8 *(Released Wed, 6 Mar 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 the not enough resources can be found to
52 allocate the instance, the temporary error code
53 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
54 retried thereafter, with or without opportunistic locking.
55 - Man pages can now be included when the documentation is built, in
56 which case the output is in ``doc/man-html``. The configure-time
57 option is ``--enable-manpages-in-doc``. Sphinx 1.0 or higher is
59 - A new htool called ``hroller`` helps scheduling parallel shutdown of
60 nodes on clusters where all instances are shut down, but without ever
61 rebooting both primary and secondary node for an instance at the same
63 - Queries not needing locks or RPC access to the node can now be
64 performed by the confd daemon, making them independent from jobs, and
65 thus faster to execute. This is selectable at configure time.
66 - The KVM hypervisor has been updated with many new hypervisor
67 parameters, including a generic one for passing arbitrary command line
68 values. See a complete list in :manpage:`gnt-instance(8)`.
69 - A new tool, called ``mon-collector``, is the stand-alone executor of
70 the data collectors for a monitoring system. As of this version, it
71 just includes the DRBD data collector, that can be executed by calling
72 ``mon-collector`` using the ``drbd`` parameter. See
73 :manpage:`mon-collector(7)`.
74 - New ``exclusive-storage`` node parameter added. When it's set to true,
75 physical disks are assigned in an exclusive fashion to instances, as
76 documented in :doc:`Partitioned Ganeti <design-partitioned>`.
77 Currently, only instances using the ``plain`` disk template are
79 - New network management functionality to support automatic allocation
80 of IP addresses and managing of network parameters. See
81 :manpage:`gnt-network(8)` for more details.
82 - New external storage backend, to allow managing arbitrary storage
83 systems external to the cluster. See
84 :manpage:`ganeti-extstorage-interface(7)`.
90 *(Released Fri, 21 Dec 2012)*
92 Important behaviour change: hbal won't rebalance anymore instances which
93 have the ``auto_balance`` attribute set to false. This was the intention
94 all along, but until now it only skipped those from the N+1 memory
95 reservation (DRBD-specific).
97 A significant number of bug fixes in this release:
99 - Fixed disk adoption interaction with ipolicy checks.
100 - Fixed networking issues when instances are started, stopped or
101 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
102 - Fixed the warning in cluster verify for shared storage instances not
104 - Fixed removal of storage directory on shared file storage (issue 262).
105 - Fixed validation of LVM volume group name in OpClusterSetParams
106 (``gnt-cluster modify``) (issue 285).
107 - Fixed runtime memory increases (``gnt-instance modify -m``).
108 - Fixed live migration under Xen's ``xl`` mode.
109 - Fixed ``gnt-instance console`` with ``xl``.
110 - Fixed building with newer Haskell compiler/libraries.
111 - Fixed PID file writing in Haskell daemons (confd); this prevents
112 restart issues if confd was launched manually (outside of
113 ``daemon-util``) while another copy of it was running
114 - Fixed a type error when doing live migrations with KVM (issue 297) and
115 the error messages for failing migrations have been improved.
116 - Fixed opcode validation for the out-of-band commands (``gnt-node
118 - Fixed a type error when unsetting OS hypervisor parameters (issue
119 311); now it's possible to unset all OS-specific hypervisor
121 - Fixed the ``dry-run`` mode for many operations: verification of
122 results was over-zealous but didn't take into account the ``dry-run``
123 operation, resulting in "wrong" failures.
124 - Fixed bash completion in ``gnt-job list`` when the job queue has
125 hundreds of entries; especially with older ``bash`` versions, this
126 results in significant CPU usage.
128 And lastly, a few other improvements have been made:
130 - Added option to force master-failover without voting (issue 282).
131 - Clarified error message on lock conflict (issue 287).
132 - Logging of newly submitted jobs has been improved (issue 290).
133 - Hostname checks have been made uniform between instance rename and
135 - The ``--submit`` option is now supported by ``gnt-debug delay``.
136 - Shutting down the master daemon by sending SIGTERM now stops it from
137 processing jobs waiting for locks; instead, those jobs will be started
138 once again after the master daemon is started the next time (issue
140 - Support for Xen's ``xl`` program has been improved (besides the fixes
142 - Reduced logging noise in the Haskell confd daemon (only show one log
143 entry for each config reload, instead of two).
144 - Several man page updates and typo fixes.
150 *(Released Fri, 12 Oct 2012)*
152 A small bugfix release. Among the bugs fixed:
154 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
155 made the command unusable.
156 - Commands that issue many jobs don't fail anymore just because some jobs
157 take so long that other jobs are archived.
158 - Failures during ``gnt-instance reinstall`` are reflected by the exit
160 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
162 - When ``always_failover`` is set, ``--allow-failover`` is not required
163 in migrate commands anymore.
164 - ``bash_completion`` works even if extglob is disabled.
165 - Fixed bug with locks that made failover for RDB-based instances fail.
166 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
167 a random node instead of one based on the allocator metric.
168 - Support for newer versions of pylint and pep8.
169 - Hail doesn't fail anymore when trying to add an instance of type
170 ``file``, ``sharedfile`` or ``rbd``.
171 - Added new Makefile target to rebuild the whole distribution, so that
172 all files are included.
178 *(Released Fri, 27 Jul 2012)*
181 .. attention:: The ``LUXI`` protocol has been made more consistent
182 regarding its handling of command arguments. This, however, leads to
183 incompatibility issues with previous versions. Please ensure that you
184 restart Ganeti daemons soon after the upgrade, otherwise most
185 ``LUXI`` calls (job submission, setting/resetting the drain flag,
186 pausing/resuming the watcher, cancelling and archiving jobs, querying
187 the cluster configuration) will fail.
196 The current ``admin_up`` field, which used to denote whether an instance
197 should be running or not, has been removed. Instead, ``admin_state`` is
198 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
200 The rational behind this is that an instance being “down” can have
203 - it could be down during a reboot
204 - it could be temporarily be down for a reinstall
205 - or it could be down because it is deprecated and kept just for its
208 The previous Boolean state was making it difficult to do capacity
209 calculations: should Ganeti reserve memory for a down instance? Now, the
210 tri-state field makes it clear:
212 - in ``up`` and ``down`` state, all resources are reserved for the
213 instance, and it can be at any time brought up if it is down
214 - in ``offline`` state, only disk space is reserved for it, but not
217 The field can have an extra use: since the transition between ``up`` and
218 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
219 transition between ``offline`` and ``down`` is done via ``gnt-instance
220 modify``, it is possible to given different rights to users. For
221 example, owners of an instance could be allowed to start/stop it, but
222 not transition it out of the offline state.
224 Instance policies and specs
225 +++++++++++++++++++++++++++
227 In previous Ganeti versions, an instance creation request was not
228 limited on the minimum size and on the maximum size just by the cluster
229 resources. As such, any policy could be implemented only in third-party
230 clients (RAPI clients, or shell wrappers over ``gnt-*``
231 tools). Furthermore, calculating cluster capacity via ``hspace`` again
232 required external input with regards to instance sizes.
234 In order to improve these workflows and to allow for example better
235 per-node group differentiation, we introduced instance specs, which
238 - minimum instance disk size, disk count, memory size, cpu count
239 - maximum values for the above metrics
240 - and “standard” values (used in ``hspace`` to calculate the standard
243 The minimum/maximum values can be also customised at node-group level,
244 for example allowing more powerful hardware to support bigger instance
247 Beside the instance specs, there are a few other settings belonging to
248 the instance policy framework. It is possible now to customise, per
249 cluster and node-group:
251 - the list of allowed disk templates
252 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
253 - the maximum ratio of instance to spindles (see below for more
254 information) for local storage
256 All these together should allow all tools that talk to Ganeti to know
257 what are the ranges of allowed values for instances and the
258 over-subscription that is allowed.
260 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
261 instance configuration, and the physical CPU configuration from the
262 node. For the spindle ratios however, we didn't track before these
263 values, so new parameters have been added:
265 - a new node parameter ``spindle_count``, defaults to 1, customisable at
266 node group or node level
267 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
269 Note that spindles in this context doesn't need to mean actual
270 mechanical hard-drives; it's just a relative number for both the node
271 I/O capacity and instance I/O consumption.
273 Instance migration behaviour
274 ++++++++++++++++++++++++++++
276 While live-migration is in general desirable over failover, it is
277 possible that for some workloads it is actually worse, due to the
278 variable time of the “suspend” phase during live migration.
280 To allow the tools to work consistently over such instances (without
281 having to hard-code instance names), a new backend parameter
282 ``always_failover`` has been added to control the migration/failover
283 behaviour. When set to True, all migration requests for an instance will
284 instead fall-back to failover.
286 Instance memory ballooning
287 ++++++++++++++++++++++++++
289 Initial support for memory ballooning has been added. The memory for an
290 instance is no longer fixed (backend parameter ``memory``), but instead
291 can vary between minimum and maximum values (backend parameters
292 ``minmem`` and ``maxmem``). Currently we only change an instance's
295 - live migrating or failing over and instance and the target node
296 doesn't have enough memory
297 - user requests changing the memory via ``gnt-instance modify
303 In order to control the use of specific CPUs by instance, support for
304 controlling CPU pinning has been added for the Xen, HVM and LXC
305 hypervisors. This is controlled by a new hypervisor parameter
306 ``cpu_mask``; details about possible values for this are in the
307 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
308 VCPU-to-CPU mapping) form will work well only when all nodes in your
309 cluster have the same amount of CPUs.
314 Another area in which Ganeti was not customisable were the parameters
315 used for storage configuration, e.g. how many stripes to use for LVM,
316 DRBD resync configuration, etc.
318 To improve this area, we've added disks parameters, which are
319 customisable at cluster and node group level, and which allow to
320 specify various parameters for disks (DRBD has the most parameters
321 currently), for example:
323 - DRBD resync algorithm and parameters (e.g. speed)
324 - the default VG for meta-data volumes for DRBD
325 - number of stripes for LVM (plain disk template)
328 These parameters can be modified via ``gnt-cluster modify -D …`` and
329 ``gnt-group modify -D …``, and are used at either instance creation (in
330 case of LVM stripes, for example) or at disk “activation” time
333 Rados block device support
334 ++++++++++++++++++++++++++
336 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
337 denoted by the ``rbd`` disk template type. This is considered
338 experimental, feedback is welcome. For details on configuring it, see
339 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
344 The existing master IP functionality works well only in simple setups (a
345 single network shared by all nodes); however, if nodes belong to
346 different networks, then the ``/32`` setup and lack of routing
347 information is not enough.
349 To allow the master IP to function well in more complex cases, the
350 system was reworked as follows:
352 - a master IP netmask setting has been added
353 - the master IP activation/turn-down code was moved from the node daemon
355 - whether to run the Ganeti-supplied master IP script or a user-supplied
356 on is a ``gnt-cluster init`` setting
358 Details about the location of the standard and custom setup scripts are
359 in the man page :manpage:`gnt-cluster(8)`; for information about the
360 setup script protocol, look at the Ganeti-supplied script.
365 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
368 It is now possible to use TLS-protected connections, and when renewing
369 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
370 it is now possible to specify spice or spice CA certificates. Also, it
371 is possible to configure a password for SPICE sessions via the
372 hypervisor parameter ``spice_password_file``.
374 There are also new parameters to control the compression and streaming
375 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
376 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
377 for the spice parameters.
379 Lastly, it is now possible to see the SPICE connection information via
380 ``gnt-instance console``.
385 A new tool (``tools/ovfconverter``) has been added that supports
386 conversion between Ganeti and the `Open Virtualization Format
387 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
390 This relies on the ``qemu-img`` tool to convert the disk formats, so the
391 actual compatibility with other virtualization solutions depends on it.
396 The configuration query daemon (``ganeti-confd``) is now optional, and
397 has been rewritten in Haskell; whether to use the daemon at all, use the
398 Python (default) or the Haskell version is selectable at configure time
399 via the ``--enable-confd`` parameter, which can take one of the
400 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
401 daemon will result in a smaller footprint; for larger systems, we
402 welcome feedback on the Haskell version which might become the default
405 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
406 daemon running. The Python version doesn't implement the new call.
409 User interface changes
410 ~~~~~~~~~~~~~~~~~~~~~~
412 We have replaced the ``--disks`` option of ``gnt-instance
413 replace-disks`` with a more flexible ``--disk`` option, which allows
414 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
415 disk size and mode can be changed upon recreation (via ``gnt-instance
416 recreate-disks``, which accepts the same ``--disk`` option).
418 As many people are used to a ``show`` command, we have added that as an
419 alias to ``info`` on all ``gnt-*`` commands.
421 The ``gnt-instance grow-disk`` command has a new mode in which it can
422 accept the target size of the disk, instead of the delta; this can be
423 more safe since two runs in absolute mode will be idempotent, and
424 sometimes it's also easier to specify the desired size directly.
426 Also the handling of instances with regard to offline secondaries has
427 been improved. Instance operations should not fail because one of it's
428 secondary nodes is offline, even though it's safe to proceed.
430 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
431 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
432 minors to instance name.
437 RAPI coverage has improved, with (for example) new resources for
438 recreate-disks, node power-cycle, etc.
443 There is partial support for ``xl`` in the Xen hypervisor; feedback is
446 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
447 whether to still support Python 2.4 or move to Python 2.6 as minimum
450 Support for Fedora has been slightly improved; the provided example
451 init.d script should work better on it and the INSTALL file should
452 document the needed dependencies.
457 The deprecated ``QueryLocks`` LUXI request has been removed. Use
458 ``Query(what=QR_LOCK, ...)`` instead.
460 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
461 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
462 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
463 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
464 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
466 RAPI client: ``CertificateError`` now derives from
467 ``GanetiApiError``. This should make it more easy to handle Ganeti
470 Deprecation warnings due to PyCrypto/paramiko import in
471 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
472 make sure to run an up-to-date paramiko version, if you use this tool.
474 The QA scripts now depend on Python 2.5 or above (the main code base
475 still works with Python 2.4).
477 The configuration file (``config.data``) is now written without
478 indentation for performance reasons; if you want to edit it, it can be
479 re-formatted via ``tools/fmtjson``.
481 A number of bugs has been fixed in the cluster merge tool.
483 ``x509`` certification verification (used in import-export) has been
484 changed to allow the same clock skew as permitted by the cluster
485 verification. This will remove some rare but hard to diagnose errors in
492 *(Released Thu, 19 Jul 2012)*
494 Very few changes from rc4 to the final release, only bugfixes:
496 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
497 instance, fix CDROM booting for KVM instances)
498 - fixed node group modification of node parameters
499 - fixed issue in LUClusterVerifyGroup with multi-group clusters
500 - fixed generation of bash completion to ensure a stable ordering
507 *(Released Fri, 13 Jul 2012)*
509 Third release candidate for 2.6. The following changes were done from
512 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
513 - Fixed an inconsistency in the LUXI protocol with the provided
514 arguments (NOT backwards compatible)
515 - Fixed a bug with node groups ipolicy where ``min`` was greater than
516 the cluster ``std`` value
517 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
518 easier instance debugging on nodes (requires ``hconfd`` to work)
524 *(Released Tue, 03 Jul 2012)*
526 Second release candidate for 2.6. The following changes were done from
529 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
531 - Fixed a RAPI regression on missing beparams/memory
532 - Fixed redistribution of files on offline nodes
533 - Added possibility to run activate-disks even though secondaries are
534 offline. With this change it relaxes also the strictness on some other
535 commands which use activate disks internally:
536 * ``gnt-instance start|reboot|rename|backup|export``
537 - Made it possible to remove safely an instance if its secondaries are
539 - Made it possible to reinstall even though secondaries are offline
545 *(Released Mon, 25 Jun 2012)*
547 First release candidate for 2.6. The following changes were done from
550 - Fixed bugs with disk parameters and ``rbd`` templates as well as
552 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
553 behaviour. It supports now the modify operation
554 - ``hcheck`` implemented to analyze cluster health and possibility of
555 improving health by rebalance
556 - ``hbal`` has been improved in dealing with split instances
562 *(Released Mon, 11 Jun 2012)*
564 Second beta release of 2.6. The following changes were done from beta2
567 - Fixed ``daemon-util`` with non-root user models
568 - Fixed creation of plain instances with ``--no-wait-for-sync``
569 - Fix wrong iv_names when running ``cfgupgrade``
570 - Export more information in RAPI group queries
571 - Fixed bug when changing instance network interfaces
572 - Extended burnin to do NIC changes
573 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
574 - Changed default for DRBD barriers
575 - Fixed DRBD error reporting for syncer rate
576 - Verify the options on disk parameters
578 And of course various fixes to documentation and improved unittests and
585 *(Released Wed, 23 May 2012)*
587 First beta release of 2.6. The following changes were done from beta1 to
590 - integrated patch for distributions without ``start-stop-daemon``
591 - adapted example init.d script to work on Fedora
592 - fixed log handling in Haskell daemons
593 - adapted checks in the watcher for pycurl linked against libnss
594 - add partial support for ``xl`` instead of ``xm`` for Xen
595 - fixed a type issue in cluster verification
596 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
599 Plus integrated fixes from the 2.5 branch:
601 - fixed ``kvm-ifup`` to use ``/bin/bash``
602 - fixed parallel build failures
603 - KVM live migration when using a custom keymap
609 *(Released Tue, 24 Jul 2012)*
611 A small bugfix release, with no new features:
613 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
614 different default shell (e.g. Debian, Ubuntu)
615 - fixed KVM startup and live migration with a custom keymap (fixes Issue
616 243 and Debian bug #650664)
617 - fixed compatibility with KVM versions that don't support multiple boot
618 devices (fixes Issue 230 and Debian bug #624256)
620 Additionally, a few fixes were done to the build system (fixed parallel
621 build failures) and to the unittests (fixed race condition in test for
622 FileID functions, and the default enable/disable mode for QA test is now
629 *(Released Fri, 11 May 2012)*
631 A small bugfix release.
633 The main issues solved are on the topic of compatibility with newer LVM
636 - fixed parsing of ``lv_attr`` field
637 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
638 the ``--force`` flag is needed
640 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
641 to accept kernel 3.x too (was hardcoded to 2.6.*).
643 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
644 of hook results and that also made display of other errors suboptimal
645 was fixed; the code behaves now like 2.4 and earlier.
647 Another change in 2.5, the cleanup of the OS scripts environment, is too
648 aggressive: it removed even the ``PATH`` variable, which requires the OS
649 scripts to *always* need to export it. Since this is a bit too strict,
650 we now export a minimal PATH, the same that we export for hooks.
652 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
653 integrated into this release.
655 Finally, a few other miscellaneous changes were done (no new features,
656 just small improvements):
658 - Fix ``gnt-group --help`` display
659 - Fix hardcoded Xen kernel path
660 - Fix grow-disk handling of invalid units
661 - Update synopsis for ``gnt-cluster repair-disk-sizes``
662 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
668 *(Released Thu, 12 Apr 2012)*
670 Incompatible/important changes and bugfixes
671 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
673 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
674 resource's ``ip_check`` parameter changed from ``True`` to ``False``
675 to match the underlying LUXI interface.
676 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
677 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
678 not maintain backwards-compatibility as the underlying operation
679 changed in an incompatible way. The RAPI client can talk to old
680 servers, but it needs to be told so as the return value changed.
681 - When creating file-based instances via RAPI, the ``file_driver``
682 parameter no longer defaults to ``loop`` and must be specified.
683 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
685 - Support for the undocumented and deprecated RAPI instance creation
686 request format version 0 has been dropped. Use version 1, supported
687 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
688 - Pyparsing 1.4.6 or above is required, see :doc:`installation
689 documentation <install>`.
690 - The "cluster-verify" hooks are now executed per group by the
691 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
692 if you just run ``gnt-cluster verify``, which generates one opcode per
694 - The environment as passed to the OS scripts is cleared, and thus no
695 environment variables defined in the node daemon's environment will be
696 inherited by the scripts.
697 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
699 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
700 support operations involving multiple node groups.
701 - Offline nodes are ignored when failing over an instance.
702 - Support for KVM version 1.0, which changed the version reporting format
704 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
706 - ``Makefile`` is now compatible with Automake 1.11.2
707 - Includes all bugfixes made in the 2.4 series
712 - The ganeti-htools project has been merged into the ganeti-core source
713 tree and will be built as part of Ganeti (see :doc:`install-quick`).
714 - Implemented support for :doc:`shared storage <design-shared-storage>`.
715 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
716 GPT-style partition tables (requires `parted
717 <http://www.gnu.org/s/parted/>`_).
718 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
719 - Allowed adding tags on instance creation.
720 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
721 - Allow instances to be started in a paused state, enabling the user to
722 see the complete console output on boot using the console.
723 - Added new hypervisor flag to control default reboot behaviour
724 (``reboot_behavior``).
725 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
726 - Improved out-of-band management support:
728 - Added ``gnt-node health`` command reporting the health status of
730 - Added ``gnt-node power`` command to manage power status of nodes.
731 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
733 - Instance migration can fall back to failover if instance is not
735 - Filters can be used when listing nodes, instances, groups and locks;
736 see :manpage:`ganeti(7)` manpage.
737 - Added post-execution status as variables to :doc:`hooks <hooks>`
739 - Instance tags are exported/imported together with the instance.
740 - When given an explicit job ID, ``gnt-job info`` will work for archived
742 - Jobs can define dependencies on other jobs (not yet supported via
743 RAPI or command line, but used by internal commands and usable via
746 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
749 - Instance failover is now available as a RAPI resource
750 (``/2/instances/[instance_name]/failover``).
751 - ``gnt-instance info`` defaults to static information if primary node
753 - Opcodes have a new ``comment`` attribute.
754 - Added basic SPICE support to KVM hypervisor.
755 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
757 Node group improvements
758 ~~~~~~~~~~~~~~~~~~~~~~~
760 - ``gnt-cluster verify`` has been modified to check groups separately,
761 thereby improving performance.
762 - Node group support has been added to ``gnt-cluster verify-disks``,
763 which now operates per node group.
764 - Watcher has been changed to work better with node groups.
766 - One process and state file per node group.
767 - Slow watcher in one group doesn't block other group's watcher.
769 - Added new command, ``gnt-group evacuate``, to move all instances in a
770 node group to other groups.
771 - Added ``gnt-instance change-group`` to move an instance to another
773 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
774 per-group operations.
775 - Node groups can be tagged.
776 - Some operations switch from an exclusive to a shared lock as soon as
778 - Instance's primary and secondary nodes' groups are now available as
779 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
780 and ``snodes.group.uuid``).
785 - Numerous updates to documentation and manpages.
787 - :doc:`RAPI <rapi>` documentation now has detailed parameter
789 - Some opcode/job results are now also documented, see :doc:`RAPI
792 - A lockset's internal lock is now also visible in lock monitor.
793 - Log messages from job queue workers now contain information about the
794 opcode they're processing.
795 - ``gnt-instance console`` no longer requires the instance lock.
796 - A short delay when waiting for job changes reduces the number of LUXI
797 requests significantly.
798 - DRBD metadata volumes are overwritten with zeros during disk creation.
799 - Out-of-band commands no longer acquire the cluster lock in exclusive
801 - ``devel/upload`` now uses correct permissions for directories.
807 *(Released Fri, 23 Mar 2012)*
809 This was the sixth release candidate of the 2.5 series.
815 *(Released Mon, 9 Jan 2012)*
817 This was the fifth release candidate of the 2.5 series.
823 *(Released Thu, 27 Oct 2011)*
825 This was the fourth release candidate of the 2.5 series.
831 *(Released Wed, 26 Oct 2011)*
833 This was the third release candidate of the 2.5 series.
839 *(Released Tue, 18 Oct 2011)*
841 This was the second release candidate of the 2.5 series.
847 *(Released Tue, 4 Oct 2011)*
849 This was the first release candidate of the 2.5 series.
855 *(Released Wed, 31 Aug 2011)*
857 This was the third beta release of the 2.5 series.
863 *(Released Mon, 22 Aug 2011)*
865 This was the second beta release of the 2.5 series.
871 *(Released Fri, 12 Aug 2011)*
873 This was the first beta release of the 2.5 series.
879 *(Released Thu, 27 Oct 2011)*
881 - Fixed bug when parsing command line parameter values ending in
883 - Fixed assertion error after unclean master shutdown
884 - Disable HTTP client pool for RPC, significantly reducing memory usage
886 - Fixed queue archive creation with wrong permissions
892 *(Released Tue, 23 Aug 2011)*
896 - Fixed documentation for importing with ``--src-dir`` option
897 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
898 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
904 *(Released Fri, 5 Aug 2011)*
906 Many bug-fixes and a few small features:
908 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
909 issues when you tried to add an instance with two MAC addresses in one
911 - KVM: fixed per-instance stored UID value
912 - KVM: configure bridged NICs at migration start
913 - KVM: Fix a bug where instance will not start with never KVM versions
915 - Added OS search path to ``gnt-cluster info``
916 - Fixed an issue with ``file_storage_dir`` where you were forced to
917 provide an absolute path, but the documentation states it is a
918 relative path, the documentation was right
919 - Added a new parameter to instance stop/start called ``--no-remember``
920 that will make the state change to not be remembered
921 - Implemented ``no_remember`` at RAPI level
922 - Improved the documentation
923 - Node evacuation: don't call IAllocator if node is already empty
924 - Fixed bug in DRBD8 replace disks on current nodes
925 - Fixed bug in recreate-disks for DRBD instances
926 - Moved assertion checking locks in ``gnt-instance replace-disks``
927 causing it to abort with not owning the right locks for some situation
928 - Job queue: Fixed potential race condition when cancelling queued jobs
929 - Fixed off-by-one bug in job serial generation
930 - ``gnt-node volumes``: Fix instance names
931 - Fixed aliases in bash completion
932 - Fixed a bug in reopening log files after being sent a SIGHUP
933 - Added a flag to burnin to allow specifying VCPU count
934 - Bugfixes to non-root Ganeti configuration
940 *(Released Thu, 12 May 2011)*
942 Many bug-fixes and a few new small features:
944 - Fixed a bug related to log opening failures
945 - Fixed a bug in instance listing with orphan instances
946 - Fixed a bug which prevented resetting the cluster-level node parameter
947 ``oob_program`` to the default
948 - Many fixes related to the ``cluster-merge`` tool
949 - Fixed a race condition in the lock monitor, which caused failures
950 during (at least) creation of many instances in parallel
951 - Improved output for gnt-job info
952 - Removed the quiet flag on some ssh calls which prevented debugging
954 - Improved the N+1 failure messages in cluster verify by actually
955 showing the memory values (needed and available)
956 - Increased lock attempt timeouts so that when executing long operations
957 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
958 too early and thus prevent the use of the 'fair' mechanism
959 - Changed instance query data (``gnt-instance info``) to not acquire
960 locks unless needed, thus allowing its use on locked instance if only
961 static information is asked for
962 - Improved behaviour with filesystems that do not support rename on an
964 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
965 kept locks on all nodes during the wipe, which is unneeded
966 - Fixed ``gnt-watcher`` handling of errors during hooks execution
967 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
968 10GiB) which caused the wipe to fail right at the end in some cases
969 - Fixed master IP activation when doing master failover with no-voting
970 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
971 the master node itself
972 - Fixed potential data-loss in under disk full conditions, where Ganeti
973 wouldn't check correctly the return code and would consider
974 partially-written files 'correct'
975 - Fixed bug related to multiple VGs and DRBD disk replacing
976 - Added new disk parameter ``metavg`` that allows placement of the meta
977 device for DRBD in a different volume group
978 - Fixed error handling in the node daemon when the system libc doesn't
979 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
980 - Fixed lock release during replace-disks, which kept cluster-wide locks
981 when doing disk replaces with an iallocator script
982 - Added check for missing bridges in cluster verify
983 - Handle EPIPE errors while writing to the terminal better, so that
984 piping the output to e.g. ``less`` doesn't cause a backtrace
985 - Fixed rare case where a ^C during Luxi calls could have been
986 interpreted as server errors, instead of simply terminating
987 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
989 - Added a few more parameters to the KVM hypervisor, allowing a second
990 CDROM, custom disk type for CDROMs and a floppy image
991 - Removed redundant message in instance rename when the name is given
993 - Added option to ``gnt-instance recreate-disks`` to allow creating the
994 disks on new nodes, allowing recreation when the original instance
995 nodes are completely gone
996 - Added option when converting disk templates to DRBD to skip waiting
997 for the resync, in order to make the instance available sooner
998 - Added two new variables to the OS scripts environment (containing the
1000 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1001 to allow use of ``pvgrub`` as bootloader
1002 - Changed the instance memory modifications to only check out-of-memory
1003 conditions on memory increases, and turned the secondary node warnings
1004 into errors (they can still be overridden via ``--force``)
1005 - Fixed the handling of a corner case when the Python installation gets
1006 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1007 to execute a command that doesn't exist
1008 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1009 node of the instance returned failures during instance shutdown; this
1010 adds the option ``--ignore-consistency`` to gnt-instance move
1012 And as usual, various improvements to the error messages, documentation
1019 *(Released Wed, 09 Mar 2011)*
1021 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1022 the RAPI users file if run twice (even with ``--dry-run``).
1024 The release fixes that bug (nothing else changed).
1030 *(Released Mon, 07 Mar 2011)*
1032 Final 2.4.0 release. Just a few small fixes:
1034 - Fixed RAPI node evacuate
1035 - Fixed the kvm-ifup script
1036 - Fixed internal error handling for special job cases
1037 - Updated man page to specify the escaping feature for options
1043 *(Released Mon, 28 Feb 2011)*
1045 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1046 that this feature wiped the disks of the wrong instance, leading to loss
1051 - Fixed title of query field containing instance name
1052 - Expanded the glossary in the documentation
1053 - Fixed one unittest (internal issue)
1059 *(Released Mon, 21 Feb 2011)*
1061 A number of bug fixes plus just a couple functionality changes.
1063 On the user-visible side, the ``gnt-* list`` command output has changed
1064 with respect to "special" field states. The current rc1 style of display
1065 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1066 the default output mode special fields states are displayed as follows:
1068 - Offline resource: ``*``
1069 - Unavailable/not applicable: ``-``
1070 - Data missing (RPC failure): ``?``
1071 - Unknown field: ``??``
1073 Another user-visible change is the addition of ``--force-join`` to
1078 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1079 - Fixed regression in RAPI/instance reinstall where all parameters were
1080 required (instead of optional)
1081 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1082 - Fixed iallocator usage (offline nodes were not considered offline)
1083 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1084 - Fixed hypervisor and OS parameter validation with respect to
1085 non-vm_capable nodes
1086 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1088 - Fixed ``tools/listrunner`` with respect to agent-based usage
1094 *(Released Fri, 4 Feb 2011)*
1096 Many changes and fixes since the beta1 release. While there were some
1097 internal changes, the code has been mostly stabilised for the RC
1100 Note: the dumb allocator was removed in this release, as it was not kept
1101 up-to-date with the IAllocator protocol changes. It is recommended to
1102 use the ``hail`` command from the ganeti-htools package.
1104 Note: the 2.4 and up versions of Ganeti are not compatible with the
1105 0.2.x branch of ganeti-htools. You need to upgrade to
1106 ganeti-htools-0.3.0 (or later).
1108 Regressions fixed from 2.3
1109 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1111 - Fixed the ``gnt-cluster verify-disks`` command
1112 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1114 - Fixed disk adoption breakage
1115 - Fixed wrong headers in instance listing for field aliases
1120 - Fixed corner case in KVM handling of NICs
1121 - Fixed many cases of wrong handling of non-vm_capable nodes
1122 - Fixed a bug where a missing instance symlink was not possible to
1123 recreate with any ``gnt-*`` command (now ``gnt-instance
1124 activate-disks`` does it)
1125 - Fixed the volume group name as reported by ``gnt-cluster
1127 - Increased timeouts for the import-export code, hopefully leading to
1128 fewer aborts due network or instance timeouts
1129 - Fixed bug in ``gnt-node list-storage``
1130 - Fixed bug where not all daemons were started on cluster
1131 initialisation, but only at the first watcher run
1132 - Fixed many bugs in the OOB implementation
1133 - Fixed watcher behaviour in presence of instances with offline
1135 - Fixed instance list output for instances running on the wrong node
1136 - a few fixes to the cluster-merge tool, but it still cannot merge
1137 multi-node groups (currently it is not recommended to use this tool)
1143 - Improved network configuration for the KVM hypervisor
1144 - Added e1000 as a supported NIC for Xen-HVM
1145 - Improved the lvmstrap tool to also be able to use partitions, as
1146 opposed to full disks
1147 - Improved speed of disk wiping (the cluster parameter
1148 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1149 of instance creations
1150 - Added documentation for the OS parameters
1151 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1152 hypervisor is not responding
1153 - Added display of blacklisted and hidden OS information in
1154 ``gnt-cluster info``
1155 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1156 NIC and node parameters, which might create problems with currently
1157 invalid (but undetected) configuration files, but prevents validation
1158 failures when unrelated parameters are modified
1159 - Changed cluster initialisation to wait for the master daemon to become
1161 - Expanded the RAPI interface:
1163 - Added config redistribution resource
1164 - Added activation/deactivation of instance disks
1165 - Added export of console information
1167 - Implemented log file reopening on SIGHUP, which allows using
1168 logrotate(8) for the Ganeti log files
1169 - Added a basic OOB helper script as an example
1175 *(Released Fri, 14 Jan 2011)*
1180 - Fixed timezone issues when formatting timestamps
1181 - Added support for node groups, available via ``gnt-group`` and other
1183 - Added out-of-band framework and management, see :doc:`design
1184 document <design-oob>`
1185 - Removed support for roman numbers from ``gnt-node list`` and
1186 ``gnt-instance list``.
1187 - Allowed modification of master network interface via ``gnt-cluster
1188 modify --master-netdev``
1189 - Accept offline secondaries while shutting down instance disks
1190 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1191 - Added support for multiple LVM volume groups
1192 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1194 - Added commands to list available fields:
1196 - ``gnt-node list-fields``
1197 - ``gnt-group list-fields``
1198 - ``gnt-instance list-fields``
1200 - Updated documentation and man pages
1205 - Moved ``rapi_users`` file into separate directory, now named
1206 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1208 - Added new tool for running commands on many machines,
1209 ``tools/ganeti-listrunner``
1210 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1211 improving the ``gnt-instance console`` output
1212 - Allowed customisation of disk index separator at ``configure`` time
1213 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1214 - Added support for non-partitioned md disks in ``lvmstrap``
1215 - Added script to gracefully power off KVM instances
1216 - Split ``utils`` module into smaller parts
1217 - Changed query operations to return more detailed information, e.g.
1218 whether an information is unavailable due to an offline node. To use
1219 this new functionality, the LUXI call ``Query`` must be used. Field
1220 information is now stored by the master daemon and can be retrieved
1221 using ``QueryFields``. Instances, nodes and groups can also be queried
1222 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1223 exposed via RAPI). The following commands make use of this
1224 infrastructure change:
1226 - ``gnt-group list``
1227 - ``gnt-group list-fields``
1229 - ``gnt-node list-fields``
1230 - ``gnt-instance list``
1231 - ``gnt-instance list-fields``
1232 - ``gnt-debug locks``
1237 - New RAPI resources (see :doc:`rapi`):
1241 - ``/2/groups/[group_name]``
1242 - ``/2/groups/[group_name]/assign-nodes``
1243 - ``/2/groups/[group_name]/modify``
1244 - ``/2/groups/[group_name]/rename``
1245 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1249 - Implemented ``no_install`` for instance creation
1250 - Implemented OS parameters for instance reinstallation, allowing
1251 use of special settings on reinstallation (e.g. for preserving data)
1256 - Added IPv6 support in import/export
1257 - Pause DRBD synchronization while wiping disks on instance creation
1258 - Updated unittests and QA scripts
1259 - Improved network parameters passed to KVM
1260 - Converted man pages from docbook to reStructuredText
1266 *(Released Mon, 20 Dec 2010)*
1268 Released version 2.3.1~rc1 without any changes.
1274 *(Released Wed, 1 Dec 2010)*
1276 - impexpd: Disable OpenSSL compression in socat if possible (backport
1277 from master, commit e90739d625b, see :doc:`installation guide
1278 <install-quick>` for details)
1279 - Changed unittest coverage report to exclude test scripts
1280 - Added script to check version format
1286 *(Released Wed, 1 Dec 2010)*
1288 Released version 2.3.0~rc1 without any changes.
1294 *(Released Fri, 19 Nov 2010)*
1296 A number of bugfixes and documentation updates:
1298 - Update ganeti-os-interface documentation
1299 - Fixed a bug related to duplicate MACs or similar items which should be
1301 - Fix breakage in OS state modify
1302 - Reinstall instance: disallow offline secondaries (fixes bug related to
1303 OS changing but reinstall failing)
1304 - plus all the other fixes between 2.2.1 and 2.2.2
1310 *(Released Tue, 2 Nov 2010)*
1312 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1313 - Fixed master failover race with watcher
1314 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1316 - Accept previously stopped instance for export with instance removal
1317 - Simplify and extend the environment variables for instance OS scripts
1318 - Added new node flags, ``master_capable`` and ``vm_capable``
1319 - Added optional instance disk wiping prior during allocation. This is a
1320 cluster-wide option and can be set/modified using
1321 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1322 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1323 :doc:`install-quick`
1324 - Added a new watcher option (``--ignore-pause``)
1325 - Added option to ignore offline node on instance start/stop
1326 (``--ignore-offline``)
1327 - Allow overriding OS parameters with ``gnt-instance reinstall``
1328 - Added ability to change node's secondary IP address using ``gnt-node
1330 - Implemented privilege separation for all daemons except
1331 ``ganeti-noded``, see ``configure`` options
1332 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1334 - Implemented job priorities (see ``ganeti(7)`` manpage)
1335 - Ignore failures while shutting down instances during failover from
1337 - Exit daemon's bootstrap process only once daemon is ready
1338 - Export more information via ``LUInstanceQuery``/remote API
1339 - Improved documentation, QA and unittests
1340 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1341 restart if the file was created or changed
1342 - Added LUXI protocol version sent with each request and response,
1343 allowing detection of server/client mismatches
1344 - Moved the Python scripts among gnt-* and ganeti-* into modules
1345 - Moved all code related to setting up SSH to an external script,
1347 - Infrastructure changes for node group support in future versions
1353 *(Released Fri, 19 Nov 2010)*
1355 A few small bugs fixed, and some improvements to the build system:
1357 - Fix documentation regarding conversion to drbd
1358 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1360 - Fix error handling in node modify with multiple changes
1361 - Allow remote imports without checked names
1367 *(Released Tue, 19 Oct 2010)*
1369 - Disable SSL session ID cache in RPC client
1375 *(Released Thu, 14 Oct 2010)*
1377 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1378 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1380 - Fix problems with interaction between Curl and Python's HTTP server,
1381 resulting in increased speed in many RPC calls
1382 - Improve our release script to prevent breakage with older aclocal and
1389 *(Released Thu, 7 Oct 2010)*
1391 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1392 value retrieved from master
1393 - Added support for blacklisted or hidden OS definitions
1394 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1395 - Added support for -mem-path in KVM hypervisor abstraction layer
1396 - Allow overriding instance parameters in tool for inter-cluster
1397 instance moves (``tools/move-instance``)
1398 - Improved opcode summaries (e.g. in ``gnt-job list``)
1399 - Improve consistency of OS listing by sorting it
1400 - Documentation updates
1406 *(Released Fri, 8 Oct 2010)*
1408 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1414 *(Released Mon, 4 Oct 2010)*
1416 - Fixed regression in ``gnt-instance rename``
1422 *(Released Wed, 22 Sep 2010)*
1424 - Fixed OS_VARIANT variable for OS scripts
1425 - Fixed cluster tag operations via RAPI
1426 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1427 - Disabled RAPI CA checks in watcher
1433 *(Released Mon, 23 Aug 2010)*
1435 - Support DRBD versions of the format "a.b.c.d"
1437 - Re-introduce support for usage from multiple threads in RAPI client
1438 - Instance renames and modify via RAPI
1439 - Work around race condition between processing and archival in job
1441 - Mark opcodes following failed one as failed, too
1442 - Job field ``lock_status`` was removed due to difficulties making it
1443 work with the changed job queue in Ganeti 2.2; a better way to monitor
1444 locks is expected for a later 2.2.x release
1445 - Fixed dry-run behaviour with many commands
1446 - Support ``ssh-agent`` again when adding nodes
1447 - Many additional bugfixes
1453 *(Released Fri, 30 Jul 2010)*
1455 Important change: the internal RPC mechanism between Ganeti nodes has
1456 changed from using a home-grown http library (based on the Python base
1457 libraries) to use the PycURL library. This requires that PycURL is
1458 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1459 against GnuTLS by default. cURL's support for GnuTLS had known issues
1460 before cURL 7.21.0 and we recommend using the latest cURL release or
1461 linking against OpenSSL. Most other distributions already link PycURL
1462 and cURL against OpenSSL. The command::
1464 python -c 'import pycurl; print pycurl.version'
1466 can be used to determine the libraries PycURL and cURL are linked
1469 Other significant changes:
1471 - Rewrote much of the internals of the job queue, in order to achieve
1472 better parallelism; this decouples job query operations from the job
1473 processing, and it should allow much nicer behaviour of the master
1474 daemon under load, and it also has uncovered some long-standing bugs
1475 related to the job serialisation (now fixed)
1476 - Added a default iallocator setting to the cluster parameters,
1477 eliminating the need to always pass nodes or an iallocator for
1478 operations that require selection of new node(s)
1479 - Added experimental support for the LXC virtualization method
1480 - Added support for OS parameters, which allows the installation of
1481 instances to pass parameter to OS scripts in order to customise the
1483 - Added a hypervisor parameter controlling the migration type (live or
1484 non-live), since hypervisors have various levels of reliability; this
1485 has renamed the 'live' parameter to 'mode'
1486 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1487 logical volumes, meaning that cluster verify will ignore them and not
1488 flag their presence as errors
1489 - The watcher will now reset the error count for failed instances after
1490 8 hours, thus allowing self-healing if the problem that caused the
1491 instances to be down/fail to start has cleared in the meantime
1492 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1493 check for, and warn, if the drbd module parameter ``usermode_helper``
1494 is not consistent with the cluster-wide setting; this is needed to
1495 make diagnose easier of failed drbd creations
1496 - Started adding base IPv6 support, but this is not yet
1497 enabled/available for use
1498 - Rename operations (cluster, instance) will now return the new name,
1499 which is especially useful if a short name was passed in
1500 - Added support for instance migration in RAPI
1501 - Added a tool to pre-configure nodes for the SSH setup, before joining
1502 them to the cluster; this will allow in the future a simplified model
1503 for node joining (but not yet fully enabled in 2.2); this needs the
1504 paramiko python library
1505 - Fixed handling of name-resolving errors
1506 - Fixed consistency of job results on the error path
1507 - Fixed master-failover race condition when executed multiple times in
1509 - Fixed many bugs related to the job queue (mostly introduced during the
1510 2.2 development cycle, so not all are impacting 2.1)
1511 - Fixed instance migration with missing disk symlinks
1512 - Fixed handling of unknown jobs in ``gnt-job archive``
1513 - And many other small fixes/improvements
1517 - Enhanced both the unittest and the QA coverage
1518 - Switched the opcode validation to a generic model, and extended the
1519 validation to all opcode parameters
1520 - Changed more parts of the code that write shell scripts to use the
1522 - Switched the master daemon to use the asyncore library for the Luxi
1529 *(Released Thu, 17 Jun 2010)*
1531 - Added tool (``move-instance``) and infrastructure to move instances
1532 between separate clusters (see :doc:`separate documentation
1533 <move-instance>` and :doc:`design document <design-2.2>`)
1534 - Added per-request RPC timeout
1535 - RAPI now requires a Content-Type header for requests with a body (e.g.
1536 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1537 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1538 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1540 - Implemented initial support for running Ganeti daemons as separate
1541 users, see configure-time flags ``--with-user-prefix`` and
1542 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1544 - Instances can be removed after export (``gnt-backup export
1545 --remove-instance``)
1546 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1547 key (instead of 1024 bit)
1548 - Added new cluster configuration file for cluster domain secret
1549 - Import/export now use SSL instead of SSH
1550 - Added support for showing estimated time when exporting an instance,
1551 see the ``ganeti-os-interface(7)`` manpage and look for
1558 *(Released Tue, 16 Nov 2010)*
1560 Some more bugfixes. Unless critical bugs occur, this will be the last
1563 - Fix case of MAC special-values
1564 - Fix mac checker regex
1565 - backend: Fix typo causing "out of range" error
1566 - Add missing --units in gnt-instance list man page
1572 *(Released Tue, 24 Aug 2010)*
1575 - Don't ignore secondary node silently on non-mirrored disk templates
1577 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1578 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1579 - Properly document the usb_mouse parameter
1580 - Fix path in ganeti-rapi(8) (issue 116)
1581 - Adjust error message when the ganeti user's .ssh directory is
1583 - Add same-node-check when changing the disk template to drbd
1589 *(Released Fri, 16 Jul 2010)*
1592 - Add an option to only select some reboot types during qa/burnin.
1593 (on some hypervisors consequent reboots are not supported)
1594 - Fix infrequent race condition in master failover. Sometimes the old
1595 master ip address would be still detected as up for a short time
1596 after it was removed, causing failover to fail.
1597 - Decrease mlockall warnings when the ctypes module is missing. On
1598 Python 2.4 we support running even if no ctypes module is installed,
1599 but we were too verbose about this issue.
1600 - Fix building on old distributions, on which man doesn't have a
1602 - Fix RAPI not to ignore the MAC address on instance creation
1603 - Implement the old instance creation format in the RAPI client.
1609 *(Released Thu, 01 Jul 2010)*
1611 A small bugfix release:
1612 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1613 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1614 a non-existing option
1615 - Fix instance create: the --force-variant option was ignored
1616 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1617 - Fix modify node storage with non-FQDN arguments
1618 - Fix RAPI client to authenticate under Python 2.6 when used
1619 for more than 5 requests needing authentication
1620 - Fix gnt-instance modify -t (storage) giving a wrong error message
1621 when converting a non-shutdown drbd instance to plain
1627 *(Released Fri, 18 Jun 2010)*
1629 A small bugfix release:
1631 - Fix live migration of KVM instances started with older Ganeti
1632 versions which had fewer hypervisor parameters
1633 - Fix gnt-instance grow-disk on down instances
1634 - Fix an error-reporting bug during instance migration
1635 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1636 silently ignoring broken ones
1637 - Fix an RPC error reporting bug affecting, for example, RAPI client
1639 - Fix bug triggered by different API version os-es on different nodes
1640 - Fix a bug in instance startup with custom hvparams: OS level
1641 parameters would fail to be applied.
1642 - Fix the RAPI client under Python 2.6 (but more work is needed to
1643 make it work completely well with OpenSSL)
1644 - Fix handling of errors when resolving names from DNS
1650 *(Released Thu, 3 Jun 2010)*
1652 A medium sized development cycle. Some new features, and some
1653 fixes/small improvements/cleanups.
1655 Significant features
1656 ~~~~~~~~~~~~~~~~~~~~
1658 The node deamon now tries to mlock itself into memory, unless the
1659 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1660 its logs, and falls back to console logging. This allows emergency
1661 features such as ``gnt-node powercycle`` to work even in the event of a
1662 broken node disk (tested offlining the disk hosting the node's
1663 filesystem and dropping its memory caches; don't try this at home)
1665 KVM: add vhost-net acceleration support. It can be tested with a new
1666 enough version of the kernel and of qemu-kvm.
1668 KVM: Add instance chrooting feature. If you use privilege dropping for
1669 your VMs you can also now force them to chroot to an empty directory,
1670 before starting the emulated guest.
1672 KVM: Add maximum migration bandwith and maximum downtime tweaking
1673 support (requires a new-enough version of qemu-kvm).
1675 Cluster verify will now warn if the master node doesn't have the master
1676 ip configured on it.
1678 Add a new (incompatible) instance creation request format to RAPI which
1679 supports all parameters (previously only a subset was supported, and it
1680 wasn't possible to extend the old format to accomodate all the new
1681 features. The old format is still supported, and a client can check for
1682 this feature, before using it, by checking for its presence in the
1683 ``features`` RAPI resource.
1685 Now with ancient latin support. Try it passing the ``--roman`` option to
1686 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1687 (requires the python-roman module to be installed, in order to work).
1692 As usual many internal code refactorings, documentation updates, and
1695 - Lots of improvements and cleanups to the experimental Remote API
1696 (RAPI) client library.
1697 - A new unit test suite for the core daemon libraries.
1698 - A fix to creating missing directories makes sure the umask is not
1699 applied anymore. This enforces the same directory permissions
1701 - Better handling terminating daemons with ctrl+c (used when running
1702 them in debugging mode).
1703 - Fix a race condition in live migrating a KVM instance, when stat()
1704 on the old proc status file returned EINVAL, which is an unexpected
1706 - Fixed manpage checking with newer man and utf-8 charachters. But now
1707 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1713 *(Released Fri, 7 May 2010)*
1715 Fix a bug which prevented untagged KVM instances from starting.
1721 *(Released Fri, 7 May 2010)*
1723 Another release with a long development cycle, during which many
1724 different features were added.
1726 Significant features
1727 ~~~~~~~~~~~~~~~~~~~~
1729 The KVM hypervisor now can run the individual instances as non-root, to
1730 reduce the impact of a VM being hijacked due to bugs in the
1731 hypervisor. It is possible to run all instances as a single (non-root)
1732 user, to manually specify a user for each instance, or to dynamically
1733 allocate a user out of a cluster-wide pool to each instance, with the
1734 guarantee that no two instances will run under the same user ID on any
1737 An experimental RAPI client library, that can be used standalone
1738 (without the other Ganeti libraries), is provided in the source tree as
1739 ``lib/rapi/client.py``. Note this client might change its interface in
1740 the future, as we iterate on its capabilities.
1742 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1743 replace the cluster's certificates and crypto keys. This might help in
1744 case they have been compromised, or have simply expired.
1746 A new disk option for instance creation has been added that allows one
1747 to "adopt" currently existing logical volumes, with data
1748 preservation. This should allow easier migration to Ganeti from
1749 unmanaged (or managed via other software) instances.
1751 Another disk improvement is the possibility to convert between redundant
1752 (DRBD) and plain (LVM) disk configuration for an instance. This should
1753 allow better scalability (starting with one node and growing the
1754 cluster, or shrinking a two-node cluster to one node).
1756 A new feature that could help with automated node failovers has been
1757 implemented: if a node sees itself as offline (by querying the master
1758 candidates), it will try to shutdown (hard) all instances and any active
1759 DRBD devices. This reduces the risk of duplicate instances if an
1760 external script automatically failovers the instances on such nodes. To
1761 enable this, the cluster parameter ``maintain_node_health`` should be
1762 enabled; in the future this option (per the name) will enable other
1763 automatic maintenance features.
1765 Instance export/import now will reuse the original instance
1766 specifications for all parameters; that means exporting an instance,
1767 deleting it and the importing it back should give an almost identical
1768 instance. Note that the default import behaviour has changed from
1769 before, where it created only one NIC; now it recreates the original
1772 Cluster verify has added a few new checks: SSL certificates validity,
1773 /etc/hosts consistency across the cluster, etc.
1778 As usual, many internal changes were done, documentation fixes,
1781 - Fixed cluster initialization with disabled cluster storage (regression
1782 introduced in 2.1.1)
1783 - File-based storage supports growing the disks
1784 - Fixed behaviour of node role changes
1785 - Fixed cluster verify for some corner cases, plus a general rewrite of
1786 cluster verify to allow future extension with more checks
1787 - Fixed log spamming by watcher and node daemon (regression introduced
1789 - Fixed possible validation issues when changing the list of enabled
1791 - Fixed cleanup of /etc/hosts during node removal
1792 - Fixed RAPI response for invalid methods
1793 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1794 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1795 from ide disks, etc.)
1796 - Allow OS changes without re-installation (to record a changed OS
1797 outside of Ganeti, or to allow OS renames)
1798 - Allow instance creation without OS installation (useful for example if
1799 the OS will be installed manually, or restored from a backup not in
1801 - Implemented option to make cluster ``copyfile`` use the replication
1803 - Added list of enabled hypervisors to ssconf (possibly useful for
1805 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1807 - A partial form of node re-IP is possible via node readd, which now
1808 allows changed node primary IP
1809 - Command line utilities now show an informational message if the job is
1811 - The logs of the master daemon now show the PID/UID/GID of the
1818 *(Released Fri, 12 Mar 2010)*
1820 During the 2.1.0 long release candidate cycle, a lot of improvements and
1821 changes have accumulated with were released later as 2.1.1.
1826 The node evacuate command (``gnt-node evacuate``) was significantly
1827 rewritten, and as such the IAllocator protocol was changed - a new
1828 request type has been added. This unfortunate change during a stable
1829 series is designed to improve performance of node evacuations; on
1830 clusters with more than about five nodes and which are well-balanced,
1831 evacuation should proceed in parallel for all instances of the node
1832 being evacuated. As such, any existing IAllocator scripts need to be
1833 updated, otherwise the above command will fail due to the unknown
1834 request. The provided "dumb" allocator has not been updated; but the
1835 ganeti-htools package supports the new protocol since version 0.2.4.
1837 Another important change is increased validation of node and instance
1838 names. This might create problems in special cases, if invalid host
1839 names are being used.
1841 Also, a new layer of hypervisor parameters has been added, that sits at
1842 OS level between the cluster defaults and the instance ones. This allows
1843 customisation of virtualization parameters depending on the installed
1844 OS. For example instances with OS 'X' may have a different KVM kernel
1845 (or any other parameter) than the cluster defaults. This is intended to
1846 help managing a multiple OSes on the same cluster, without manual
1847 modification of each instance's parameters.
1849 A tool for merging clusters, ``cluster-merge``, has been added in the
1850 tools sub-directory.
1855 - Improved the int/float conversions that should make the code more
1856 robust in face of errors from the node daemons
1857 - Fixed the remove node code in case of internal configuration errors
1858 - Fixed the node daemon behaviour in face of inconsistent queue
1859 directory (e.g. read-only file-system where we can't open the files
1861 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1862 now it either aborts cleanly or, if given the new "auto_promote"
1863 parameter, will automatically promote other nodes as needed
1864 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1865 completely prevent Ganeti from working
1866 - Fixed bug for instance export when not all disks were successfully
1868 - Fixed behaviour of node add when the new node is slow in starting up
1870 - Fixed handling of signals in the LUXI client, which should improve
1871 behaviour of command-line scripts
1872 - Added checks for invalid node/instance names in the configuration (now
1873 flagged during cluster verify)
1874 - Fixed watcher behaviour for disk activation errors
1875 - Fixed two potentially endless loops in http library, which led to the
1876 RAPI daemon hanging and consuming 100% CPU in some cases
1877 - Fixed bug in RAPI daemon related to hashed passwords
1878 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1880 - Enhanced compatibility with non-Debian OSes, but not using absolute
1881 path in some commands and allowing customisation of the ssh
1882 configuration directory
1883 - Fixed possible future issue with new Python versions by abiding to the
1884 proper use of ``__slots__`` attribute on classes
1885 - Added checks that should prevent directory traversal attacks
1886 - Many documentation fixes based on feedback from users
1891 - Added an "early_release" more for instance replace disks and node
1892 evacuate, where we release locks earlier and thus allow higher
1893 parallelism within the cluster
1894 - Added watcher hooks, intended to allow the watcher to restart other
1895 daemons (e.g. from the ganeti-nbma project), but they can be used of
1896 course for any other purpose
1897 - Added a compile-time disable for DRBD barriers, to increase
1898 performance if the administrator trusts the power supply or the
1899 storage system to not lose writes
1900 - Added the option of using syslog for logging instead of, or in
1901 addition to, Ganeti's own log files
1902 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1903 can indeed boot from a paravirtual NIC
1904 - Added a generic debug level for many operations; while this is not
1905 used widely yet, it allows one to pass the debug value all the way to
1907 - Enhanced the hooks environment for instance moves (failovers,
1908 migrations) where the primary/secondary nodes changed during the
1909 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1910 - Enhanced data validations for many user-supplied values; one important
1911 item is the restrictions imposed on instance and node names, which
1912 might reject some (invalid) host names
1913 - Add a configure-time option to disable file-based storage, if it's not
1914 needed; this allows greater security separation between the master
1915 node and the other nodes from the point of view of the inter-node RPC
1917 - Added user notification in interactive tools if job is waiting in the
1918 job queue or trying to acquire locks
1919 - Added log messages when a job is waiting for locks
1920 - Added filtering by node tags in instance operations which admit
1921 multiple instances (start, stop, reboot, reinstall)
1922 - Added a new tool for cluster mergers, ``cluster-merge``
1923 - Parameters from command line which are of the form ``a=b,c=d`` can now
1924 use backslash escapes to pass in values which contain commas,
1925 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1927 - For KVM, the instance name is the first parameter passed to KVM, so
1928 that it's more visible in the process list
1934 *(Released Tue, 2 Mar 2010)*
1936 Ganeti 2.1 brings many improvements with it. Major changes:
1938 - Added infrastructure to ease automated disk repairs
1939 - Added new daemon to export configuration data in a cheaper way than
1940 using the remote API
1941 - Instance NICs can now be routed instead of being associated with a
1943 - Improved job locking logic to reduce impact of jobs acquiring multiple
1944 locks waiting for other long-running jobs
1946 In-depth implementation details can be found in the Ganeti 2.1 design
1952 - Added chroot hypervisor
1953 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1955 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1956 ``bootloader_path`` and ``bootloader_args``)
1957 - Added the ``use_localtime`` option for the xen-hvm and kvm
1958 hypervisors, and the default value for this has changed to false (in
1959 2.0 xen-hvm always enabled it)
1960 - Added luxi call to submit multiple jobs in one go
1961 - Added cluster initialization option to not modify ``/etc/hosts``
1963 - Added network interface parameters
1964 - Added dry run mode to some LUs
1965 - Added RAPI resources:
1967 - ``/2/instances/[instance_name]/info``
1968 - ``/2/instances/[instance_name]/replace-disks``
1969 - ``/2/nodes/[node_name]/evacuate``
1970 - ``/2/nodes/[node_name]/migrate``
1971 - ``/2/nodes/[node_name]/role``
1972 - ``/2/nodes/[node_name]/storage``
1973 - ``/2/nodes/[node_name]/storage/modify``
1974 - ``/2/nodes/[node_name]/storage/repair``
1976 - Added OpCodes to evacuate or migrate all instances on a node
1977 - Added new command to list storage elements on nodes (``gnt-node
1978 list-storage``) and modify them (``gnt-node modify-storage``)
1979 - Added new ssconf files with master candidate IP address
1980 (``ssconf_master_candidates_ips``), node primary IP address
1981 (``ssconf_node_primary_ips``) and node secondary IP address
1982 (``ssconf_node_secondary_ips``)
1983 - Added ``ganeti-confd`` and a client library to query the Ganeti
1984 configuration via UDP
1985 - Added ability to run hooks after cluster initialization and before
1987 - Added automatic mode for disk replace (``gnt-instance replace-disks
1989 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1990 after catastrophic data-loss
1991 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1993 - Added ``gnt-instance move`` command to move instances
1994 - Added ``gnt-cluster watcher`` command to control watcher
1995 - Added ``gnt-node powercycle`` command to powercycle nodes
1996 - Added new job status field ``lock_status``
1997 - Added parseable error codes to cluster verification (``gnt-cluster
1998 verify --error-codes``) and made output less verbose (use
1999 ``--verbose`` to restore previous behaviour)
2000 - Added UUIDs to the main config entities (cluster, nodes, instances)
2001 - Added support for OS variants
2002 - Added support for hashed passwords in the Ganeti remote API users file
2004 - Added option to specify maximum timeout on instance shutdown
2005 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2006 - Added new helper script to start and stop Ganeti daemons
2007 (``daemon-util``), with the intent to reduce the work necessary to
2008 adjust Ganeti for non-Debian distributions and to start/stop daemons
2010 - Added more unittests
2011 - Fixed critical bug in ganeti-masterd startup
2012 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2013 now customisable at the cluster level for both the KVM and Xen
2014 hypervisors using the new ``migration_port`` parameter
2015 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2016 reinstalling an instance
2017 - Allowed ``@`` in tag names
2018 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2019 - Many documentation updates
2020 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2021 - ``gnt-instance reinstall`` can now reinstall multiple instances
2022 - Updated many command line parameters
2023 - Introduced new OS API version 15
2024 - No longer support a default hypervisor
2025 - Treat virtual LVs as inexistent
2026 - Improved job locking logic to reduce lock contention
2027 - Match instance and node names case insensitively
2028 - Reimplemented bash completion script to be more complete
2035 *(Released Thu, 4 Feb 2010)*
2037 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2038 - Fix a string formatting bug
2039 - Improve safety of the code in some error paths
2040 - Improve data validation in the master of values returned from nodes
2046 *(Released Thu, 17 Dec 2009)*
2048 - Fix security issue due to missing validation of iallocator names; this
2049 allows local and remote execution of arbitrary executables
2050 - Fix failure of gnt-node list during instance removal
2051 - Ship the RAPI documentation in the archive
2057 *(Released Wed, 30 Sep 2009)*
2059 - Fixed many wrong messages
2060 - Fixed a few bugs related to the locking library
2061 - Fixed MAC checking at instance creation time
2062 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2063 - Fixed a few issues related to signal handling in both daemons and
2065 - Fixed the example startup script provided
2066 - Fixed insserv dependencies in the example startup script (patch from
2068 - Fixed handling of drained nodes in the iallocator framework
2069 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2071 - Fixed error related to invalid job IDs in job polling
2072 - Fixed job/opcode persistence on unclean master shutdown
2073 - Fixed handling of partial job processing after unclean master
2075 - Fixed error reporting from LUs, previously all errors were converted
2076 into execution errors
2077 - Fixed error reporting from burnin
2078 - Decreased significantly the memory usage of the job queue
2079 - Optimised slightly multi-job submission
2080 - Optimised slightly opcode loading
2081 - Backported the multi-job submit framework from the development
2082 branch; multi-instance start and stop should be faster
2083 - Added script to clean archived jobs after 21 days; this will reduce
2084 the size of the queue directory
2085 - Added some extra checks in disk size tracking
2086 - Added an example ethers hook script
2087 - Added a cluster parameter that prevents Ganeti from modifying of
2089 - Added more node information to RAPI responses
2090 - Added a ``gnt-job watch`` command that allows following the ouput of a
2092 - Added a bind-address option to ganeti-rapi
2093 - Added more checks to the configuration verify
2094 - Enhanced the burnin script such that some operations can be retried
2096 - Converted instance reinstall to multi-instance model
2102 *(Released Fri, 7 Aug 2009)*
2104 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2105 to allow using the pre-2.0.2 behaviour in activation, if any existing
2106 instances have mismatched disk sizes in the configuration
2107 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2108 any configuration mismatches for disk sizes
2109 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2110 failover to work on two-node clusters
2111 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2113 - Fixed detection of OS script errors in ``gnt-backup export``
2114 - Fixed exit code of ``gnt-backup export``
2120 *(Released Fri, 17 Jul 2009)*
2122 - Added experimental support for stripped logical volumes; this should
2123 enhance performance but comes with a higher complexity in the block
2124 device handling; stripping is only enabled when passing
2125 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2126 affected even in the non-stripped mode
2127 - Improved resiliency against transient failures at the end of DRBD
2128 resyncs, and in general of DRBD resync checks
2129 - Fixed a couple of issues with exports and snapshot errors
2130 - Fixed a couple of issues in instance listing
2131 - Added display of the disk size in ``gnt-instance info``
2132 - Fixed checking for valid OSes in instance creation
2133 - Fixed handling of the "vcpus" parameter in instance listing and in
2134 general of invalid parameters
2135 - Fixed http server library, and thus RAPI, to handle invalid
2136 username/password combinations correctly; this means that now they
2137 report unauthorized for queries too, not only for modifications,
2138 allowing earlier detect of configuration problems
2139 - Added a new "role" node list field, equivalent to the master/master
2140 candidate/drained/offline flags combinations
2141 - Fixed cluster modify and changes of candidate pool size
2142 - Fixed cluster verify error messages for wrong files on regular nodes
2143 - Fixed a couple of issues with node demotion from master candidate role
2144 - Fixed node readd issues
2145 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2146 - Added a new ``--no-voting`` option for masterfailover to fix failover
2147 on two-nodes clusters when the former master node is unreachable
2148 - Added instance reinstall over RAPI
2154 *(Released Tue, 16 Jun 2009)*
2156 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2157 allow re-adding the start in single-user option (regression from 1.2)
2158 - the watcher writes the instance status to a file, to allow monitoring
2159 to report the instance status (from the master) based on cached
2160 results of the watcher's queries; while this can get stale if the
2161 watcher is being locked due to other work on the cluster, this is
2162 still an improvement
2163 - the watcher now also restarts the node daemon and the rapi daemon if
2165 - fixed the watcher to handle full and drained queue cases
2166 - hooks export more instance data in the environment, which helps if
2167 hook scripts need to take action based on the instance's properties
2168 (no longer need to query back into ganeti)
2169 - instance failovers when the instance is stopped do not check for free
2170 RAM, so that failing over a stopped instance is possible in low memory
2172 - rapi uses queries for tags instead of jobs (for less job traffic), and
2173 for cluster tags it won't talk to masterd at all but read them from
2175 - a couple of error handling fixes in RAPI
2176 - drbd handling: improved the error handling of inconsistent disks after
2177 resync to reduce the frequency of "there are some degraded disks for
2178 this instance" messages
2179 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2180 error handling path called a wrong function name)
2186 *(Released Wed, 27 May 2009)*
2188 - no changes from rc5
2194 *(Released Wed, 20 May 2009)*
2196 - fix a couple of bugs (validation, argument checks)
2197 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2198 - some small improvements to RAPI and IAllocator
2199 - make watcher automatically start the master daemon if down
2205 *(Released Mon, 27 Apr 2009)*
2207 - change the OS list to not require locks; this helps with big clusters
2208 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2209 volume group is broken
2210 - ``gnt-instance info``, without any arguments, doesn't run for all
2211 instances anymore; either pass ``--all`` or pass the desired
2212 instances; this helps against mistakes on big clusters where listing
2213 the information for all instances takes a long time
2214 - miscellaneous doc and man pages fixes
2220 *(Released Wed, 8 Apr 2009)*
2222 - Change the internal locking model of some ``gnt-node`` commands, in
2223 order to reduce contention (and blocking of master daemon) when
2224 batching many creation/reinstall jobs
2225 - Fixes to Xen soft reboot
2226 - No longer build documentation at build time, instead distribute it in
2227 the archive, in order to reduce the need for the whole docbook/rst
2234 *(Released Fri, 27 Mar 2009)*
2236 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2237 - Fix watcher startup sequence, improves the behaviour of busy clusters
2238 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2239 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2240 bind address and other places
2241 - Some documentation fixes and updates
2247 *(Released Mon, 2 Mar 2009)*
2249 - More documentation updates, now all docs should be more-or-less
2251 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2253 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2255 - Fix an issue related to $libdir/run/ganeti and cluster creation
2261 *(Released Thu, 19 Feb 2009)*
2263 - Xen PVM and KVM have switched the default value for the instance root
2264 disk to the first partition on the first drive, instead of the whole
2265 drive; this means that the OS installation scripts must be changed
2267 - Man pages have been updated
2268 - RAPI has been switched by default to HTTPS, and the exported functions
2269 should all work correctly
2270 - RAPI v1 has been removed
2271 - Many improvements to the KVM hypervisor
2272 - Block device errors are now better reported
2273 - Many other bugfixes and small improvements
2279 *(Released Mon, 26 Jan 2009)*
2281 - Version 2 is a general rewrite of the code and therefore the
2282 differences are too many to list, see the design document for 2.0 in
2283 the ``doc/`` subdirectory for more details
2284 - In this beta version there is not yet a migration path from 1.2 (there
2285 will be one in the final 2.0 release)
2286 - A few significant changes are:
2288 - all commands are executed by a daemon (``ganeti-masterd``) and the
2289 various ``gnt-*`` commands are just front-ends to it
2290 - all the commands are entered into, and executed from a job queue,
2291 see the ``gnt-job(8)`` manpage
2292 - the RAPI daemon supports read-write operations, secured by basic
2293 HTTP authentication on top of HTTPS
2294 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2295 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2296 to migrate to DRBD 8 first while still running Ganeti 1.2)
2297 - DRBD devices are using statically allocated minor numbers, which
2298 will be assigned to existing instances during the migration process
2299 - there is support for both Xen PVM and Xen HVM instances running on
2301 - KVM virtualization is supported too
2302 - file-based storage has been implemented, which means that it is
2303 possible to run the cluster without LVM and DRBD storage, for
2304 example using a shared filesystem exported from shared storage (and
2305 still have live migration)
2311 *(Released Tue, 13 Jan 2009)*
2313 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2314 - Reuse the old instance mac address by default on instance import, if
2315 the instance name is the same.
2316 - Handle situations in which the node info rpc returns incomplete
2318 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2319 - Improved version of batcher:
2321 - state file support
2322 - instance mac address support
2323 - support for HVM clusters/instances
2325 - Add an option to show the number of cpu sockets and nodes in
2327 - Support OSes that handle more than one version of the OS api (but do
2328 not change the current API in any other way)
2329 - Fix ``gnt-node migrate``
2330 - ``gnt-debug`` man page
2331 - Fixes various more typos and small issues
2332 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2338 *(Released Wed, 24 Sep 2008)*
2340 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2341 type of disk exported to fully virtualized instances.
2342 - provide access to the serial console of HVM instances
2343 - instance auto_balance flag, set by default. If turned off it will
2344 avoid warnings on cluster verify if there is not enough memory to fail
2345 over an instance. in the future it will prevent automatically failing
2346 it over when we will support that.
2347 - batcher tool for instance creation, see ``tools/README.batcher``
2348 - ``gnt-instance reinstall --select-os`` to interactively select a new
2349 operating system when reinstalling an instance.
2350 - when changing the memory amount on instance modify a check has been
2351 added that the instance will be able to start. also warnings are
2352 emitted if the instance will not be able to fail over, if auto_balance
2354 - documentation fixes
2355 - sync fields between ``gnt-instance list/modify/add/import``
2356 - fix a race condition in drbd when the sync speed was set after giving
2357 the device a remote peer.
2363 *(Released Tue, 22 Jul 2008)*
2365 - note: the allowed size and number of tags per object were reduced
2366 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2367 - fixed twisted 8.x compatibility
2368 - fixed ``gnt-instance replace-disks`` with iallocator
2369 - add TCP keepalives on twisted connections to detect restarted nodes
2370 - disk increase support, see ``gnt-instance grow-disk``
2371 - implement bulk node/instance query for RAPI
2372 - add tags in node/instance listing (optional)
2373 - experimental migration (and live migration) support, read the man page
2374 for ``gnt-instance migrate``
2375 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2376 has some small improvements in handling its state file
2382 *(Released Fri, 13 Jun 2008)*
2384 - Experimental readonly, REST-based remote API implementation;
2385 automatically started on master node, TCP port 5080, if enabled by
2386 ``--enable-rapi`` parameter to configure script.
2387 - Instance allocator support. Add and import instance accept a
2388 ``--iallocator`` parameter, and call that instance allocator to decide
2389 which node to use for the instance. The iallocator document describes
2390 what's expected from an allocator script.
2391 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2392 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2393 node is lost there is still enough memory to fail over the instances
2395 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2396 ``gnt-cluster verify``, to check for site-specific compliance. All the
2397 hooks will run, and their output, if any, will be displayed. Any
2398 failing hook will make the verification return an error value.
2399 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2400 primary and secondary interfaces
2401 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2402 that is still declared as part of the cluster and has failed.
2403 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2404 specifying output fields, that just adds the chosen fields to the
2406 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2407 export from the filesystem.
2408 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2409 have been added. Using them you can enable/disable acpi and pae
2410 support, and specify a path for a cd image to be exported to the
2411 instance. These parameters as the name suggest only work on HVM
2413 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2414 PAE support will be set to the previously hardcoded values, but the
2415 (previously hardcoded) path to the CDROM ISO image will be unset and
2416 if required, needs to be set manually with ``gnt-instance modify``
2418 - The address to which an instance's VNC console is bound is now
2419 selectable per-instance, rather than being cluster wide. Of course
2420 this only applies to instances controlled via VNC, so currently just
2421 applies to HVM clusters.
2427 *(Released Mon, 18 Feb 2008)*
2429 - more tweaks to the disk activation code (especially helpful for DRBD)
2430 - change the default ``gnt-instance list`` output format, now there is
2431 one combined status field (see the manpage for the exact values this
2433 - some more fixes for the mac export to hooks change
2434 - make Ganeti not break with DRBD 8.2.x (which changed the version
2435 format in ``/proc/drbd``) (issue 24)
2436 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2437 template, allowing migration from DRBD0.7+MD to DRBD8
2443 *(Released Wed, 30 Jan 2008)*
2445 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2447 - add command aliases infrastructure and a few aliases
2448 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2449 man pages and the ``--help`` option of ``gnt-node
2450 list``/``gnt-instance list``
2451 - fix ``gnt-backup list`` with down nodes (issue 21)
2452 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2453 - fix the dist archive and add a check for including svn/git files in
2455 - some developer-related changes: improve the burnin and the QA suite,
2456 add an upload script for testing during development
2462 *(Released Wed, 16 Jan 2008)*
2464 - experimental HVM support, read the install document, section
2465 "Initializing the cluster"
2466 - allow for the PVM hypervisor per-instance kernel and initrd paths
2467 - add a new command ``gnt-cluster verify-disks`` which uses a new
2468 algorithm to improve the reconnection of the DRBD pairs if the device
2469 on the secondary node has gone away
2470 - make logical volume code auto-activate LVs at disk activation time
2471 - slightly improve the speed of activating disks
2472 - allow specification of the MAC address at instance creation time, and
2473 changing it later via ``gnt-instance modify``
2474 - fix handling of external commands that generate lots of output on
2476 - update documentation with regard to minimum version of DRBD8 supported
2482 *(Released Tue, 4 Dec 2007)*
2484 - Log the ``xm create`` output to the node daemon log on failure (to
2485 help diagnosing the error)
2486 - In debug mode, log all external commands output if failed to the logs
2487 - Change parsing of lvm commands to ignore stderr
2493 *(Released Wed, 28 Nov 2007)*
2495 - Another round of updates to the DRBD 8 code to deal with more failures
2496 in the replace secondary node operation
2497 - Some more logging of failures in disk operations (lvm, drbd)
2498 - A few documentation updates
2505 *(Released Tue, 13 Nov 2007)*
2507 - Change configuration file format from Python's Pickle to JSON.
2508 Upgrading is possible using the cfgupgrade utility.
2509 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2510 faster replace disks and is more stable (DRBD 8 has many improvements
2511 compared to DRBD 0.7)
2512 - Added command line tags support (see man pages for ``gnt-instance``,
2513 ``gnt-node``, ``gnt-cluster``)
2514 - Added instance rename support
2515 - Added multi-instance startup/shutdown
2516 - Added cluster rename support
2517 - Added ``gnt-node evacuate`` to simplify some node operations
2518 - Added instance reboot operation that can speedup reboot as compared to
2520 - Soften the requirement that hostnames are in FQDN format
2521 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2523 - Removed dependency on debian's patched fping that uses the
2524 non-standard ``-S`` option
2525 - Now the OS definitions are searched for in multiple, configurable
2526 paths (easier for distros to package)
2527 - Some changes to the hooks infrastructure (especially the new
2528 post-configuration update hook)
2529 - Other small bugfixes
2531 .. vim: set textwidth=72 syntax=rst :