10 - The minimum Python version needed to run Ganeti is now 2.6.
11 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
12 to always require authentication, ``--require-authentication``. It can
13 be specified in ``$sysconfdir/default/ganeti``.
19 *(Released Wed, 6 Feb 2013)*
21 - ``gnt-instance batch-create`` has been changed to use the bulk create
22 opcode from Ganeti. This lead to incompatible changes in the format of
23 the JSON file. It's now not a custom dict anymore but a dict
24 compatible with the ``OpInstanceCreate`` opcode.
25 - Parent directories for file storage need to be listed in
26 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
27 write the file automatically based on old configuration values, but it
28 can not distribute it across all nodes and the file contents should be
29 verified. Use ``gnt-cluster copyfile
30 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
31 upgraded. The reason for requiring this list of paths now is that
32 before it would have been possible to inject new paths via RPC,
33 allowing files to be created in arbitrary locations. The RPC protocol
34 is protected using SSL/X.509 certificates, but as a design principle
35 Ganeti does not permit arbitrary paths to be passed.
36 - The parsing of the variants file for OSes (see
37 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
38 lines and comment lines (starting with ``#``) are ignored for better
40 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
41 longer available. ``gnt-node add`` now invokes a new tool on the
42 destination node, named ``prepare-node-join``, to configure the SSH
43 daemon. Paramiko is no longer necessary to configure nodes' SSH
44 daemons via ``gnt-node add``.
45 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
46 for RAPI users. It allows granting permissions to query for
47 information to a specific user without giving
48 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
49 - A new tool named ``node-cleanup`` has been added. It cleans remains of
50 a cluster from a machine by stopping all daemons, removing
51 certificates and ssconf files. Unless the ``--no-backup`` option is
52 given, copies of the certificates are made.
53 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
54 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
55 the flag is not reset after a master failover.
56 - Instance creations now support the use of opportunistic locking,
57 potentially speeding up the (parallel) creation of multiple instances.
58 This feature is currently only available via the :doc:`RAPI
59 <rapi>` interface and when an instance allocator is used. If the
60 ``opportunistic_locking`` parameter is set the opcode will try to
61 acquire as many locks as possible, but will not wait for any locks
62 held by other opcodes. If not enough resources can be found to
63 allocate the instance, the temporary error code
64 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
65 retried thereafter, with or without opportunistic locking.
66 - The functionality for allocating multiple instances at once has been
67 overhauled and is now also available through :doc:`RAPI <rapi>`.
68 - Man pages can now be included when the documentation is built, in
69 which case the output is in ``doc/man-html``. The configure-time
70 option is ``--enable-manpages-in-doc``. Sphinx 1.0 or higher is
72 - A new htool called ``hroller`` helps scheduling parallel shutdown of
73 nodes on clusters where all instances are shut down, but without ever
74 rebooting both primary and secondary node for an instance at the same
76 - Queries not needing locks or RPC access to the node can now be
77 performed by the confd daemon, making them independent from jobs, and
78 thus faster to execute. This is selectable at configure time.
79 - The KVM hypervisor has been updated with many new hypervisor
80 parameters, including a generic one for passing arbitrary command line
81 values. See a complete list in :manpage:`gnt-instance(8)`.
82 - A new tool, called ``mon-collector``, is the stand-alone executor of
83 the data collectors for a monitoring system. As of this version, it
84 just includes the DRBD data collector, that can be executed by calling
85 ``mon-collector`` using the ``drbd`` parameter. See
86 :manpage:`mon-collector(7)`.
87 - New ``exclusive-storage`` node parameter added. When it's set to true,
88 physical disks are assigned in an exclusive fashion to instances, as
89 documented in :doc:`Partitioned Ganeti <design-partitioned>`.
90 Currently, only instances using the ``plain`` disk template are
92 - New network management functionality to support automatic allocation
93 of IP addresses and managing of network parameters. See
94 :manpage:`gnt-network(8)` for more details.
95 - New external storage backend, to allow managing arbitrary storage
96 systems external to the cluster. See
97 :manpage:`ganeti-extstorage-interface(7)`.
98 - Instance renames of LVM-based instances will now update the LV tags
99 (which can be used to recover the instance-to-LV mapping in case of
106 *(Released Fri, 21 Dec 2012)*
108 Important behaviour change: hbal won't rebalance anymore instances which
109 have the ``auto_balance`` attribute set to false. This was the intention
110 all along, but until now it only skipped those from the N+1 memory
111 reservation (DRBD-specific).
113 A significant number of bug fixes in this release:
115 - Fixed disk adoption interaction with ipolicy checks.
116 - Fixed networking issues when instances are started, stopped or
117 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
118 - Fixed the warning in cluster verify for shared storage instances not
120 - Fixed removal of storage directory on shared file storage (issue 262).
121 - Fixed validation of LVM volume group name in OpClusterSetParams
122 (``gnt-cluster modify``) (issue 285).
123 - Fixed runtime memory increases (``gnt-instance modify -m``).
124 - Fixed live migration under Xen's ``xl`` mode.
125 - Fixed ``gnt-instance console`` with ``xl``.
126 - Fixed building with newer Haskell compiler/libraries.
127 - Fixed PID file writing in Haskell daemons (confd); this prevents
128 restart issues if confd was launched manually (outside of
129 ``daemon-util``) while another copy of it was running
130 - Fixed a type error when doing live migrations with KVM (issue 297) and
131 the error messages for failing migrations have been improved.
132 - Fixed opcode validation for the out-of-band commands (``gnt-node
134 - Fixed a type error when unsetting OS hypervisor parameters (issue
135 311); now it's possible to unset all OS-specific hypervisor
137 - Fixed the ``dry-run`` mode for many operations: verification of
138 results was over-zealous but didn't take into account the ``dry-run``
139 operation, resulting in "wrong" failures.
140 - Fixed bash completion in ``gnt-job list`` when the job queue has
141 hundreds of entries; especially with older ``bash`` versions, this
142 results in significant CPU usage.
144 And lastly, a few other improvements have been made:
146 - Added option to force master-failover without voting (issue 282).
147 - Clarified error message on lock conflict (issue 287).
148 - Logging of newly submitted jobs has been improved (issue 290).
149 - Hostname checks have been made uniform between instance rename and
151 - The ``--submit`` option is now supported by ``gnt-debug delay``.
152 - Shutting down the master daemon by sending SIGTERM now stops it from
153 processing jobs waiting for locks; instead, those jobs will be started
154 once again after the master daemon is started the next time (issue
156 - Support for Xen's ``xl`` program has been improved (besides the fixes
158 - Reduced logging noise in the Haskell confd daemon (only show one log
159 entry for each config reload, instead of two).
160 - Several man page updates and typo fixes.
166 *(Released Fri, 12 Oct 2012)*
168 A small bugfix release. Among the bugs fixed:
170 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
171 made the command unusable.
172 - Commands that issue many jobs don't fail anymore just because some jobs
173 take so long that other jobs are archived.
174 - Failures during ``gnt-instance reinstall`` are reflected by the exit
176 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
178 - When ``always_failover`` is set, ``--allow-failover`` is not required
179 in migrate commands anymore.
180 - ``bash_completion`` works even if extglob is disabled.
181 - Fixed bug with locks that made failover for RDB-based instances fail.
182 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
183 a random node instead of one based on the allocator metric.
184 - Support for newer versions of pylint and pep8.
185 - Hail doesn't fail anymore when trying to add an instance of type
186 ``file``, ``sharedfile`` or ``rbd``.
187 - Added new Makefile target to rebuild the whole distribution, so that
188 all files are included.
194 *(Released Fri, 27 Jul 2012)*
197 .. attention:: The ``LUXI`` protocol has been made more consistent
198 regarding its handling of command arguments. This, however, leads to
199 incompatibility issues with previous versions. Please ensure that you
200 restart Ganeti daemons soon after the upgrade, otherwise most
201 ``LUXI`` calls (job submission, setting/resetting the drain flag,
202 pausing/resuming the watcher, cancelling and archiving jobs, querying
203 the cluster configuration) will fail.
212 The current ``admin_up`` field, which used to denote whether an instance
213 should be running or not, has been removed. Instead, ``admin_state`` is
214 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
216 The rational behind this is that an instance being “down” can have
219 - it could be down during a reboot
220 - it could be temporarily be down for a reinstall
221 - or it could be down because it is deprecated and kept just for its
224 The previous Boolean state was making it difficult to do capacity
225 calculations: should Ganeti reserve memory for a down instance? Now, the
226 tri-state field makes it clear:
228 - in ``up`` and ``down`` state, all resources are reserved for the
229 instance, and it can be at any time brought up if it is down
230 - in ``offline`` state, only disk space is reserved for it, but not
233 The field can have an extra use: since the transition between ``up`` and
234 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
235 transition between ``offline`` and ``down`` is done via ``gnt-instance
236 modify``, it is possible to given different rights to users. For
237 example, owners of an instance could be allowed to start/stop it, but
238 not transition it out of the offline state.
240 Instance policies and specs
241 +++++++++++++++++++++++++++
243 In previous Ganeti versions, an instance creation request was not
244 limited on the minimum size and on the maximum size just by the cluster
245 resources. As such, any policy could be implemented only in third-party
246 clients (RAPI clients, or shell wrappers over ``gnt-*``
247 tools). Furthermore, calculating cluster capacity via ``hspace`` again
248 required external input with regards to instance sizes.
250 In order to improve these workflows and to allow for example better
251 per-node group differentiation, we introduced instance specs, which
254 - minimum instance disk size, disk count, memory size, cpu count
255 - maximum values for the above metrics
256 - and “standard” values (used in ``hspace`` to calculate the standard
259 The minimum/maximum values can be also customised at node-group level,
260 for example allowing more powerful hardware to support bigger instance
263 Beside the instance specs, there are a few other settings belonging to
264 the instance policy framework. It is possible now to customise, per
265 cluster and node-group:
267 - the list of allowed disk templates
268 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
269 - the maximum ratio of instance to spindles (see below for more
270 information) for local storage
272 All these together should allow all tools that talk to Ganeti to know
273 what are the ranges of allowed values for instances and the
274 over-subscription that is allowed.
276 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
277 instance configuration, and the physical CPU configuration from the
278 node. For the spindle ratios however, we didn't track before these
279 values, so new parameters have been added:
281 - a new node parameter ``spindle_count``, defaults to 1, customisable at
282 node group or node level
283 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
285 Note that spindles in this context doesn't need to mean actual
286 mechanical hard-drives; it's just a relative number for both the node
287 I/O capacity and instance I/O consumption.
289 Instance migration behaviour
290 ++++++++++++++++++++++++++++
292 While live-migration is in general desirable over failover, it is
293 possible that for some workloads it is actually worse, due to the
294 variable time of the “suspend” phase during live migration.
296 To allow the tools to work consistently over such instances (without
297 having to hard-code instance names), a new backend parameter
298 ``always_failover`` has been added to control the migration/failover
299 behaviour. When set to True, all migration requests for an instance will
300 instead fall-back to failover.
302 Instance memory ballooning
303 ++++++++++++++++++++++++++
305 Initial support for memory ballooning has been added. The memory for an
306 instance is no longer fixed (backend parameter ``memory``), but instead
307 can vary between minimum and maximum values (backend parameters
308 ``minmem`` and ``maxmem``). Currently we only change an instance's
311 - live migrating or failing over and instance and the target node
312 doesn't have enough memory
313 - user requests changing the memory via ``gnt-instance modify
319 In order to control the use of specific CPUs by instance, support for
320 controlling CPU pinning has been added for the Xen, HVM and LXC
321 hypervisors. This is controlled by a new hypervisor parameter
322 ``cpu_mask``; details about possible values for this are in the
323 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
324 VCPU-to-CPU mapping) form will work well only when all nodes in your
325 cluster have the same amount of CPUs.
330 Another area in which Ganeti was not customisable were the parameters
331 used for storage configuration, e.g. how many stripes to use for LVM,
332 DRBD resync configuration, etc.
334 To improve this area, we've added disks parameters, which are
335 customisable at cluster and node group level, and which allow to
336 specify various parameters for disks (DRBD has the most parameters
337 currently), for example:
339 - DRBD resync algorithm and parameters (e.g. speed)
340 - the default VG for meta-data volumes for DRBD
341 - number of stripes for LVM (plain disk template)
344 These parameters can be modified via ``gnt-cluster modify -D …`` and
345 ``gnt-group modify -D …``, and are used at either instance creation (in
346 case of LVM stripes, for example) or at disk “activation” time
349 Rados block device support
350 ++++++++++++++++++++++++++
352 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
353 denoted by the ``rbd`` disk template type. This is considered
354 experimental, feedback is welcome. For details on configuring it, see
355 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
360 The existing master IP functionality works well only in simple setups (a
361 single network shared by all nodes); however, if nodes belong to
362 different networks, then the ``/32`` setup and lack of routing
363 information is not enough.
365 To allow the master IP to function well in more complex cases, the
366 system was reworked as follows:
368 - a master IP netmask setting has been added
369 - the master IP activation/turn-down code was moved from the node daemon
371 - whether to run the Ganeti-supplied master IP script or a user-supplied
372 on is a ``gnt-cluster init`` setting
374 Details about the location of the standard and custom setup scripts are
375 in the man page :manpage:`gnt-cluster(8)`; for information about the
376 setup script protocol, look at the Ganeti-supplied script.
381 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
384 It is now possible to use TLS-protected connections, and when renewing
385 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
386 it is now possible to specify spice or spice CA certificates. Also, it
387 is possible to configure a password for SPICE sessions via the
388 hypervisor parameter ``spice_password_file``.
390 There are also new parameters to control the compression and streaming
391 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
392 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
393 for the spice parameters.
395 Lastly, it is now possible to see the SPICE connection information via
396 ``gnt-instance console``.
401 A new tool (``tools/ovfconverter``) has been added that supports
402 conversion between Ganeti and the `Open Virtualization Format
403 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
406 This relies on the ``qemu-img`` tool to convert the disk formats, so the
407 actual compatibility with other virtualization solutions depends on it.
412 The configuration query daemon (``ganeti-confd``) is now optional, and
413 has been rewritten in Haskell; whether to use the daemon at all, use the
414 Python (default) or the Haskell version is selectable at configure time
415 via the ``--enable-confd`` parameter, which can take one of the
416 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
417 daemon will result in a smaller footprint; for larger systems, we
418 welcome feedback on the Haskell version which might become the default
421 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
422 daemon running. The Python version doesn't implement the new call.
425 User interface changes
426 ~~~~~~~~~~~~~~~~~~~~~~
428 We have replaced the ``--disks`` option of ``gnt-instance
429 replace-disks`` with a more flexible ``--disk`` option, which allows
430 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
431 disk size and mode can be changed upon recreation (via ``gnt-instance
432 recreate-disks``, which accepts the same ``--disk`` option).
434 As many people are used to a ``show`` command, we have added that as an
435 alias to ``info`` on all ``gnt-*`` commands.
437 The ``gnt-instance grow-disk`` command has a new mode in which it can
438 accept the target size of the disk, instead of the delta; this can be
439 more safe since two runs in absolute mode will be idempotent, and
440 sometimes it's also easier to specify the desired size directly.
442 Also the handling of instances with regard to offline secondaries has
443 been improved. Instance operations should not fail because one of it's
444 secondary nodes is offline, even though it's safe to proceed.
446 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
447 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
448 minors to instance name.
453 RAPI coverage has improved, with (for example) new resources for
454 recreate-disks, node power-cycle, etc.
459 There is partial support for ``xl`` in the Xen hypervisor; feedback is
462 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
463 whether to still support Python 2.4 or move to Python 2.6 as minimum
466 Support for Fedora has been slightly improved; the provided example
467 init.d script should work better on it and the INSTALL file should
468 document the needed dependencies.
473 The deprecated ``QueryLocks`` LUXI request has been removed. Use
474 ``Query(what=QR_LOCK, ...)`` instead.
476 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
477 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
478 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
479 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
480 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
482 RAPI client: ``CertificateError`` now derives from
483 ``GanetiApiError``. This should make it more easy to handle Ganeti
486 Deprecation warnings due to PyCrypto/paramiko import in
487 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
488 make sure to run an up-to-date paramiko version, if you use this tool.
490 The QA scripts now depend on Python 2.5 or above (the main code base
491 still works with Python 2.4).
493 The configuration file (``config.data``) is now written without
494 indentation for performance reasons; if you want to edit it, it can be
495 re-formatted via ``tools/fmtjson``.
497 A number of bugs has been fixed in the cluster merge tool.
499 ``x509`` certification verification (used in import-export) has been
500 changed to allow the same clock skew as permitted by the cluster
501 verification. This will remove some rare but hard to diagnose errors in
508 *(Released Thu, 19 Jul 2012)*
510 Very few changes from rc4 to the final release, only bugfixes:
512 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
513 instance, fix CDROM booting for KVM instances)
514 - fixed node group modification of node parameters
515 - fixed issue in LUClusterVerifyGroup with multi-group clusters
516 - fixed generation of bash completion to ensure a stable ordering
523 *(Released Fri, 13 Jul 2012)*
525 Third release candidate for 2.6. The following changes were done from
528 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
529 - Fixed an inconsistency in the LUXI protocol with the provided
530 arguments (NOT backwards compatible)
531 - Fixed a bug with node groups ipolicy where ``min`` was greater than
532 the cluster ``std`` value
533 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
534 easier instance debugging on nodes (requires ``hconfd`` to work)
540 *(Released Tue, 03 Jul 2012)*
542 Second release candidate for 2.6. The following changes were done from
545 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
547 - Fixed a RAPI regression on missing beparams/memory
548 - Fixed redistribution of files on offline nodes
549 - Added possibility to run activate-disks even though secondaries are
550 offline. With this change it relaxes also the strictness on some other
551 commands which use activate disks internally:
552 * ``gnt-instance start|reboot|rename|backup|export``
553 - Made it possible to remove safely an instance if its secondaries are
555 - Made it possible to reinstall even though secondaries are offline
561 *(Released Mon, 25 Jun 2012)*
563 First release candidate for 2.6. The following changes were done from
566 - Fixed bugs with disk parameters and ``rbd`` templates as well as
568 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
569 behaviour. It supports now the modify operation
570 - ``hcheck`` implemented to analyze cluster health and possibility of
571 improving health by rebalance
572 - ``hbal`` has been improved in dealing with split instances
578 *(Released Mon, 11 Jun 2012)*
580 Second beta release of 2.6. The following changes were done from beta2
583 - Fixed ``daemon-util`` with non-root user models
584 - Fixed creation of plain instances with ``--no-wait-for-sync``
585 - Fix wrong iv_names when running ``cfgupgrade``
586 - Export more information in RAPI group queries
587 - Fixed bug when changing instance network interfaces
588 - Extended burnin to do NIC changes
589 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
590 - Changed default for DRBD barriers
591 - Fixed DRBD error reporting for syncer rate
592 - Verify the options on disk parameters
594 And of course various fixes to documentation and improved unittests and
601 *(Released Wed, 23 May 2012)*
603 First beta release of 2.6. The following changes were done from beta1 to
606 - integrated patch for distributions without ``start-stop-daemon``
607 - adapted example init.d script to work on Fedora
608 - fixed log handling in Haskell daemons
609 - adapted checks in the watcher for pycurl linked against libnss
610 - add partial support for ``xl`` instead of ``xm`` for Xen
611 - fixed a type issue in cluster verification
612 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
615 Plus integrated fixes from the 2.5 branch:
617 - fixed ``kvm-ifup`` to use ``/bin/bash``
618 - fixed parallel build failures
619 - KVM live migration when using a custom keymap
625 *(Released Tue, 24 Jul 2012)*
627 A small bugfix release, with no new features:
629 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
630 different default shell (e.g. Debian, Ubuntu)
631 - fixed KVM startup and live migration with a custom keymap (fixes Issue
632 243 and Debian bug #650664)
633 - fixed compatibility with KVM versions that don't support multiple boot
634 devices (fixes Issue 230 and Debian bug #624256)
636 Additionally, a few fixes were done to the build system (fixed parallel
637 build failures) and to the unittests (fixed race condition in test for
638 FileID functions, and the default enable/disable mode for QA test is now
645 *(Released Fri, 11 May 2012)*
647 A small bugfix release.
649 The main issues solved are on the topic of compatibility with newer LVM
652 - fixed parsing of ``lv_attr`` field
653 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
654 the ``--force`` flag is needed
656 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
657 to accept kernel 3.x too (was hardcoded to 2.6.*).
659 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
660 of hook results and that also made display of other errors suboptimal
661 was fixed; the code behaves now like 2.4 and earlier.
663 Another change in 2.5, the cleanup of the OS scripts environment, is too
664 aggressive: it removed even the ``PATH`` variable, which requires the OS
665 scripts to *always* need to export it. Since this is a bit too strict,
666 we now export a minimal PATH, the same that we export for hooks.
668 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
669 integrated into this release.
671 Finally, a few other miscellaneous changes were done (no new features,
672 just small improvements):
674 - Fix ``gnt-group --help`` display
675 - Fix hardcoded Xen kernel path
676 - Fix grow-disk handling of invalid units
677 - Update synopsis for ``gnt-cluster repair-disk-sizes``
678 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
684 *(Released Thu, 12 Apr 2012)*
686 Incompatible/important changes and bugfixes
687 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
689 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
690 resource's ``ip_check`` parameter changed from ``True`` to ``False``
691 to match the underlying LUXI interface.
692 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
693 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
694 not maintain backwards-compatibility as the underlying operation
695 changed in an incompatible way. The RAPI client can talk to old
696 servers, but it needs to be told so as the return value changed.
697 - When creating file-based instances via RAPI, the ``file_driver``
698 parameter no longer defaults to ``loop`` and must be specified.
699 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
701 - Support for the undocumented and deprecated RAPI instance creation
702 request format version 0 has been dropped. Use version 1, supported
703 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
704 - Pyparsing 1.4.6 or above is required, see :doc:`installation
705 documentation <install>`.
706 - The "cluster-verify" hooks are now executed per group by the
707 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
708 if you just run ``gnt-cluster verify``, which generates one opcode per
710 - The environment as passed to the OS scripts is cleared, and thus no
711 environment variables defined in the node daemon's environment will be
712 inherited by the scripts.
713 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
715 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
716 support operations involving multiple node groups.
717 - Offline nodes are ignored when failing over an instance.
718 - Support for KVM version 1.0, which changed the version reporting format
720 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
722 - ``Makefile`` is now compatible with Automake 1.11.2
723 - Includes all bugfixes made in the 2.4 series
728 - The ganeti-htools project has been merged into the ganeti-core source
729 tree and will be built as part of Ganeti (see :doc:`install-quick`).
730 - Implemented support for :doc:`shared storage <design-shared-storage>`.
731 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
732 GPT-style partition tables (requires `parted
733 <http://www.gnu.org/s/parted/>`_).
734 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
735 - Allowed adding tags on instance creation.
736 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
737 - Allow instances to be started in a paused state, enabling the user to
738 see the complete console output on boot using the console.
739 - Added new hypervisor flag to control default reboot behaviour
740 (``reboot_behavior``).
741 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
742 - Improved out-of-band management support:
744 - Added ``gnt-node health`` command reporting the health status of
746 - Added ``gnt-node power`` command to manage power status of nodes.
747 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
749 - Instance migration can fall back to failover if instance is not
751 - Filters can be used when listing nodes, instances, groups and locks;
752 see :manpage:`ganeti(7)` manpage.
753 - Added post-execution status as variables to :doc:`hooks <hooks>`
755 - Instance tags are exported/imported together with the instance.
756 - When given an explicit job ID, ``gnt-job info`` will work for archived
758 - Jobs can define dependencies on other jobs (not yet supported via
759 RAPI or command line, but used by internal commands and usable via
762 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
765 - Instance failover is now available as a RAPI resource
766 (``/2/instances/[instance_name]/failover``).
767 - ``gnt-instance info`` defaults to static information if primary node
769 - Opcodes have a new ``comment`` attribute.
770 - Added basic SPICE support to KVM hypervisor.
771 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
773 Node group improvements
774 ~~~~~~~~~~~~~~~~~~~~~~~
776 - ``gnt-cluster verify`` has been modified to check groups separately,
777 thereby improving performance.
778 - Node group support has been added to ``gnt-cluster verify-disks``,
779 which now operates per node group.
780 - Watcher has been changed to work better with node groups.
782 - One process and state file per node group.
783 - Slow watcher in one group doesn't block other group's watcher.
785 - Added new command, ``gnt-group evacuate``, to move all instances in a
786 node group to other groups.
787 - Added ``gnt-instance change-group`` to move an instance to another
789 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
790 per-group operations.
791 - Node groups can be tagged.
792 - Some operations switch from an exclusive to a shared lock as soon as
794 - Instance's primary and secondary nodes' groups are now available as
795 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
796 and ``snodes.group.uuid``).
801 - Numerous updates to documentation and manpages.
803 - :doc:`RAPI <rapi>` documentation now has detailed parameter
805 - Some opcode/job results are now also documented, see :doc:`RAPI
808 - A lockset's internal lock is now also visible in lock monitor.
809 - Log messages from job queue workers now contain information about the
810 opcode they're processing.
811 - ``gnt-instance console`` no longer requires the instance lock.
812 - A short delay when waiting for job changes reduces the number of LUXI
813 requests significantly.
814 - DRBD metadata volumes are overwritten with zeros during disk creation.
815 - Out-of-band commands no longer acquire the cluster lock in exclusive
817 - ``devel/upload`` now uses correct permissions for directories.
823 *(Released Fri, 23 Mar 2012)*
825 This was the sixth release candidate of the 2.5 series.
831 *(Released Mon, 9 Jan 2012)*
833 This was the fifth release candidate of the 2.5 series.
839 *(Released Thu, 27 Oct 2011)*
841 This was the fourth release candidate of the 2.5 series.
847 *(Released Wed, 26 Oct 2011)*
849 This was the third release candidate of the 2.5 series.
855 *(Released Tue, 18 Oct 2011)*
857 This was the second release candidate of the 2.5 series.
863 *(Released Tue, 4 Oct 2011)*
865 This was the first release candidate of the 2.5 series.
871 *(Released Wed, 31 Aug 2011)*
873 This was the third beta release of the 2.5 series.
879 *(Released Mon, 22 Aug 2011)*
881 This was the second beta release of the 2.5 series.
887 *(Released Fri, 12 Aug 2011)*
889 This was the first beta release of the 2.5 series.
895 *(Released Thu, 27 Oct 2011)*
897 - Fixed bug when parsing command line parameter values ending in
899 - Fixed assertion error after unclean master shutdown
900 - Disable HTTP client pool for RPC, significantly reducing memory usage
902 - Fixed queue archive creation with wrong permissions
908 *(Released Tue, 23 Aug 2011)*
912 - Fixed documentation for importing with ``--src-dir`` option
913 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
914 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
920 *(Released Fri, 5 Aug 2011)*
922 Many bug-fixes and a few small features:
924 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
925 issues when you tried to add an instance with two MAC addresses in one
927 - KVM: fixed per-instance stored UID value
928 - KVM: configure bridged NICs at migration start
929 - KVM: Fix a bug where instance will not start with never KVM versions
931 - Added OS search path to ``gnt-cluster info``
932 - Fixed an issue with ``file_storage_dir`` where you were forced to
933 provide an absolute path, but the documentation states it is a
934 relative path, the documentation was right
935 - Added a new parameter to instance stop/start called ``--no-remember``
936 that will make the state change to not be remembered
937 - Implemented ``no_remember`` at RAPI level
938 - Improved the documentation
939 - Node evacuation: don't call IAllocator if node is already empty
940 - Fixed bug in DRBD8 replace disks on current nodes
941 - Fixed bug in recreate-disks for DRBD instances
942 - Moved assertion checking locks in ``gnt-instance replace-disks``
943 causing it to abort with not owning the right locks for some situation
944 - Job queue: Fixed potential race condition when cancelling queued jobs
945 - Fixed off-by-one bug in job serial generation
946 - ``gnt-node volumes``: Fix instance names
947 - Fixed aliases in bash completion
948 - Fixed a bug in reopening log files after being sent a SIGHUP
949 - Added a flag to burnin to allow specifying VCPU count
950 - Bugfixes to non-root Ganeti configuration
956 *(Released Thu, 12 May 2011)*
958 Many bug-fixes and a few new small features:
960 - Fixed a bug related to log opening failures
961 - Fixed a bug in instance listing with orphan instances
962 - Fixed a bug which prevented resetting the cluster-level node parameter
963 ``oob_program`` to the default
964 - Many fixes related to the ``cluster-merge`` tool
965 - Fixed a race condition in the lock monitor, which caused failures
966 during (at least) creation of many instances in parallel
967 - Improved output for gnt-job info
968 - Removed the quiet flag on some ssh calls which prevented debugging
970 - Improved the N+1 failure messages in cluster verify by actually
971 showing the memory values (needed and available)
972 - Increased lock attempt timeouts so that when executing long operations
973 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
974 too early and thus prevent the use of the 'fair' mechanism
975 - Changed instance query data (``gnt-instance info``) to not acquire
976 locks unless needed, thus allowing its use on locked instance if only
977 static information is asked for
978 - Improved behaviour with filesystems that do not support rename on an
980 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
981 kept locks on all nodes during the wipe, which is unneeded
982 - Fixed ``gnt-watcher`` handling of errors during hooks execution
983 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
984 10GiB) which caused the wipe to fail right at the end in some cases
985 - Fixed master IP activation when doing master failover with no-voting
986 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
987 the master node itself
988 - Fixed potential data-loss in under disk full conditions, where Ganeti
989 wouldn't check correctly the return code and would consider
990 partially-written files 'correct'
991 - Fixed bug related to multiple VGs and DRBD disk replacing
992 - Added new disk parameter ``metavg`` that allows placement of the meta
993 device for DRBD in a different volume group
994 - Fixed error handling in the node daemon when the system libc doesn't
995 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
996 - Fixed lock release during replace-disks, which kept cluster-wide locks
997 when doing disk replaces with an iallocator script
998 - Added check for missing bridges in cluster verify
999 - Handle EPIPE errors while writing to the terminal better, so that
1000 piping the output to e.g. ``less`` doesn't cause a backtrace
1001 - Fixed rare case where a ^C during Luxi calls could have been
1002 interpreted as server errors, instead of simply terminating
1003 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1005 - Added a few more parameters to the KVM hypervisor, allowing a second
1006 CDROM, custom disk type for CDROMs and a floppy image
1007 - Removed redundant message in instance rename when the name is given
1009 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1010 disks on new nodes, allowing recreation when the original instance
1011 nodes are completely gone
1012 - Added option when converting disk templates to DRBD to skip waiting
1013 for the resync, in order to make the instance available sooner
1014 - Added two new variables to the OS scripts environment (containing the
1016 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1017 to allow use of ``pvgrub`` as bootloader
1018 - Changed the instance memory modifications to only check out-of-memory
1019 conditions on memory increases, and turned the secondary node warnings
1020 into errors (they can still be overridden via ``--force``)
1021 - Fixed the handling of a corner case when the Python installation gets
1022 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1023 to execute a command that doesn't exist
1024 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1025 node of the instance returned failures during instance shutdown; this
1026 adds the option ``--ignore-consistency`` to gnt-instance move
1028 And as usual, various improvements to the error messages, documentation
1035 *(Released Wed, 09 Mar 2011)*
1037 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1038 the RAPI users file if run twice (even with ``--dry-run``).
1040 The release fixes that bug (nothing else changed).
1046 *(Released Mon, 07 Mar 2011)*
1048 Final 2.4.0 release. Just a few small fixes:
1050 - Fixed RAPI node evacuate
1051 - Fixed the kvm-ifup script
1052 - Fixed internal error handling for special job cases
1053 - Updated man page to specify the escaping feature for options
1059 *(Released Mon, 28 Feb 2011)*
1061 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1062 that this feature wiped the disks of the wrong instance, leading to loss
1067 - Fixed title of query field containing instance name
1068 - Expanded the glossary in the documentation
1069 - Fixed one unittest (internal issue)
1075 *(Released Mon, 21 Feb 2011)*
1077 A number of bug fixes plus just a couple functionality changes.
1079 On the user-visible side, the ``gnt-* list`` command output has changed
1080 with respect to "special" field states. The current rc1 style of display
1081 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1082 the default output mode special fields states are displayed as follows:
1084 - Offline resource: ``*``
1085 - Unavailable/not applicable: ``-``
1086 - Data missing (RPC failure): ``?``
1087 - Unknown field: ``??``
1089 Another user-visible change is the addition of ``--force-join`` to
1094 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1095 - Fixed regression in RAPI/instance reinstall where all parameters were
1096 required (instead of optional)
1097 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1098 - Fixed iallocator usage (offline nodes were not considered offline)
1099 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1100 - Fixed hypervisor and OS parameter validation with respect to
1101 non-vm_capable nodes
1102 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1104 - Fixed ``tools/listrunner`` with respect to agent-based usage
1110 *(Released Fri, 4 Feb 2011)*
1112 Many changes and fixes since the beta1 release. While there were some
1113 internal changes, the code has been mostly stabilised for the RC
1116 Note: the dumb allocator was removed in this release, as it was not kept
1117 up-to-date with the IAllocator protocol changes. It is recommended to
1118 use the ``hail`` command from the ganeti-htools package.
1120 Note: the 2.4 and up versions of Ganeti are not compatible with the
1121 0.2.x branch of ganeti-htools. You need to upgrade to
1122 ganeti-htools-0.3.0 (or later).
1124 Regressions fixed from 2.3
1125 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1127 - Fixed the ``gnt-cluster verify-disks`` command
1128 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1130 - Fixed disk adoption breakage
1131 - Fixed wrong headers in instance listing for field aliases
1136 - Fixed corner case in KVM handling of NICs
1137 - Fixed many cases of wrong handling of non-vm_capable nodes
1138 - Fixed a bug where a missing instance symlink was not possible to
1139 recreate with any ``gnt-*`` command (now ``gnt-instance
1140 activate-disks`` does it)
1141 - Fixed the volume group name as reported by ``gnt-cluster
1143 - Increased timeouts for the import-export code, hopefully leading to
1144 fewer aborts due network or instance timeouts
1145 - Fixed bug in ``gnt-node list-storage``
1146 - Fixed bug where not all daemons were started on cluster
1147 initialisation, but only at the first watcher run
1148 - Fixed many bugs in the OOB implementation
1149 - Fixed watcher behaviour in presence of instances with offline
1151 - Fixed instance list output for instances running on the wrong node
1152 - a few fixes to the cluster-merge tool, but it still cannot merge
1153 multi-node groups (currently it is not recommended to use this tool)
1159 - Improved network configuration for the KVM hypervisor
1160 - Added e1000 as a supported NIC for Xen-HVM
1161 - Improved the lvmstrap tool to also be able to use partitions, as
1162 opposed to full disks
1163 - Improved speed of disk wiping (the cluster parameter
1164 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1165 of instance creations
1166 - Added documentation for the OS parameters
1167 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1168 hypervisor is not responding
1169 - Added display of blacklisted and hidden OS information in
1170 ``gnt-cluster info``
1171 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1172 NIC and node parameters, which might create problems with currently
1173 invalid (but undetected) configuration files, but prevents validation
1174 failures when unrelated parameters are modified
1175 - Changed cluster initialisation to wait for the master daemon to become
1177 - Expanded the RAPI interface:
1179 - Added config redistribution resource
1180 - Added activation/deactivation of instance disks
1181 - Added export of console information
1183 - Implemented log file reopening on SIGHUP, which allows using
1184 logrotate(8) for the Ganeti log files
1185 - Added a basic OOB helper script as an example
1191 *(Released Fri, 14 Jan 2011)*
1196 - Fixed timezone issues when formatting timestamps
1197 - Added support for node groups, available via ``gnt-group`` and other
1199 - Added out-of-band framework and management, see :doc:`design
1200 document <design-oob>`
1201 - Removed support for roman numbers from ``gnt-node list`` and
1202 ``gnt-instance list``.
1203 - Allowed modification of master network interface via ``gnt-cluster
1204 modify --master-netdev``
1205 - Accept offline secondaries while shutting down instance disks
1206 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1207 - Added support for multiple LVM volume groups
1208 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1210 - Added commands to list available fields:
1212 - ``gnt-node list-fields``
1213 - ``gnt-group list-fields``
1214 - ``gnt-instance list-fields``
1216 - Updated documentation and man pages
1221 - Moved ``rapi_users`` file into separate directory, now named
1222 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1224 - Added new tool for running commands on many machines,
1225 ``tools/ganeti-listrunner``
1226 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1227 improving the ``gnt-instance console`` output
1228 - Allowed customisation of disk index separator at ``configure`` time
1229 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1230 - Added support for non-partitioned md disks in ``lvmstrap``
1231 - Added script to gracefully power off KVM instances
1232 - Split ``utils`` module into smaller parts
1233 - Changed query operations to return more detailed information, e.g.
1234 whether an information is unavailable due to an offline node. To use
1235 this new functionality, the LUXI call ``Query`` must be used. Field
1236 information is now stored by the master daemon and can be retrieved
1237 using ``QueryFields``. Instances, nodes and groups can also be queried
1238 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1239 exposed via RAPI). The following commands make use of this
1240 infrastructure change:
1242 - ``gnt-group list``
1243 - ``gnt-group list-fields``
1245 - ``gnt-node list-fields``
1246 - ``gnt-instance list``
1247 - ``gnt-instance list-fields``
1248 - ``gnt-debug locks``
1253 - New RAPI resources (see :doc:`rapi`):
1257 - ``/2/groups/[group_name]``
1258 - ``/2/groups/[group_name]/assign-nodes``
1259 - ``/2/groups/[group_name]/modify``
1260 - ``/2/groups/[group_name]/rename``
1261 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1265 - Implemented ``no_install`` for instance creation
1266 - Implemented OS parameters for instance reinstallation, allowing
1267 use of special settings on reinstallation (e.g. for preserving data)
1272 - Added IPv6 support in import/export
1273 - Pause DRBD synchronization while wiping disks on instance creation
1274 - Updated unittests and QA scripts
1275 - Improved network parameters passed to KVM
1276 - Converted man pages from docbook to reStructuredText
1282 *(Released Mon, 20 Dec 2010)*
1284 Released version 2.3.1~rc1 without any changes.
1290 *(Released Wed, 1 Dec 2010)*
1292 - impexpd: Disable OpenSSL compression in socat if possible (backport
1293 from master, commit e90739d625b, see :doc:`installation guide
1294 <install-quick>` for details)
1295 - Changed unittest coverage report to exclude test scripts
1296 - Added script to check version format
1302 *(Released Wed, 1 Dec 2010)*
1304 Released version 2.3.0~rc1 without any changes.
1310 *(Released Fri, 19 Nov 2010)*
1312 A number of bugfixes and documentation updates:
1314 - Update ganeti-os-interface documentation
1315 - Fixed a bug related to duplicate MACs or similar items which should be
1317 - Fix breakage in OS state modify
1318 - Reinstall instance: disallow offline secondaries (fixes bug related to
1319 OS changing but reinstall failing)
1320 - plus all the other fixes between 2.2.1 and 2.2.2
1326 *(Released Tue, 2 Nov 2010)*
1328 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1329 - Fixed master failover race with watcher
1330 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1332 - Accept previously stopped instance for export with instance removal
1333 - Simplify and extend the environment variables for instance OS scripts
1334 - Added new node flags, ``master_capable`` and ``vm_capable``
1335 - Added optional instance disk wiping prior during allocation. This is a
1336 cluster-wide option and can be set/modified using
1337 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1338 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1339 :doc:`install-quick`
1340 - Added a new watcher option (``--ignore-pause``)
1341 - Added option to ignore offline node on instance start/stop
1342 (``--ignore-offline``)
1343 - Allow overriding OS parameters with ``gnt-instance reinstall``
1344 - Added ability to change node's secondary IP address using ``gnt-node
1346 - Implemented privilege separation for all daemons except
1347 ``ganeti-noded``, see ``configure`` options
1348 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1350 - Implemented job priorities (see ``ganeti(7)`` manpage)
1351 - Ignore failures while shutting down instances during failover from
1353 - Exit daemon's bootstrap process only once daemon is ready
1354 - Export more information via ``LUInstanceQuery``/remote API
1355 - Improved documentation, QA and unittests
1356 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1357 restart if the file was created or changed
1358 - Added LUXI protocol version sent with each request and response,
1359 allowing detection of server/client mismatches
1360 - Moved the Python scripts among gnt-* and ganeti-* into modules
1361 - Moved all code related to setting up SSH to an external script,
1363 - Infrastructure changes for node group support in future versions
1369 *(Released Fri, 19 Nov 2010)*
1371 A few small bugs fixed, and some improvements to the build system:
1373 - Fix documentation regarding conversion to drbd
1374 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1376 - Fix error handling in node modify with multiple changes
1377 - Allow remote imports without checked names
1383 *(Released Tue, 19 Oct 2010)*
1385 - Disable SSL session ID cache in RPC client
1391 *(Released Thu, 14 Oct 2010)*
1393 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1394 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1396 - Fix problems with interaction between Curl and Python's HTTP server,
1397 resulting in increased speed in many RPC calls
1398 - Improve our release script to prevent breakage with older aclocal and
1405 *(Released Thu, 7 Oct 2010)*
1407 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1408 value retrieved from master
1409 - Added support for blacklisted or hidden OS definitions
1410 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1411 - Added support for -mem-path in KVM hypervisor abstraction layer
1412 - Allow overriding instance parameters in tool for inter-cluster
1413 instance moves (``tools/move-instance``)
1414 - Improved opcode summaries (e.g. in ``gnt-job list``)
1415 - Improve consistency of OS listing by sorting it
1416 - Documentation updates
1422 *(Released Fri, 8 Oct 2010)*
1424 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1430 *(Released Mon, 4 Oct 2010)*
1432 - Fixed regression in ``gnt-instance rename``
1438 *(Released Wed, 22 Sep 2010)*
1440 - Fixed OS_VARIANT variable for OS scripts
1441 - Fixed cluster tag operations via RAPI
1442 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1443 - Disabled RAPI CA checks in watcher
1449 *(Released Mon, 23 Aug 2010)*
1451 - Support DRBD versions of the format "a.b.c.d"
1453 - Re-introduce support for usage from multiple threads in RAPI client
1454 - Instance renames and modify via RAPI
1455 - Work around race condition between processing and archival in job
1457 - Mark opcodes following failed one as failed, too
1458 - Job field ``lock_status`` was removed due to difficulties making it
1459 work with the changed job queue in Ganeti 2.2; a better way to monitor
1460 locks is expected for a later 2.2.x release
1461 - Fixed dry-run behaviour with many commands
1462 - Support ``ssh-agent`` again when adding nodes
1463 - Many additional bugfixes
1469 *(Released Fri, 30 Jul 2010)*
1471 Important change: the internal RPC mechanism between Ganeti nodes has
1472 changed from using a home-grown http library (based on the Python base
1473 libraries) to use the PycURL library. This requires that PycURL is
1474 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1475 against GnuTLS by default. cURL's support for GnuTLS had known issues
1476 before cURL 7.21.0 and we recommend using the latest cURL release or
1477 linking against OpenSSL. Most other distributions already link PycURL
1478 and cURL against OpenSSL. The command::
1480 python -c 'import pycurl; print pycurl.version'
1482 can be used to determine the libraries PycURL and cURL are linked
1485 Other significant changes:
1487 - Rewrote much of the internals of the job queue, in order to achieve
1488 better parallelism; this decouples job query operations from the job
1489 processing, and it should allow much nicer behaviour of the master
1490 daemon under load, and it also has uncovered some long-standing bugs
1491 related to the job serialisation (now fixed)
1492 - Added a default iallocator setting to the cluster parameters,
1493 eliminating the need to always pass nodes or an iallocator for
1494 operations that require selection of new node(s)
1495 - Added experimental support for the LXC virtualization method
1496 - Added support for OS parameters, which allows the installation of
1497 instances to pass parameter to OS scripts in order to customise the
1499 - Added a hypervisor parameter controlling the migration type (live or
1500 non-live), since hypervisors have various levels of reliability; this
1501 has renamed the 'live' parameter to 'mode'
1502 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1503 logical volumes, meaning that cluster verify will ignore them and not
1504 flag their presence as errors
1505 - The watcher will now reset the error count for failed instances after
1506 8 hours, thus allowing self-healing if the problem that caused the
1507 instances to be down/fail to start has cleared in the meantime
1508 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1509 check for, and warn, if the drbd module parameter ``usermode_helper``
1510 is not consistent with the cluster-wide setting; this is needed to
1511 make diagnose easier of failed drbd creations
1512 - Started adding base IPv6 support, but this is not yet
1513 enabled/available for use
1514 - Rename operations (cluster, instance) will now return the new name,
1515 which is especially useful if a short name was passed in
1516 - Added support for instance migration in RAPI
1517 - Added a tool to pre-configure nodes for the SSH setup, before joining
1518 them to the cluster; this will allow in the future a simplified model
1519 for node joining (but not yet fully enabled in 2.2); this needs the
1520 paramiko python library
1521 - Fixed handling of name-resolving errors
1522 - Fixed consistency of job results on the error path
1523 - Fixed master-failover race condition when executed multiple times in
1525 - Fixed many bugs related to the job queue (mostly introduced during the
1526 2.2 development cycle, so not all are impacting 2.1)
1527 - Fixed instance migration with missing disk symlinks
1528 - Fixed handling of unknown jobs in ``gnt-job archive``
1529 - And many other small fixes/improvements
1533 - Enhanced both the unittest and the QA coverage
1534 - Switched the opcode validation to a generic model, and extended the
1535 validation to all opcode parameters
1536 - Changed more parts of the code that write shell scripts to use the
1538 - Switched the master daemon to use the asyncore library for the Luxi
1545 *(Released Thu, 17 Jun 2010)*
1547 - Added tool (``move-instance``) and infrastructure to move instances
1548 between separate clusters (see :doc:`separate documentation
1549 <move-instance>` and :doc:`design document <design-2.2>`)
1550 - Added per-request RPC timeout
1551 - RAPI now requires a Content-Type header for requests with a body (e.g.
1552 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1553 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1554 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1556 - Implemented initial support for running Ganeti daemons as separate
1557 users, see configure-time flags ``--with-user-prefix`` and
1558 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1560 - Instances can be removed after export (``gnt-backup export
1561 --remove-instance``)
1562 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1563 key (instead of 1024 bit)
1564 - Added new cluster configuration file for cluster domain secret
1565 - Import/export now use SSL instead of SSH
1566 - Added support for showing estimated time when exporting an instance,
1567 see the ``ganeti-os-interface(7)`` manpage and look for
1574 *(Released Tue, 16 Nov 2010)*
1576 Some more bugfixes. Unless critical bugs occur, this will be the last
1579 - Fix case of MAC special-values
1580 - Fix mac checker regex
1581 - backend: Fix typo causing "out of range" error
1582 - Add missing --units in gnt-instance list man page
1588 *(Released Tue, 24 Aug 2010)*
1591 - Don't ignore secondary node silently on non-mirrored disk templates
1593 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1594 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1595 - Properly document the usb_mouse parameter
1596 - Fix path in ganeti-rapi(8) (issue 116)
1597 - Adjust error message when the ganeti user's .ssh directory is
1599 - Add same-node-check when changing the disk template to drbd
1605 *(Released Fri, 16 Jul 2010)*
1608 - Add an option to only select some reboot types during qa/burnin.
1609 (on some hypervisors consequent reboots are not supported)
1610 - Fix infrequent race condition in master failover. Sometimes the old
1611 master ip address would be still detected as up for a short time
1612 after it was removed, causing failover to fail.
1613 - Decrease mlockall warnings when the ctypes module is missing. On
1614 Python 2.4 we support running even if no ctypes module is installed,
1615 but we were too verbose about this issue.
1616 - Fix building on old distributions, on which man doesn't have a
1618 - Fix RAPI not to ignore the MAC address on instance creation
1619 - Implement the old instance creation format in the RAPI client.
1625 *(Released Thu, 01 Jul 2010)*
1627 A small bugfix release:
1628 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1629 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1630 a non-existing option
1631 - Fix instance create: the --force-variant option was ignored
1632 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1633 - Fix modify node storage with non-FQDN arguments
1634 - Fix RAPI client to authenticate under Python 2.6 when used
1635 for more than 5 requests needing authentication
1636 - Fix gnt-instance modify -t (storage) giving a wrong error message
1637 when converting a non-shutdown drbd instance to plain
1643 *(Released Fri, 18 Jun 2010)*
1645 A small bugfix release:
1647 - Fix live migration of KVM instances started with older Ganeti
1648 versions which had fewer hypervisor parameters
1649 - Fix gnt-instance grow-disk on down instances
1650 - Fix an error-reporting bug during instance migration
1651 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1652 silently ignoring broken ones
1653 - Fix an RPC error reporting bug affecting, for example, RAPI client
1655 - Fix bug triggered by different API version os-es on different nodes
1656 - Fix a bug in instance startup with custom hvparams: OS level
1657 parameters would fail to be applied.
1658 - Fix the RAPI client under Python 2.6 (but more work is needed to
1659 make it work completely well with OpenSSL)
1660 - Fix handling of errors when resolving names from DNS
1666 *(Released Thu, 3 Jun 2010)*
1668 A medium sized development cycle. Some new features, and some
1669 fixes/small improvements/cleanups.
1671 Significant features
1672 ~~~~~~~~~~~~~~~~~~~~
1674 The node deamon now tries to mlock itself into memory, unless the
1675 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1676 its logs, and falls back to console logging. This allows emergency
1677 features such as ``gnt-node powercycle`` to work even in the event of a
1678 broken node disk (tested offlining the disk hosting the node's
1679 filesystem and dropping its memory caches; don't try this at home)
1681 KVM: add vhost-net acceleration support. It can be tested with a new
1682 enough version of the kernel and of qemu-kvm.
1684 KVM: Add instance chrooting feature. If you use privilege dropping for
1685 your VMs you can also now force them to chroot to an empty directory,
1686 before starting the emulated guest.
1688 KVM: Add maximum migration bandwith and maximum downtime tweaking
1689 support (requires a new-enough version of qemu-kvm).
1691 Cluster verify will now warn if the master node doesn't have the master
1692 ip configured on it.
1694 Add a new (incompatible) instance creation request format to RAPI which
1695 supports all parameters (previously only a subset was supported, and it
1696 wasn't possible to extend the old format to accomodate all the new
1697 features. The old format is still supported, and a client can check for
1698 this feature, before using it, by checking for its presence in the
1699 ``features`` RAPI resource.
1701 Now with ancient latin support. Try it passing the ``--roman`` option to
1702 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1703 (requires the python-roman module to be installed, in order to work).
1708 As usual many internal code refactorings, documentation updates, and
1711 - Lots of improvements and cleanups to the experimental Remote API
1712 (RAPI) client library.
1713 - A new unit test suite for the core daemon libraries.
1714 - A fix to creating missing directories makes sure the umask is not
1715 applied anymore. This enforces the same directory permissions
1717 - Better handling terminating daemons with ctrl+c (used when running
1718 them in debugging mode).
1719 - Fix a race condition in live migrating a KVM instance, when stat()
1720 on the old proc status file returned EINVAL, which is an unexpected
1722 - Fixed manpage checking with newer man and utf-8 charachters. But now
1723 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1729 *(Released Fri, 7 May 2010)*
1731 Fix a bug which prevented untagged KVM instances from starting.
1737 *(Released Fri, 7 May 2010)*
1739 Another release with a long development cycle, during which many
1740 different features were added.
1742 Significant features
1743 ~~~~~~~~~~~~~~~~~~~~
1745 The KVM hypervisor now can run the individual instances as non-root, to
1746 reduce the impact of a VM being hijacked due to bugs in the
1747 hypervisor. It is possible to run all instances as a single (non-root)
1748 user, to manually specify a user for each instance, or to dynamically
1749 allocate a user out of a cluster-wide pool to each instance, with the
1750 guarantee that no two instances will run under the same user ID on any
1753 An experimental RAPI client library, that can be used standalone
1754 (without the other Ganeti libraries), is provided in the source tree as
1755 ``lib/rapi/client.py``. Note this client might change its interface in
1756 the future, as we iterate on its capabilities.
1758 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1759 replace the cluster's certificates and crypto keys. This might help in
1760 case they have been compromised, or have simply expired.
1762 A new disk option for instance creation has been added that allows one
1763 to "adopt" currently existing logical volumes, with data
1764 preservation. This should allow easier migration to Ganeti from
1765 unmanaged (or managed via other software) instances.
1767 Another disk improvement is the possibility to convert between redundant
1768 (DRBD) and plain (LVM) disk configuration for an instance. This should
1769 allow better scalability (starting with one node and growing the
1770 cluster, or shrinking a two-node cluster to one node).
1772 A new feature that could help with automated node failovers has been
1773 implemented: if a node sees itself as offline (by querying the master
1774 candidates), it will try to shutdown (hard) all instances and any active
1775 DRBD devices. This reduces the risk of duplicate instances if an
1776 external script automatically failovers the instances on such nodes. To
1777 enable this, the cluster parameter ``maintain_node_health`` should be
1778 enabled; in the future this option (per the name) will enable other
1779 automatic maintenance features.
1781 Instance export/import now will reuse the original instance
1782 specifications for all parameters; that means exporting an instance,
1783 deleting it and the importing it back should give an almost identical
1784 instance. Note that the default import behaviour has changed from
1785 before, where it created only one NIC; now it recreates the original
1788 Cluster verify has added a few new checks: SSL certificates validity,
1789 /etc/hosts consistency across the cluster, etc.
1794 As usual, many internal changes were done, documentation fixes,
1797 - Fixed cluster initialization with disabled cluster storage (regression
1798 introduced in 2.1.1)
1799 - File-based storage supports growing the disks
1800 - Fixed behaviour of node role changes
1801 - Fixed cluster verify for some corner cases, plus a general rewrite of
1802 cluster verify to allow future extension with more checks
1803 - Fixed log spamming by watcher and node daemon (regression introduced
1805 - Fixed possible validation issues when changing the list of enabled
1807 - Fixed cleanup of /etc/hosts during node removal
1808 - Fixed RAPI response for invalid methods
1809 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1810 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1811 from ide disks, etc.)
1812 - Allow OS changes without re-installation (to record a changed OS
1813 outside of Ganeti, or to allow OS renames)
1814 - Allow instance creation without OS installation (useful for example if
1815 the OS will be installed manually, or restored from a backup not in
1817 - Implemented option to make cluster ``copyfile`` use the replication
1819 - Added list of enabled hypervisors to ssconf (possibly useful for
1821 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1823 - A partial form of node re-IP is possible via node readd, which now
1824 allows changed node primary IP
1825 - Command line utilities now show an informational message if the job is
1827 - The logs of the master daemon now show the PID/UID/GID of the
1834 *(Released Fri, 12 Mar 2010)*
1836 During the 2.1.0 long release candidate cycle, a lot of improvements and
1837 changes have accumulated with were released later as 2.1.1.
1842 The node evacuate command (``gnt-node evacuate``) was significantly
1843 rewritten, and as such the IAllocator protocol was changed - a new
1844 request type has been added. This unfortunate change during a stable
1845 series is designed to improve performance of node evacuations; on
1846 clusters with more than about five nodes and which are well-balanced,
1847 evacuation should proceed in parallel for all instances of the node
1848 being evacuated. As such, any existing IAllocator scripts need to be
1849 updated, otherwise the above command will fail due to the unknown
1850 request. The provided "dumb" allocator has not been updated; but the
1851 ganeti-htools package supports the new protocol since version 0.2.4.
1853 Another important change is increased validation of node and instance
1854 names. This might create problems in special cases, if invalid host
1855 names are being used.
1857 Also, a new layer of hypervisor parameters has been added, that sits at
1858 OS level between the cluster defaults and the instance ones. This allows
1859 customisation of virtualization parameters depending on the installed
1860 OS. For example instances with OS 'X' may have a different KVM kernel
1861 (or any other parameter) than the cluster defaults. This is intended to
1862 help managing a multiple OSes on the same cluster, without manual
1863 modification of each instance's parameters.
1865 A tool for merging clusters, ``cluster-merge``, has been added in the
1866 tools sub-directory.
1871 - Improved the int/float conversions that should make the code more
1872 robust in face of errors from the node daemons
1873 - Fixed the remove node code in case of internal configuration errors
1874 - Fixed the node daemon behaviour in face of inconsistent queue
1875 directory (e.g. read-only file-system where we can't open the files
1877 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1878 now it either aborts cleanly or, if given the new "auto_promote"
1879 parameter, will automatically promote other nodes as needed
1880 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1881 completely prevent Ganeti from working
1882 - Fixed bug for instance export when not all disks were successfully
1884 - Fixed behaviour of node add when the new node is slow in starting up
1886 - Fixed handling of signals in the LUXI client, which should improve
1887 behaviour of command-line scripts
1888 - Added checks for invalid node/instance names in the configuration (now
1889 flagged during cluster verify)
1890 - Fixed watcher behaviour for disk activation errors
1891 - Fixed two potentially endless loops in http library, which led to the
1892 RAPI daemon hanging and consuming 100% CPU in some cases
1893 - Fixed bug in RAPI daemon related to hashed passwords
1894 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1896 - Enhanced compatibility with non-Debian OSes, but not using absolute
1897 path in some commands and allowing customisation of the ssh
1898 configuration directory
1899 - Fixed possible future issue with new Python versions by abiding to the
1900 proper use of ``__slots__`` attribute on classes
1901 - Added checks that should prevent directory traversal attacks
1902 - Many documentation fixes based on feedback from users
1907 - Added an "early_release" more for instance replace disks and node
1908 evacuate, where we release locks earlier and thus allow higher
1909 parallelism within the cluster
1910 - Added watcher hooks, intended to allow the watcher to restart other
1911 daemons (e.g. from the ganeti-nbma project), but they can be used of
1912 course for any other purpose
1913 - Added a compile-time disable for DRBD barriers, to increase
1914 performance if the administrator trusts the power supply or the
1915 storage system to not lose writes
1916 - Added the option of using syslog for logging instead of, or in
1917 addition to, Ganeti's own log files
1918 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1919 can indeed boot from a paravirtual NIC
1920 - Added a generic debug level for many operations; while this is not
1921 used widely yet, it allows one to pass the debug value all the way to
1923 - Enhanced the hooks environment for instance moves (failovers,
1924 migrations) where the primary/secondary nodes changed during the
1925 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1926 - Enhanced data validations for many user-supplied values; one important
1927 item is the restrictions imposed on instance and node names, which
1928 might reject some (invalid) host names
1929 - Add a configure-time option to disable file-based storage, if it's not
1930 needed; this allows greater security separation between the master
1931 node and the other nodes from the point of view of the inter-node RPC
1933 - Added user notification in interactive tools if job is waiting in the
1934 job queue or trying to acquire locks
1935 - Added log messages when a job is waiting for locks
1936 - Added filtering by node tags in instance operations which admit
1937 multiple instances (start, stop, reboot, reinstall)
1938 - Added a new tool for cluster mergers, ``cluster-merge``
1939 - Parameters from command line which are of the form ``a=b,c=d`` can now
1940 use backslash escapes to pass in values which contain commas,
1941 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1943 - For KVM, the instance name is the first parameter passed to KVM, so
1944 that it's more visible in the process list
1950 *(Released Tue, 2 Mar 2010)*
1952 Ganeti 2.1 brings many improvements with it. Major changes:
1954 - Added infrastructure to ease automated disk repairs
1955 - Added new daemon to export configuration data in a cheaper way than
1956 using the remote API
1957 - Instance NICs can now be routed instead of being associated with a
1959 - Improved job locking logic to reduce impact of jobs acquiring multiple
1960 locks waiting for other long-running jobs
1962 In-depth implementation details can be found in the Ganeti 2.1 design
1968 - Added chroot hypervisor
1969 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1971 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1972 ``bootloader_path`` and ``bootloader_args``)
1973 - Added the ``use_localtime`` option for the xen-hvm and kvm
1974 hypervisors, and the default value for this has changed to false (in
1975 2.0 xen-hvm always enabled it)
1976 - Added luxi call to submit multiple jobs in one go
1977 - Added cluster initialization option to not modify ``/etc/hosts``
1979 - Added network interface parameters
1980 - Added dry run mode to some LUs
1981 - Added RAPI resources:
1983 - ``/2/instances/[instance_name]/info``
1984 - ``/2/instances/[instance_name]/replace-disks``
1985 - ``/2/nodes/[node_name]/evacuate``
1986 - ``/2/nodes/[node_name]/migrate``
1987 - ``/2/nodes/[node_name]/role``
1988 - ``/2/nodes/[node_name]/storage``
1989 - ``/2/nodes/[node_name]/storage/modify``
1990 - ``/2/nodes/[node_name]/storage/repair``
1992 - Added OpCodes to evacuate or migrate all instances on a node
1993 - Added new command to list storage elements on nodes (``gnt-node
1994 list-storage``) and modify them (``gnt-node modify-storage``)
1995 - Added new ssconf files with master candidate IP address
1996 (``ssconf_master_candidates_ips``), node primary IP address
1997 (``ssconf_node_primary_ips``) and node secondary IP address
1998 (``ssconf_node_secondary_ips``)
1999 - Added ``ganeti-confd`` and a client library to query the Ganeti
2000 configuration via UDP
2001 - Added ability to run hooks after cluster initialization and before
2003 - Added automatic mode for disk replace (``gnt-instance replace-disks
2005 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2006 after catastrophic data-loss
2007 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2009 - Added ``gnt-instance move`` command to move instances
2010 - Added ``gnt-cluster watcher`` command to control watcher
2011 - Added ``gnt-node powercycle`` command to powercycle nodes
2012 - Added new job status field ``lock_status``
2013 - Added parseable error codes to cluster verification (``gnt-cluster
2014 verify --error-codes``) and made output less verbose (use
2015 ``--verbose`` to restore previous behaviour)
2016 - Added UUIDs to the main config entities (cluster, nodes, instances)
2017 - Added support for OS variants
2018 - Added support for hashed passwords in the Ganeti remote API users file
2020 - Added option to specify maximum timeout on instance shutdown
2021 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2022 - Added new helper script to start and stop Ganeti daemons
2023 (``daemon-util``), with the intent to reduce the work necessary to
2024 adjust Ganeti for non-Debian distributions and to start/stop daemons
2026 - Added more unittests
2027 - Fixed critical bug in ganeti-masterd startup
2028 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2029 now customisable at the cluster level for both the KVM and Xen
2030 hypervisors using the new ``migration_port`` parameter
2031 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2032 reinstalling an instance
2033 - Allowed ``@`` in tag names
2034 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2035 - Many documentation updates
2036 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2037 - ``gnt-instance reinstall`` can now reinstall multiple instances
2038 - Updated many command line parameters
2039 - Introduced new OS API version 15
2040 - No longer support a default hypervisor
2041 - Treat virtual LVs as inexistent
2042 - Improved job locking logic to reduce lock contention
2043 - Match instance and node names case insensitively
2044 - Reimplemented bash completion script to be more complete
2051 *(Released Thu, 4 Feb 2010)*
2053 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2054 - Fix a string formatting bug
2055 - Improve safety of the code in some error paths
2056 - Improve data validation in the master of values returned from nodes
2062 *(Released Thu, 17 Dec 2009)*
2064 - Fix security issue due to missing validation of iallocator names; this
2065 allows local and remote execution of arbitrary executables
2066 - Fix failure of gnt-node list during instance removal
2067 - Ship the RAPI documentation in the archive
2073 *(Released Wed, 30 Sep 2009)*
2075 - Fixed many wrong messages
2076 - Fixed a few bugs related to the locking library
2077 - Fixed MAC checking at instance creation time
2078 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2079 - Fixed a few issues related to signal handling in both daemons and
2081 - Fixed the example startup script provided
2082 - Fixed insserv dependencies in the example startup script (patch from
2084 - Fixed handling of drained nodes in the iallocator framework
2085 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2087 - Fixed error related to invalid job IDs in job polling
2088 - Fixed job/opcode persistence on unclean master shutdown
2089 - Fixed handling of partial job processing after unclean master
2091 - Fixed error reporting from LUs, previously all errors were converted
2092 into execution errors
2093 - Fixed error reporting from burnin
2094 - Decreased significantly the memory usage of the job queue
2095 - Optimised slightly multi-job submission
2096 - Optimised slightly opcode loading
2097 - Backported the multi-job submit framework from the development
2098 branch; multi-instance start and stop should be faster
2099 - Added script to clean archived jobs after 21 days; this will reduce
2100 the size of the queue directory
2101 - Added some extra checks in disk size tracking
2102 - Added an example ethers hook script
2103 - Added a cluster parameter that prevents Ganeti from modifying of
2105 - Added more node information to RAPI responses
2106 - Added a ``gnt-job watch`` command that allows following the ouput of a
2108 - Added a bind-address option to ganeti-rapi
2109 - Added more checks to the configuration verify
2110 - Enhanced the burnin script such that some operations can be retried
2112 - Converted instance reinstall to multi-instance model
2118 *(Released Fri, 7 Aug 2009)*
2120 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2121 to allow using the pre-2.0.2 behaviour in activation, if any existing
2122 instances have mismatched disk sizes in the configuration
2123 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2124 any configuration mismatches for disk sizes
2125 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2126 failover to work on two-node clusters
2127 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2129 - Fixed detection of OS script errors in ``gnt-backup export``
2130 - Fixed exit code of ``gnt-backup export``
2136 *(Released Fri, 17 Jul 2009)*
2138 - Added experimental support for stripped logical volumes; this should
2139 enhance performance but comes with a higher complexity in the block
2140 device handling; stripping is only enabled when passing
2141 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2142 affected even in the non-stripped mode
2143 - Improved resiliency against transient failures at the end of DRBD
2144 resyncs, and in general of DRBD resync checks
2145 - Fixed a couple of issues with exports and snapshot errors
2146 - Fixed a couple of issues in instance listing
2147 - Added display of the disk size in ``gnt-instance info``
2148 - Fixed checking for valid OSes in instance creation
2149 - Fixed handling of the "vcpus" parameter in instance listing and in
2150 general of invalid parameters
2151 - Fixed http server library, and thus RAPI, to handle invalid
2152 username/password combinations correctly; this means that now they
2153 report unauthorized for queries too, not only for modifications,
2154 allowing earlier detect of configuration problems
2155 - Added a new "role" node list field, equivalent to the master/master
2156 candidate/drained/offline flags combinations
2157 - Fixed cluster modify and changes of candidate pool size
2158 - Fixed cluster verify error messages for wrong files on regular nodes
2159 - Fixed a couple of issues with node demotion from master candidate role
2160 - Fixed node readd issues
2161 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2162 - Added a new ``--no-voting`` option for masterfailover to fix failover
2163 on two-nodes clusters when the former master node is unreachable
2164 - Added instance reinstall over RAPI
2170 *(Released Tue, 16 Jun 2009)*
2172 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2173 allow re-adding the start in single-user option (regression from 1.2)
2174 - the watcher writes the instance status to a file, to allow monitoring
2175 to report the instance status (from the master) based on cached
2176 results of the watcher's queries; while this can get stale if the
2177 watcher is being locked due to other work on the cluster, this is
2178 still an improvement
2179 - the watcher now also restarts the node daemon and the rapi daemon if
2181 - fixed the watcher to handle full and drained queue cases
2182 - hooks export more instance data in the environment, which helps if
2183 hook scripts need to take action based on the instance's properties
2184 (no longer need to query back into ganeti)
2185 - instance failovers when the instance is stopped do not check for free
2186 RAM, so that failing over a stopped instance is possible in low memory
2188 - rapi uses queries for tags instead of jobs (for less job traffic), and
2189 for cluster tags it won't talk to masterd at all but read them from
2191 - a couple of error handling fixes in RAPI
2192 - drbd handling: improved the error handling of inconsistent disks after
2193 resync to reduce the frequency of "there are some degraded disks for
2194 this instance" messages
2195 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2196 error handling path called a wrong function name)
2202 *(Released Wed, 27 May 2009)*
2204 - no changes from rc5
2210 *(Released Wed, 20 May 2009)*
2212 - fix a couple of bugs (validation, argument checks)
2213 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2214 - some small improvements to RAPI and IAllocator
2215 - make watcher automatically start the master daemon if down
2221 *(Released Mon, 27 Apr 2009)*
2223 - change the OS list to not require locks; this helps with big clusters
2224 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2225 volume group is broken
2226 - ``gnt-instance info``, without any arguments, doesn't run for all
2227 instances anymore; either pass ``--all`` or pass the desired
2228 instances; this helps against mistakes on big clusters where listing
2229 the information for all instances takes a long time
2230 - miscellaneous doc and man pages fixes
2236 *(Released Wed, 8 Apr 2009)*
2238 - Change the internal locking model of some ``gnt-node`` commands, in
2239 order to reduce contention (and blocking of master daemon) when
2240 batching many creation/reinstall jobs
2241 - Fixes to Xen soft reboot
2242 - No longer build documentation at build time, instead distribute it in
2243 the archive, in order to reduce the need for the whole docbook/rst
2250 *(Released Fri, 27 Mar 2009)*
2252 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2253 - Fix watcher startup sequence, improves the behaviour of busy clusters
2254 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2255 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2256 bind address and other places
2257 - Some documentation fixes and updates
2263 *(Released Mon, 2 Mar 2009)*
2265 - More documentation updates, now all docs should be more-or-less
2267 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2269 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2271 - Fix an issue related to $libdir/run/ganeti and cluster creation
2277 *(Released Thu, 19 Feb 2009)*
2279 - Xen PVM and KVM have switched the default value for the instance root
2280 disk to the first partition on the first drive, instead of the whole
2281 drive; this means that the OS installation scripts must be changed
2283 - Man pages have been updated
2284 - RAPI has been switched by default to HTTPS, and the exported functions
2285 should all work correctly
2286 - RAPI v1 has been removed
2287 - Many improvements to the KVM hypervisor
2288 - Block device errors are now better reported
2289 - Many other bugfixes and small improvements
2295 *(Released Mon, 26 Jan 2009)*
2297 - Version 2 is a general rewrite of the code and therefore the
2298 differences are too many to list, see the design document for 2.0 in
2299 the ``doc/`` subdirectory for more details
2300 - In this beta version there is not yet a migration path from 1.2 (there
2301 will be one in the final 2.0 release)
2302 - A few significant changes are:
2304 - all commands are executed by a daemon (``ganeti-masterd``) and the
2305 various ``gnt-*`` commands are just front-ends to it
2306 - all the commands are entered into, and executed from a job queue,
2307 see the ``gnt-job(8)`` manpage
2308 - the RAPI daemon supports read-write operations, secured by basic
2309 HTTP authentication on top of HTTPS
2310 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2311 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2312 to migrate to DRBD 8 first while still running Ganeti 1.2)
2313 - DRBD devices are using statically allocated minor numbers, which
2314 will be assigned to existing instances during the migration process
2315 - there is support for both Xen PVM and Xen HVM instances running on
2317 - KVM virtualization is supported too
2318 - file-based storage has been implemented, which means that it is
2319 possible to run the cluster without LVM and DRBD storage, for
2320 example using a shared filesystem exported from shared storage (and
2321 still have live migration)
2327 *(Released Tue, 13 Jan 2009)*
2329 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2330 - Reuse the old instance mac address by default on instance import, if
2331 the instance name is the same.
2332 - Handle situations in which the node info rpc returns incomplete
2334 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2335 - Improved version of batcher:
2337 - state file support
2338 - instance mac address support
2339 - support for HVM clusters/instances
2341 - Add an option to show the number of cpu sockets and nodes in
2343 - Support OSes that handle more than one version of the OS api (but do
2344 not change the current API in any other way)
2345 - Fix ``gnt-node migrate``
2346 - ``gnt-debug`` man page
2347 - Fixes various more typos and small issues
2348 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2354 *(Released Wed, 24 Sep 2008)*
2356 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2357 type of disk exported to fully virtualized instances.
2358 - provide access to the serial console of HVM instances
2359 - instance auto_balance flag, set by default. If turned off it will
2360 avoid warnings on cluster verify if there is not enough memory to fail
2361 over an instance. in the future it will prevent automatically failing
2362 it over when we will support that.
2363 - batcher tool for instance creation, see ``tools/README.batcher``
2364 - ``gnt-instance reinstall --select-os`` to interactively select a new
2365 operating system when reinstalling an instance.
2366 - when changing the memory amount on instance modify a check has been
2367 added that the instance will be able to start. also warnings are
2368 emitted if the instance will not be able to fail over, if auto_balance
2370 - documentation fixes
2371 - sync fields between ``gnt-instance list/modify/add/import``
2372 - fix a race condition in drbd when the sync speed was set after giving
2373 the device a remote peer.
2379 *(Released Tue, 22 Jul 2008)*
2381 - note: the allowed size and number of tags per object were reduced
2382 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2383 - fixed twisted 8.x compatibility
2384 - fixed ``gnt-instance replace-disks`` with iallocator
2385 - add TCP keepalives on twisted connections to detect restarted nodes
2386 - disk increase support, see ``gnt-instance grow-disk``
2387 - implement bulk node/instance query for RAPI
2388 - add tags in node/instance listing (optional)
2389 - experimental migration (and live migration) support, read the man page
2390 for ``gnt-instance migrate``
2391 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2392 has some small improvements in handling its state file
2398 *(Released Fri, 13 Jun 2008)*
2400 - Experimental readonly, REST-based remote API implementation;
2401 automatically started on master node, TCP port 5080, if enabled by
2402 ``--enable-rapi`` parameter to configure script.
2403 - Instance allocator support. Add and import instance accept a
2404 ``--iallocator`` parameter, and call that instance allocator to decide
2405 which node to use for the instance. The iallocator document describes
2406 what's expected from an allocator script.
2407 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2408 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2409 node is lost there is still enough memory to fail over the instances
2411 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2412 ``gnt-cluster verify``, to check for site-specific compliance. All the
2413 hooks will run, and their output, if any, will be displayed. Any
2414 failing hook will make the verification return an error value.
2415 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2416 primary and secondary interfaces
2417 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2418 that is still declared as part of the cluster and has failed.
2419 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2420 specifying output fields, that just adds the chosen fields to the
2422 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2423 export from the filesystem.
2424 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2425 have been added. Using them you can enable/disable acpi and pae
2426 support, and specify a path for a cd image to be exported to the
2427 instance. These parameters as the name suggest only work on HVM
2429 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2430 PAE support will be set to the previously hardcoded values, but the
2431 (previously hardcoded) path to the CDROM ISO image will be unset and
2432 if required, needs to be set manually with ``gnt-instance modify``
2434 - The address to which an instance's VNC console is bound is now
2435 selectable per-instance, rather than being cluster wide. Of course
2436 this only applies to instances controlled via VNC, so currently just
2437 applies to HVM clusters.
2443 *(Released Mon, 18 Feb 2008)*
2445 - more tweaks to the disk activation code (especially helpful for DRBD)
2446 - change the default ``gnt-instance list`` output format, now there is
2447 one combined status field (see the manpage for the exact values this
2449 - some more fixes for the mac export to hooks change
2450 - make Ganeti not break with DRBD 8.2.x (which changed the version
2451 format in ``/proc/drbd``) (issue 24)
2452 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2453 template, allowing migration from DRBD0.7+MD to DRBD8
2459 *(Released Wed, 30 Jan 2008)*
2461 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2463 - add command aliases infrastructure and a few aliases
2464 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2465 man pages and the ``--help`` option of ``gnt-node
2466 list``/``gnt-instance list``
2467 - fix ``gnt-backup list`` with down nodes (issue 21)
2468 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2469 - fix the dist archive and add a check for including svn/git files in
2471 - some developer-related changes: improve the burnin and the QA suite,
2472 add an upload script for testing during development
2478 *(Released Wed, 16 Jan 2008)*
2480 - experimental HVM support, read the install document, section
2481 "Initializing the cluster"
2482 - allow for the PVM hypervisor per-instance kernel and initrd paths
2483 - add a new command ``gnt-cluster verify-disks`` which uses a new
2484 algorithm to improve the reconnection of the DRBD pairs if the device
2485 on the secondary node has gone away
2486 - make logical volume code auto-activate LVs at disk activation time
2487 - slightly improve the speed of activating disks
2488 - allow specification of the MAC address at instance creation time, and
2489 changing it later via ``gnt-instance modify``
2490 - fix handling of external commands that generate lots of output on
2492 - update documentation with regard to minimum version of DRBD8 supported
2498 *(Released Tue, 4 Dec 2007)*
2500 - Log the ``xm create`` output to the node daemon log on failure (to
2501 help diagnosing the error)
2502 - In debug mode, log all external commands output if failed to the logs
2503 - Change parsing of lvm commands to ignore stderr
2509 *(Released Wed, 28 Nov 2007)*
2511 - Another round of updates to the DRBD 8 code to deal with more failures
2512 in the replace secondary node operation
2513 - Some more logging of failures in disk operations (lvm, drbd)
2514 - A few documentation updates
2521 *(Released Tue, 13 Nov 2007)*
2523 - Change configuration file format from Python's Pickle to JSON.
2524 Upgrading is possible using the cfgupgrade utility.
2525 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2526 faster replace disks and is more stable (DRBD 8 has many improvements
2527 compared to DRBD 0.7)
2528 - Added command line tags support (see man pages for ``gnt-instance``,
2529 ``gnt-node``, ``gnt-cluster``)
2530 - Added instance rename support
2531 - Added multi-instance startup/shutdown
2532 - Added cluster rename support
2533 - Added ``gnt-node evacuate`` to simplify some node operations
2534 - Added instance reboot operation that can speedup reboot as compared to
2536 - Soften the requirement that hostnames are in FQDN format
2537 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2539 - Removed dependency on debian's patched fping that uses the
2540 non-standard ``-S`` option
2541 - Now the OS definitions are searched for in multiple, configurable
2542 paths (easier for distros to package)
2543 - Some changes to the hooks infrastructure (especially the new
2544 post-configuration update hook)
2545 - Other small bugfixes
2547 .. vim: set textwidth=72 syntax=rst :