8 *(Released Thu, 6 Feb 2013)*
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(7)`) 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 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.
55 - The functionality for allocating multiple instances at once has been
56 overhauled and is now also available through :doc:`RAPI <rapi>`.
57 - Man pages can now be included when the documentation is built, in
58 which case the output is in ``doc/man-html``. The configure-time
59 option is ``--enable-manpages-in-doc``. Sphinx 1.0 or higher is
61 - A new htool called ``hroller`` helps scheduling parallel shutdown of
62 nodes on clusters where all instances are shut down, but without ever
63 rebooting both primary and secondary node for an instance at the same
65 - Queries not needing locks or RPC access to the node can now be
66 performed by the confd daemon, making them independent from jobs, and
67 thus faster to execute. This is selectable at configure time.
68 - The KVM hypervisor has been updated with many new hypervisor
69 parameters, including a generic one for passing arbitrary command line
70 values. See a complete list in :manpage:`gnt-instance(8)`.
71 - A new tool, called ``mon-collector``, is the stand-alone executor of
72 the data collectors for a monitoring system. As of this version, it
73 just includes the DRBD data collector, that can be executed by calling
74 ``mon-collector`` using the ``drbd`` parameter. See
75 :manpage:`mon-collector(7)`.
76 - New ``exclusive-storage`` node parameter added. When it's set to true,
77 physical disks are assigned in an exclusive fashion to instances, as
78 documented in :doc:`Partitioned Ganeti <design-partitioned>`.
79 Currently, only instances using the ``plain`` disk template are
81 - New network management functionality to support automatic allocation
82 of IP addresses and managing of network parameters. See
83 :manpage:`gnt-network(8)` for more details.
84 - New external storage backend, to allow managing arbitrary storage
85 systems external to the cluster. See
86 :manpage:`ganeti-extstorage-interface(7)`.
87 - Instance renames of LVM-based instances will now update the LV tags
88 (which can be used to recover the instance-to-LV mapping in case of
95 *(Released Fri, 21 Dec 2012)*
97 Important behaviour change: hbal won't rebalance anymore instances which
98 have the ``auto_balance`` attribute set to false. This was the intention
99 all along, but until now it only skipped those from the N+1 memory
100 reservation (DRBD-specific).
102 A significant number of bug fixes in this release:
104 - Fixed disk adoption interaction with ipolicy checks.
105 - Fixed networking issues when instances are started, stopped or
106 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
107 - Fixed the warning in cluster verify for shared storage instances not
109 - Fixed removal of storage directory on shared file storage (issue 262).
110 - Fixed validation of LVM volume group name in OpClusterSetParams
111 (``gnt-cluster modify``) (issue 285).
112 - Fixed runtime memory increases (``gnt-instance modify -m``).
113 - Fixed live migration under Xen's ``xl`` mode.
114 - Fixed ``gnt-instance console`` with ``xl``.
115 - Fixed building with newer Haskell compiler/libraries.
116 - Fixed PID file writing in Haskell daemons (confd); this prevents
117 restart issues if confd was launched manually (outside of
118 ``daemon-util``) while another copy of it was running
119 - Fixed a type error when doing live migrations with KVM (issue 297) and
120 the error messages for failing migrations have been improved.
121 - Fixed opcode validation for the out-of-band commands (``gnt-node
123 - Fixed a type error when unsetting OS hypervisor parameters (issue
124 311); now it's possible to unset all OS-specific hypervisor
126 - Fixed the ``dry-run`` mode for many operations: verification of
127 results was over-zealous but didn't take into account the ``dry-run``
128 operation, resulting in "wrong" failures.
129 - Fixed bash completion in ``gnt-job list`` when the job queue has
130 hundreds of entries; especially with older ``bash`` versions, this
131 results in significant CPU usage.
133 And lastly, a few other improvements have been made:
135 - Added option to force master-failover without voting (issue 282).
136 - Clarified error message on lock conflict (issue 287).
137 - Logging of newly submitted jobs has been improved (issue 290).
138 - Hostname checks have been made uniform between instance rename and
140 - The ``--submit`` option is now supported by ``gnt-debug delay``.
141 - Shutting down the master daemon by sending SIGTERM now stops it from
142 processing jobs waiting for locks; instead, those jobs will be started
143 once again after the master daemon is started the next time (issue
145 - Support for Xen's ``xl`` program has been improved (besides the fixes
147 - Reduced logging noise in the Haskell confd daemon (only show one log
148 entry for each config reload, instead of two).
149 - Several man page updates and typo fixes.
155 *(Released Fri, 12 Oct 2012)*
157 A small bugfix release. Among the bugs fixed:
159 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
160 made the command unusable.
161 - Commands that issue many jobs don't fail anymore just because some jobs
162 take so long that other jobs are archived.
163 - Failures during ``gnt-instance reinstall`` are reflected by the exit
165 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
167 - When ``always_failover`` is set, ``--allow-failover`` is not required
168 in migrate commands anymore.
169 - ``bash_completion`` works even if extglob is disabled.
170 - Fixed bug with locks that made failover for RDB-based instances fail.
171 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
172 a random node instead of one based on the allocator metric.
173 - Support for newer versions of pylint and pep8.
174 - Hail doesn't fail anymore when trying to add an instance of type
175 ``file``, ``sharedfile`` or ``rbd``.
176 - Added new Makefile target to rebuild the whole distribution, so that
177 all files are included.
183 *(Released Fri, 27 Jul 2012)*
186 .. attention:: The ``LUXI`` protocol has been made more consistent
187 regarding its handling of command arguments. This, however, leads to
188 incompatibility issues with previous versions. Please ensure that you
189 restart Ganeti daemons soon after the upgrade, otherwise most
190 ``LUXI`` calls (job submission, setting/resetting the drain flag,
191 pausing/resuming the watcher, cancelling and archiving jobs, querying
192 the cluster configuration) will fail.
201 The current ``admin_up`` field, which used to denote whether an instance
202 should be running or not, has been removed. Instead, ``admin_state`` is
203 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
205 The rational behind this is that an instance being “down” can have
208 - it could be down during a reboot
209 - it could be temporarily be down for a reinstall
210 - or it could be down because it is deprecated and kept just for its
213 The previous Boolean state was making it difficult to do capacity
214 calculations: should Ganeti reserve memory for a down instance? Now, the
215 tri-state field makes it clear:
217 - in ``up`` and ``down`` state, all resources are reserved for the
218 instance, and it can be at any time brought up if it is down
219 - in ``offline`` state, only disk space is reserved for it, but not
222 The field can have an extra use: since the transition between ``up`` and
223 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
224 transition between ``offline`` and ``down`` is done via ``gnt-instance
225 modify``, it is possible to given different rights to users. For
226 example, owners of an instance could be allowed to start/stop it, but
227 not transition it out of the offline state.
229 Instance policies and specs
230 +++++++++++++++++++++++++++
232 In previous Ganeti versions, an instance creation request was not
233 limited on the minimum size and on the maximum size just by the cluster
234 resources. As such, any policy could be implemented only in third-party
235 clients (RAPI clients, or shell wrappers over ``gnt-*``
236 tools). Furthermore, calculating cluster capacity via ``hspace`` again
237 required external input with regards to instance sizes.
239 In order to improve these workflows and to allow for example better
240 per-node group differentiation, we introduced instance specs, which
243 - minimum instance disk size, disk count, memory size, cpu count
244 - maximum values for the above metrics
245 - and “standard” values (used in ``hspace`` to calculate the standard
248 The minimum/maximum values can be also customised at node-group level,
249 for example allowing more powerful hardware to support bigger instance
252 Beside the instance specs, there are a few other settings belonging to
253 the instance policy framework. It is possible now to customise, per
254 cluster and node-group:
256 - the list of allowed disk templates
257 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
258 - the maximum ratio of instance to spindles (see below for more
259 information) for local storage
261 All these together should allow all tools that talk to Ganeti to know
262 what are the ranges of allowed values for instances and the
263 over-subscription that is allowed.
265 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
266 instance configuration, and the physical CPU configuration from the
267 node. For the spindle ratios however, we didn't track before these
268 values, so new parameters have been added:
270 - a new node parameter ``spindle_count``, defaults to 1, customisable at
271 node group or node level
272 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
274 Note that spindles in this context doesn't need to mean actual
275 mechanical hard-drives; it's just a relative number for both the node
276 I/O capacity and instance I/O consumption.
278 Instance migration behaviour
279 ++++++++++++++++++++++++++++
281 While live-migration is in general desirable over failover, it is
282 possible that for some workloads it is actually worse, due to the
283 variable time of the “suspend” phase during live migration.
285 To allow the tools to work consistently over such instances (without
286 having to hard-code instance names), a new backend parameter
287 ``always_failover`` has been added to control the migration/failover
288 behaviour. When set to True, all migration requests for an instance will
289 instead fall-back to failover.
291 Instance memory ballooning
292 ++++++++++++++++++++++++++
294 Initial support for memory ballooning has been added. The memory for an
295 instance is no longer fixed (backend parameter ``memory``), but instead
296 can vary between minimum and maximum values (backend parameters
297 ``minmem`` and ``maxmem``). Currently we only change an instance's
300 - live migrating or failing over and instance and the target node
301 doesn't have enough memory
302 - user requests changing the memory via ``gnt-instance modify
308 In order to control the use of specific CPUs by instance, support for
309 controlling CPU pinning has been added for the Xen, HVM and LXC
310 hypervisors. This is controlled by a new hypervisor parameter
311 ``cpu_mask``; details about possible values for this are in the
312 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
313 VCPU-to-CPU mapping) form will work well only when all nodes in your
314 cluster have the same amount of CPUs.
319 Another area in which Ganeti was not customisable were the parameters
320 used for storage configuration, e.g. how many stripes to use for LVM,
321 DRBD resync configuration, etc.
323 To improve this area, we've added disks parameters, which are
324 customisable at cluster and node group level, and which allow to
325 specify various parameters for disks (DRBD has the most parameters
326 currently), for example:
328 - DRBD resync algorithm and parameters (e.g. speed)
329 - the default VG for meta-data volumes for DRBD
330 - number of stripes for LVM (plain disk template)
333 These parameters can be modified via ``gnt-cluster modify -D …`` and
334 ``gnt-group modify -D …``, and are used at either instance creation (in
335 case of LVM stripes, for example) or at disk “activation” time
338 Rados block device support
339 ++++++++++++++++++++++++++
341 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
342 denoted by the ``rbd`` disk template type. This is considered
343 experimental, feedback is welcome. For details on configuring it, see
344 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
349 The existing master IP functionality works well only in simple setups (a
350 single network shared by all nodes); however, if nodes belong to
351 different networks, then the ``/32`` setup and lack of routing
352 information is not enough.
354 To allow the master IP to function well in more complex cases, the
355 system was reworked as follows:
357 - a master IP netmask setting has been added
358 - the master IP activation/turn-down code was moved from the node daemon
360 - whether to run the Ganeti-supplied master IP script or a user-supplied
361 on is a ``gnt-cluster init`` setting
363 Details about the location of the standard and custom setup scripts are
364 in the man page :manpage:`gnt-cluster(8)`; for information about the
365 setup script protocol, look at the Ganeti-supplied script.
370 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
373 It is now possible to use TLS-protected connections, and when renewing
374 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
375 it is now possible to specify spice or spice CA certificates. Also, it
376 is possible to configure a password for SPICE sessions via the
377 hypervisor parameter ``spice_password_file``.
379 There are also new parameters to control the compression and streaming
380 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
381 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
382 for the spice parameters.
384 Lastly, it is now possible to see the SPICE connection information via
385 ``gnt-instance console``.
390 A new tool (``tools/ovfconverter``) has been added that supports
391 conversion between Ganeti and the `Open Virtualization Format
392 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
395 This relies on the ``qemu-img`` tool to convert the disk formats, so the
396 actual compatibility with other virtualization solutions depends on it.
401 The configuration query daemon (``ganeti-confd``) is now optional, and
402 has been rewritten in Haskell; whether to use the daemon at all, use the
403 Python (default) or the Haskell version is selectable at configure time
404 via the ``--enable-confd`` parameter, which can take one of the
405 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
406 daemon will result in a smaller footprint; for larger systems, we
407 welcome feedback on the Haskell version which might become the default
410 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
411 daemon running. The Python version doesn't implement the new call.
414 User interface changes
415 ~~~~~~~~~~~~~~~~~~~~~~
417 We have replaced the ``--disks`` option of ``gnt-instance
418 replace-disks`` with a more flexible ``--disk`` option, which allows
419 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
420 disk size and mode can be changed upon recreation (via ``gnt-instance
421 recreate-disks``, which accepts the same ``--disk`` option).
423 As many people are used to a ``show`` command, we have added that as an
424 alias to ``info`` on all ``gnt-*`` commands.
426 The ``gnt-instance grow-disk`` command has a new mode in which it can
427 accept the target size of the disk, instead of the delta; this can be
428 more safe since two runs in absolute mode will be idempotent, and
429 sometimes it's also easier to specify the desired size directly.
431 Also the handling of instances with regard to offline secondaries has
432 been improved. Instance operations should not fail because one of it's
433 secondary nodes is offline, even though it's safe to proceed.
435 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
436 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
437 minors to instance name.
442 RAPI coverage has improved, with (for example) new resources for
443 recreate-disks, node power-cycle, etc.
448 There is partial support for ``xl`` in the Xen hypervisor; feedback is
451 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
452 whether to still support Python 2.4 or move to Python 2.6 as minimum
455 Support for Fedora has been slightly improved; the provided example
456 init.d script should work better on it and the INSTALL file should
457 document the needed dependencies.
462 The deprecated ``QueryLocks`` LUXI request has been removed. Use
463 ``Query(what=QR_LOCK, ...)`` instead.
465 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
466 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
467 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
468 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
469 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
471 RAPI client: ``CertificateError`` now derives from
472 ``GanetiApiError``. This should make it more easy to handle Ganeti
475 Deprecation warnings due to PyCrypto/paramiko import in
476 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
477 make sure to run an up-to-date paramiko version, if you use this tool.
479 The QA scripts now depend on Python 2.5 or above (the main code base
480 still works with Python 2.4).
482 The configuration file (``config.data``) is now written without
483 indentation for performance reasons; if you want to edit it, it can be
484 re-formatted via ``tools/fmtjson``.
486 A number of bugs has been fixed in the cluster merge tool.
488 ``x509`` certification verification (used in import-export) has been
489 changed to allow the same clock skew as permitted by the cluster
490 verification. This will remove some rare but hard to diagnose errors in
497 *(Released Thu, 19 Jul 2012)*
499 Very few changes from rc4 to the final release, only bugfixes:
501 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
502 instance, fix CDROM booting for KVM instances)
503 - fixed node group modification of node parameters
504 - fixed issue in LUClusterVerifyGroup with multi-group clusters
505 - fixed generation of bash completion to ensure a stable ordering
512 *(Released Fri, 13 Jul 2012)*
514 Third release candidate for 2.6. The following changes were done from
517 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
518 - Fixed an inconsistency in the LUXI protocol with the provided
519 arguments (NOT backwards compatible)
520 - Fixed a bug with node groups ipolicy where ``min`` was greater than
521 the cluster ``std`` value
522 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
523 easier instance debugging on nodes (requires ``hconfd`` to work)
529 *(Released Tue, 03 Jul 2012)*
531 Second release candidate for 2.6. The following changes were done from
534 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
536 - Fixed a RAPI regression on missing beparams/memory
537 - Fixed redistribution of files on offline nodes
538 - Added possibility to run activate-disks even though secondaries are
539 offline. With this change it relaxes also the strictness on some other
540 commands which use activate disks internally:
541 * ``gnt-instance start|reboot|rename|backup|export``
542 - Made it possible to remove safely an instance if its secondaries are
544 - Made it possible to reinstall even though secondaries are offline
550 *(Released Mon, 25 Jun 2012)*
552 First release candidate for 2.6. The following changes were done from
555 - Fixed bugs with disk parameters and ``rbd`` templates as well as
557 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
558 behaviour. It supports now the modify operation
559 - ``hcheck`` implemented to analyze cluster health and possibility of
560 improving health by rebalance
561 - ``hbal`` has been improved in dealing with split instances
567 *(Released Mon, 11 Jun 2012)*
569 Second beta release of 2.6. The following changes were done from beta2
572 - Fixed ``daemon-util`` with non-root user models
573 - Fixed creation of plain instances with ``--no-wait-for-sync``
574 - Fix wrong iv_names when running ``cfgupgrade``
575 - Export more information in RAPI group queries
576 - Fixed bug when changing instance network interfaces
577 - Extended burnin to do NIC changes
578 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
579 - Changed default for DRBD barriers
580 - Fixed DRBD error reporting for syncer rate
581 - Verify the options on disk parameters
583 And of course various fixes to documentation and improved unittests and
590 *(Released Wed, 23 May 2012)*
592 First beta release of 2.6. The following changes were done from beta1 to
595 - integrated patch for distributions without ``start-stop-daemon``
596 - adapted example init.d script to work on Fedora
597 - fixed log handling in Haskell daemons
598 - adapted checks in the watcher for pycurl linked against libnss
599 - add partial support for ``xl`` instead of ``xm`` for Xen
600 - fixed a type issue in cluster verification
601 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
604 Plus integrated fixes from the 2.5 branch:
606 - fixed ``kvm-ifup`` to use ``/bin/bash``
607 - fixed parallel build failures
608 - KVM live migration when using a custom keymap
614 *(Released Tue, 24 Jul 2012)*
616 A small bugfix release, with no new features:
618 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
619 different default shell (e.g. Debian, Ubuntu)
620 - fixed KVM startup and live migration with a custom keymap (fixes Issue
621 243 and Debian bug #650664)
622 - fixed compatibility with KVM versions that don't support multiple boot
623 devices (fixes Issue 230 and Debian bug #624256)
625 Additionally, a few fixes were done to the build system (fixed parallel
626 build failures) and to the unittests (fixed race condition in test for
627 FileID functions, and the default enable/disable mode for QA test is now
634 *(Released Fri, 11 May 2012)*
636 A small bugfix release.
638 The main issues solved are on the topic of compatibility with newer LVM
641 - fixed parsing of ``lv_attr`` field
642 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
643 the ``--force`` flag is needed
645 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
646 to accept kernel 3.x too (was hardcoded to 2.6.*).
648 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
649 of hook results and that also made display of other errors suboptimal
650 was fixed; the code behaves now like 2.4 and earlier.
652 Another change in 2.5, the cleanup of the OS scripts environment, is too
653 aggressive: it removed even the ``PATH`` variable, which requires the OS
654 scripts to *always* need to export it. Since this is a bit too strict,
655 we now export a minimal PATH, the same that we export for hooks.
657 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
658 integrated into this release.
660 Finally, a few other miscellaneous changes were done (no new features,
661 just small improvements):
663 - Fix ``gnt-group --help`` display
664 - Fix hardcoded Xen kernel path
665 - Fix grow-disk handling of invalid units
666 - Update synopsis for ``gnt-cluster repair-disk-sizes``
667 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
673 *(Released Thu, 12 Apr 2012)*
675 Incompatible/important changes and bugfixes
676 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
678 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
679 resource's ``ip_check`` parameter changed from ``True`` to ``False``
680 to match the underlying LUXI interface.
681 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
682 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
683 not maintain backwards-compatibility as the underlying operation
684 changed in an incompatible way. The RAPI client can talk to old
685 servers, but it needs to be told so as the return value changed.
686 - When creating file-based instances via RAPI, the ``file_driver``
687 parameter no longer defaults to ``loop`` and must be specified.
688 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
690 - Support for the undocumented and deprecated RAPI instance creation
691 request format version 0 has been dropped. Use version 1, supported
692 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
693 - Pyparsing 1.4.6 or above is required, see :doc:`installation
694 documentation <install>`.
695 - The "cluster-verify" hooks are now executed per group by the
696 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
697 if you just run ``gnt-cluster verify``, which generates one opcode per
699 - The environment as passed to the OS scripts is cleared, and thus no
700 environment variables defined in the node daemon's environment will be
701 inherited by the scripts.
702 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
704 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
705 support operations involving multiple node groups.
706 - Offline nodes are ignored when failing over an instance.
707 - Support for KVM version 1.0, which changed the version reporting format
709 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
711 - ``Makefile`` is now compatible with Automake 1.11.2
712 - Includes all bugfixes made in the 2.4 series
717 - The ganeti-htools project has been merged into the ganeti-core source
718 tree and will be built as part of Ganeti (see :doc:`install-quick`).
719 - Implemented support for :doc:`shared storage <design-shared-storage>`.
720 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
721 GPT-style partition tables (requires `parted
722 <http://www.gnu.org/s/parted/>`_).
723 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
724 - Allowed adding tags on instance creation.
725 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
726 - Allow instances to be started in a paused state, enabling the user to
727 see the complete console output on boot using the console.
728 - Added new hypervisor flag to control default reboot behaviour
729 (``reboot_behavior``).
730 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
731 - Improved out-of-band management support:
733 - Added ``gnt-node health`` command reporting the health status of
735 - Added ``gnt-node power`` command to manage power status of nodes.
736 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
738 - Instance migration can fall back to failover if instance is not
740 - Filters can be used when listing nodes, instances, groups and locks;
741 see :manpage:`ganeti(7)` manpage.
742 - Added post-execution status as variables to :doc:`hooks <hooks>`
744 - Instance tags are exported/imported together with the instance.
745 - When given an explicit job ID, ``gnt-job info`` will work for archived
747 - Jobs can define dependencies on other jobs (not yet supported via
748 RAPI or command line, but used by internal commands and usable via
751 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
754 - Instance failover is now available as a RAPI resource
755 (``/2/instances/[instance_name]/failover``).
756 - ``gnt-instance info`` defaults to static information if primary node
758 - Opcodes have a new ``comment`` attribute.
759 - Added basic SPICE support to KVM hypervisor.
760 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
762 Node group improvements
763 ~~~~~~~~~~~~~~~~~~~~~~~
765 - ``gnt-cluster verify`` has been modified to check groups separately,
766 thereby improving performance.
767 - Node group support has been added to ``gnt-cluster verify-disks``,
768 which now operates per node group.
769 - Watcher has been changed to work better with node groups.
771 - One process and state file per node group.
772 - Slow watcher in one group doesn't block other group's watcher.
774 - Added new command, ``gnt-group evacuate``, to move all instances in a
775 node group to other groups.
776 - Added ``gnt-instance change-group`` to move an instance to another
778 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
779 per-group operations.
780 - Node groups can be tagged.
781 - Some operations switch from an exclusive to a shared lock as soon as
783 - Instance's primary and secondary nodes' groups are now available as
784 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
785 and ``snodes.group.uuid``).
790 - Numerous updates to documentation and manpages.
792 - :doc:`RAPI <rapi>` documentation now has detailed parameter
794 - Some opcode/job results are now also documented, see :doc:`RAPI
797 - A lockset's internal lock is now also visible in lock monitor.
798 - Log messages from job queue workers now contain information about the
799 opcode they're processing.
800 - ``gnt-instance console`` no longer requires the instance lock.
801 - A short delay when waiting for job changes reduces the number of LUXI
802 requests significantly.
803 - DRBD metadata volumes are overwritten with zeros during disk creation.
804 - Out-of-band commands no longer acquire the cluster lock in exclusive
806 - ``devel/upload`` now uses correct permissions for directories.
812 *(Released Fri, 23 Mar 2012)*
814 This was the sixth release candidate of the 2.5 series.
820 *(Released Mon, 9 Jan 2012)*
822 This was the fifth release candidate of the 2.5 series.
828 *(Released Thu, 27 Oct 2011)*
830 This was the fourth release candidate of the 2.5 series.
836 *(Released Wed, 26 Oct 2011)*
838 This was the third release candidate of the 2.5 series.
844 *(Released Tue, 18 Oct 2011)*
846 This was the second release candidate of the 2.5 series.
852 *(Released Tue, 4 Oct 2011)*
854 This was the first release candidate of the 2.5 series.
860 *(Released Wed, 31 Aug 2011)*
862 This was the third beta release of the 2.5 series.
868 *(Released Mon, 22 Aug 2011)*
870 This was the second beta release of the 2.5 series.
876 *(Released Fri, 12 Aug 2011)*
878 This was the first beta release of the 2.5 series.
884 *(Released Thu, 27 Oct 2011)*
886 - Fixed bug when parsing command line parameter values ending in
888 - Fixed assertion error after unclean master shutdown
889 - Disable HTTP client pool for RPC, significantly reducing memory usage
891 - Fixed queue archive creation with wrong permissions
897 *(Released Tue, 23 Aug 2011)*
901 - Fixed documentation for importing with ``--src-dir`` option
902 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
903 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
909 *(Released Fri, 5 Aug 2011)*
911 Many bug-fixes and a few small features:
913 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
914 issues when you tried to add an instance with two MAC addresses in one
916 - KVM: fixed per-instance stored UID value
917 - KVM: configure bridged NICs at migration start
918 - KVM: Fix a bug where instance will not start with never KVM versions
920 - Added OS search path to ``gnt-cluster info``
921 - Fixed an issue with ``file_storage_dir`` where you were forced to
922 provide an absolute path, but the documentation states it is a
923 relative path, the documentation was right
924 - Added a new parameter to instance stop/start called ``--no-remember``
925 that will make the state change to not be remembered
926 - Implemented ``no_remember`` at RAPI level
927 - Improved the documentation
928 - Node evacuation: don't call IAllocator if node is already empty
929 - Fixed bug in DRBD8 replace disks on current nodes
930 - Fixed bug in recreate-disks for DRBD instances
931 - Moved assertion checking locks in ``gnt-instance replace-disks``
932 causing it to abort with not owning the right locks for some situation
933 - Job queue: Fixed potential race condition when cancelling queued jobs
934 - Fixed off-by-one bug in job serial generation
935 - ``gnt-node volumes``: Fix instance names
936 - Fixed aliases in bash completion
937 - Fixed a bug in reopening log files after being sent a SIGHUP
938 - Added a flag to burnin to allow specifying VCPU count
939 - Bugfixes to non-root Ganeti configuration
945 *(Released Thu, 12 May 2011)*
947 Many bug-fixes and a few new small features:
949 - Fixed a bug related to log opening failures
950 - Fixed a bug in instance listing with orphan instances
951 - Fixed a bug which prevented resetting the cluster-level node parameter
952 ``oob_program`` to the default
953 - Many fixes related to the ``cluster-merge`` tool
954 - Fixed a race condition in the lock monitor, which caused failures
955 during (at least) creation of many instances in parallel
956 - Improved output for gnt-job info
957 - Removed the quiet flag on some ssh calls which prevented debugging
959 - Improved the N+1 failure messages in cluster verify by actually
960 showing the memory values (needed and available)
961 - Increased lock attempt timeouts so that when executing long operations
962 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
963 too early and thus prevent the use of the 'fair' mechanism
964 - Changed instance query data (``gnt-instance info``) to not acquire
965 locks unless needed, thus allowing its use on locked instance if only
966 static information is asked for
967 - Improved behaviour with filesystems that do not support rename on an
969 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
970 kept locks on all nodes during the wipe, which is unneeded
971 - Fixed ``gnt-watcher`` handling of errors during hooks execution
972 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
973 10GiB) which caused the wipe to fail right at the end in some cases
974 - Fixed master IP activation when doing master failover with no-voting
975 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
976 the master node itself
977 - Fixed potential data-loss in under disk full conditions, where Ganeti
978 wouldn't check correctly the return code and would consider
979 partially-written files 'correct'
980 - Fixed bug related to multiple VGs and DRBD disk replacing
981 - Added new disk parameter ``metavg`` that allows placement of the meta
982 device for DRBD in a different volume group
983 - Fixed error handling in the node daemon when the system libc doesn't
984 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
985 - Fixed lock release during replace-disks, which kept cluster-wide locks
986 when doing disk replaces with an iallocator script
987 - Added check for missing bridges in cluster verify
988 - Handle EPIPE errors while writing to the terminal better, so that
989 piping the output to e.g. ``less`` doesn't cause a backtrace
990 - Fixed rare case where a ^C during Luxi calls could have been
991 interpreted as server errors, instead of simply terminating
992 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
994 - Added a few more parameters to the KVM hypervisor, allowing a second
995 CDROM, custom disk type for CDROMs and a floppy image
996 - Removed redundant message in instance rename when the name is given
998 - Added option to ``gnt-instance recreate-disks`` to allow creating the
999 disks on new nodes, allowing recreation when the original instance
1000 nodes are completely gone
1001 - Added option when converting disk templates to DRBD to skip waiting
1002 for the resync, in order to make the instance available sooner
1003 - Added two new variables to the OS scripts environment (containing the
1005 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1006 to allow use of ``pvgrub`` as bootloader
1007 - Changed the instance memory modifications to only check out-of-memory
1008 conditions on memory increases, and turned the secondary node warnings
1009 into errors (they can still be overridden via ``--force``)
1010 - Fixed the handling of a corner case when the Python installation gets
1011 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1012 to execute a command that doesn't exist
1013 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1014 node of the instance returned failures during instance shutdown; this
1015 adds the option ``--ignore-consistency`` to gnt-instance move
1017 And as usual, various improvements to the error messages, documentation
1024 *(Released Wed, 09 Mar 2011)*
1026 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1027 the RAPI users file if run twice (even with ``--dry-run``).
1029 The release fixes that bug (nothing else changed).
1035 *(Released Mon, 07 Mar 2011)*
1037 Final 2.4.0 release. Just a few small fixes:
1039 - Fixed RAPI node evacuate
1040 - Fixed the kvm-ifup script
1041 - Fixed internal error handling for special job cases
1042 - Updated man page to specify the escaping feature for options
1048 *(Released Mon, 28 Feb 2011)*
1050 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1051 that this feature wiped the disks of the wrong instance, leading to loss
1056 - Fixed title of query field containing instance name
1057 - Expanded the glossary in the documentation
1058 - Fixed one unittest (internal issue)
1064 *(Released Mon, 21 Feb 2011)*
1066 A number of bug fixes plus just a couple functionality changes.
1068 On the user-visible side, the ``gnt-* list`` command output has changed
1069 with respect to "special" field states. The current rc1 style of display
1070 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1071 the default output mode special fields states are displayed as follows:
1073 - Offline resource: ``*``
1074 - Unavailable/not applicable: ``-``
1075 - Data missing (RPC failure): ``?``
1076 - Unknown field: ``??``
1078 Another user-visible change is the addition of ``--force-join`` to
1083 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1084 - Fixed regression in RAPI/instance reinstall where all parameters were
1085 required (instead of optional)
1086 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1087 - Fixed iallocator usage (offline nodes were not considered offline)
1088 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1089 - Fixed hypervisor and OS parameter validation with respect to
1090 non-vm_capable nodes
1091 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1093 - Fixed ``tools/listrunner`` with respect to agent-based usage
1099 *(Released Fri, 4 Feb 2011)*
1101 Many changes and fixes since the beta1 release. While there were some
1102 internal changes, the code has been mostly stabilised for the RC
1105 Note: the dumb allocator was removed in this release, as it was not kept
1106 up-to-date with the IAllocator protocol changes. It is recommended to
1107 use the ``hail`` command from the ganeti-htools package.
1109 Note: the 2.4 and up versions of Ganeti are not compatible with the
1110 0.2.x branch of ganeti-htools. You need to upgrade to
1111 ganeti-htools-0.3.0 (or later).
1113 Regressions fixed from 2.3
1114 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1116 - Fixed the ``gnt-cluster verify-disks`` command
1117 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1119 - Fixed disk adoption breakage
1120 - Fixed wrong headers in instance listing for field aliases
1125 - Fixed corner case in KVM handling of NICs
1126 - Fixed many cases of wrong handling of non-vm_capable nodes
1127 - Fixed a bug where a missing instance symlink was not possible to
1128 recreate with any ``gnt-*`` command (now ``gnt-instance
1129 activate-disks`` does it)
1130 - Fixed the volume group name as reported by ``gnt-cluster
1132 - Increased timeouts for the import-export code, hopefully leading to
1133 fewer aborts due network or instance timeouts
1134 - Fixed bug in ``gnt-node list-storage``
1135 - Fixed bug where not all daemons were started on cluster
1136 initialisation, but only at the first watcher run
1137 - Fixed many bugs in the OOB implementation
1138 - Fixed watcher behaviour in presence of instances with offline
1140 - Fixed instance list output for instances running on the wrong node
1141 - a few fixes to the cluster-merge tool, but it still cannot merge
1142 multi-node groups (currently it is not recommended to use this tool)
1148 - Improved network configuration for the KVM hypervisor
1149 - Added e1000 as a supported NIC for Xen-HVM
1150 - Improved the lvmstrap tool to also be able to use partitions, as
1151 opposed to full disks
1152 - Improved speed of disk wiping (the cluster parameter
1153 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1154 of instance creations
1155 - Added documentation for the OS parameters
1156 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1157 hypervisor is not responding
1158 - Added display of blacklisted and hidden OS information in
1159 ``gnt-cluster info``
1160 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1161 NIC and node parameters, which might create problems with currently
1162 invalid (but undetected) configuration files, but prevents validation
1163 failures when unrelated parameters are modified
1164 - Changed cluster initialisation to wait for the master daemon to become
1166 - Expanded the RAPI interface:
1168 - Added config redistribution resource
1169 - Added activation/deactivation of instance disks
1170 - Added export of console information
1172 - Implemented log file reopening on SIGHUP, which allows using
1173 logrotate(8) for the Ganeti log files
1174 - Added a basic OOB helper script as an example
1180 *(Released Fri, 14 Jan 2011)*
1185 - Fixed timezone issues when formatting timestamps
1186 - Added support for node groups, available via ``gnt-group`` and other
1188 - Added out-of-band framework and management, see :doc:`design
1189 document <design-oob>`
1190 - Removed support for roman numbers from ``gnt-node list`` and
1191 ``gnt-instance list``.
1192 - Allowed modification of master network interface via ``gnt-cluster
1193 modify --master-netdev``
1194 - Accept offline secondaries while shutting down instance disks
1195 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1196 - Added support for multiple LVM volume groups
1197 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1199 - Added commands to list available fields:
1201 - ``gnt-node list-fields``
1202 - ``gnt-group list-fields``
1203 - ``gnt-instance list-fields``
1205 - Updated documentation and man pages
1210 - Moved ``rapi_users`` file into separate directory, now named
1211 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1213 - Added new tool for running commands on many machines,
1214 ``tools/ganeti-listrunner``
1215 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1216 improving the ``gnt-instance console`` output
1217 - Allowed customisation of disk index separator at ``configure`` time
1218 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1219 - Added support for non-partitioned md disks in ``lvmstrap``
1220 - Added script to gracefully power off KVM instances
1221 - Split ``utils`` module into smaller parts
1222 - Changed query operations to return more detailed information, e.g.
1223 whether an information is unavailable due to an offline node. To use
1224 this new functionality, the LUXI call ``Query`` must be used. Field
1225 information is now stored by the master daemon and can be retrieved
1226 using ``QueryFields``. Instances, nodes and groups can also be queried
1227 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1228 exposed via RAPI). The following commands make use of this
1229 infrastructure change:
1231 - ``gnt-group list``
1232 - ``gnt-group list-fields``
1234 - ``gnt-node list-fields``
1235 - ``gnt-instance list``
1236 - ``gnt-instance list-fields``
1237 - ``gnt-debug locks``
1242 - New RAPI resources (see :doc:`rapi`):
1246 - ``/2/groups/[group_name]``
1247 - ``/2/groups/[group_name]/assign-nodes``
1248 - ``/2/groups/[group_name]/modify``
1249 - ``/2/groups/[group_name]/rename``
1250 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1254 - Implemented ``no_install`` for instance creation
1255 - Implemented OS parameters for instance reinstallation, allowing
1256 use of special settings on reinstallation (e.g. for preserving data)
1261 - Added IPv6 support in import/export
1262 - Pause DRBD synchronization while wiping disks on instance creation
1263 - Updated unittests and QA scripts
1264 - Improved network parameters passed to KVM
1265 - Converted man pages from docbook to reStructuredText
1271 *(Released Mon, 20 Dec 2010)*
1273 Released version 2.3.1~rc1 without any changes.
1279 *(Released Wed, 1 Dec 2010)*
1281 - impexpd: Disable OpenSSL compression in socat if possible (backport
1282 from master, commit e90739d625b, see :doc:`installation guide
1283 <install-quick>` for details)
1284 - Changed unittest coverage report to exclude test scripts
1285 - Added script to check version format
1291 *(Released Wed, 1 Dec 2010)*
1293 Released version 2.3.0~rc1 without any changes.
1299 *(Released Fri, 19 Nov 2010)*
1301 A number of bugfixes and documentation updates:
1303 - Update ganeti-os-interface documentation
1304 - Fixed a bug related to duplicate MACs or similar items which should be
1306 - Fix breakage in OS state modify
1307 - Reinstall instance: disallow offline secondaries (fixes bug related to
1308 OS changing but reinstall failing)
1309 - plus all the other fixes between 2.2.1 and 2.2.2
1315 *(Released Tue, 2 Nov 2010)*
1317 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1318 - Fixed master failover race with watcher
1319 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1321 - Accept previously stopped instance for export with instance removal
1322 - Simplify and extend the environment variables for instance OS scripts
1323 - Added new node flags, ``master_capable`` and ``vm_capable``
1324 - Added optional instance disk wiping prior during allocation. This is a
1325 cluster-wide option and can be set/modified using
1326 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1327 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1328 :doc:`install-quick`
1329 - Added a new watcher option (``--ignore-pause``)
1330 - Added option to ignore offline node on instance start/stop
1331 (``--ignore-offline``)
1332 - Allow overriding OS parameters with ``gnt-instance reinstall``
1333 - Added ability to change node's secondary IP address using ``gnt-node
1335 - Implemented privilege separation for all daemons except
1336 ``ganeti-noded``, see ``configure`` options
1337 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1339 - Implemented job priorities (see ``ganeti(7)`` manpage)
1340 - Ignore failures while shutting down instances during failover from
1342 - Exit daemon's bootstrap process only once daemon is ready
1343 - Export more information via ``LUInstanceQuery``/remote API
1344 - Improved documentation, QA and unittests
1345 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1346 restart if the file was created or changed
1347 - Added LUXI protocol version sent with each request and response,
1348 allowing detection of server/client mismatches
1349 - Moved the Python scripts among gnt-* and ganeti-* into modules
1350 - Moved all code related to setting up SSH to an external script,
1352 - Infrastructure changes for node group support in future versions
1358 *(Released Fri, 19 Nov 2010)*
1360 A few small bugs fixed, and some improvements to the build system:
1362 - Fix documentation regarding conversion to drbd
1363 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1365 - Fix error handling in node modify with multiple changes
1366 - Allow remote imports without checked names
1372 *(Released Tue, 19 Oct 2010)*
1374 - Disable SSL session ID cache in RPC client
1380 *(Released Thu, 14 Oct 2010)*
1382 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1383 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1385 - Fix problems with interaction between Curl and Python's HTTP server,
1386 resulting in increased speed in many RPC calls
1387 - Improve our release script to prevent breakage with older aclocal and
1394 *(Released Thu, 7 Oct 2010)*
1396 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1397 value retrieved from master
1398 - Added support for blacklisted or hidden OS definitions
1399 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1400 - Added support for -mem-path in KVM hypervisor abstraction layer
1401 - Allow overriding instance parameters in tool for inter-cluster
1402 instance moves (``tools/move-instance``)
1403 - Improved opcode summaries (e.g. in ``gnt-job list``)
1404 - Improve consistency of OS listing by sorting it
1405 - Documentation updates
1411 *(Released Fri, 8 Oct 2010)*
1413 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1419 *(Released Mon, 4 Oct 2010)*
1421 - Fixed regression in ``gnt-instance rename``
1427 *(Released Wed, 22 Sep 2010)*
1429 - Fixed OS_VARIANT variable for OS scripts
1430 - Fixed cluster tag operations via RAPI
1431 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1432 - Disabled RAPI CA checks in watcher
1438 *(Released Mon, 23 Aug 2010)*
1440 - Support DRBD versions of the format "a.b.c.d"
1442 - Re-introduce support for usage from multiple threads in RAPI client
1443 - Instance renames and modify via RAPI
1444 - Work around race condition between processing and archival in job
1446 - Mark opcodes following failed one as failed, too
1447 - Job field ``lock_status`` was removed due to difficulties making it
1448 work with the changed job queue in Ganeti 2.2; a better way to monitor
1449 locks is expected for a later 2.2.x release
1450 - Fixed dry-run behaviour with many commands
1451 - Support ``ssh-agent`` again when adding nodes
1452 - Many additional bugfixes
1458 *(Released Fri, 30 Jul 2010)*
1460 Important change: the internal RPC mechanism between Ganeti nodes has
1461 changed from using a home-grown http library (based on the Python base
1462 libraries) to use the PycURL library. This requires that PycURL is
1463 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1464 against GnuTLS by default. cURL's support for GnuTLS had known issues
1465 before cURL 7.21.0 and we recommend using the latest cURL release or
1466 linking against OpenSSL. Most other distributions already link PycURL
1467 and cURL against OpenSSL. The command::
1469 python -c 'import pycurl; print pycurl.version'
1471 can be used to determine the libraries PycURL and cURL are linked
1474 Other significant changes:
1476 - Rewrote much of the internals of the job queue, in order to achieve
1477 better parallelism; this decouples job query operations from the job
1478 processing, and it should allow much nicer behaviour of the master
1479 daemon under load, and it also has uncovered some long-standing bugs
1480 related to the job serialisation (now fixed)
1481 - Added a default iallocator setting to the cluster parameters,
1482 eliminating the need to always pass nodes or an iallocator for
1483 operations that require selection of new node(s)
1484 - Added experimental support for the LXC virtualization method
1485 - Added support for OS parameters, which allows the installation of
1486 instances to pass parameter to OS scripts in order to customise the
1488 - Added a hypervisor parameter controlling the migration type (live or
1489 non-live), since hypervisors have various levels of reliability; this
1490 has renamed the 'live' parameter to 'mode'
1491 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1492 logical volumes, meaning that cluster verify will ignore them and not
1493 flag their presence as errors
1494 - The watcher will now reset the error count for failed instances after
1495 8 hours, thus allowing self-healing if the problem that caused the
1496 instances to be down/fail to start has cleared in the meantime
1497 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1498 check for, and warn, if the drbd module parameter ``usermode_helper``
1499 is not consistent with the cluster-wide setting; this is needed to
1500 make diagnose easier of failed drbd creations
1501 - Started adding base IPv6 support, but this is not yet
1502 enabled/available for use
1503 - Rename operations (cluster, instance) will now return the new name,
1504 which is especially useful if a short name was passed in
1505 - Added support for instance migration in RAPI
1506 - Added a tool to pre-configure nodes for the SSH setup, before joining
1507 them to the cluster; this will allow in the future a simplified model
1508 for node joining (but not yet fully enabled in 2.2); this needs the
1509 paramiko python library
1510 - Fixed handling of name-resolving errors
1511 - Fixed consistency of job results on the error path
1512 - Fixed master-failover race condition when executed multiple times in
1514 - Fixed many bugs related to the job queue (mostly introduced during the
1515 2.2 development cycle, so not all are impacting 2.1)
1516 - Fixed instance migration with missing disk symlinks
1517 - Fixed handling of unknown jobs in ``gnt-job archive``
1518 - And many other small fixes/improvements
1522 - Enhanced both the unittest and the QA coverage
1523 - Switched the opcode validation to a generic model, and extended the
1524 validation to all opcode parameters
1525 - Changed more parts of the code that write shell scripts to use the
1527 - Switched the master daemon to use the asyncore library for the Luxi
1534 *(Released Thu, 17 Jun 2010)*
1536 - Added tool (``move-instance``) and infrastructure to move instances
1537 between separate clusters (see :doc:`separate documentation
1538 <move-instance>` and :doc:`design document <design-2.2>`)
1539 - Added per-request RPC timeout
1540 - RAPI now requires a Content-Type header for requests with a body (e.g.
1541 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1542 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1543 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1545 - Implemented initial support for running Ganeti daemons as separate
1546 users, see configure-time flags ``--with-user-prefix`` and
1547 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1549 - Instances can be removed after export (``gnt-backup export
1550 --remove-instance``)
1551 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1552 key (instead of 1024 bit)
1553 - Added new cluster configuration file for cluster domain secret
1554 - Import/export now use SSL instead of SSH
1555 - Added support for showing estimated time when exporting an instance,
1556 see the ``ganeti-os-interface(7)`` manpage and look for
1563 *(Released Tue, 16 Nov 2010)*
1565 Some more bugfixes. Unless critical bugs occur, this will be the last
1568 - Fix case of MAC special-values
1569 - Fix mac checker regex
1570 - backend: Fix typo causing "out of range" error
1571 - Add missing --units in gnt-instance list man page
1577 *(Released Tue, 24 Aug 2010)*
1580 - Don't ignore secondary node silently on non-mirrored disk templates
1582 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1583 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1584 - Properly document the usb_mouse parameter
1585 - Fix path in ganeti-rapi(8) (issue 116)
1586 - Adjust error message when the ganeti user's .ssh directory is
1588 - Add same-node-check when changing the disk template to drbd
1594 *(Released Fri, 16 Jul 2010)*
1597 - Add an option to only select some reboot types during qa/burnin.
1598 (on some hypervisors consequent reboots are not supported)
1599 - Fix infrequent race condition in master failover. Sometimes the old
1600 master ip address would be still detected as up for a short time
1601 after it was removed, causing failover to fail.
1602 - Decrease mlockall warnings when the ctypes module is missing. On
1603 Python 2.4 we support running even if no ctypes module is installed,
1604 but we were too verbose about this issue.
1605 - Fix building on old distributions, on which man doesn't have a
1607 - Fix RAPI not to ignore the MAC address on instance creation
1608 - Implement the old instance creation format in the RAPI client.
1614 *(Released Thu, 01 Jul 2010)*
1616 A small bugfix release:
1617 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1618 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1619 a non-existing option
1620 - Fix instance create: the --force-variant option was ignored
1621 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1622 - Fix modify node storage with non-FQDN arguments
1623 - Fix RAPI client to authenticate under Python 2.6 when used
1624 for more than 5 requests needing authentication
1625 - Fix gnt-instance modify -t (storage) giving a wrong error message
1626 when converting a non-shutdown drbd instance to plain
1632 *(Released Fri, 18 Jun 2010)*
1634 A small bugfix release:
1636 - Fix live migration of KVM instances started with older Ganeti
1637 versions which had fewer hypervisor parameters
1638 - Fix gnt-instance grow-disk on down instances
1639 - Fix an error-reporting bug during instance migration
1640 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1641 silently ignoring broken ones
1642 - Fix an RPC error reporting bug affecting, for example, RAPI client
1644 - Fix bug triggered by different API version os-es on different nodes
1645 - Fix a bug in instance startup with custom hvparams: OS level
1646 parameters would fail to be applied.
1647 - Fix the RAPI client under Python 2.6 (but more work is needed to
1648 make it work completely well with OpenSSL)
1649 - Fix handling of errors when resolving names from DNS
1655 *(Released Thu, 3 Jun 2010)*
1657 A medium sized development cycle. Some new features, and some
1658 fixes/small improvements/cleanups.
1660 Significant features
1661 ~~~~~~~~~~~~~~~~~~~~
1663 The node deamon now tries to mlock itself into memory, unless the
1664 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1665 its logs, and falls back to console logging. This allows emergency
1666 features such as ``gnt-node powercycle`` to work even in the event of a
1667 broken node disk (tested offlining the disk hosting the node's
1668 filesystem and dropping its memory caches; don't try this at home)
1670 KVM: add vhost-net acceleration support. It can be tested with a new
1671 enough version of the kernel and of qemu-kvm.
1673 KVM: Add instance chrooting feature. If you use privilege dropping for
1674 your VMs you can also now force them to chroot to an empty directory,
1675 before starting the emulated guest.
1677 KVM: Add maximum migration bandwith and maximum downtime tweaking
1678 support (requires a new-enough version of qemu-kvm).
1680 Cluster verify will now warn if the master node doesn't have the master
1681 ip configured on it.
1683 Add a new (incompatible) instance creation request format to RAPI which
1684 supports all parameters (previously only a subset was supported, and it
1685 wasn't possible to extend the old format to accomodate all the new
1686 features. The old format is still supported, and a client can check for
1687 this feature, before using it, by checking for its presence in the
1688 ``features`` RAPI resource.
1690 Now with ancient latin support. Try it passing the ``--roman`` option to
1691 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1692 (requires the python-roman module to be installed, in order to work).
1697 As usual many internal code refactorings, documentation updates, and
1700 - Lots of improvements and cleanups to the experimental Remote API
1701 (RAPI) client library.
1702 - A new unit test suite for the core daemon libraries.
1703 - A fix to creating missing directories makes sure the umask is not
1704 applied anymore. This enforces the same directory permissions
1706 - Better handling terminating daemons with ctrl+c (used when running
1707 them in debugging mode).
1708 - Fix a race condition in live migrating a KVM instance, when stat()
1709 on the old proc status file returned EINVAL, which is an unexpected
1711 - Fixed manpage checking with newer man and utf-8 charachters. But now
1712 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1718 *(Released Fri, 7 May 2010)*
1720 Fix a bug which prevented untagged KVM instances from starting.
1726 *(Released Fri, 7 May 2010)*
1728 Another release with a long development cycle, during which many
1729 different features were added.
1731 Significant features
1732 ~~~~~~~~~~~~~~~~~~~~
1734 The KVM hypervisor now can run the individual instances as non-root, to
1735 reduce the impact of a VM being hijacked due to bugs in the
1736 hypervisor. It is possible to run all instances as a single (non-root)
1737 user, to manually specify a user for each instance, or to dynamically
1738 allocate a user out of a cluster-wide pool to each instance, with the
1739 guarantee that no two instances will run under the same user ID on any
1742 An experimental RAPI client library, that can be used standalone
1743 (without the other Ganeti libraries), is provided in the source tree as
1744 ``lib/rapi/client.py``. Note this client might change its interface in
1745 the future, as we iterate on its capabilities.
1747 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1748 replace the cluster's certificates and crypto keys. This might help in
1749 case they have been compromised, or have simply expired.
1751 A new disk option for instance creation has been added that allows one
1752 to "adopt" currently existing logical volumes, with data
1753 preservation. This should allow easier migration to Ganeti from
1754 unmanaged (or managed via other software) instances.
1756 Another disk improvement is the possibility to convert between redundant
1757 (DRBD) and plain (LVM) disk configuration for an instance. This should
1758 allow better scalability (starting with one node and growing the
1759 cluster, or shrinking a two-node cluster to one node).
1761 A new feature that could help with automated node failovers has been
1762 implemented: if a node sees itself as offline (by querying the master
1763 candidates), it will try to shutdown (hard) all instances and any active
1764 DRBD devices. This reduces the risk of duplicate instances if an
1765 external script automatically failovers the instances on such nodes. To
1766 enable this, the cluster parameter ``maintain_node_health`` should be
1767 enabled; in the future this option (per the name) will enable other
1768 automatic maintenance features.
1770 Instance export/import now will reuse the original instance
1771 specifications for all parameters; that means exporting an instance,
1772 deleting it and the importing it back should give an almost identical
1773 instance. Note that the default import behaviour has changed from
1774 before, where it created only one NIC; now it recreates the original
1777 Cluster verify has added a few new checks: SSL certificates validity,
1778 /etc/hosts consistency across the cluster, etc.
1783 As usual, many internal changes were done, documentation fixes,
1786 - Fixed cluster initialization with disabled cluster storage (regression
1787 introduced in 2.1.1)
1788 - File-based storage supports growing the disks
1789 - Fixed behaviour of node role changes
1790 - Fixed cluster verify for some corner cases, plus a general rewrite of
1791 cluster verify to allow future extension with more checks
1792 - Fixed log spamming by watcher and node daemon (regression introduced
1794 - Fixed possible validation issues when changing the list of enabled
1796 - Fixed cleanup of /etc/hosts during node removal
1797 - Fixed RAPI response for invalid methods
1798 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1799 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1800 from ide disks, etc.)
1801 - Allow OS changes without re-installation (to record a changed OS
1802 outside of Ganeti, or to allow OS renames)
1803 - Allow instance creation without OS installation (useful for example if
1804 the OS will be installed manually, or restored from a backup not in
1806 - Implemented option to make cluster ``copyfile`` use the replication
1808 - Added list of enabled hypervisors to ssconf (possibly useful for
1810 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1812 - A partial form of node re-IP is possible via node readd, which now
1813 allows changed node primary IP
1814 - Command line utilities now show an informational message if the job is
1816 - The logs of the master daemon now show the PID/UID/GID of the
1823 *(Released Fri, 12 Mar 2010)*
1825 During the 2.1.0 long release candidate cycle, a lot of improvements and
1826 changes have accumulated with were released later as 2.1.1.
1831 The node evacuate command (``gnt-node evacuate``) was significantly
1832 rewritten, and as such the IAllocator protocol was changed - a new
1833 request type has been added. This unfortunate change during a stable
1834 series is designed to improve performance of node evacuations; on
1835 clusters with more than about five nodes and which are well-balanced,
1836 evacuation should proceed in parallel for all instances of the node
1837 being evacuated. As such, any existing IAllocator scripts need to be
1838 updated, otherwise the above command will fail due to the unknown
1839 request. The provided "dumb" allocator has not been updated; but the
1840 ganeti-htools package supports the new protocol since version 0.2.4.
1842 Another important change is increased validation of node and instance
1843 names. This might create problems in special cases, if invalid host
1844 names are being used.
1846 Also, a new layer of hypervisor parameters has been added, that sits at
1847 OS level between the cluster defaults and the instance ones. This allows
1848 customisation of virtualization parameters depending on the installed
1849 OS. For example instances with OS 'X' may have a different KVM kernel
1850 (or any other parameter) than the cluster defaults. This is intended to
1851 help managing a multiple OSes on the same cluster, without manual
1852 modification of each instance's parameters.
1854 A tool for merging clusters, ``cluster-merge``, has been added in the
1855 tools sub-directory.
1860 - Improved the int/float conversions that should make the code more
1861 robust in face of errors from the node daemons
1862 - Fixed the remove node code in case of internal configuration errors
1863 - Fixed the node daemon behaviour in face of inconsistent queue
1864 directory (e.g. read-only file-system where we can't open the files
1866 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1867 now it either aborts cleanly or, if given the new "auto_promote"
1868 parameter, will automatically promote other nodes as needed
1869 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1870 completely prevent Ganeti from working
1871 - Fixed bug for instance export when not all disks were successfully
1873 - Fixed behaviour of node add when the new node is slow in starting up
1875 - Fixed handling of signals in the LUXI client, which should improve
1876 behaviour of command-line scripts
1877 - Added checks for invalid node/instance names in the configuration (now
1878 flagged during cluster verify)
1879 - Fixed watcher behaviour for disk activation errors
1880 - Fixed two potentially endless loops in http library, which led to the
1881 RAPI daemon hanging and consuming 100% CPU in some cases
1882 - Fixed bug in RAPI daemon related to hashed passwords
1883 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1885 - Enhanced compatibility with non-Debian OSes, but not using absolute
1886 path in some commands and allowing customisation of the ssh
1887 configuration directory
1888 - Fixed possible future issue with new Python versions by abiding to the
1889 proper use of ``__slots__`` attribute on classes
1890 - Added checks that should prevent directory traversal attacks
1891 - Many documentation fixes based on feedback from users
1896 - Added an "early_release" more for instance replace disks and node
1897 evacuate, where we release locks earlier and thus allow higher
1898 parallelism within the cluster
1899 - Added watcher hooks, intended to allow the watcher to restart other
1900 daemons (e.g. from the ganeti-nbma project), but they can be used of
1901 course for any other purpose
1902 - Added a compile-time disable for DRBD barriers, to increase
1903 performance if the administrator trusts the power supply or the
1904 storage system to not lose writes
1905 - Added the option of using syslog for logging instead of, or in
1906 addition to, Ganeti's own log files
1907 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1908 can indeed boot from a paravirtual NIC
1909 - Added a generic debug level for many operations; while this is not
1910 used widely yet, it allows one to pass the debug value all the way to
1912 - Enhanced the hooks environment for instance moves (failovers,
1913 migrations) where the primary/secondary nodes changed during the
1914 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1915 - Enhanced data validations for many user-supplied values; one important
1916 item is the restrictions imposed on instance and node names, which
1917 might reject some (invalid) host names
1918 - Add a configure-time option to disable file-based storage, if it's not
1919 needed; this allows greater security separation between the master
1920 node and the other nodes from the point of view of the inter-node RPC
1922 - Added user notification in interactive tools if job is waiting in the
1923 job queue or trying to acquire locks
1924 - Added log messages when a job is waiting for locks
1925 - Added filtering by node tags in instance operations which admit
1926 multiple instances (start, stop, reboot, reinstall)
1927 - Added a new tool for cluster mergers, ``cluster-merge``
1928 - Parameters from command line which are of the form ``a=b,c=d`` can now
1929 use backslash escapes to pass in values which contain commas,
1930 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1932 - For KVM, the instance name is the first parameter passed to KVM, so
1933 that it's more visible in the process list
1939 *(Released Tue, 2 Mar 2010)*
1941 Ganeti 2.1 brings many improvements with it. Major changes:
1943 - Added infrastructure to ease automated disk repairs
1944 - Added new daemon to export configuration data in a cheaper way than
1945 using the remote API
1946 - Instance NICs can now be routed instead of being associated with a
1948 - Improved job locking logic to reduce impact of jobs acquiring multiple
1949 locks waiting for other long-running jobs
1951 In-depth implementation details can be found in the Ganeti 2.1 design
1957 - Added chroot hypervisor
1958 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1960 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1961 ``bootloader_path`` and ``bootloader_args``)
1962 - Added the ``use_localtime`` option for the xen-hvm and kvm
1963 hypervisors, and the default value for this has changed to false (in
1964 2.0 xen-hvm always enabled it)
1965 - Added luxi call to submit multiple jobs in one go
1966 - Added cluster initialization option to not modify ``/etc/hosts``
1968 - Added network interface parameters
1969 - Added dry run mode to some LUs
1970 - Added RAPI resources:
1972 - ``/2/instances/[instance_name]/info``
1973 - ``/2/instances/[instance_name]/replace-disks``
1974 - ``/2/nodes/[node_name]/evacuate``
1975 - ``/2/nodes/[node_name]/migrate``
1976 - ``/2/nodes/[node_name]/role``
1977 - ``/2/nodes/[node_name]/storage``
1978 - ``/2/nodes/[node_name]/storage/modify``
1979 - ``/2/nodes/[node_name]/storage/repair``
1981 - Added OpCodes to evacuate or migrate all instances on a node
1982 - Added new command to list storage elements on nodes (``gnt-node
1983 list-storage``) and modify them (``gnt-node modify-storage``)
1984 - Added new ssconf files with master candidate IP address
1985 (``ssconf_master_candidates_ips``), node primary IP address
1986 (``ssconf_node_primary_ips``) and node secondary IP address
1987 (``ssconf_node_secondary_ips``)
1988 - Added ``ganeti-confd`` and a client library to query the Ganeti
1989 configuration via UDP
1990 - Added ability to run hooks after cluster initialization and before
1992 - Added automatic mode for disk replace (``gnt-instance replace-disks
1994 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1995 after catastrophic data-loss
1996 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1998 - Added ``gnt-instance move`` command to move instances
1999 - Added ``gnt-cluster watcher`` command to control watcher
2000 - Added ``gnt-node powercycle`` command to powercycle nodes
2001 - Added new job status field ``lock_status``
2002 - Added parseable error codes to cluster verification (``gnt-cluster
2003 verify --error-codes``) and made output less verbose (use
2004 ``--verbose`` to restore previous behaviour)
2005 - Added UUIDs to the main config entities (cluster, nodes, instances)
2006 - Added support for OS variants
2007 - Added support for hashed passwords in the Ganeti remote API users file
2009 - Added option to specify maximum timeout on instance shutdown
2010 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2011 - Added new helper script to start and stop Ganeti daemons
2012 (``daemon-util``), with the intent to reduce the work necessary to
2013 adjust Ganeti for non-Debian distributions and to start/stop daemons
2015 - Added more unittests
2016 - Fixed critical bug in ganeti-masterd startup
2017 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2018 now customisable at the cluster level for both the KVM and Xen
2019 hypervisors using the new ``migration_port`` parameter
2020 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2021 reinstalling an instance
2022 - Allowed ``@`` in tag names
2023 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2024 - Many documentation updates
2025 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2026 - ``gnt-instance reinstall`` can now reinstall multiple instances
2027 - Updated many command line parameters
2028 - Introduced new OS API version 15
2029 - No longer support a default hypervisor
2030 - Treat virtual LVs as inexistent
2031 - Improved job locking logic to reduce lock contention
2032 - Match instance and node names case insensitively
2033 - Reimplemented bash completion script to be more complete
2040 *(Released Thu, 4 Feb 2010)*
2042 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2043 - Fix a string formatting bug
2044 - Improve safety of the code in some error paths
2045 - Improve data validation in the master of values returned from nodes
2051 *(Released Thu, 17 Dec 2009)*
2053 - Fix security issue due to missing validation of iallocator names; this
2054 allows local and remote execution of arbitrary executables
2055 - Fix failure of gnt-node list during instance removal
2056 - Ship the RAPI documentation in the archive
2062 *(Released Wed, 30 Sep 2009)*
2064 - Fixed many wrong messages
2065 - Fixed a few bugs related to the locking library
2066 - Fixed MAC checking at instance creation time
2067 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2068 - Fixed a few issues related to signal handling in both daemons and
2070 - Fixed the example startup script provided
2071 - Fixed insserv dependencies in the example startup script (patch from
2073 - Fixed handling of drained nodes in the iallocator framework
2074 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2076 - Fixed error related to invalid job IDs in job polling
2077 - Fixed job/opcode persistence on unclean master shutdown
2078 - Fixed handling of partial job processing after unclean master
2080 - Fixed error reporting from LUs, previously all errors were converted
2081 into execution errors
2082 - Fixed error reporting from burnin
2083 - Decreased significantly the memory usage of the job queue
2084 - Optimised slightly multi-job submission
2085 - Optimised slightly opcode loading
2086 - Backported the multi-job submit framework from the development
2087 branch; multi-instance start and stop should be faster
2088 - Added script to clean archived jobs after 21 days; this will reduce
2089 the size of the queue directory
2090 - Added some extra checks in disk size tracking
2091 - Added an example ethers hook script
2092 - Added a cluster parameter that prevents Ganeti from modifying of
2094 - Added more node information to RAPI responses
2095 - Added a ``gnt-job watch`` command that allows following the ouput of a
2097 - Added a bind-address option to ganeti-rapi
2098 - Added more checks to the configuration verify
2099 - Enhanced the burnin script such that some operations can be retried
2101 - Converted instance reinstall to multi-instance model
2107 *(Released Fri, 7 Aug 2009)*
2109 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2110 to allow using the pre-2.0.2 behaviour in activation, if any existing
2111 instances have mismatched disk sizes in the configuration
2112 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2113 any configuration mismatches for disk sizes
2114 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2115 failover to work on two-node clusters
2116 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2118 - Fixed detection of OS script errors in ``gnt-backup export``
2119 - Fixed exit code of ``gnt-backup export``
2125 *(Released Fri, 17 Jul 2009)*
2127 - Added experimental support for stripped logical volumes; this should
2128 enhance performance but comes with a higher complexity in the block
2129 device handling; stripping is only enabled when passing
2130 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2131 affected even in the non-stripped mode
2132 - Improved resiliency against transient failures at the end of DRBD
2133 resyncs, and in general of DRBD resync checks
2134 - Fixed a couple of issues with exports and snapshot errors
2135 - Fixed a couple of issues in instance listing
2136 - Added display of the disk size in ``gnt-instance info``
2137 - Fixed checking for valid OSes in instance creation
2138 - Fixed handling of the "vcpus" parameter in instance listing and in
2139 general of invalid parameters
2140 - Fixed http server library, and thus RAPI, to handle invalid
2141 username/password combinations correctly; this means that now they
2142 report unauthorized for queries too, not only for modifications,
2143 allowing earlier detect of configuration problems
2144 - Added a new "role" node list field, equivalent to the master/master
2145 candidate/drained/offline flags combinations
2146 - Fixed cluster modify and changes of candidate pool size
2147 - Fixed cluster verify error messages for wrong files on regular nodes
2148 - Fixed a couple of issues with node demotion from master candidate role
2149 - Fixed node readd issues
2150 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2151 - Added a new ``--no-voting`` option for masterfailover to fix failover
2152 on two-nodes clusters when the former master node is unreachable
2153 - Added instance reinstall over RAPI
2159 *(Released Tue, 16 Jun 2009)*
2161 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2162 allow re-adding the start in single-user option (regression from 1.2)
2163 - the watcher writes the instance status to a file, to allow monitoring
2164 to report the instance status (from the master) based on cached
2165 results of the watcher's queries; while this can get stale if the
2166 watcher is being locked due to other work on the cluster, this is
2167 still an improvement
2168 - the watcher now also restarts the node daemon and the rapi daemon if
2170 - fixed the watcher to handle full and drained queue cases
2171 - hooks export more instance data in the environment, which helps if
2172 hook scripts need to take action based on the instance's properties
2173 (no longer need to query back into ganeti)
2174 - instance failovers when the instance is stopped do not check for free
2175 RAM, so that failing over a stopped instance is possible in low memory
2177 - rapi uses queries for tags instead of jobs (for less job traffic), and
2178 for cluster tags it won't talk to masterd at all but read them from
2180 - a couple of error handling fixes in RAPI
2181 - drbd handling: improved the error handling of inconsistent disks after
2182 resync to reduce the frequency of "there are some degraded disks for
2183 this instance" messages
2184 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2185 error handling path called a wrong function name)
2191 *(Released Wed, 27 May 2009)*
2193 - no changes from rc5
2199 *(Released Wed, 20 May 2009)*
2201 - fix a couple of bugs (validation, argument checks)
2202 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2203 - some small improvements to RAPI and IAllocator
2204 - make watcher automatically start the master daemon if down
2210 *(Released Mon, 27 Apr 2009)*
2212 - change the OS list to not require locks; this helps with big clusters
2213 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2214 volume group is broken
2215 - ``gnt-instance info``, without any arguments, doesn't run for all
2216 instances anymore; either pass ``--all`` or pass the desired
2217 instances; this helps against mistakes on big clusters where listing
2218 the information for all instances takes a long time
2219 - miscellaneous doc and man pages fixes
2225 *(Released Wed, 8 Apr 2009)*
2227 - Change the internal locking model of some ``gnt-node`` commands, in
2228 order to reduce contention (and blocking of master daemon) when
2229 batching many creation/reinstall jobs
2230 - Fixes to Xen soft reboot
2231 - No longer build documentation at build time, instead distribute it in
2232 the archive, in order to reduce the need for the whole docbook/rst
2239 *(Released Fri, 27 Mar 2009)*
2241 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2242 - Fix watcher startup sequence, improves the behaviour of busy clusters
2243 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2244 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2245 bind address and other places
2246 - Some documentation fixes and updates
2252 *(Released Mon, 2 Mar 2009)*
2254 - More documentation updates, now all docs should be more-or-less
2256 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2258 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2260 - Fix an issue related to $libdir/run/ganeti and cluster creation
2266 *(Released Thu, 19 Feb 2009)*
2268 - Xen PVM and KVM have switched the default value for the instance root
2269 disk to the first partition on the first drive, instead of the whole
2270 drive; this means that the OS installation scripts must be changed
2272 - Man pages have been updated
2273 - RAPI has been switched by default to HTTPS, and the exported functions
2274 should all work correctly
2275 - RAPI v1 has been removed
2276 - Many improvements to the KVM hypervisor
2277 - Block device errors are now better reported
2278 - Many other bugfixes and small improvements
2284 *(Released Mon, 26 Jan 2009)*
2286 - Version 2 is a general rewrite of the code and therefore the
2287 differences are too many to list, see the design document for 2.0 in
2288 the ``doc/`` subdirectory for more details
2289 - In this beta version there is not yet a migration path from 1.2 (there
2290 will be one in the final 2.0 release)
2291 - A few significant changes are:
2293 - all commands are executed by a daemon (``ganeti-masterd``) and the
2294 various ``gnt-*`` commands are just front-ends to it
2295 - all the commands are entered into, and executed from a job queue,
2296 see the ``gnt-job(8)`` manpage
2297 - the RAPI daemon supports read-write operations, secured by basic
2298 HTTP authentication on top of HTTPS
2299 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2300 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2301 to migrate to DRBD 8 first while still running Ganeti 1.2)
2302 - DRBD devices are using statically allocated minor numbers, which
2303 will be assigned to existing instances during the migration process
2304 - there is support for both Xen PVM and Xen HVM instances running on
2306 - KVM virtualization is supported too
2307 - file-based storage has been implemented, which means that it is
2308 possible to run the cluster without LVM and DRBD storage, for
2309 example using a shared filesystem exported from shared storage (and
2310 still have live migration)
2316 *(Released Tue, 13 Jan 2009)*
2318 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2319 - Reuse the old instance mac address by default on instance import, if
2320 the instance name is the same.
2321 - Handle situations in which the node info rpc returns incomplete
2323 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2324 - Improved version of batcher:
2326 - state file support
2327 - instance mac address support
2328 - support for HVM clusters/instances
2330 - Add an option to show the number of cpu sockets and nodes in
2332 - Support OSes that handle more than one version of the OS api (but do
2333 not change the current API in any other way)
2334 - Fix ``gnt-node migrate``
2335 - ``gnt-debug`` man page
2336 - Fixes various more typos and small issues
2337 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2343 *(Released Wed, 24 Sep 2008)*
2345 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2346 type of disk exported to fully virtualized instances.
2347 - provide access to the serial console of HVM instances
2348 - instance auto_balance flag, set by default. If turned off it will
2349 avoid warnings on cluster verify if there is not enough memory to fail
2350 over an instance. in the future it will prevent automatically failing
2351 it over when we will support that.
2352 - batcher tool for instance creation, see ``tools/README.batcher``
2353 - ``gnt-instance reinstall --select-os`` to interactively select a new
2354 operating system when reinstalling an instance.
2355 - when changing the memory amount on instance modify a check has been
2356 added that the instance will be able to start. also warnings are
2357 emitted if the instance will not be able to fail over, if auto_balance
2359 - documentation fixes
2360 - sync fields between ``gnt-instance list/modify/add/import``
2361 - fix a race condition in drbd when the sync speed was set after giving
2362 the device a remote peer.
2368 *(Released Tue, 22 Jul 2008)*
2370 - note: the allowed size and number of tags per object were reduced
2371 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2372 - fixed twisted 8.x compatibility
2373 - fixed ``gnt-instance replace-disks`` with iallocator
2374 - add TCP keepalives on twisted connections to detect restarted nodes
2375 - disk increase support, see ``gnt-instance grow-disk``
2376 - implement bulk node/instance query for RAPI
2377 - add tags in node/instance listing (optional)
2378 - experimental migration (and live migration) support, read the man page
2379 for ``gnt-instance migrate``
2380 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2381 has some small improvements in handling its state file
2387 *(Released Fri, 13 Jun 2008)*
2389 - Experimental readonly, REST-based remote API implementation;
2390 automatically started on master node, TCP port 5080, if enabled by
2391 ``--enable-rapi`` parameter to configure script.
2392 - Instance allocator support. Add and import instance accept a
2393 ``--iallocator`` parameter, and call that instance allocator to decide
2394 which node to use for the instance. The iallocator document describes
2395 what's expected from an allocator script.
2396 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2397 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2398 node is lost there is still enough memory to fail over the instances
2400 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2401 ``gnt-cluster verify``, to check for site-specific compliance. All the
2402 hooks will run, and their output, if any, will be displayed. Any
2403 failing hook will make the verification return an error value.
2404 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2405 primary and secondary interfaces
2406 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2407 that is still declared as part of the cluster and has failed.
2408 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2409 specifying output fields, that just adds the chosen fields to the
2411 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2412 export from the filesystem.
2413 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2414 have been added. Using them you can enable/disable acpi and pae
2415 support, and specify a path for a cd image to be exported to the
2416 instance. These parameters as the name suggest only work on HVM
2418 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2419 PAE support will be set to the previously hardcoded values, but the
2420 (previously hardcoded) path to the CDROM ISO image will be unset and
2421 if required, needs to be set manually with ``gnt-instance modify``
2423 - The address to which an instance's VNC console is bound is now
2424 selectable per-instance, rather than being cluster wide. Of course
2425 this only applies to instances controlled via VNC, so currently just
2426 applies to HVM clusters.
2432 *(Released Mon, 18 Feb 2008)*
2434 - more tweaks to the disk activation code (especially helpful for DRBD)
2435 - change the default ``gnt-instance list`` output format, now there is
2436 one combined status field (see the manpage for the exact values this
2438 - some more fixes for the mac export to hooks change
2439 - make Ganeti not break with DRBD 8.2.x (which changed the version
2440 format in ``/proc/drbd``) (issue 24)
2441 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2442 template, allowing migration from DRBD0.7+MD to DRBD8
2448 *(Released Wed, 30 Jan 2008)*
2450 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2452 - add command aliases infrastructure and a few aliases
2453 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2454 man pages and the ``--help`` option of ``gnt-node
2455 list``/``gnt-instance list``
2456 - fix ``gnt-backup list`` with down nodes (issue 21)
2457 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2458 - fix the dist archive and add a check for including svn/git files in
2460 - some developer-related changes: improve the burnin and the QA suite,
2461 add an upload script for testing during development
2467 *(Released Wed, 16 Jan 2008)*
2469 - experimental HVM support, read the install document, section
2470 "Initializing the cluster"
2471 - allow for the PVM hypervisor per-instance kernel and initrd paths
2472 - add a new command ``gnt-cluster verify-disks`` which uses a new
2473 algorithm to improve the reconnection of the DRBD pairs if the device
2474 on the secondary node has gone away
2475 - make logical volume code auto-activate LVs at disk activation time
2476 - slightly improve the speed of activating disks
2477 - allow specification of the MAC address at instance creation time, and
2478 changing it later via ``gnt-instance modify``
2479 - fix handling of external commands that generate lots of output on
2481 - update documentation with regard to minimum version of DRBD8 supported
2487 *(Released Tue, 4 Dec 2007)*
2489 - Log the ``xm create`` output to the node daemon log on failure (to
2490 help diagnosing the error)
2491 - In debug mode, log all external commands output if failed to the logs
2492 - Change parsing of lvm commands to ignore stderr
2498 *(Released Wed, 28 Nov 2007)*
2500 - Another round of updates to the DRBD 8 code to deal with more failures
2501 in the replace secondary node operation
2502 - Some more logging of failures in disk operations (lvm, drbd)
2503 - A few documentation updates
2510 *(Released Tue, 13 Nov 2007)*
2512 - Change configuration file format from Python's Pickle to JSON.
2513 Upgrading is possible using the cfgupgrade utility.
2514 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2515 faster replace disks and is more stable (DRBD 8 has many improvements
2516 compared to DRBD 0.7)
2517 - Added command line tags support (see man pages for ``gnt-instance``,
2518 ``gnt-node``, ``gnt-cluster``)
2519 - Added instance rename support
2520 - Added multi-instance startup/shutdown
2521 - Added cluster rename support
2522 - Added ``gnt-node evacuate`` to simplify some node operations
2523 - Added instance reboot operation that can speedup reboot as compared to
2525 - Soften the requirement that hostnames are in FQDN format
2526 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2528 - Removed dependency on debian's patched fping that uses the
2529 non-standard ``-S`` option
2530 - Now the OS definitions are searched for in multiple, configurable
2531 paths (easier for distros to package)
2532 - Some changes to the hooks infrastructure (especially the new
2533 post-configuration update hook)
2534 - Other small bugfixes
2536 .. vim: set textwidth=72 syntax=rst :