8 *(Released Fri, 27 Jul 2012)*
11 .. attention:: The ``LUXI`` protocol has been made more consistent
12 regarding its handling of command arguments. This, however, leads to
13 incompatibility issues with previous versions. Please ensure that you
14 restart Ganeti daemons soon after the upgrade, otherwise most
15 ``LUXI`` calls (job submission, setting/resetting the drain flag,
16 pausing/resuming the watcher, cancelling and archiving jobs, querying
17 the cluster configuration) will fail.
26 The current ``admin_up`` field, which used to denote whether an instance
27 should be running or not, has been removed. Instead, ``admin_state`` is
28 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
30 The rational behind this is that an instance being “down” can have
33 - it could be down during a reboot
34 - it could be temporarily be down for a reinstall
35 - or it could be down because it is deprecated and kept just for its
38 The previous Boolean state was making it difficult to do capacity
39 calculations: should Ganeti reserve memory for a down instance? Now, the
40 tri-state field makes it clear:
42 - in ``up`` and ``down`` state, all resources are reserved for the
43 instance, and it can be at any time brought up if it is down
44 - in ``offline`` state, only disk space is reserved for it, but not
47 The field can have an extra use: since the transition between ``up`` and
48 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
49 transition between ``offline`` and ``down`` is done via ``gnt-instance
50 modify``, it is possible to given different rights to users. For
51 example, owners of an instance could be allowed to start/stop it, but
52 not transition it out of the offline state.
54 Instance policies and specs
55 +++++++++++++++++++++++++++
57 In previous Ganeti versions, an instance creation request was not
58 limited on the minimum size and on the maximum size just by the cluster
59 resources. As such, any policy could be implemented only in third-party
60 clients (RAPI clients, or shell wrappers over ``gnt-*``
61 tools). Furthermore, calculating cluster capacity via ``hspace`` again
62 required external input with regards to instance sizes.
64 In order to improve these workflows and to allow for example better
65 per-node group differentiation, we introduced instance specs, which
68 - minimum instance disk size, disk count, memory size, cpu count
69 - maximum values for the above metrics
70 - and “standard” values (used in ``hspace`` to calculate the standard
73 The minimum/maximum values can be also customised at node-group level,
74 for example allowing more powerful hardware to support bigger instance
77 Beside the instance specs, there are a few other settings belonging to
78 the instance policy framework. It is possible now to customise, per
79 cluster and node-group:
81 - the list of allowed disk templates
82 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
83 - the maximum ratio of instance to spindles (see below for more
84 information) for local storage
86 All these together should allow all tools that talk to Ganeti to know
87 what are the ranges of allowed values for instances and the
88 over-subscription that is allowed.
90 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
91 instance configuration, and the physical CPU configuration from the
92 node. For the spindle ratios however, we didn't track before these
93 values, so new parameters have been added:
95 - a new node parameter ``spindle_count``, defaults to 1, customisable at
96 node group or node level
97 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
99 Note that spindles in this context doesn't need to mean actual
100 mechanical hard-drives; it's just a relative number for both the node
101 I/O capacity and instance I/O consumption.
103 Instance migration behaviour
104 ++++++++++++++++++++++++++++
106 While live-migration is in general desirable over failover, it is
107 possible that for some workloads it is actually worse, due to the
108 variable time of the “suspend” phase during live migration.
110 To allow the tools to work consistently over such instances (without
111 having to hard-code instance names), a new backend parameter
112 ``always_failover`` has been added to control the migration/failover
113 behaviour. When set to True, all migration requests for an instance will
114 instead fall-back to failover.
116 Instance memory ballooning
117 ++++++++++++++++++++++++++
119 Initial support for memory ballooning has been added. The memory for an
120 instance is no longer fixed (backend parameter ``memory``), but instead
121 can vary between minimum and maximum values (backend parameters
122 ``minmem`` and ``maxmem``). Currently we only change an instance's
125 - live migrating or failing over and instance and the target node
126 doesn't have enough memory
127 - user requests changing the memory via ``gnt-instance modify
133 In order to control the use of specific CPUs by instance, support for
134 controlling CPU pinning has been added for the Xen, HVM and LXC
135 hypervisors. This is controlled by a new hypervisor parameter
136 ``cpu_mask``; details about possible values for this are in the
137 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
138 VCPU-to-CPU mapping) form will work well only when all nodes in your
139 cluster have the same amount of CPUs.
144 Another area in which Ganeti was not customisable were the parameters
145 used for storage configuration, e.g. how many stripes to use for LVM,
146 DRBD resync configuration, etc.
148 To improve this area, we've added disks parameters, which are
149 customisable at cluster and node group level, and which allow to
150 specify various parameters for disks (DRBD has the most parameters
151 currently), for example:
153 - DRBD resync algorithm and parameters (e.g. speed)
154 - the default VG for meta-data volumes for DRBD
155 - number of stripes for LVM (plain disk template)
158 These parameters can be modified via ``gnt-cluster modify -D …`` and
159 ``gnt-group modify -D …``, and are used at either instance creation (in
160 case of LVM stripes, for example) or at disk “activation” time
163 Rados block device support
164 ++++++++++++++++++++++++++
166 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
167 denoted by the ``rbd`` disk template type. This is considered
168 experimental, feedback is welcome. For details on configuring it, see
169 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
174 The existing master IP functionality works well only in simple setups (a
175 single network shared by all nodes); however, if nodes belong to
176 different networks, then the ``/32`` setup and lack of routing
177 information is not enough.
179 To allow the master IP to function well in more complex cases, the
180 system was reworked as follows:
182 - a master IP netmask setting has been added
183 - the master IP activation/turn-down code was moved from the node daemon
185 - whether to run the Ganeti-supplied master IP script or a user-supplied
186 on is a ``gnt-cluster init`` setting
188 Details about the location of the standard and custom setup scripts are
189 in the man page :manpage:`gnt-cluster(8)`; for information about the
190 setup script protocol, look at the Ganeti-supplied script.
195 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
198 It is now possible to use TLS-protected connections, and when renewing
199 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
200 it is now possible to specify spice or spice CA certificates. Also, it
201 is possible to configure a password for SPICE sessions via the
202 hypervisor parameter ``spice_password_file``.
204 There are also new parameters to control the compression and streaming
205 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
206 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
207 for the spice parameters.
209 Lastly, it is now possible to see the SPICE connection information via
210 ``gnt-instance console``.
215 A new tool (``tools/ovfconverter``) has been added that supports
216 conversion between Ganeti and the `Open Virtualization Format
217 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
220 This relies on the ``qemu-img`` tool to convert the disk formats, so the
221 actual compatibility with other virtualization solutions depends on it.
226 The configuration query daemon (``ganeti-confd``) is now optional, and
227 has been rewritten in Haskell; whether to use the daemon at all, use the
228 Python (default) or the Haskell version is selectable at configure time
229 via the ``--enable-confd`` parameter, which can take one of the
230 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
231 daemon will result in a smaller footprint; for larger systems, we
232 welcome feedback on the Haskell version which might become the default
235 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
236 daemon running. The Python version doesn't implement the new call.
239 User interface changes
240 ~~~~~~~~~~~~~~~~~~~~~~
242 We have replaced the ``--disks`` option of ``gnt-instance
243 replace-disks`` with a more flexible ``--disk`` option, which allows
244 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
245 disk size and mode can be changed upon recreation (via ``gnt-instance
246 recreate-disks``, which accepts the same ``--disk`` option).
248 As many people are used to a ``show`` command, we have added that as an
249 alias to ``info`` on all ``gnt-*`` commands.
251 The ``gnt-instance grow-disk`` command has a new mode in which it can
252 accept the target size of the disk, instead of the delta; this can be
253 more safe since two runs in absolute mode will be idempotent, and
254 sometimes it's also easier to specify the desired size directly.
256 Also the handling of instances with regard to offline secondaries has
257 been improved. Instance operations should not fail because one of it's
258 secondary nodes is offline, even though it's safe to proceed.
260 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
261 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
262 minors to instance name.
267 RAPI coverage has improved, with (for example) new resources for
268 recreate-disks, node power-cycle, etc.
273 There is partial support for ``xl`` in the Xen hypervisor; feedback is
276 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
277 whether to still support Python 2.4 or move to Python 2.6 as minimum
280 Support for Fedora has been slightly improved; the provided example
281 init.d script should work better on it and the INSTALL file should
282 document the needed dependencies.
287 The deprecated ``QueryLocks`` LUXI request has been removed. Use
288 ``Query(what=QR_LOCK, ...)`` instead.
290 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
291 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
292 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
293 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
294 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
296 RAPI client: ``CertificateError`` now derives from
297 ``GanetiApiError``. This should make it more easy to handle Ganeti
300 Deprecation warnings due to PyCrypto/paramiko import in
301 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
302 make sure to run an up-to-date paramiko version, if you use this tool.
304 The QA scripts now depend on Python 2.5 or above (the main code base
305 still works with Python 2.4).
307 The configuration file (``config.data``) is now written without
308 indentation for performance reasons; if you want to edit it, it can be
309 re-formatted via ``tools/fmtjson``.
311 A number of bugs has been fixed in the cluster merge tool.
313 ``x509`` certification verification (used in import-export) has been
314 changed to allow the same clock skew as permitted by the cluster
315 verification. This will remove some rare but hard to diagnose errors in
322 *(Released Thu, 19 Jul 2012)*
324 Very few changes from rc4 to the final release, only bugfixes:
326 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
327 instance, fix CDROM booting for KVM instances)
328 - fixed node group modification of node parameters
329 - fixed issue in LUClusterVerifyGroup with multi-group clusters
330 - fixed generation of bash completion to ensure a stable ordering
337 *(Released Fri, 13 Jul 2012)*
339 Third release candidate for 2.6. The following changes were done from
342 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
343 - Fixed an inconsistency in the LUXI protocol with the provided
344 arguments (NOT backwards compatible)
345 - Fixed a bug with node groups ipolicy where ``min`` was greater than
346 the cluster ``std`` value
347 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
348 easier instance debugging on nodes (requires ``hconfd`` to work)
354 *(Released Tue, 03 Jul 2012)*
356 Second release candidate for 2.6. The following changes were done from
359 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
361 - Fixed a RAPI regression on missing beparams/memory
362 - Fixed redistribution of files on offline nodes
363 - Added possibility to run activate-disks even though secondaries are
364 offline. With this change it relaxes also the strictness on some other
365 commands which use activate disks internally:
366 * ``gnt-instance start|reboot|rename|backup|export``
367 - Made it possible to remove safely an instance if its secondaries are
369 - Made it possible to reinstall even though secondaries are offline
375 *(Released Mon, 25 Jun 2012)*
377 First release candidate for 2.6. The following changes were done from
380 - Fixed bugs with disk parameters and ``rbd`` templates as well as
382 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
383 behaviour. It supports now the modify operation
384 - ``hcheck`` implemented to analyze cluster health and possibility of
385 improving health by rebalance
386 - ``hbal`` has been improved in dealing with split instances
392 *(Released Mon, 11 Jun 2012)*
394 Second beta release of 2.6. The following changes were done from beta2
397 - Fixed ``daemon-util`` with non-root user models
398 - Fixed creation of plain instances with ``--no-wait-for-sync``
399 - Fix wrong iv_names when running ``cfgupgrade``
400 - Export more information in RAPI group queries
401 - Fixed bug when changing instance network interfaces
402 - Extended burnin to do NIC changes
403 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
404 - Changed default for DRBD barriers
405 - Fixed DRBD error reporting for syncer rate
406 - Verify the options on disk parameters
408 And of course various fixes to documentation and improved unittests and
415 *(Released Wed, 23 May 2012)*
417 First beta release of 2.6. The following changes were done from beta1 to
420 - integrated patch for distributions without ``start-stop-daemon``
421 - adapted example init.d script to work on Fedora
422 - fixed log handling in Haskell daemons
423 - adapted checks in the watcher for pycurl linked against libnss
424 - add partial support for ``xl`` instead of ``xm`` for Xen
425 - fixed a type issue in cluster verification
426 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
429 Plus integrated fixes from the 2.5 branch:
431 - fixed ``kvm-ifup`` to use ``/bin/bash``
432 - fixed parallel build failures
433 - KVM live migration when using a custom keymap
439 *(Released Tue, 24 Jul 2012)*
441 A small bugfix release, with no new features:
443 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
444 different default shell (e.g. Debian, Ubuntu)
445 - fixed KVM startup and live migration with a custom keymap (fixes Issue
446 243 and Debian bug #650664)
447 - fixed compatibility with KVM versions that don't support multiple boot
448 devices (fixes Issue 230 and Debian bug #624256)
450 Additionally, a few fixes were done to the build system (fixed parallel
451 build failures) and to the unittests (fixed race condition in test for
452 FileID functions, and the default enable/disable mode for QA test is now
459 *(Released Fri, 11 May 2012)*
461 A small bugfix release.
463 The main issues solved are on the topic of compatibility with newer LVM
466 - fixed parsing of ``lv_attr`` field
467 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
468 the ``--force`` flag is needed
470 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
471 to accept kernel 3.x too (was hardcoded to 2.6.*).
473 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
474 of hook results and that also made display of other errors suboptimal
475 was fixed; the code behaves now like 2.4 and earlier.
477 Another change in 2.5, the cleanup of the OS scripts environment, is too
478 aggressive: it removed even the ``PATH`` variable, which requires the OS
479 scripts to *always* need to export it. Since this is a bit too strict,
480 we now export a minimal PATH, the same that we export for hooks.
482 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
483 integrated into this release.
485 Finally, a few other miscellaneous changes were done (no new features,
486 just small improvements):
488 - Fix ``gnt-group --help`` display
489 - Fix hardcoded Xen kernel path
490 - Fix grow-disk handling of invalid units
491 - Update synopsis for ``gnt-cluster repair-disk-sizes``
492 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
498 *(Released Thu, 12 Apr 2012)*
500 Incompatible/important changes and bugfixes
501 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
503 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
504 resource's ``ip_check`` parameter changed from ``True`` to ``False``
505 to match the underlying LUXI interface.
506 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
507 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
508 not maintain backwards-compatibility as the underlying operation
509 changed in an incompatible way. The RAPI client can talk to old
510 servers, but it needs to be told so as the return value changed.
511 - When creating file-based instances via RAPI, the ``file_driver``
512 parameter no longer defaults to ``loop`` and must be specified.
513 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
515 - Support for the undocumented and deprecated RAPI instance creation
516 request format version 0 has been dropped. Use version 1, supported
517 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
518 - Pyparsing 1.4.6 or above is required, see :doc:`installation
519 documentation <install>`.
520 - The "cluster-verify" hooks are now executed per group by the
521 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
522 if you just run ``gnt-cluster verify``, which generates one opcode per
524 - The environment as passed to the OS scripts is cleared, and thus no
525 environment variables defined in the node daemon's environment will be
526 inherited by the scripts.
527 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
529 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
530 support operations involving multiple node groups.
531 - Offline nodes are ignored when failing over an instance.
532 - Support for KVM version 1.0, which changed the version reporting format
534 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
536 - ``Makefile`` is now compatible with Automake 1.11.2
537 - Includes all bugfixes made in the 2.4 series
542 - The ganeti-htools project has been merged into the ganeti-core source
543 tree and will be built as part of Ganeti (see :doc:`install-quick`).
544 - Implemented support for :doc:`shared storage <design-shared-storage>`.
545 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
546 GPT-style partition tables (requires `parted
547 <http://www.gnu.org/s/parted/>`_).
548 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
549 - Allowed adding tags on instance creation.
550 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
551 - Allow instances to be started in a paused state, enabling the user to
552 see the complete console output on boot using the console.
553 - Added new hypervisor flag to control default reboot behaviour
554 (``reboot_behavior``).
555 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
556 - Improved out-of-band management support:
558 - Added ``gnt-node health`` command reporting the health status of
560 - Added ``gnt-node power`` command to manage power status of nodes.
561 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
563 - Instance migration can fall back to failover if instance is not
565 - Filters can be used when listing nodes, instances, groups and locks;
566 see *ganeti(7)* manpage.
567 - Added post-execution status as variables to :doc:`hooks <hooks>`
569 - Instance tags are exported/imported together with the instance.
570 - When given an explicit job ID, ``gnt-job info`` will work for archived
572 - Jobs can define dependencies on other jobs (not yet supported via
573 RAPI or command line, but used by internal commands and usable via
576 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
579 - Instance failover is now available as a RAPI resource
580 (``/2/instances/[instance_name]/failover``).
581 - ``gnt-instance info`` defaults to static information if primary node
583 - Opcodes have a new ``comment`` attribute.
584 - Added basic SPICE support to KVM hypervisor.
585 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
587 Node group improvements
588 ~~~~~~~~~~~~~~~~~~~~~~~
590 - ``gnt-cluster verify`` has been modified to check groups separately,
591 thereby improving performance.
592 - Node group support has been added to ``gnt-cluster verify-disks``,
593 which now operates per node group.
594 - Watcher has been changed to work better with node groups.
596 - One process and state file per node group.
597 - Slow watcher in one group doesn't block other group's watcher.
599 - Added new command, ``gnt-group evacuate``, to move all instances in a
600 node group to other groups.
601 - Added ``gnt-instance change-group`` to move an instance to another
603 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
604 per-group operations.
605 - Node groups can be tagged.
606 - Some operations switch from an exclusive to a shared lock as soon as
608 - Instance's primary and secondary nodes' groups are now available as
609 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
610 and ``snodes.group.uuid``).
615 - Numerous updates to documentation and manpages.
617 - :doc:`RAPI <rapi>` documentation now has detailed parameter
619 - Some opcode/job results are now also documented, see :doc:`RAPI
622 - A lockset's internal lock is now also visible in lock monitor.
623 - Log messages from job queue workers now contain information about the
624 opcode they're processing.
625 - ``gnt-instance console`` no longer requires the instance lock.
626 - A short delay when waiting for job changes reduces the number of LUXI
627 requests significantly.
628 - DRBD metadata volumes are overwritten with zeros during disk creation.
629 - Out-of-band commands no longer acquire the cluster lock in exclusive
631 - ``devel/upload`` now uses correct permissions for directories.
637 *(Released Fri, 23 Mar 2012)*
639 This was the sixth release candidate of the 2.5 series.
645 *(Released Mon, 9 Jan 2012)*
647 This was the fifth release candidate of the 2.5 series.
653 *(Released Thu, 27 Oct 2011)*
655 This was the fourth release candidate of the 2.5 series.
661 *(Released Wed, 26 Oct 2011)*
663 This was the third release candidate of the 2.5 series.
669 *(Released Tue, 18 Oct 2011)*
671 This was the second release candidate of the 2.5 series.
677 *(Released Tue, 4 Oct 2011)*
679 This was the first release candidate of the 2.5 series.
685 *(Released Wed, 31 Aug 2011)*
687 This was the third beta release of the 2.5 series.
693 *(Released Mon, 22 Aug 2011)*
695 This was the second beta release of the 2.5 series.
701 *(Released Fri, 12 Aug 2011)*
703 This was the first beta release of the 2.5 series.
709 *(Released Thu, 27 Oct 2011)*
711 - Fixed bug when parsing command line parameter values ending in
713 - Fixed assertion error after unclean master shutdown
714 - Disable HTTP client pool for RPC, significantly reducing memory usage
716 - Fixed queue archive creation with wrong permissions
722 *(Released Tue, 23 Aug 2011)*
726 - Fixed documentation for importing with ``--src-dir`` option
727 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
728 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
734 *(Released Fri, 5 Aug 2011)*
736 Many bug-fixes and a few small features:
738 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
739 issues when you tried to add an instance with two MAC addresses in one
741 - KVM: fixed per-instance stored UID value
742 - KVM: configure bridged NICs at migration start
743 - KVM: Fix a bug where instance will not start with never KVM versions
745 - Added OS search path to ``gnt-cluster info``
746 - Fixed an issue with ``file_storage_dir`` where you were forced to
747 provide an absolute path, but the documentation states it is a
748 relative path, the documentation was right
749 - Added a new parameter to instance stop/start called ``--no-remember``
750 that will make the state change to not be remembered
751 - Implemented ``no_remember`` at RAPI level
752 - Improved the documentation
753 - Node evacuation: don't call IAllocator if node is already empty
754 - Fixed bug in DRBD8 replace disks on current nodes
755 - Fixed bug in recreate-disks for DRBD instances
756 - Moved assertion checking locks in ``gnt-instance replace-disks``
757 causing it to abort with not owning the right locks for some situation
758 - Job queue: Fixed potential race condition when cancelling queued jobs
759 - Fixed off-by-one bug in job serial generation
760 - ``gnt-node volumes``: Fix instance names
761 - Fixed aliases in bash completion
762 - Fixed a bug in reopening log files after being sent a SIGHUP
763 - Added a flag to burnin to allow specifying VCPU count
764 - Bugfixes to non-root Ganeti configuration
770 *(Released Thu, 12 May 2011)*
772 Many bug-fixes and a few new small features:
774 - Fixed a bug related to log opening failures
775 - Fixed a bug in instance listing with orphan instances
776 - Fixed a bug which prevented resetting the cluster-level node parameter
777 ``oob_program`` to the default
778 - Many fixes related to the ``cluster-merge`` tool
779 - Fixed a race condition in the lock monitor, which caused failures
780 during (at least) creation of many instances in parallel
781 - Improved output for gnt-job info
782 - Removed the quiet flag on some ssh calls which prevented debugging
784 - Improved the N+1 failure messages in cluster verify by actually
785 showing the memory values (needed and available)
786 - Increased lock attempt timeouts so that when executing long operations
787 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
788 too early and thus prevent the use of the 'fair' mechanism
789 - Changed instance query data (``gnt-instance info``) to not acquire
790 locks unless needed, thus allowing its use on locked instance if only
791 static information is asked for
792 - Improved behaviour with filesystems that do not support rename on an
794 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
795 kept locks on all nodes during the wipe, which is unneeded
796 - Fixed ``gnt-watcher`` handling of errors during hooks execution
797 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
798 10GiB) which caused the wipe to fail right at the end in some cases
799 - Fixed master IP activation when doing master failover with no-voting
800 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
801 the master node itself
802 - Fixed potential data-loss in under disk full conditions, where Ganeti
803 wouldn't check correctly the return code and would consider
804 partially-written files 'correct'
805 - Fixed bug related to multiple VGs and DRBD disk replacing
806 - Added new disk parameter ``metavg`` that allows placement of the meta
807 device for DRBD in a different volume group
808 - Fixed error handling in the node daemon when the system libc doesn't
809 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
810 - Fixed lock release during replace-disks, which kept cluster-wide locks
811 when doing disk replaces with an iallocator script
812 - Added check for missing bridges in cluster verify
813 - Handle EPIPE errors while writing to the terminal better, so that
814 piping the output to e.g. ``less`` doesn't cause a backtrace
815 - Fixed rare case where a ^C during Luxi calls could have been
816 interpreted as server errors, instead of simply terminating
817 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
819 - Added a few more parameters to the KVM hypervisor, allowing a second
820 CDROM, custom disk type for CDROMs and a floppy image
821 - Removed redundant message in instance rename when the name is given
823 - Added option to ``gnt-instance recreate-disks`` to allow creating the
824 disks on new nodes, allowing recreation when the original instance
825 nodes are completely gone
826 - Added option when converting disk templates to DRBD to skip waiting
827 for the resync, in order to make the instance available sooner
828 - Added two new variables to the OS scripts environment (containing the
830 - Made the root_path and optional parameter for the xen-pvm hypervisor,
831 to allow use of ``pvgrub`` as bootloader
832 - Changed the instance memory modifications to only check out-of-memory
833 conditions on memory increases, and turned the secondary node warnings
834 into errors (they can still be overridden via ``--force``)
835 - Fixed the handling of a corner case when the Python installation gets
836 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
837 to execute a command that doesn't exist
838 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
839 node of the instance returned failures during instance shutdown; this
840 adds the option ``--ignore-consistency`` to gnt-instance move
842 And as usual, various improvements to the error messages, documentation
849 *(Released Wed, 09 Mar 2011)*
851 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
852 the RAPI users file if run twice (even with ``--dry-run``).
854 The release fixes that bug (nothing else changed).
860 *(Released Mon, 07 Mar 2011)*
862 Final 2.4.0 release. Just a few small fixes:
864 - Fixed RAPI node evacuate
865 - Fixed the kvm-ifup script
866 - Fixed internal error handling for special job cases
867 - Updated man page to specify the escaping feature for options
873 *(Released Mon, 28 Feb 2011)*
875 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
876 that this feature wiped the disks of the wrong instance, leading to loss
881 - Fixed title of query field containing instance name
882 - Expanded the glossary in the documentation
883 - Fixed one unittest (internal issue)
889 *(Released Mon, 21 Feb 2011)*
891 A number of bug fixes plus just a couple functionality changes.
893 On the user-visible side, the ``gnt-* list`` command output has changed
894 with respect to "special" field states. The current rc1 style of display
895 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
896 the default output mode special fields states are displayed as follows:
898 - Offline resource: ``*``
899 - Unavailable/not applicable: ``-``
900 - Data missing (RPC failure): ``?``
901 - Unknown field: ``??``
903 Another user-visible change is the addition of ``--force-join`` to
908 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
909 - Fixed regression in RAPI/instance reinstall where all parameters were
910 required (instead of optional)
911 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
912 - Fixed iallocator usage (offline nodes were not considered offline)
913 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
914 - Fixed hypervisor and OS parameter validation with respect to
916 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
918 - Fixed ``tools/listrunner`` with respect to agent-based usage
924 *(Released Fri, 4 Feb 2011)*
926 Many changes and fixes since the beta1 release. While there were some
927 internal changes, the code has been mostly stabilised for the RC
930 Note: the dumb allocator was removed in this release, as it was not kept
931 up-to-date with the IAllocator protocol changes. It is recommended to
932 use the ``hail`` command from the ganeti-htools package.
934 Note: the 2.4 and up versions of Ganeti are not compatible with the
935 0.2.x branch of ganeti-htools. You need to upgrade to
936 ganeti-htools-0.3.0 (or later).
938 Regressions fixed from 2.3
939 ~~~~~~~~~~~~~~~~~~~~~~~~~~
941 - Fixed the ``gnt-cluster verify-disks`` command
942 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
944 - Fixed disk adoption breakage
945 - Fixed wrong headers in instance listing for field aliases
950 - Fixed corner case in KVM handling of NICs
951 - Fixed many cases of wrong handling of non-vm_capable nodes
952 - Fixed a bug where a missing instance symlink was not possible to
953 recreate with any ``gnt-*`` command (now ``gnt-instance
954 activate-disks`` does it)
955 - Fixed the volume group name as reported by ``gnt-cluster
957 - Increased timeouts for the import-export code, hopefully leading to
958 fewer aborts due network or instance timeouts
959 - Fixed bug in ``gnt-node list-storage``
960 - Fixed bug where not all daemons were started on cluster
961 initialisation, but only at the first watcher run
962 - Fixed many bugs in the OOB implementation
963 - Fixed watcher behaviour in presence of instances with offline
965 - Fixed instance list output for instances running on the wrong node
966 - a few fixes to the cluster-merge tool, but it still cannot merge
967 multi-node groups (currently it is not recommended to use this tool)
973 - Improved network configuration for the KVM hypervisor
974 - Added e1000 as a supported NIC for Xen-HVM
975 - Improved the lvmstrap tool to also be able to use partitions, as
976 opposed to full disks
977 - Improved speed of disk wiping (the cluster parameter
978 ``prealloc_wipe_disks``, so that it has a low impact on the total time
979 of instance creations
980 - Added documentation for the OS parameters
981 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
982 hypervisor is not responding
983 - Added display of blacklisted and hidden OS information in
985 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
986 NIC and node parameters, which might create problems with currently
987 invalid (but undetected) configuration files, but prevents validation
988 failures when unrelated parameters are modified
989 - Changed cluster initialisation to wait for the master daemon to become
991 - Expanded the RAPI interface:
993 - Added config redistribution resource
994 - Added activation/deactivation of instance disks
995 - Added export of console information
997 - Implemented log file reopening on SIGHUP, which allows using
998 logrotate(8) for the Ganeti log files
999 - Added a basic OOB helper script as an example
1005 *(Released Fri, 14 Jan 2011)*
1010 - Fixed timezone issues when formatting timestamps
1011 - Added support for node groups, available via ``gnt-group`` and other
1013 - Added out-of-band framework and management, see :doc:`design
1014 document <design-oob>`
1015 - Removed support for roman numbers from ``gnt-node list`` and
1016 ``gnt-instance list``.
1017 - Allowed modification of master network interface via ``gnt-cluster
1018 modify --master-netdev``
1019 - Accept offline secondaries while shutting down instance disks
1020 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1021 - Added support for multiple LVM volume groups
1022 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1024 - Added commands to list available fields:
1026 - ``gnt-node list-fields``
1027 - ``gnt-group list-fields``
1028 - ``gnt-instance list-fields``
1030 - Updated documentation and man pages
1035 - Moved ``rapi_users`` file into separate directory, now named
1036 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1038 - Added new tool for running commands on many machines,
1039 ``tools/ganeti-listrunner``
1040 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1041 improving the ``gnt-instance console`` output
1042 - Allowed customisation of disk index separator at ``configure`` time
1043 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1044 - Added support for non-partitioned md disks in ``lvmstrap``
1045 - Added script to gracefully power off KVM instances
1046 - Split ``utils`` module into smaller parts
1047 - Changed query operations to return more detailed information, e.g.
1048 whether an information is unavailable due to an offline node. To use
1049 this new functionality, the LUXI call ``Query`` must be used. Field
1050 information is now stored by the master daemon and can be retrieved
1051 using ``QueryFields``. Instances, nodes and groups can also be queried
1052 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1053 exposed via RAPI). The following commands make use of this
1054 infrastructure change:
1056 - ``gnt-group list``
1057 - ``gnt-group list-fields``
1059 - ``gnt-node list-fields``
1060 - ``gnt-instance list``
1061 - ``gnt-instance list-fields``
1062 - ``gnt-debug locks``
1067 - New RAPI resources (see :doc:`rapi`):
1071 - ``/2/groups/[group_name]``
1072 - ``/2/groups/[group_name]/assign-nodes``
1073 - ``/2/groups/[group_name]/modify``
1074 - ``/2/groups/[group_name]/rename``
1075 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1079 - Implemented ``no_install`` for instance creation
1080 - Implemented OS parameters for instance reinstallation, allowing
1081 use of special settings on reinstallation (e.g. for preserving data)
1086 - Added IPv6 support in import/export
1087 - Pause DRBD synchronization while wiping disks on instance creation
1088 - Updated unittests and QA scripts
1089 - Improved network parameters passed to KVM
1090 - Converted man pages from docbook to reStructuredText
1096 *(Released Mon, 20 Dec 2010)*
1098 Released version 2.3.1~rc1 without any changes.
1104 *(Released Wed, 1 Dec 2010)*
1106 - impexpd: Disable OpenSSL compression in socat if possible (backport
1107 from master, commit e90739d625b, see :doc:`installation guide
1108 <install-quick>` for details)
1109 - Changed unittest coverage report to exclude test scripts
1110 - Added script to check version format
1116 *(Released Wed, 1 Dec 2010)*
1118 Released version 2.3.0~rc1 without any changes.
1124 *(Released Fri, 19 Nov 2010)*
1126 A number of bugfixes and documentation updates:
1128 - Update ganeti-os-interface documentation
1129 - Fixed a bug related to duplicate MACs or similar items which should be
1131 - Fix breakage in OS state modify
1132 - Reinstall instance: disallow offline secondaries (fixes bug related to
1133 OS changing but reinstall failing)
1134 - plus all the other fixes between 2.2.1 and 2.2.2
1140 *(Released Tue, 2 Nov 2010)*
1142 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1143 - Fixed master failover race with watcher
1144 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1146 - Accept previously stopped instance for export with instance removal
1147 - Simplify and extend the environment variables for instance OS scripts
1148 - Added new node flags, ``master_capable`` and ``vm_capable``
1149 - Added optional instance disk wiping prior during allocation. This is a
1150 cluster-wide option and can be set/modified using
1151 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1152 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1153 :doc:`install-quick`
1154 - Added a new watcher option (``--ignore-pause``)
1155 - Added option to ignore offline node on instance start/stop
1156 (``--ignore-offline``)
1157 - Allow overriding OS parameters with ``gnt-instance reinstall``
1158 - Added ability to change node's secondary IP address using ``gnt-node
1160 - Implemented privilege separation for all daemons except
1161 ``ganeti-noded``, see ``configure`` options
1162 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1164 - Implemented job priorities (see ``ganeti(7)`` manpage)
1165 - Ignore failures while shutting down instances during failover from
1167 - Exit daemon's bootstrap process only once daemon is ready
1168 - Export more information via ``LUInstanceQuery``/remote API
1169 - Improved documentation, QA and unittests
1170 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1171 restart if the file was created or changed
1172 - Added LUXI protocol version sent with each request and response,
1173 allowing detection of server/client mismatches
1174 - Moved the Python scripts among gnt-* and ganeti-* into modules
1175 - Moved all code related to setting up SSH to an external script,
1177 - Infrastructure changes for node group support in future versions
1183 *(Released Fri, 19 Nov 2010)*
1185 A few small bugs fixed, and some improvements to the build system:
1187 - Fix documentation regarding conversion to drbd
1188 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1190 - Fix error handling in node modify with multiple changes
1191 - Allow remote imports without checked names
1197 *(Released Tue, 19 Oct 2010)*
1199 - Disable SSL session ID cache in RPC client
1205 *(Released Thu, 14 Oct 2010)*
1207 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1208 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1210 - Fix problems with interaction between Curl and Python's HTTP server,
1211 resulting in increased speed in many RPC calls
1212 - Improve our release script to prevent breakage with older aclocal and
1219 *(Released Thu, 7 Oct 2010)*
1221 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1222 value retrieved from master
1223 - Added support for blacklisted or hidden OS definitions
1224 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1225 - Added support for -mem-path in KVM hypervisor abstraction layer
1226 - Allow overriding instance parameters in tool for inter-cluster
1227 instance moves (``tools/move-instance``)
1228 - Improved opcode summaries (e.g. in ``gnt-job list``)
1229 - Improve consistency of OS listing by sorting it
1230 - Documentation updates
1236 *(Released Fri, 8 Oct 2010)*
1238 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1244 *(Released Mon, 4 Oct 2010)*
1246 - Fixed regression in ``gnt-instance rename``
1252 *(Released Wed, 22 Sep 2010)*
1254 - Fixed OS_VARIANT variable for OS scripts
1255 - Fixed cluster tag operations via RAPI
1256 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1257 - Disabled RAPI CA checks in watcher
1263 *(Released Mon, 23 Aug 2010)*
1265 - Support DRBD versions of the format "a.b.c.d"
1267 - Re-introduce support for usage from multiple threads in RAPI client
1268 - Instance renames and modify via RAPI
1269 - Work around race condition between processing and archival in job
1271 - Mark opcodes following failed one as failed, too
1272 - Job field ``lock_status`` was removed due to difficulties making it
1273 work with the changed job queue in Ganeti 2.2; a better way to monitor
1274 locks is expected for a later 2.2.x release
1275 - Fixed dry-run behaviour with many commands
1276 - Support ``ssh-agent`` again when adding nodes
1277 - Many additional bugfixes
1283 *(Released Fri, 30 Jul 2010)*
1285 Important change: the internal RPC mechanism between Ganeti nodes has
1286 changed from using a home-grown http library (based on the Python base
1287 libraries) to use the PycURL library. This requires that PycURL is
1288 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1289 against GnuTLS by default. cURL's support for GnuTLS had known issues
1290 before cURL 7.21.0 and we recommend using the latest cURL release or
1291 linking against OpenSSL. Most other distributions already link PycURL
1292 and cURL against OpenSSL. The command::
1294 python -c 'import pycurl; print pycurl.version'
1296 can be used to determine the libraries PycURL and cURL are linked
1299 Other significant changes:
1301 - Rewrote much of the internals of the job queue, in order to achieve
1302 better parallelism; this decouples job query operations from the job
1303 processing, and it should allow much nicer behaviour of the master
1304 daemon under load, and it also has uncovered some long-standing bugs
1305 related to the job serialisation (now fixed)
1306 - Added a default iallocator setting to the cluster parameters,
1307 eliminating the need to always pass nodes or an iallocator for
1308 operations that require selection of new node(s)
1309 - Added experimental support for the LXC virtualization method
1310 - Added support for OS parameters, which allows the installation of
1311 instances to pass parameter to OS scripts in order to customise the
1313 - Added a hypervisor parameter controlling the migration type (live or
1314 non-live), since hypervisors have various levels of reliability; this
1315 has renamed the 'live' parameter to 'mode'
1316 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1317 logical volumes, meaning that cluster verify will ignore them and not
1318 flag their presence as errors
1319 - The watcher will now reset the error count for failed instances after
1320 8 hours, thus allowing self-healing if the problem that caused the
1321 instances to be down/fail to start has cleared in the meantime
1322 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1323 check for, and warn, if the drbd module parameter ``usermode_helper``
1324 is not consistent with the cluster-wide setting; this is needed to
1325 make diagnose easier of failed drbd creations
1326 - Started adding base IPv6 support, but this is not yet
1327 enabled/available for use
1328 - Rename operations (cluster, instance) will now return the new name,
1329 which is especially useful if a short name was passed in
1330 - Added support for instance migration in RAPI
1331 - Added a tool to pre-configure nodes for the SSH setup, before joining
1332 them to the cluster; this will allow in the future a simplified model
1333 for node joining (but not yet fully enabled in 2.2); this needs the
1334 paramiko python library
1335 - Fixed handling of name-resolving errors
1336 - Fixed consistency of job results on the error path
1337 - Fixed master-failover race condition when executed multiple times in
1339 - Fixed many bugs related to the job queue (mostly introduced during the
1340 2.2 development cycle, so not all are impacting 2.1)
1341 - Fixed instance migration with missing disk symlinks
1342 - Fixed handling of unknown jobs in ``gnt-job archive``
1343 - And many other small fixes/improvements
1347 - Enhanced both the unittest and the QA coverage
1348 - Switched the opcode validation to a generic model, and extended the
1349 validation to all opcode parameters
1350 - Changed more parts of the code that write shell scripts to use the
1352 - Switched the master daemon to use the asyncore library for the Luxi
1359 *(Released Thu, 17 Jun 2010)*
1361 - Added tool (``move-instance``) and infrastructure to move instances
1362 between separate clusters (see :doc:`separate documentation
1363 <move-instance>` and :doc:`design document <design-2.2>`)
1364 - Added per-request RPC timeout
1365 - RAPI now requires a Content-Type header for requests with a body (e.g.
1366 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1367 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1368 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1370 - Implemented initial support for running Ganeti daemons as separate
1371 users, see configure-time flags ``--with-user-prefix`` and
1372 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1374 - Instances can be removed after export (``gnt-backup export
1375 --remove-instance``)
1376 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1377 key (instead of 1024 bit)
1378 - Added new cluster configuration file for cluster domain secret
1379 - Import/export now use SSL instead of SSH
1380 - Added support for showing estimated time when exporting an instance,
1381 see the ``ganeti-os-interface(7)`` manpage and look for
1388 *(Released Tue, 16 Nov 2010)*
1390 Some more bugfixes. Unless critical bugs occur, this will be the last
1393 - Fix case of MAC special-values
1394 - Fix mac checker regex
1395 - backend: Fix typo causing "out of range" error
1396 - Add missing --units in gnt-instance list man page
1402 *(Released Tue, 24 Aug 2010)*
1405 - Don't ignore secondary node silently on non-mirrored disk templates
1407 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1408 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1409 - Properly document the usb_mouse parameter
1410 - Fix path in ganeti-rapi(8) (issue 116)
1411 - Adjust error message when the ganeti user's .ssh directory is
1413 - Add same-node-check when changing the disk template to drbd
1419 *(Released Fri, 16 Jul 2010)*
1422 - Add an option to only select some reboot types during qa/burnin.
1423 (on some hypervisors consequent reboots are not supported)
1424 - Fix infrequent race condition in master failover. Sometimes the old
1425 master ip address would be still detected as up for a short time
1426 after it was removed, causing failover to fail.
1427 - Decrease mlockall warnings when the ctypes module is missing. On
1428 Python 2.4 we support running even if no ctypes module is installed,
1429 but we were too verbose about this issue.
1430 - Fix building on old distributions, on which man doesn't have a
1432 - Fix RAPI not to ignore the MAC address on instance creation
1433 - Implement the old instance creation format in the RAPI client.
1439 *(Released Thu, 01 Jul 2010)*
1441 A small bugfix release:
1442 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1443 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1444 a non-existing option
1445 - Fix instance create: the --force-variant option was ignored
1446 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1447 - Fix modify node storage with non-FQDN arguments
1448 - Fix RAPI client to authenticate under Python 2.6 when used
1449 for more than 5 requests needing authentication
1450 - Fix gnt-instance modify -t (storage) giving a wrong error message
1451 when converting a non-shutdown drbd instance to plain
1457 *(Released Fri, 18 Jun 2010)*
1459 A small bugfix release:
1461 - Fix live migration of KVM instances started with older Ganeti
1462 versions which had fewer hypervisor parameters
1463 - Fix gnt-instance grow-disk on down instances
1464 - Fix an error-reporting bug during instance migration
1465 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1466 silently ignoring broken ones
1467 - Fix an RPC error reporting bug affecting, for example, RAPI client
1469 - Fix bug triggered by different API version os-es on different nodes
1470 - Fix a bug in instance startup with custom hvparams: OS level
1471 parameters would fail to be applied.
1472 - Fix the RAPI client under Python 2.6 (but more work is needed to
1473 make it work completely well with OpenSSL)
1474 - Fix handling of errors when resolving names from DNS
1480 *(Released Thu, 3 Jun 2010)*
1482 A medium sized development cycle. Some new features, and some
1483 fixes/small improvements/cleanups.
1485 Significant features
1486 ~~~~~~~~~~~~~~~~~~~~
1488 The node deamon now tries to mlock itself into memory, unless the
1489 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1490 its logs, and falls back to console logging. This allows emergency
1491 features such as ``gnt-node powercycle`` to work even in the event of a
1492 broken node disk (tested offlining the disk hosting the node's
1493 filesystem and dropping its memory caches; don't try this at home)
1495 KVM: add vhost-net acceleration support. It can be tested with a new
1496 enough version of the kernel and of qemu-kvm.
1498 KVM: Add instance chrooting feature. If you use privilege dropping for
1499 your VMs you can also now force them to chroot to an empty directory,
1500 before starting the emulated guest.
1502 KVM: Add maximum migration bandwith and maximum downtime tweaking
1503 support (requires a new-enough version of qemu-kvm).
1505 Cluster verify will now warn if the master node doesn't have the master
1506 ip configured on it.
1508 Add a new (incompatible) instance creation request format to RAPI which
1509 supports all parameters (previously only a subset was supported, and it
1510 wasn't possible to extend the old format to accomodate all the new
1511 features. The old format is still supported, and a client can check for
1512 this feature, before using it, by checking for its presence in the
1513 ``features`` RAPI resource.
1515 Now with ancient latin support. Try it passing the ``--roman`` option to
1516 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1517 (requires the python-roman module to be installed, in order to work).
1522 As usual many internal code refactorings, documentation updates, and
1525 - Lots of improvements and cleanups to the experimental Remote API
1526 (RAPI) client library.
1527 - A new unit test suite for the core daemon libraries.
1528 - A fix to creating missing directories makes sure the umask is not
1529 applied anymore. This enforces the same directory permissions
1531 - Better handling terminating daemons with ctrl+c (used when running
1532 them in debugging mode).
1533 - Fix a race condition in live migrating a KVM instance, when stat()
1534 on the old proc status file returned EINVAL, which is an unexpected
1536 - Fixed manpage checking with newer man and utf-8 charachters. But now
1537 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1543 *(Released Fri, 7 May 2010)*
1545 Fix a bug which prevented untagged KVM instances from starting.
1551 *(Released Fri, 7 May 2010)*
1553 Another release with a long development cycle, during which many
1554 different features were added.
1556 Significant features
1557 ~~~~~~~~~~~~~~~~~~~~
1559 The KVM hypervisor now can run the individual instances as non-root, to
1560 reduce the impact of a VM being hijacked due to bugs in the
1561 hypervisor. It is possible to run all instances as a single (non-root)
1562 user, to manually specify a user for each instance, or to dynamically
1563 allocate a user out of a cluster-wide pool to each instance, with the
1564 guarantee that no two instances will run under the same user ID on any
1567 An experimental RAPI client library, that can be used standalone
1568 (without the other Ganeti libraries), is provided in the source tree as
1569 ``lib/rapi/client.py``. Note this client might change its interface in
1570 the future, as we iterate on its capabilities.
1572 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1573 replace the cluster's certificates and crypto keys. This might help in
1574 case they have been compromised, or have simply expired.
1576 A new disk option for instance creation has been added that allows one
1577 to "adopt" currently existing logical volumes, with data
1578 preservation. This should allow easier migration to Ganeti from
1579 unmanaged (or managed via other software) instances.
1581 Another disk improvement is the possibility to convert between redundant
1582 (DRBD) and plain (LVM) disk configuration for an instance. This should
1583 allow better scalability (starting with one node and growing the
1584 cluster, or shrinking a two-node cluster to one node).
1586 A new feature that could help with automated node failovers has been
1587 implemented: if a node sees itself as offline (by querying the master
1588 candidates), it will try to shutdown (hard) all instances and any active
1589 DRBD devices. This reduces the risk of duplicate instances if an
1590 external script automatically failovers the instances on such nodes. To
1591 enable this, the cluster parameter ``maintain_node_health`` should be
1592 enabled; in the future this option (per the name) will enable other
1593 automatic maintenance features.
1595 Instance export/import now will reuse the original instance
1596 specifications for all parameters; that means exporting an instance,
1597 deleting it and the importing it back should give an almost identical
1598 instance. Note that the default import behaviour has changed from
1599 before, where it created only one NIC; now it recreates the original
1602 Cluster verify has added a few new checks: SSL certificates validity,
1603 /etc/hosts consistency across the cluster, etc.
1608 As usual, many internal changes were done, documentation fixes,
1611 - Fixed cluster initialization with disabled cluster storage (regression
1612 introduced in 2.1.1)
1613 - File-based storage supports growing the disks
1614 - Fixed behaviour of node role changes
1615 - Fixed cluster verify for some corner cases, plus a general rewrite of
1616 cluster verify to allow future extension with more checks
1617 - Fixed log spamming by watcher and node daemon (regression introduced
1619 - Fixed possible validation issues when changing the list of enabled
1621 - Fixed cleanup of /etc/hosts during node removal
1622 - Fixed RAPI response for invalid methods
1623 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1624 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1625 from ide disks, etc.)
1626 - Allow OS changes without re-installation (to record a changed OS
1627 outside of Ganeti, or to allow OS renames)
1628 - Allow instance creation without OS installation (useful for example if
1629 the OS will be installed manually, or restored from a backup not in
1631 - Implemented option to make cluster ``copyfile`` use the replication
1633 - Added list of enabled hypervisors to ssconf (possibly useful for
1635 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1637 - A partial form of node re-IP is possible via node readd, which now
1638 allows changed node primary IP
1639 - Command line utilities now show an informational message if the job is
1641 - The logs of the master daemon now show the PID/UID/GID of the
1648 *(Released Fri, 12 Mar 2010)*
1650 During the 2.1.0 long release candidate cycle, a lot of improvements and
1651 changes have accumulated with were released later as 2.1.1.
1656 The node evacuate command (``gnt-node evacuate``) was significantly
1657 rewritten, and as such the IAllocator protocol was changed - a new
1658 request type has been added. This unfortunate change during a stable
1659 series is designed to improve performance of node evacuations; on
1660 clusters with more than about five nodes and which are well-balanced,
1661 evacuation should proceed in parallel for all instances of the node
1662 being evacuated. As such, any existing IAllocator scripts need to be
1663 updated, otherwise the above command will fail due to the unknown
1664 request. The provided "dumb" allocator has not been updated; but the
1665 ganeti-htools package supports the new protocol since version 0.2.4.
1667 Another important change is increased validation of node and instance
1668 names. This might create problems in special cases, if invalid host
1669 names are being used.
1671 Also, a new layer of hypervisor parameters has been added, that sits at
1672 OS level between the cluster defaults and the instance ones. This allows
1673 customisation of virtualization parameters depending on the installed
1674 OS. For example instances with OS 'X' may have a different KVM kernel
1675 (or any other parameter) than the cluster defaults. This is intended to
1676 help managing a multiple OSes on the same cluster, without manual
1677 modification of each instance's parameters.
1679 A tool for merging clusters, ``cluster-merge``, has been added in the
1680 tools sub-directory.
1685 - Improved the int/float conversions that should make the code more
1686 robust in face of errors from the node daemons
1687 - Fixed the remove node code in case of internal configuration errors
1688 - Fixed the node daemon behaviour in face of inconsistent queue
1689 directory (e.g. read-only file-system where we can't open the files
1691 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1692 now it either aborts cleanly or, if given the new "auto_promote"
1693 parameter, will automatically promote other nodes as needed
1694 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1695 completely prevent Ganeti from working
1696 - Fixed bug for instance export when not all disks were successfully
1698 - Fixed behaviour of node add when the new node is slow in starting up
1700 - Fixed handling of signals in the LUXI client, which should improve
1701 behaviour of command-line scripts
1702 - Added checks for invalid node/instance names in the configuration (now
1703 flagged during cluster verify)
1704 - Fixed watcher behaviour for disk activation errors
1705 - Fixed two potentially endless loops in http library, which led to the
1706 RAPI daemon hanging and consuming 100% CPU in some cases
1707 - Fixed bug in RAPI daemon related to hashed passwords
1708 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1710 - Enhanced compatibility with non-Debian OSes, but not using absolute
1711 path in some commands and allowing customisation of the ssh
1712 configuration directory
1713 - Fixed possible future issue with new Python versions by abiding to the
1714 proper use of ``__slots__`` attribute on classes
1715 - Added checks that should prevent directory traversal attacks
1716 - Many documentation fixes based on feedback from users
1721 - Added an "early_release" more for instance replace disks and node
1722 evacuate, where we release locks earlier and thus allow higher
1723 parallelism within the cluster
1724 - Added watcher hooks, intended to allow the watcher to restart other
1725 daemons (e.g. from the ganeti-nbma project), but they can be used of
1726 course for any other purpose
1727 - Added a compile-time disable for DRBD barriers, to increase
1728 performance if the administrator trusts the power supply or the
1729 storage system to not lose writes
1730 - Added the option of using syslog for logging instead of, or in
1731 addition to, Ganeti's own log files
1732 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1733 can indeed boot from a paravirtual NIC
1734 - Added a generic debug level for many operations; while this is not
1735 used widely yet, it allows one to pass the debug value all the way to
1737 - Enhanced the hooks environment for instance moves (failovers,
1738 migrations) where the primary/secondary nodes changed during the
1739 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1740 - Enhanced data validations for many user-supplied values; one important
1741 item is the restrictions imposed on instance and node names, which
1742 might reject some (invalid) host names
1743 - Add a configure-time option to disable file-based storage, if it's not
1744 needed; this allows greater security separation between the master
1745 node and the other nodes from the point of view of the inter-node RPC
1747 - Added user notification in interactive tools if job is waiting in the
1748 job queue or trying to acquire locks
1749 - Added log messages when a job is waiting for locks
1750 - Added filtering by node tags in instance operations which admit
1751 multiple instances (start, stop, reboot, reinstall)
1752 - Added a new tool for cluster mergers, ``cluster-merge``
1753 - Parameters from command line which are of the form ``a=b,c=d`` can now
1754 use backslash escapes to pass in values which contain commas,
1755 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1757 - For KVM, the instance name is the first parameter passed to KVM, so
1758 that it's more visible in the process list
1764 *(Released Tue, 2 Mar 2010)*
1766 Ganeti 2.1 brings many improvements with it. Major changes:
1768 - Added infrastructure to ease automated disk repairs
1769 - Added new daemon to export configuration data in a cheaper way than
1770 using the remote API
1771 - Instance NICs can now be routed instead of being associated with a
1773 - Improved job locking logic to reduce impact of jobs acquiring multiple
1774 locks waiting for other long-running jobs
1776 In-depth implementation details can be found in the Ganeti 2.1 design
1782 - Added chroot hypervisor
1783 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1785 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1786 ``bootloader_path`` and ``bootloader_args``)
1787 - Added the ``use_localtime`` option for the xen-hvm and kvm
1788 hypervisors, and the default value for this has changed to false (in
1789 2.0 xen-hvm always enabled it)
1790 - Added luxi call to submit multiple jobs in one go
1791 - Added cluster initialization option to not modify ``/etc/hosts``
1793 - Added network interface parameters
1794 - Added dry run mode to some LUs
1795 - Added RAPI resources:
1797 - ``/2/instances/[instance_name]/info``
1798 - ``/2/instances/[instance_name]/replace-disks``
1799 - ``/2/nodes/[node_name]/evacuate``
1800 - ``/2/nodes/[node_name]/migrate``
1801 - ``/2/nodes/[node_name]/role``
1802 - ``/2/nodes/[node_name]/storage``
1803 - ``/2/nodes/[node_name]/storage/modify``
1804 - ``/2/nodes/[node_name]/storage/repair``
1806 - Added OpCodes to evacuate or migrate all instances on a node
1807 - Added new command to list storage elements on nodes (``gnt-node
1808 list-storage``) and modify them (``gnt-node modify-storage``)
1809 - Added new ssconf files with master candidate IP address
1810 (``ssconf_master_candidates_ips``), node primary IP address
1811 (``ssconf_node_primary_ips``) and node secondary IP address
1812 (``ssconf_node_secondary_ips``)
1813 - Added ``ganeti-confd`` and a client library to query the Ganeti
1814 configuration via UDP
1815 - Added ability to run hooks after cluster initialization and before
1817 - Added automatic mode for disk replace (``gnt-instance replace-disks
1819 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1820 after catastrophic data-loss
1821 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1823 - Added ``gnt-instance move`` command to move instances
1824 - Added ``gnt-cluster watcher`` command to control watcher
1825 - Added ``gnt-node powercycle`` command to powercycle nodes
1826 - Added new job status field ``lock_status``
1827 - Added parseable error codes to cluster verification (``gnt-cluster
1828 verify --error-codes``) and made output less verbose (use
1829 ``--verbose`` to restore previous behaviour)
1830 - Added UUIDs to the main config entities (cluster, nodes, instances)
1831 - Added support for OS variants
1832 - Added support for hashed passwords in the Ganeti remote API users file
1834 - Added option to specify maximum timeout on instance shutdown
1835 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1836 - Added new helper script to start and stop Ganeti daemons
1837 (``daemon-util``), with the intent to reduce the work necessary to
1838 adjust Ganeti for non-Debian distributions and to start/stop daemons
1840 - Added more unittests
1841 - Fixed critical bug in ganeti-masterd startup
1842 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1843 now customisable at the cluster level for both the KVM and Xen
1844 hypervisors using the new ``migration_port`` parameter
1845 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1846 reinstalling an instance
1847 - Allowed ``@`` in tag names
1848 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1849 - Many documentation updates
1850 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1851 - ``gnt-instance reinstall`` can now reinstall multiple instances
1852 - Updated many command line parameters
1853 - Introduced new OS API version 15
1854 - No longer support a default hypervisor
1855 - Treat virtual LVs as inexistent
1856 - Improved job locking logic to reduce lock contention
1857 - Match instance and node names case insensitively
1858 - Reimplemented bash completion script to be more complete
1865 *(Released Thu, 4 Feb 2010)*
1867 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1868 - Fix a string formatting bug
1869 - Improve safety of the code in some error paths
1870 - Improve data validation in the master of values returned from nodes
1876 *(Released Thu, 17 Dec 2009)*
1878 - Fix security issue due to missing validation of iallocator names; this
1879 allows local and remote execution of arbitrary executables
1880 - Fix failure of gnt-node list during instance removal
1881 - Ship the RAPI documentation in the archive
1887 *(Released Wed, 30 Sep 2009)*
1889 - Fixed many wrong messages
1890 - Fixed a few bugs related to the locking library
1891 - Fixed MAC checking at instance creation time
1892 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1893 - Fixed a few issues related to signal handling in both daemons and
1895 - Fixed the example startup script provided
1896 - Fixed insserv dependencies in the example startup script (patch from
1898 - Fixed handling of drained nodes in the iallocator framework
1899 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1901 - Fixed error related to invalid job IDs in job polling
1902 - Fixed job/opcode persistence on unclean master shutdown
1903 - Fixed handling of partial job processing after unclean master
1905 - Fixed error reporting from LUs, previously all errors were converted
1906 into execution errors
1907 - Fixed error reporting from burnin
1908 - Decreased significantly the memory usage of the job queue
1909 - Optimised slightly multi-job submission
1910 - Optimised slightly opcode loading
1911 - Backported the multi-job submit framework from the development
1912 branch; multi-instance start and stop should be faster
1913 - Added script to clean archived jobs after 21 days; this will reduce
1914 the size of the queue directory
1915 - Added some extra checks in disk size tracking
1916 - Added an example ethers hook script
1917 - Added a cluster parameter that prevents Ganeti from modifying of
1919 - Added more node information to RAPI responses
1920 - Added a ``gnt-job watch`` command that allows following the ouput of a
1922 - Added a bind-address option to ganeti-rapi
1923 - Added more checks to the configuration verify
1924 - Enhanced the burnin script such that some operations can be retried
1926 - Converted instance reinstall to multi-instance model
1932 *(Released Fri, 7 Aug 2009)*
1934 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1935 to allow using the pre-2.0.2 behaviour in activation, if any existing
1936 instances have mismatched disk sizes in the configuration
1937 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
1938 any configuration mismatches for disk sizes
1939 - Added ``gnt-master cluste-failover --no-voting`` to allow master
1940 failover to work on two-node clusters
1941 - Fixed the ``--net`` option of ``gnt-backup import``, which was
1943 - Fixed detection of OS script errors in ``gnt-backup export``
1944 - Fixed exit code of ``gnt-backup export``
1950 *(Released Fri, 17 Jul 2009)*
1952 - Added experimental support for stripped logical volumes; this should
1953 enhance performance but comes with a higher complexity in the block
1954 device handling; stripping is only enabled when passing
1955 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
1956 affected even in the non-stripped mode
1957 - Improved resiliency against transient failures at the end of DRBD
1958 resyncs, and in general of DRBD resync checks
1959 - Fixed a couple of issues with exports and snapshot errors
1960 - Fixed a couple of issues in instance listing
1961 - Added display of the disk size in ``gnt-instance info``
1962 - Fixed checking for valid OSes in instance creation
1963 - Fixed handling of the "vcpus" parameter in instance listing and in
1964 general of invalid parameters
1965 - Fixed http server library, and thus RAPI, to handle invalid
1966 username/password combinations correctly; this means that now they
1967 report unauthorized for queries too, not only for modifications,
1968 allowing earlier detect of configuration problems
1969 - Added a new "role" node list field, equivalent to the master/master
1970 candidate/drained/offline flags combinations
1971 - Fixed cluster modify and changes of candidate pool size
1972 - Fixed cluster verify error messages for wrong files on regular nodes
1973 - Fixed a couple of issues with node demotion from master candidate role
1974 - Fixed node readd issues
1975 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
1976 - Added a new ``--no-voting`` option for masterfailover to fix failover
1977 on two-nodes clusters when the former master node is unreachable
1978 - Added instance reinstall over RAPI
1984 *(Released Tue, 16 Jun 2009)*
1986 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
1987 allow re-adding the start in single-user option (regression from 1.2)
1988 - the watcher writes the instance status to a file, to allow monitoring
1989 to report the instance status (from the master) based on cached
1990 results of the watcher's queries; while this can get stale if the
1991 watcher is being locked due to other work on the cluster, this is
1992 still an improvement
1993 - the watcher now also restarts the node daemon and the rapi daemon if
1995 - fixed the watcher to handle full and drained queue cases
1996 - hooks export more instance data in the environment, which helps if
1997 hook scripts need to take action based on the instance's properties
1998 (no longer need to query back into ganeti)
1999 - instance failovers when the instance is stopped do not check for free
2000 RAM, so that failing over a stopped instance is possible in low memory
2002 - rapi uses queries for tags instead of jobs (for less job traffic), and
2003 for cluster tags it won't talk to masterd at all but read them from
2005 - a couple of error handling fixes in RAPI
2006 - drbd handling: improved the error handling of inconsistent disks after
2007 resync to reduce the frequency of "there are some degraded disks for
2008 this instance" messages
2009 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2010 error handling path called a wrong function name)
2016 *(Released Wed, 27 May 2009)*
2018 - no changes from rc5
2024 *(Released Wed, 20 May 2009)*
2026 - fix a couple of bugs (validation, argument checks)
2027 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2028 - some small improvements to RAPI and IAllocator
2029 - make watcher automatically start the master daemon if down
2035 *(Released Mon, 27 Apr 2009)*
2037 - change the OS list to not require locks; this helps with big clusters
2038 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2039 volume group is broken
2040 - ``gnt-instance info``, without any arguments, doesn't run for all
2041 instances anymore; either pass ``--all`` or pass the desired
2042 instances; this helps against mistakes on big clusters where listing
2043 the information for all instances takes a long time
2044 - miscellaneous doc and man pages fixes
2050 *(Released Wed, 8 Apr 2009)*
2052 - Change the internal locking model of some ``gnt-node`` commands, in
2053 order to reduce contention (and blocking of master daemon) when
2054 batching many creation/reinstall jobs
2055 - Fixes to Xen soft reboot
2056 - No longer build documentation at build time, instead distribute it in
2057 the archive, in order to reduce the need for the whole docbook/rst
2064 *(Released Fri, 27 Mar 2009)*
2066 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2067 - Fix watcher startup sequence, improves the behaviour of busy clusters
2068 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2069 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2070 bind address and other places
2071 - Some documentation fixes and updates
2077 *(Released Mon, 2 Mar 2009)*
2079 - More documentation updates, now all docs should be more-or-less
2081 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2083 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2085 - Fix an issue related to $libdir/run/ganeti and cluster creation
2091 *(Released Thu, 19 Feb 2009)*
2093 - Xen PVM and KVM have switched the default value for the instance root
2094 disk to the first partition on the first drive, instead of the whole
2095 drive; this means that the OS installation scripts must be changed
2097 - Man pages have been updated
2098 - RAPI has been switched by default to HTTPS, and the exported functions
2099 should all work correctly
2100 - RAPI v1 has been removed
2101 - Many improvements to the KVM hypervisor
2102 - Block device errors are now better reported
2103 - Many other bugfixes and small improvements
2109 *(Released Mon, 26 Jan 2009)*
2111 - Version 2 is a general rewrite of the code and therefore the
2112 differences are too many to list, see the design document for 2.0 in
2113 the ``doc/`` subdirectory for more details
2114 - In this beta version there is not yet a migration path from 1.2 (there
2115 will be one in the final 2.0 release)
2116 - A few significant changes are:
2118 - all commands are executed by a daemon (``ganeti-masterd``) and the
2119 various ``gnt-*`` commands are just front-ends to it
2120 - all the commands are entered into, and executed from a job queue,
2121 see the ``gnt-job(8)`` manpage
2122 - the RAPI daemon supports read-write operations, secured by basic
2123 HTTP authentication on top of HTTPS
2124 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2125 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2126 to migrate to DRBD 8 first while still running Ganeti 1.2)
2127 - DRBD devices are using statically allocated minor numbers, which
2128 will be assigned to existing instances during the migration process
2129 - there is support for both Xen PVM and Xen HVM instances running on
2131 - KVM virtualization is supported too
2132 - file-based storage has been implemented, which means that it is
2133 possible to run the cluster without LVM and DRBD storage, for
2134 example using a shared filesystem exported from shared storage (and
2135 still have live migration)
2141 *(Released Tue, 13 Jan 2009)*
2143 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2144 - Reuse the old instance mac address by default on instance import, if
2145 the instance name is the same.
2146 - Handle situations in which the node info rpc returns incomplete
2148 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2149 - Improved version of batcher:
2151 - state file support
2152 - instance mac address support
2153 - support for HVM clusters/instances
2155 - Add an option to show the number of cpu sockets and nodes in
2157 - Support OSes that handle more than one version of the OS api (but do
2158 not change the current API in any other way)
2159 - Fix ``gnt-node migrate``
2160 - ``gnt-debug`` man page
2161 - Fixes various more typos and small issues
2162 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2168 *(Released Wed, 24 Sep 2008)*
2170 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2171 type of disk exported to fully virtualized instances.
2172 - provide access to the serial console of HVM instances
2173 - instance auto_balance flag, set by default. If turned off it will
2174 avoid warnings on cluster verify if there is not enough memory to fail
2175 over an instance. in the future it will prevent automatically failing
2176 it over when we will support that.
2177 - batcher tool for instance creation, see ``tools/README.batcher``
2178 - ``gnt-instance reinstall --select-os`` to interactively select a new
2179 operating system when reinstalling an instance.
2180 - when changing the memory amount on instance modify a check has been
2181 added that the instance will be able to start. also warnings are
2182 emitted if the instance will not be able to fail over, if auto_balance
2184 - documentation fixes
2185 - sync fields between ``gnt-instance list/modify/add/import``
2186 - fix a race condition in drbd when the sync speed was set after giving
2187 the device a remote peer.
2193 *(Released Tue, 22 Jul 2008)*
2195 - note: the allowed size and number of tags per object were reduced
2196 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2197 - fixed twisted 8.x compatibility
2198 - fixed ``gnt-instance replace-disks`` with iallocator
2199 - add TCP keepalives on twisted connections to detect restarted nodes
2200 - disk increase support, see ``gnt-instance grow-disk``
2201 - implement bulk node/instance query for RAPI
2202 - add tags in node/instance listing (optional)
2203 - experimental migration (and live migration) support, read the man page
2204 for ``gnt-instance migrate``
2205 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2206 has some small improvements in handling its state file
2212 *(Released Fri, 13 Jun 2008)*
2214 - Experimental readonly, REST-based remote API implementation;
2215 automatically started on master node, TCP port 5080, if enabled by
2216 ``--enable-rapi`` parameter to configure script.
2217 - Instance allocator support. Add and import instance accept a
2218 ``--iallocator`` parameter, and call that instance allocator to decide
2219 which node to use for the instance. The iallocator document describes
2220 what's expected from an allocator script.
2221 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2222 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2223 node is lost there is still enough memory to fail over the instances
2225 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2226 ``gnt-cluster verify``, to check for site-specific compliance. All the
2227 hooks will run, and their output, if any, will be displayed. Any
2228 failing hook will make the verification return an error value.
2229 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2230 primary and secondary interfaces
2231 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2232 that is still declared as part of the cluster and has failed.
2233 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2234 specifying output fields, that just adds the chosen fields to the
2236 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2237 export from the filesystem.
2238 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2239 have been added. Using them you can enable/disable acpi and pae
2240 support, and specify a path for a cd image to be exported to the
2241 instance. These parameters as the name suggest only work on HVM
2243 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2244 PAE support will be set to the previously hardcoded values, but the
2245 (previously hardcoded) path to the CDROM ISO image will be unset and
2246 if required, needs to be set manually with ``gnt-instance modify``
2248 - The address to which an instance's VNC console is bound is now
2249 selectable per-instance, rather than being cluster wide. Of course
2250 this only applies to instances controlled via VNC, so currently just
2251 applies to HVM clusters.
2257 *(Released Mon, 18 Feb 2008)*
2259 - more tweaks to the disk activation code (especially helpful for DRBD)
2260 - change the default ``gnt-instance list`` output format, now there is
2261 one combined status field (see the manpage for the exact values this
2263 - some more fixes for the mac export to hooks change
2264 - make Ganeti not break with DRBD 8.2.x (which changed the version
2265 format in ``/proc/drbd``) (issue 24)
2266 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2267 template, allowing migration from DRBD0.7+MD to DRBD8
2273 *(Released Wed, 30 Jan 2008)*
2275 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2277 - add command aliases infrastructure and a few aliases
2278 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2279 man pages and the ``--help`` option of ``gnt-node
2280 list``/``gnt-instance list``
2281 - fix ``gnt-backup list`` with down nodes (issue 21)
2282 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2283 - fix the dist archive and add a check for including svn/git files in
2285 - some developer-related changes: improve the burnin and the QA suite,
2286 add an upload script for testing during development
2292 *(Released Wed, 16 Jan 2008)*
2294 - experimental HVM support, read the install document, section
2295 "Initializing the cluster"
2296 - allow for the PVM hypervisor per-instance kernel and initrd paths
2297 - add a new command ``gnt-cluster verify-disks`` which uses a new
2298 algorithm to improve the reconnection of the DRBD pairs if the device
2299 on the secondary node has gone away
2300 - make logical volume code auto-activate LVs at disk activation time
2301 - slightly improve the speed of activating disks
2302 - allow specification of the MAC address at instance creation time, and
2303 changing it later via ``gnt-instance modify``
2304 - fix handling of external commands that generate lots of output on
2306 - update documentation with regard to minimum version of DRBD8 supported
2312 *(Released Tue, 4 Dec 2007)*
2314 - Log the ``xm create`` output to the node daemon log on failure (to
2315 help diagnosing the error)
2316 - In debug mode, log all external commands output if failed to the logs
2317 - Change parsing of lvm commands to ignore stderr
2323 *(Released Wed, 28 Nov 2007)*
2325 - Another round of updates to the DRBD 8 code to deal with more failures
2326 in the replace secondary node operation
2327 - Some more logging of failures in disk operations (lvm, drbd)
2328 - A few documentation updates
2335 *(Released Tue, 13 Nov 2007)*
2337 - Change configuration file format from Python's Pickle to JSON.
2338 Upgrading is possible using the cfgupgrade utility.
2339 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2340 faster replace disks and is more stable (DRBD 8 has many improvements
2341 compared to DRBD 0.7)
2342 - Added command line tags support (see man pages for ``gnt-instance``,
2343 ``gnt-node``, ``gnt-cluster``)
2344 - Added instance rename support
2345 - Added multi-instance startup/shutdown
2346 - Added cluster rename support
2347 - Added ``gnt-node evacuate`` to simplify some node operations
2348 - Added instance reboot operation that can speedup reboot as compared to
2350 - Soften the requirement that hostnames are in FQDN format
2351 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2353 - Removed dependency on debian's patched fping that uses the
2354 non-standard ``-S`` option
2355 - Now the OS definitions are searched for in multiple, configurable
2356 paths (easier for distros to package)
2357 - Some changes to the hooks infrastructure (especially the new
2358 post-configuration update hook)
2359 - Other small bugfixes
2361 .. vim: set textwidth=72 syntax=rst :