broken
2180 - ``gnt-instance info``, without any arguments, doesn't run for all
2181 instances anymore; either pass ``--all`` or pass the desired
2182 instances; this helps against mistakes on big clusters where listing
2183 the information for all instances takes a long time
2184 - miscellaneous doc and man pages fixes
2190 *(Released Wed, 8 Apr 2009)*
2192 - Change the internal locking model of some ``gnt-node`` commands, in
2193 order to reduce contention (and blocking of master daemon) when
2194 batching many creation/reinstall jobs
2195 - Fixes to Xen soft reboot
2196 - No longer build documentation at build time, instead distribute it in
2197 the archive, in order to reduce the need for the whole docbook/rst
2204 *(Released Fri, 27 Mar 2009)*
2206 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2207 - Fix watcher startup sequence, improves the behaviour of busy clusters
2208 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2209 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2210 bind address and other places
2211 - Some documentation fixes and updates
2217 *(Released Mon, 2 Mar 2009)*
2219 - More documentation updates, now all docs should be more-or-less
2221 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2223 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2225 - Fix an issue related to $libdir/run/ganeti and cluster creation
2231 *(Released Thu, 19 Feb 2009)*
2233 - Xen PVM and KVM have switched the default value for the instance root
2234 disk to the first partition on the first drive, instead of the whole
2235 drive; this means that the OS installation scripts must be changed
2237 - Man pages have been updated
2238 - RAPI has been switched by default to HTTPS, and the exported functions
2239 should all work correctly
2240 - RAPI v1 has been removed
2241 - Many improvements to the KVM hypervisor
2242 - Block device errors are now better reported
2243 - Many other bugfixes and small improvements
2249 *(Released Mon, 26 Jan 2009)*
2251 - Version 2 is a general rewrite of the code and therefore the
2252 differences are too many to list, see the design document for 2.0 in
2253 the ``doc/`` subdirectory for more details
2254 - In this beta version there is not yet a migration path from 1.2 (there
2255 will be one in the final 2.0 release)
2256 - A few significant changes are:
2258 - all commands are executed by a daemon (``ganeti-masterd``) and the
2259 various ``gnt-*`` commands are just front-ends to it
2260 - all the commands are entered into, and executed from a job queue,
2261 see the ``gnt-job(8)`` manpage
2262 - the RAPI daemon supports read-write operations, secured by basic
2263 HTTP authentication on top of HTTPS
2264 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2265 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2266 to migrate to DRBD 8 first while still running Ganeti 1.2)
2267 - DRBD devices are using statically allocated minor numbers, which
2268 will be assigned to existing instances during the migration process
2269 - there is support for both Xen PVM and Xen HVM instances running on
2271 - KVM virtualization is supported too
2272 - file-based storage has been implemented, which means that it is
2273 possible to run the cluster without LVM and DRBD storage, for
2274 example using a shared filesystem exported from shared storage (and
2275 still have live migration)
2281 *(Released Tue, 13 Jan 2009)*
2283 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2284 - Reuse the old instance mac address by default on instance import, if
2285 the instance name is the same.
2286 - Handle situations in which the node info rpc returns incomplete
2288 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2289 - Improved version of batcher:
2291 - state file support
2292 - instance mac address support
2293 - support for HVM clusters/instances
2295 - Add an option to show the number of cpu sockets and nodes in
2297 - Support OSes that handle more than one version of the OS api (but do
2298 not change the current API in any other way)
2299 - Fix ``gnt-node migrate``
2300 - ``gnt-debug`` man page
2301 - Fixes various more typos and small issues
2302 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2308 *(Released Wed, 24 Sep 2008)*
2310 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2311 type of disk exported to fully virtualized instances.
2312 - provide access to the serial console of HVM instances
2313 - instance auto_balance flag, set by default. If turned off it will
2314 avoid warnings on cluster verify if there is not enough memory to fail
2315 over an instance. in the future it will prevent automatically failing
2316 it over when we will support that.
2317 - batcher tool for instance creation, see ``tools/README.batcher``
2318 - ``gnt-instance reinstall --select-os`` to interactively select a new
2319 operating system when reinstalling an instance.
2320 - when changing the memory amount on instance modify a check has been
2321 added that the instance will be able to start. also warnings are
2322 emitted if the instance will not be able to fail over, if auto_balance
2324 - documentation fixes
2325 - sync fields between ``gnt-instance list/modify/add/import``
2326 - fix a race condition in drbd when the sync speed was set after giving
2327 the device a remote peer.
2333 *(Released Tue, 22 Jul 2008)*
2335 - note: the allowed size and number of tags per object were reduced
2336 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2337 - fixed twisted 8.x compatibility
2338 - fixed ``gnt-instance replace-disks`` with iallocator
2339 - add TCP keepalives on twisted connections to detect restarted nodes
2340 - disk increase support, see ``gnt-instance grow-disk``
2341 - implement bulk node/instance query for RAPI
2342 - add tags in node/instance listing (optional)
2343 - experimental migration (and live migration) support, read the man page
2344 for ``gnt-instance migrate``
2345 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2346 has some small improvements in handling its state file
2352 *(Released Fri, 13 Jun 2008)*
2354 - Experimental readonly, REST-based remote API implementation;
2355 automatically started on master node, TCP port 5080, if enabled by
2356 ``--enable-rapi`` parameter to configure script.
2357 - Instance allocator support. Add and import instance accept a
2358 ``--iallocator`` parameter, and call that instance allocator to decide
2359 which node to use for the instance. The iallocator document describes
2360 what's expected from an allocator script.
2361 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2362 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2363 node is lost there is still enough memory to fail over the instances
2365 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2366 ``gnt-cluster verify``, to check for site-specific compliance. All the
2367 hooks will run, and their output, if any, will be displayed. Any
2368 failing hook will make the verification return an error value.
2369 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2370 primary and secondary interfaces
2371 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2372 that is still declared as part of the cluster and has failed.
2373 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2374 specifying output fields, that just adds the chosen fields to the
2376 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2377 export from the filesystem.
2378 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2379 have been added. Using them you can enable/disable acpi and pae
2380 support, and specify a path for a cd image to be exported to the
2381 instance. These parameters as the name suggest only work on HVM
2383 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2384 PAE support will be set to the previously hardcoded values, but the
2385 (previously hardcoded) path to the CDROM ISO image will be unset and
2386 if required, needs to be set manually with ``gnt-instance modify``
2388 - The address to which an instance's VNC console is bound is now
2389 selectable per-instance, rather than being cluster wide. Of course
2390 this only applies to instances controlled via VNC, so currently just
2391 applies to HVM clusters.
2397 *(Released Mon, 18 Feb 2008)*
2399 - more tweaks to the disk activation code (especially helpful for DRBD)
2400 - change the default ``gnt-instance list`` output format, now there is
2401 one combined status field (see the manpage for the exact values this
2403 - some more fixes for the mac export to hooks change
2404 - make Ganeti not break with DRBD 8.2.x (which changed the version
2405 format in ``/proc/drbd``) (issue 24)
2406 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2407 template, allowing migration from DRBD0.7+MD to DRBD8
2413 *(Released Wed, 30 Jan 2008)*
2415 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2417 - add command aliases infrastructure and a few aliases
2418 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2419 man pages and the ``--help`` option of ``gnt-node
2420 list``/``gnt-instance list``
2421 - fix ``gnt-backup list`` with down nodes (issue 21)
2422 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2423 - fix the dist archive and add a check for including svn/git files in
2425 - some developer-related changes: improve the burnin and the QA suite,
2426 add an upload script for testing during development
2432 *(Released Wed, 16 Jan 2008)*
2434 - experimental HVM support, read the install document, section
2435 "Initializing the cluster"
2436 - allow for the PVM hypervisor per-instance kernel and initrd paths
2437 - add a new command ``gnt-cluster verify-disks`` which uses a new
2438 algorithm to improve the reconnection of the DRBD pairs if the device
2439 on the secondary node has gone away
2440 - make logical volume code auto-activate LVs at disk activation time
2441 - slightly improve the speed of activating disks
2442 - allow specification of the MAC address at instance creation time, and
2443 changing it later via ``gnt-instance modify``
2444 - fix handling of external commands that generate lots of output on
2446 - update documentation with regard to minimum version of DRBD8 supported
2452 *(Released Tue, 4 Dec 2007)*
2454 - Log the ``xm create`` output to the node daemon log on failure (to
2455 help diagnosing the error)
2456 - In debug mode, log all external commands output if failed to the logs
2457 - Change parsing of lvm commands to ignore stderr
2463 *(Released Wed, 28 Nov 2007)*
2465 - Another round of updates to the DRBD 8 code to deal with more failures
2466 in the replace secondary node operation
2467 - Some more logging of failures in disk operations (lvm, drbd)
2468 - A few documentation updates
2475 *(Released Tue, 13 Nov 2007)*
2477 - Change configuration file format from Python's Pickle to JSON.
2478 Upgrading is possible using the cfgupgrade utility.
2479 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2480 faster replace disks and is more stable (DRBD 8 has many improvements
2481 compared to DRBD 0.7)
2482 - Added command line tags support (see man pages for ``gnt-instance``,
2483 ``gnt-node``, ``gnt-cluster``)
2484 - Added instance rename support
2485 - Added multi-instance startup/shutdown
2486 - Added cluster rename support
2487 - Added ``gnt-node evacuate`` to simplify some node operations
2488 - Added instance reboot operation that can speedup reboot as compared to
2490 - Soften the requirement that hostnames are in FQDN format
2491 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2493 - Removed dependency on debian's patched fping that uses the
2494 non-standard ``-S`` option
2495 - Now the OS definitions are searched for in multiple, configurable
2496 paths (easier for distros to package)
2497 - Some changes to the hooks infrastructure (especially the new
2498 post-configuration update hook)
2499 - Other small bugfixes
2501 .. vim: set textwidth=72 syntax=rst :