10 - ``gnt-instance batch-create`` has been changed to use the bulk create
11 opcode from Ganeti. This lead to incompatible changes in the format of
12 the JSON file. It's now not a custom dict anymore but a dict
13 compatible with the ``OpInstanceCreate`` opcode.
19 *(Released Fri, 27 Jul 2012)*
22 .. attention:: The ``LUXI`` protocol has been made more consistent
23 regarding its handling of command arguments. This, however, leads to
24 incompatibility issues with previous versions. Please ensure that you
25 restart Ganeti daemons soon after the upgrade, otherwise most
26 ``LUXI`` calls (job submission, setting/resetting the drain flag,
27 pausing/resuming the watcher, cancelling and archiving jobs, querying
28 the cluster configuration) will fail.
37 The current ``admin_up`` field, which used to denote whether an instance
38 should be running or not, has been removed. Instead, ``admin_state`` is
39 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
41 The rational behind this is that an instance being “down” can have
44 - it could be down during a reboot
45 - it could be temporarily be down for a reinstall
46 - or it could be down because it is deprecated and kept just for its
49 The previous Boolean state was making it difficult to do capacity
50 calculations: should Ganeti reserve memory for a down instance? Now, the
51 tri-state field makes it clear:
53 - in ``up`` and ``down`` state, all resources are reserved for the
54 instance, and it can be at any time brought up if it is down
55 - in ``offline`` state, only disk space is reserved for it, but not
58 The field can have an extra use: since the transition between ``up`` and
59 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
60 transition between ``offline`` and ``down`` is done via ``gnt-instance
61 modify``, it is possible to given different rights to users. For
62 example, owners of an instance could be allowed to start/stop it, but
63 not transition it out of the offline state.
65 Instance policies and specs
66 +++++++++++++++++++++++++++
68 In previous Ganeti versions, an instance creation request was not
69 limited on the minimum size and on the maximum size just by the cluster
70 resources. As such, any policy could be implemented only in third-party
71 clients (RAPI clients, or shell wrappers over ``gnt-*``
72 tools). Furthermore, calculating cluster capacity via ``hspace`` again
73 required external input with regards to instance sizes.
75 In order to improve these workflows and to allow for example better
76 per-node group differentiation, we introduced instance specs, which
79 - minimum instance disk size, disk count, memory size, cpu count
80 - maximum values for the above metrics
81 - and “standard” values (used in ``hspace`` to calculate the standard
84 The minimum/maximum values can be also customised at node-group level,
85 for example allowing more powerful hardware to support bigger instance
88 Beside the instance specs, there are a few other settings belonging to
89 the instance policy framework. It is possible now to customise, per
90 cluster and node-group:
92 - the list of allowed disk templates
93 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
94 - the maximum ratio of instance to spindles (see below for more
95 information) for local storage
97 All these together should allow all tools that talk to Ganeti to know
98 what are the ranges of allowed values for instances and the
99 over-subscription that is allowed.
101 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
102 instance configuration, and the physical CPU configuration from the
103 node. For the spindle ratios however, we didn't track before these
104 values, so new parameters have been added:
106 - a new node parameter ``spindle_count``, defaults to 1, customisable at
107 node group or node level
108 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
110 Note that spindles in this context doesn't need to mean actual
111 mechanical hard-drives; it's just a relative number for both the node
112 I/O capacity and instance I/O consumption.
114 Instance migration behaviour
115 ++++++++++++++++++++++++++++
117 While live-migration is in general desirable over failover, it is
118 possible that for some workloads it is actually worse, due to the
119 variable time of the “suspend” phase during live migration.
121 To allow the tools to work consistently over such instances (without
122 having to hard-code instance names), a new backend parameter
123 ``always_failover`` has been added to control the migration/failover
124 behaviour. When set to True, all migration requests for an instance will
125 instead fall-back to failover.
127 Instance memory ballooning
128 ++++++++++++++++++++++++++
130 Initial support for memory ballooning has been added. The memory for an
131 instance is no longer fixed (backend parameter ``memory``), but instead
132 can vary between minimum and maximum values (backend parameters
133 ``minmem`` and ``maxmem``). Currently we only change an instance's
136 - live migrating or failing over and instance and the target node
137 doesn't have enough memory
138 - user requests changing the memory via ``gnt-instance modify
144 In order to control the use of specific CPUs by instance, support for
145 controlling CPU pinning has been added for the Xen, HVM and LXC
146 hypervisors. This is controlled by a new hypervisor parameter
147 ``cpu_mask``; details about possible values for this are in the
148 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
149 VCPU-to-CPU mapping) form will work well only when all nodes in your
150 cluster have the same amount of CPUs.
155 Another area in which Ganeti was not customisable were the parameters
156 used for storage configuration, e.g. how many stripes to use for LVM,
157 DRBD resync configuration, etc.
159 To improve this area, we've added disks parameters, which are
160 customisable at cluster and node group level, and which allow to
161 specify various parameters for disks (DRBD has the most parameters
162 currently), for example:
164 - DRBD resync algorithm and parameters (e.g. speed)
165 - the default VG for meta-data volumes for DRBD
166 - number of stripes for LVM (plain disk template)
169 These parameters can be modified via ``gnt-cluster modify -D …`` and
170 ``gnt-group modify -D …``, and are used at either instance creation (in
171 case of LVM stripes, for example) or at disk “activation” time
174 Rados block device support
175 ++++++++++++++++++++++++++
177 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
178 denoted by the ``rbd`` disk template type. This is considered
179 experimental, feedback is welcome. For details on configuring it, see
180 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
185 The existing master IP functionality works well only in simple setups (a
186 single network shared by all nodes); however, if nodes belong to
187 different networks, then the ``/32`` setup and lack of routing
188 information is not enough.
190 To allow the master IP to function well in more complex cases, the
191 system was reworked as follows:
193 - a master IP netmask setting has been added
194 - the master IP activation/turn-down code was moved from the node daemon
196 - whether to run the Ganeti-supplied master IP script or a user-supplied
197 on is a ``gnt-cluster init`` setting
199 Details about the location of the standard and custom setup scripts are
200 in the man page :manpage:`gnt-cluster(8)`; for information about the
201 setup script protocol, look at the Ganeti-supplied script.
206 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
209 It is now possible to use TLS-protected connections, and when renewing
210 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
211 it is now possible to specify spice or spice CA certificates. Also, it
212 is possible to configure a password for SPICE sessions via the
213 hypervisor parameter ``spice_password_file``.
215 There are also new parameters to control the compression and streaming
216 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
217 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
218 for the spice parameters.
220 Lastly, it is now possible to see the SPICE connection information via
221 ``gnt-instance console``.
226 A new tool (``tools/ovfconverter``) has been added that supports
227 conversion between Ganeti and the `Open Virtualization Format
228 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
231 This relies on the ``qemu-img`` tool to convert the disk formats, so the
232 actual compatibility with other virtualization solutions depends on it.
237 The configuration query daemon (``ganeti-confd``) is now optional, and
238 has been rewritten in Haskell; whether to use the daemon at all, use the
239 Python (default) or the Haskell version is selectable at configure time
240 via the ``--enable-confd`` parameter, which can take one of the
241 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
242 daemon will result in a smaller footprint; for larger systems, we
243 welcome feedback on the Haskell version which might become the default
246 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
247 daemon running. The Python version doesn't implement the new call.
250 User interface changes
251 ~~~~~~~~~~~~~~~~~~~~~~
253 We have replaced the ``--disks`` option of ``gnt-instance
254 replace-disks`` with a more flexible ``--disk`` option, which allows
255 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
256 disk size and mode can be changed upon recreation (via ``gnt-instance
257 recreate-disks``, which accepts the same ``--disk`` option).
259 As many people are used to a ``show`` command, we have added that as an
260 alias to ``info`` on all ``gnt-*`` commands.
262 The ``gnt-instance grow-disk`` command has a new mode in which it can
263 accept the target size of the disk, instead of the delta; this can be
264 more safe since two runs in absolute mode will be idempotent, and
265 sometimes it's also easier to specify the desired size directly.
267 Also the handling of instances with regard to offline secondaries has
268 been improved. Instance operations should not fail because one of it's
269 secondary nodes is offline, even though it's safe to proceed.
271 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
272 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
273 minors to instance name.
278 RAPI coverage has improved, with (for example) new resources for
279 recreate-disks, node power-cycle, etc.
284 There is partial support for ``xl`` in the Xen hypervisor; feedback is
287 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
288 whether to still support Python 2.4 or move to Python 2.6 as minimum
291 Support for Fedora has been slightly improved; the provided example
292 init.d script should work better on it and the INSTALL file should
293 document the needed dependencies.
298 The deprecated ``QueryLocks`` LUXI request has been removed. Use
299 ``Query(what=QR_LOCK, ...)`` instead.
301 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
302 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
303 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
304 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
305 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
307 RAPI client: ``CertificateError`` now derives from
308 ``GanetiApiError``. This should make it more easy to handle Ganeti
311 Deprecation warnings due to PyCrypto/paramiko import in
312 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
313 make sure to run an up-to-date paramiko version, if you use this tool.
315 The QA scripts now depend on Python 2.5 or above (the main code base
316 still works with Python 2.4).
318 The configuration file (``config.data``) is now written without
319 indentation for performance reasons; if you want to edit it, it can be
320 re-formatted via ``tools/fmtjson``.
322 A number of bugs has been fixed in the cluster merge tool.
324 ``x509`` certification verification (used in import-export) has been
325 changed to allow the same clock skew as permitted by the cluster
326 verification. This will remove some rare but hard to diagnose errors in
333 *(Released Thu, 19 Jul 2012)*
335 Very few changes from rc4 to the final release, only bugfixes:
337 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
338 instance, fix CDROM booting for KVM instances)
339 - fixed node group modification of node parameters
340 - fixed issue in LUClusterVerifyGroup with multi-group clusters
341 - fixed generation of bash completion to ensure a stable ordering
348 *(Released Fri, 13 Jul 2012)*
350 Third release candidate for 2.6. The following changes were done from
353 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
354 - Fixed an inconsistency in the LUXI protocol with the provided
355 arguments (NOT backwards compatible)
356 - Fixed a bug with node groups ipolicy where ``min`` was greater than
357 the cluster ``std`` value
358 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
359 easier instance debugging on nodes (requires ``hconfd`` to work)
365 *(Released Tue, 03 Jul 2012)*
367 Second release candidate for 2.6. The following changes were done from
370 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
372 - Fixed a RAPI regression on missing beparams/memory
373 - Fixed redistribution of files on offline nodes
374 - Added possibility to run activate-disks even though secondaries are
375 offline. With this change it relaxes also the strictness on some other
376 commands which use activate disks internally:
377 * ``gnt-instance start|reboot|rename|backup|export``
378 - Made it possible to remove safely an instance if its secondaries are
380 - Made it possible to reinstall even though secondaries are offline
386 *(Released Mon, 25 Jun 2012)*
388 First release candidate for 2.6. The following changes were done from
391 - Fixed bugs with disk parameters and ``rbd`` templates as well as
393 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
394 behaviour. It supports now the modify operation
395 - ``hcheck`` implemented to analyze cluster health and possibility of
396 improving health by rebalance
397 - ``hbal`` has been improved in dealing with split instances
403 *(Released Mon, 11 Jun 2012)*
405 Second beta release of 2.6. The following changes were done from beta2
408 - Fixed ``daemon-util`` with non-root user models
409 - Fixed creation of plain instances with ``--no-wait-for-sync``
410 - Fix wrong iv_names when running ``cfgupgrade``
411 - Export more information in RAPI group queries
412 - Fixed bug when changing instance network interfaces
413 - Extended burnin to do NIC changes
414 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
415 - Changed default for DRBD barriers
416 - Fixed DRBD error reporting for syncer rate
417 - Verify the options on disk parameters
419 And of course various fixes to documentation and improved unittests and
426 *(Released Wed, 23 May 2012)*
428 First beta release of 2.6. The following changes were done from beta1 to
431 - integrated patch for distributions without ``start-stop-daemon``
432 - adapted example init.d script to work on Fedora
433 - fixed log handling in Haskell daemons
434 - adapted checks in the watcher for pycurl linked against libnss
435 - add partial support for ``xl`` instead of ``xm`` for Xen
436 - fixed a type issue in cluster verification
437 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
440 Plus integrated fixes from the 2.5 branch:
442 - fixed ``kvm-ifup`` to use ``/bin/bash``
443 - fixed parallel build failures
444 - KVM live migration when using a custom keymap
450 *(Released Tue, 24 Jul 2012)*
452 A small bugfix release, with no new features:
454 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
455 different default shell (e.g. Debian, Ubuntu)
456 - fixed KVM startup and live migration with a custom keymap (fixes Issue
457 243 and Debian bug #650664)
458 - fixed compatibility with KVM versions that don't support multiple boot
459 devices (fixes Issue 230 and Debian bug #624256)
461 Additionally, a few fixes were done to the build system (fixed parallel
462 build failures) and to the unittests (fixed race condition in test for
463 FileID functions, and the default enable/disable mode for QA test is now
470 *(Released Fri, 11 May 2012)*
472 A small bugfix release.
474 The main issues solved are on the topic of compatibility with newer LVM
477 - fixed parsing of ``lv_attr`` field
478 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
479 the ``--force`` flag is needed
481 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
482 to accept kernel 3.x too (was hardcoded to 2.6.*).
484 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
485 of hook results and that also made display of other errors suboptimal
486 was fixed; the code behaves now like 2.4 and earlier.
488 Another change in 2.5, the cleanup of the OS scripts environment, is too
489 aggressive: it removed even the ``PATH`` variable, which requires the OS
490 scripts to *always* need to export it. Since this is a bit too strict,
491 we now export a minimal PATH, the same that we export for hooks.
493 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
494 integrated into this release.
496 Finally, a few other miscellaneous changes were done (no new features,
497 just small improvements):
499 - Fix ``gnt-group --help`` display
500 - Fix hardcoded Xen kernel path
501 - Fix grow-disk handling of invalid units
502 - Update synopsis for ``gnt-cluster repair-disk-sizes``
503 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
509 *(Released Thu, 12 Apr 2012)*
511 Incompatible/important changes and bugfixes
512 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
514 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
515 resource's ``ip_check`` parameter changed from ``True`` to ``False``
516 to match the underlying LUXI interface.
517 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
518 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
519 not maintain backwards-compatibility as the underlying operation
520 changed in an incompatible way. The RAPI client can talk to old
521 servers, but it needs to be told so as the return value changed.
522 - When creating file-based instances via RAPI, the ``file_driver``
523 parameter no longer defaults to ``loop`` and must be specified.
524 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
526 - Support for the undocumented and deprecated RAPI instance creation
527 request format version 0 has been dropped. Use version 1, supported
528 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
529 - Pyparsing 1.4.6 or above is required, see :doc:`installation
530 documentation <install>`.
531 - The "cluster-verify" hooks are now executed per group by the
532 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
533 if you just run ``gnt-cluster verify``, which generates one opcode per
535 - The environment as passed to the OS scripts is cleared, and thus no
536 environment variables defined in the node daemon's environment will be
537 inherited by the scripts.
538 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
540 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
541 support operations involving multiple node groups.
542 - Offline nodes are ignored when failing over an instance.
543 - Support for KVM version 1.0, which changed the version reporting format
545 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
547 - ``Makefile`` is now compatible with Automake 1.11.2
548 - Includes all bugfixes made in the 2.4 series
553 - The ganeti-htools project has been merged into the ganeti-core source
554 tree and will be built as part of Ganeti (see :doc:`install-quick`).
555 - Implemented support for :doc:`shared storage <design-shared-storage>`.
556 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
557 GPT-style partition tables (requires `parted
558 <http://www.gnu.org/s/parted/>`_).
559 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
560 - Allowed adding tags on instance creation.
561 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
562 - Allow instances to be started in a paused state, enabling the user to
563 see the complete console output on boot using the console.
564 - Added new hypervisor flag to control default reboot behaviour
565 (``reboot_behavior``).
566 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
567 - Improved out-of-band management support:
569 - Added ``gnt-node health`` command reporting the health status of
571 - Added ``gnt-node power`` command to manage power status of nodes.
572 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
574 - Instance migration can fall back to failover if instance is not
576 - Filters can be used when listing nodes, instances, groups and locks;
577 see *ganeti(7)* manpage.
578 - Added post-execution status as variables to :doc:`hooks <hooks>`
580 - Instance tags are exported/imported together with the instance.
581 - When given an explicit job ID, ``gnt-job info`` will work for archived
583 - Jobs can define dependencies on other jobs (not yet supported via
584 RAPI or command line, but used by internal commands and usable via
587 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
590 - Instance failover is now available as a RAPI resource
591 (``/2/instances/[instance_name]/failover``).
592 - ``gnt-instance info`` defaults to static information if primary node
594 - Opcodes have a new ``comment`` attribute.
595 - Added basic SPICE support to KVM hypervisor.
596 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
598 Node group improvements
599 ~~~~~~~~~~~~~~~~~~~~~~~
601 - ``gnt-cluster verify`` has been modified to check groups separately,
602 thereby improving performance.
603 - Node group support has been added to ``gnt-cluster verify-disks``,
604 which now operates per node group.
605 - Watcher has been changed to work better with node groups.
607 - One process and state file per node group.
608 - Slow watcher in one group doesn't block other group's watcher.
610 - Added new command, ``gnt-group evacuate``, to move all instances in a
611 node group to other groups.
612 - Added ``gnt-instance change-group`` to move an instance to another
614 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
615 per-group operations.
616 - Node groups can be tagged.
617 - Some operations switch from an exclusive to a shared lock as soon as
619 - Instance's primary and secondary nodes' groups are now available as
620 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
621 and ``snodes.group.uuid``).
626 - Numerous updates to documentation and manpages.
628 - :doc:`RAPI <rapi>` documentation now has detailed parameter
630 - Some opcode/job results are now also documented, see :doc:`RAPI
633 - A lockset's internal lock is now also visible in lock monitor.
634 - Log messages from job queue workers now contain information about the
635 opcode they're processing.
636 - ``gnt-instance console`` no longer requires the instance lock.
637 - A short delay when waiting for job changes reduces the number of LUXI
638 requests significantly.
639 - DRBD metadata volumes are overwritten with zeros during disk creation.
640 - Out-of-band commands no longer acquire the cluster lock in exclusive
642 - ``devel/upload`` now uses correct permissions for directories.
648 *(Released Fri, 23 Mar 2012)*
650 This was the sixth release candidate of the 2.5 series.
656 *(Released Mon, 9 Jan 2012)*
658 This was the fifth release candidate of the 2.5 series.
664 *(Released Thu, 27 Oct 2011)*
666 This was the fourth release candidate of the 2.5 series.
672 *(Released Wed, 26 Oct 2011)*
674 This was the third release candidate of the 2.5 series.
680 *(Released Tue, 18 Oct 2011)*
682 This was the second release candidate of the 2.5 series.
688 *(Released Tue, 4 Oct 2011)*
690 This was the first release candidate of the 2.5 series.
696 *(Released Wed, 31 Aug 2011)*
698 This was the third beta release of the 2.5 series.
704 *(Released Mon, 22 Aug 2011)*
706 This was the second beta release of the 2.5 series.
712 *(Released Fri, 12 Aug 2011)*
714 This was the first beta release of the 2.5 series.
720 *(Released Thu, 27 Oct 2011)*
722 - Fixed bug when parsing command line parameter values ending in
724 - Fixed assertion error after unclean master shutdown
725 - Disable HTTP client pool for RPC, significantly reducing memory usage
727 - Fixed queue archive creation with wrong permissions
733 *(Released Tue, 23 Aug 2011)*
737 - Fixed documentation for importing with ``--src-dir`` option
738 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
739 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
745 *(Released Fri, 5 Aug 2011)*
747 Many bug-fixes and a few small features:
749 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
750 issues when you tried to add an instance with two MAC addresses in one
752 - KVM: fixed per-instance stored UID value
753 - KVM: configure bridged NICs at migration start
754 - KVM: Fix a bug where instance will not start with never KVM versions
756 - Added OS search path to ``gnt-cluster info``
757 - Fixed an issue with ``file_storage_dir`` where you were forced to
758 provide an absolute path, but the documentation states it is a
759 relative path, the documentation was right
760 - Added a new parameter to instance stop/start called ``--no-remember``
761 that will make the state change to not be remembered
762 - Implemented ``no_remember`` at RAPI level
763 - Improved the documentation
764 - Node evacuation: don't call IAllocator if node is already empty
765 - Fixed bug in DRBD8 replace disks on current nodes
766 - Fixed bug in recreate-disks for DRBD instances
767 - Moved assertion checking locks in ``gnt-instance replace-disks``
768 causing it to abort with not owning the right locks for some situation
769 - Job queue: Fixed potential race condition when cancelling queued jobs
770 - Fixed off-by-one bug in job serial generation
771 - ``gnt-node volumes``: Fix instance names
772 - Fixed aliases in bash completion
773 - Fixed a bug in reopening log files after being sent a SIGHUP
774 - Added a flag to burnin to allow specifying VCPU count
775 - Bugfixes to non-root Ganeti configuration
781 *(Released Thu, 12 May 2011)*
783 Many bug-fixes and a few new small features:
785 - Fixed a bug related to log opening failures
786 - Fixed a bug in instance listing with orphan instances
787 - Fixed a bug which prevented resetting the cluster-level node parameter
788 ``oob_program`` to the default
789 - Many fixes related to the ``cluster-merge`` tool
790 - Fixed a race condition in the lock monitor, which caused failures
791 during (at least) creation of many instances in parallel
792 - Improved output for gnt-job info
793 - Removed the quiet flag on some ssh calls which prevented debugging
795 - Improved the N+1 failure messages in cluster verify by actually
796 showing the memory values (needed and available)
797 - Increased lock attempt timeouts so that when executing long operations
798 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
799 too early and thus prevent the use of the 'fair' mechanism
800 - Changed instance query data (``gnt-instance info``) to not acquire
801 locks unless needed, thus allowing its use on locked instance if only
802 static information is asked for
803 - Improved behaviour with filesystems that do not support rename on an
805 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
806 kept locks on all nodes during the wipe, which is unneeded
807 - Fixed ``gnt-watcher`` handling of errors during hooks execution
808 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
809 10GiB) which caused the wipe to fail right at the end in some cases
810 - Fixed master IP activation when doing master failover with no-voting
811 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
812 the master node itself
813 - Fixed potential data-loss in under disk full conditions, where Ganeti
814 wouldn't check correctly the return code and would consider
815 partially-written files 'correct'
816 - Fixed bug related to multiple VGs and DRBD disk replacing
817 - Added new disk parameter ``metavg`` that allows placement of the meta
818 device for DRBD in a different volume group
819 - Fixed error handling in the node daemon when the system libc doesn't
820 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
821 - Fixed lock release during replace-disks, which kept cluster-wide locks
822 when doing disk replaces with an iallocator script
823 - Added check for missing bridges in cluster verify
824 - Handle EPIPE errors while writing to the terminal better, so that
825 piping the output to e.g. ``less`` doesn't cause a backtrace
826 - Fixed rare case where a ^C during Luxi calls could have been
827 interpreted as server errors, instead of simply terminating
828 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
830 - Added a few more parameters to the KVM hypervisor, allowing a second
831 CDROM, custom disk type for CDROMs and a floppy image
832 - Removed redundant message in instance rename when the name is given
834 - Added option to ``gnt-instance recreate-disks`` to allow creating the
835 disks on new nodes, allowing recreation when the original instance
836 nodes are completely gone
837 - Added option when converting disk templates to DRBD to skip waiting
838 for the resync, in order to make the instance available sooner
839 - Added two new variables to the OS scripts environment (containing the
841 - Made the root_path and optional parameter for the xen-pvm hypervisor,
842 to allow use of ``pvgrub`` as bootloader
843 - Changed the instance memory modifications to only check out-of-memory
844 conditions on memory increases, and turned the secondary node warnings
845 into errors (they can still be overridden via ``--force``)
846 - Fixed the handling of a corner case when the Python installation gets
847 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
848 to execute a command that doesn't exist
849 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
850 node of the instance returned failures during instance shutdown; this
851 adds the option ``--ignore-consistency`` to gnt-instance move
853 And as usual, various improvements to the error messages, documentation
860 *(Released Wed, 09 Mar 2011)*
862 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
863 the RAPI users file if run twice (even with ``--dry-run``).
865 The release fixes that bug (nothing else changed).
871 *(Released Mon, 07 Mar 2011)*
873 Final 2.4.0 release. Just a few small fixes:
875 - Fixed RAPI node evacuate
876 - Fixed the kvm-ifup script
877 - Fixed internal error handling for special job cases
878 - Updated man page to specify the escaping feature for options
884 *(Released Mon, 28 Feb 2011)*
886 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
887 that this feature wiped the disks of the wrong instance, leading to loss
892 - Fixed title of query field containing instance name
893 - Expanded the glossary in the documentation
894 - Fixed one unittest (internal issue)
900 *(Released Mon, 21 Feb 2011)*
902 A number of bug fixes plus just a couple functionality changes.
904 On the user-visible side, the ``gnt-* list`` command output has changed
905 with respect to "special" field states. The current rc1 style of display
906 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
907 the default output mode special fields states are displayed as follows:
909 - Offline resource: ``*``
910 - Unavailable/not applicable: ``-``
911 - Data missing (RPC failure): ``?``
912 - Unknown field: ``??``
914 Another user-visible change is the addition of ``--force-join`` to
919 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
920 - Fixed regression in RAPI/instance reinstall where all parameters were
921 required (instead of optional)
922 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
923 - Fixed iallocator usage (offline nodes were not considered offline)
924 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
925 - Fixed hypervisor and OS parameter validation with respect to
927 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
929 - Fixed ``tools/listrunner`` with respect to agent-based usage
935 *(Released Fri, 4 Feb 2011)*
937 Many changes and fixes since the beta1 release. While there were some
938 internal changes, the code has been mostly stabilised for the RC
941 Note: the dumb allocator was removed in this release, as it was not kept
942 up-to-date with the IAllocator protocol changes. It is recommended to
943 use the ``hail`` command from the ganeti-htools package.
945 Note: the 2.4 and up versions of Ganeti are not compatible with the
946 0.2.x branch of ganeti-htools. You need to upgrade to
947 ganeti-htools-0.3.0 (or later).
949 Regressions fixed from 2.3
950 ~~~~~~~~~~~~~~~~~~~~~~~~~~
952 - Fixed the ``gnt-cluster verify-disks`` command
953 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
955 - Fixed disk adoption breakage
956 - Fixed wrong headers in instance listing for field aliases
961 - Fixed corner case in KVM handling of NICs
962 - Fixed many cases of wrong handling of non-vm_capable nodes
963 - Fixed a bug where a missing instance symlink was not possible to
964 recreate with any ``gnt-*`` command (now ``gnt-instance
965 activate-disks`` does it)
966 - Fixed the volume group name as reported by ``gnt-cluster
968 - Increased timeouts for the import-export code, hopefully leading to
969 fewer aborts due network or instance timeouts
970 - Fixed bug in ``gnt-node list-storage``
971 - Fixed bug where not all daemons were started on cluster
972 initialisation, but only at the first watcher run
973 - Fixed many bugs in the OOB implementation
974 - Fixed watcher behaviour in presence of instances with offline
976 - Fixed instance list output for instances running on the wrong node
977 - a few fixes to the cluster-merge tool, but it still cannot merge
978 multi-node groups (currently it is not recommended to use this tool)
984 - Improved network configuration for the KVM hypervisor
985 - Added e1000 as a supported NIC for Xen-HVM
986 - Improved the lvmstrap tool to also be able to use partitions, as
987 opposed to full disks
988 - Improved speed of disk wiping (the cluster parameter
989 ``prealloc_wipe_disks``, so that it has a low impact on the total time
990 of instance creations
991 - Added documentation for the OS parameters
992 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
993 hypervisor is not responding
994 - Added display of blacklisted and hidden OS information in
996 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
997 NIC and node parameters, which might create problems with currently
998 invalid (but undetected) configuration files, but prevents validation
999 failures when unrelated parameters are modified
1000 - Changed cluster initialisation to wait for the master daemon to become
1002 - Expanded the RAPI interface:
1004 - Added config redistribution resource
1005 - Added activation/deactivation of instance disks
1006 - Added export of console information
1008 - Implemented log file reopening on SIGHUP, which allows using
1009 logrotate(8) for the Ganeti log files
1010 - Added a basic OOB helper script as an example
1016 *(Released Fri, 14 Jan 2011)*
1021 - Fixed timezone issues when formatting timestamps
1022 - Added support for node groups, available via ``gnt-group`` and other
1024 - Added out-of-band framework and management, see :doc:`design
1025 document <design-oob>`
1026 - Removed support for roman numbers from ``gnt-node list`` and
1027 ``gnt-instance list``.
1028 - Allowed modification of master network interface via ``gnt-cluster
1029 modify --master-netdev``
1030 - Accept offline secondaries while shutting down instance disks
1031 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1032 - Added support for multiple LVM volume groups
1033 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1035 - Added commands to list available fields:
1037 - ``gnt-node list-fields``
1038 - ``gnt-group list-fields``
1039 - ``gnt-instance list-fields``
1041 - Updated documentation and man pages
1046 - Moved ``rapi_users`` file into separate directory, now named
1047 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1049 - Added new tool for running commands on many machines,
1050 ``tools/ganeti-listrunner``
1051 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1052 improving the ``gnt-instance console`` output
1053 - Allowed customisation of disk index separator at ``configure`` time
1054 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1055 - Added support for non-partitioned md disks in ``lvmstrap``
1056 - Added script to gracefully power off KVM instances
1057 - Split ``utils`` module into smaller parts
1058 - Changed query operations to return more detailed information, e.g.
1059 whether an information is unavailable due to an offline node. To use
1060 this new functionality, the LUXI call ``Query`` must be used. Field
1061 information is now stored by the master daemon and can be retrieved
1062 using ``QueryFields``. Instances, nodes and groups can also be queried
1063 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1064 exposed via RAPI). The following commands make use of this
1065 infrastructure change:
1067 - ``gnt-group list``
1068 - ``gnt-group list-fields``
1070 - ``gnt-node list-fields``
1071 - ``gnt-instance list``
1072 - ``gnt-instance list-fields``
1073 - ``gnt-debug locks``
1078 - New RAPI resources (see :doc:`rapi`):
1082 - ``/2/groups/[group_name]``
1083 - ``/2/groups/[group_name]/assign-nodes``
1084 - ``/2/groups/[group_name]/modify``
1085 - ``/2/groups/[group_name]/rename``
1086 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1090 - Implemented ``no_install`` for instance creation
1091 - Implemented OS parameters for instance reinstallation, allowing
1092 use of special settings on reinstallation (e.g. for preserving data)
1097 - Added IPv6 support in import/export
1098 - Pause DRBD synchronization while wiping disks on instance creation
1099 - Updated unittests and QA scripts
1100 - Improved network parameters passed to KVM
1101 - Converted man pages from docbook to reStructuredText
1107 *(Released Mon, 20 Dec 2010)*
1109 Released version 2.3.1~rc1 without any changes.
1115 *(Released Wed, 1 Dec 2010)*
1117 - impexpd: Disable OpenSSL compression in socat if possible (backport
1118 from master, commit e90739d625b, see :doc:`installation guide
1119 <install-quick>` for details)
1120 - Changed unittest coverage report to exclude test scripts
1121 - Added script to check version format
1127 *(Released Wed, 1 Dec 2010)*
1129 Released version 2.3.0~rc1 without any changes.
1135 *(Released Fri, 19 Nov 2010)*
1137 A number of bugfixes and documentation updates:
1139 - Update ganeti-os-interface documentation
1140 - Fixed a bug related to duplicate MACs or similar items which should be
1142 - Fix breakage in OS state modify
1143 - Reinstall instance: disallow offline secondaries (fixes bug related to
1144 OS changing but reinstall failing)
1145 - plus all the other fixes between 2.2.1 and 2.2.2
1151 *(Released Tue, 2 Nov 2010)*
1153 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1154 - Fixed master failover race with watcher
1155 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1157 - Accept previously stopped instance for export with instance removal
1158 - Simplify and extend the environment variables for instance OS scripts
1159 - Added new node flags, ``master_capable`` and ``vm_capable``
1160 - Added optional instance disk wiping prior during allocation. This is a
1161 cluster-wide option and can be set/modified using
1162 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1163 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1164 :doc:`install-quick`
1165 - Added a new watcher option (``--ignore-pause``)
1166 - Added option to ignore offline node on instance start/stop
1167 (``--ignore-offline``)
1168 - Allow overriding OS parameters with ``gnt-instance reinstall``
1169 - Added ability to change node's secondary IP address using ``gnt-node
1171 - Implemented privilege separation for all daemons except
1172 ``ganeti-noded``, see ``configure`` options
1173 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1175 - Implemented job priorities (see ``ganeti(7)`` manpage)
1176 - Ignore failures while shutting down instances during failover from
1178 - Exit daemon's bootstrap process only once daemon is ready
1179 - Export more information via ``LUInstanceQuery``/remote API
1180 - Improved documentation, QA and unittests
1181 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1182 restart if the file was created or changed
1183 - Added LUXI protocol version sent with each request and response,
1184 allowing detection of server/client mismatches
1185 - Moved the Python scripts among gnt-* and ganeti-* into modules
1186 - Moved all code related to setting up SSH to an external script,
1188 - Infrastructure changes for node group support in future versions
1194 *(Released Fri, 19 Nov 2010)*
1196 A few small bugs fixed, and some improvements to the build system:
1198 - Fix documentation regarding conversion to drbd
1199 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1201 - Fix error handling in node modify with multiple changes
1202 - Allow remote imports without checked names
1208 *(Released Tue, 19 Oct 2010)*
1210 - Disable SSL session ID cache in RPC client
1216 *(Released Thu, 14 Oct 2010)*
1218 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1219 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1221 - Fix problems with interaction between Curl and Python's HTTP server,
1222 resulting in increased speed in many RPC calls
1223 - Improve our release script to prevent breakage with older aclocal and
1230 *(Released Thu, 7 Oct 2010)*
1232 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1233 value retrieved from master
1234 - Added support for blacklisted or hidden OS definitions
1235 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1236 - Added support for -mem-path in KVM hypervisor abstraction layer
1237 - Allow overriding instance parameters in tool for inter-cluster
1238 instance moves (``tools/move-instance``)
1239 - Improved opcode summaries (e.g. in ``gnt-job list``)
1240 - Improve consistency of OS listing by sorting it
1241 - Documentation updates
1247 *(Released Fri, 8 Oct 2010)*
1249 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1255 *(Released Mon, 4 Oct 2010)*
1257 - Fixed regression in ``gnt-instance rename``
1263 *(Released Wed, 22 Sep 2010)*
1265 - Fixed OS_VARIANT variable for OS scripts
1266 - Fixed cluster tag operations via RAPI
1267 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1268 - Disabled RAPI CA checks in watcher
1274 *(Released Mon, 23 Aug 2010)*
1276 - Support DRBD versions of the format "a.b.c.d"
1278 - Re-introduce support for usage from multiple threads in RAPI client
1279 - Instance renames and modify via RAPI
1280 - Work around race condition between processing and archival in job
1282 - Mark opcodes following failed one as failed, too
1283 - Job field ``lock_status`` was removed due to difficulties making it
1284 work with the changed job queue in Ganeti 2.2; a better way to monitor
1285 locks is expected for a later 2.2.x release
1286 - Fixed dry-run behaviour with many commands
1287 - Support ``ssh-agent`` again when adding nodes
1288 - Many additional bugfixes
1294 *(Released Fri, 30 Jul 2010)*
1296 Important change: the internal RPC mechanism between Ganeti nodes has
1297 changed from using a home-grown http library (based on the Python base
1298 libraries) to use the PycURL library. This requires that PycURL is
1299 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1300 against GnuTLS by default. cURL's support for GnuTLS had known issues
1301 before cURL 7.21.0 and we recommend using the latest cURL release or
1302 linking against OpenSSL. Most other distributions already link PycURL
1303 and cURL against OpenSSL. The command::
1305 python -c 'import pycurl; print pycurl.version'
1307 can be used to determine the libraries PycURL and cURL are linked
1310 Other significant changes:
1312 - Rewrote much of the internals of the job queue, in order to achieve
1313 better parallelism; this decouples job query operations from the job
1314 processing, and it should allow much nicer behaviour of the master
1315 daemon under load, and it also has uncovered some long-standing bugs
1316 related to the job serialisation (now fixed)
1317 - Added a default iallocator setting to the cluster parameters,
1318 eliminating the need to always pass nodes or an iallocator for
1319 operations that require selection of new node(s)
1320 - Added experimental support for the LXC virtualization method
1321 - Added support for OS parameters, which allows the installation of
1322 instances to pass parameter to OS scripts in order to customise the
1324 - Added a hypervisor parameter controlling the migration type (live or
1325 non-live), since hypervisors have various levels of reliability; this
1326 has renamed the 'live' parameter to 'mode'
1327 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1328 logical volumes, meaning that cluster verify will ignore them and not
1329 flag their presence as errors
1330 - The watcher will now reset the error count for failed instances after
1331 8 hours, thus allowing self-healing if the problem that caused the
1332 instances to be down/fail to start has cleared in the meantime
1333 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1334 check for, and warn, if the drbd module parameter ``usermode_helper``
1335 is not consistent with the cluster-wide setting; this is needed to
1336 make diagnose easier of failed drbd creations
1337 - Started adding base IPv6 support, but this is not yet
1338 enabled/available for use
1339 - Rename operations (cluster, instance) will now return the new name,
1340 which is especially useful if a short name was passed in
1341 - Added support for instance migration in RAPI
1342 - Added a tool to pre-configure nodes for the SSH setup, before joining
1343 them to the cluster; this will allow in the future a simplified model
1344 for node joining (but not yet fully enabled in 2.2); this needs the
1345 paramiko python library
1346 - Fixed handling of name-resolving errors
1347 - Fixed consistency of job results on the error path
1348 - Fixed master-failover race condition when executed multiple times in
1350 - Fixed many bugs related to the job queue (mostly introduced during the
1351 2.2 development cycle, so not all are impacting 2.1)
1352 - Fixed instance migration with missing disk symlinks
1353 - Fixed handling of unknown jobs in ``gnt-job archive``
1354 - And many other small fixes/improvements
1358 - Enhanced both the unittest and the QA coverage
1359 - Switched the opcode validation to a generic model, and extended the
1360 validation to all opcode parameters
1361 - Changed more parts of the code that write shell scripts to use the
1363 - Switched the master daemon to use the asyncore library for the Luxi
1370 *(Released Thu, 17 Jun 2010)*
1372 - Added tool (``move-instance``) and infrastructure to move instances
1373 between separate clusters (see :doc:`separate documentation
1374 <move-instance>` and :doc:`design document <design-2.2>`)
1375 - Added per-request RPC timeout
1376 - RAPI now requires a Content-Type header for requests with a body (e.g.
1377 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1378 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1379 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1381 - Implemented initial support for running Ganeti daemons as separate
1382 users, see configure-time flags ``--with-user-prefix`` and
1383 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1385 - Instances can be removed after export (``gnt-backup export
1386 --remove-instance``)
1387 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1388 key (instead of 1024 bit)
1389 - Added new cluster configuration file for cluster domain secret
1390 - Import/export now use SSL instead of SSH
1391 - Added support for showing estimated time when exporting an instance,
1392 see the ``ganeti-os-interface(7)`` manpage and look for
1399 *(Released Tue, 16 Nov 2010)*
1401 Some more bugfixes. Unless critical bugs occur, this will be the last
1404 - Fix case of MAC special-values
1405 - Fix mac checker regex
1406 - backend: Fix typo causing "out of range" error
1407 - Add missing --units in gnt-instance list man page
1413 *(Released Tue, 24 Aug 2010)*
1416 - Don't ignore secondary node silently on non-mirrored disk templates
1418 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1419 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1420 - Properly document the usb_mouse parameter
1421 - Fix path in ganeti-rapi(8) (issue 116)
1422 - Adjust error message when the ganeti user's .ssh directory is
1424 - Add same-node-check when changing the disk template to drbd
1430 *(Released Fri, 16 Jul 2010)*
1433 - Add an option to only select some reboot types during qa/burnin.
1434 (on some hypervisors consequent reboots are not supported)
1435 - Fix infrequent race condition in master failover. Sometimes the old
1436 master ip address would be still detected as up for a short time
1437 after it was removed, causing failover to fail.
1438 - Decrease mlockall warnings when the ctypes module is missing. On
1439 Python 2.4 we support running even if no ctypes module is installed,
1440 but we were too verbose about this issue.
1441 - Fix building on old distributions, on which man doesn't have a
1443 - Fix RAPI not to ignore the MAC address on instance creation
1444 - Implement the old instance creation format in the RAPI client.
1450 *(Released Thu, 01 Jul 2010)*
1452 A small bugfix release:
1453 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1454 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1455 a non-existing option
1456 - Fix instance create: the --force-variant option was ignored
1457 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1458 - Fix modify node storage with non-FQDN arguments
1459 - Fix RAPI client to authenticate under Python 2.6 when used
1460 for more than 5 requests needing authentication
1461 - Fix gnt-instance modify -t (storage) giving a wrong error message
1462 when converting a non-shutdown drbd instance to plain
1468 *(Released Fri, 18 Jun 2010)*
1470 A small bugfix release:
1472 - Fix live migration of KVM instances started with older Ganeti
1473 versions which had fewer hypervisor parameters
1474 - Fix gnt-instance grow-disk on down instances
1475 - Fix an error-reporting bug during instance migration
1476 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1477 silently ignoring broken ones
1478 - Fix an RPC error reporting bug affecting, for example, RAPI client
1480 - Fix bug triggered by different API version os-es on different nodes
1481 - Fix a bug in instance startup with custom hvparams: OS level
1482 parameters would fail to be applied.
1483 - Fix the RAPI client under Python 2.6 (but more work is needed to
1484 make it work completely well with OpenSSL)
1485 - Fix handling of errors when resolving names from DNS
1491 *(Released Thu, 3 Jun 2010)*
1493 A medium sized development cycle. Some new features, and some
1494 fixes/small improvements/cleanups.
1496 Significant features
1497 ~~~~~~~~~~~~~~~~~~~~
1499 The node deamon now tries to mlock itself into memory, unless the
1500 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1501 its logs, and falls back to console logging. This allows emergency
1502 features such as ``gnt-node powercycle`` to work even in the event of a
1503 broken node disk (tested offlining the disk hosting the node's
1504 filesystem and dropping its memory caches; don't try this at home)
1506 KVM: add vhost-net acceleration support. It can be tested with a new
1507 enough version of the kernel and of qemu-kvm.
1509 KVM: Add instance chrooting feature. If you use privilege dropping for
1510 your VMs you can also now force them to chroot to an empty directory,
1511 before starting the emulated guest.
1513 KVM: Add maximum migration bandwith and maximum downtime tweaking
1514 support (requires a new-enough version of qemu-kvm).
1516 Cluster verify will now warn if the master node doesn't have the master
1517 ip configured on it.
1519 Add a new (incompatible) instance creation request format to RAPI which
1520 supports all parameters (previously only a subset was supported, and it
1521 wasn't possible to extend the old format to accomodate all the new
1522 features. The old format is still supported, and a client can check for
1523 this feature, before using it, by checking for its presence in the
1524 ``features`` RAPI resource.
1526 Now with ancient latin support. Try it passing the ``--roman`` option to
1527 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1528 (requires the python-roman module to be installed, in order to work).
1533 As usual many internal code refactorings, documentation updates, and
1536 - Lots of improvements and cleanups to the experimental Remote API
1537 (RAPI) client library.
1538 - A new unit test suite for the core daemon libraries.
1539 - A fix to creating missing directories makes sure the umask is not
1540 applied anymore. This enforces the same directory permissions
1542 - Better handling terminating daemons with ctrl+c (used when running
1543 them in debugging mode).
1544 - Fix a race condition in live migrating a KVM instance, when stat()
1545 on the old proc status file returned EINVAL, which is an unexpected
1547 - Fixed manpage checking with newer man and utf-8 charachters. But now
1548 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1554 *(Released Fri, 7 May 2010)*
1556 Fix a bug which prevented untagged KVM instances from starting.
1562 *(Released Fri, 7 May 2010)*
1564 Another release with a long development cycle, during which many
1565 different features were added.
1567 Significant features
1568 ~~~~~~~~~~~~~~~~~~~~
1570 The KVM hypervisor now can run the individual instances as non-root, to
1571 reduce the impact of a VM being hijacked due to bugs in the
1572 hypervisor. It is possible to run all instances as a single (non-root)
1573 user, to manually specify a user for each instance, or to dynamically
1574 allocate a user out of a cluster-wide pool to each instance, with the
1575 guarantee that no two instances will run under the same user ID on any
1578 An experimental RAPI client library, that can be used standalone
1579 (without the other Ganeti libraries), is provided in the source tree as
1580 ``lib/rapi/client.py``. Note this client might change its interface in
1581 the future, as we iterate on its capabilities.
1583 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1584 replace the cluster's certificates and crypto keys. This might help in
1585 case they have been compromised, or have simply expired.
1587 A new disk option for instance creation has been added that allows one
1588 to "adopt" currently existing logical volumes, with data
1589 preservation. This should allow easier migration to Ganeti from
1590 unmanaged (or managed via other software) instances.
1592 Another disk improvement is the possibility to convert between redundant
1593 (DRBD) and plain (LVM) disk configuration for an instance. This should
1594 allow better scalability (starting with one node and growing the
1595 cluster, or shrinking a two-node cluster to one node).
1597 A new feature that could help with automated node failovers has been
1598 implemented: if a node sees itself as offline (by querying the master
1599 candidates), it will try to shutdown (hard) all instances and any active
1600 DRBD devices. This reduces the risk of duplicate instances if an
1601 external script automatically failovers the instances on such nodes. To
1602 enable this, the cluster parameter ``maintain_node_health`` should be
1603 enabled; in the future this option (per the name) will enable other
1604 automatic maintenance features.
1606 Instance export/import now will reuse the original instance
1607 specifications for all parameters; that means exporting an instance,
1608 deleting it and the importing it back should give an almost identical
1609 instance. Note that the default import behaviour has changed from
1610 before, where it created only one NIC; now it recreates the original
1613 Cluster verify has added a few new checks: SSL certificates validity,
1614 /etc/hosts consistency across the cluster, etc.
1619 As usual, many internal changes were done, documentation fixes,
1622 - Fixed cluster initialization with disabled cluster storage (regression
1623 introduced in 2.1.1)
1624 - File-based storage supports growing the disks
1625 - Fixed behaviour of node role changes
1626 - Fixed cluster verify for some corner cases, plus a general rewrite of
1627 cluster verify to allow future extension with more checks
1628 - Fixed log spamming by watcher and node daemon (regression introduced
1630 - Fixed possible validation issues when changing the list of enabled
1632 - Fixed cleanup of /etc/hosts during node removal
1633 - Fixed RAPI response for invalid methods
1634 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1635 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1636 from ide disks, etc.)
1637 - Allow OS changes without re-installation (to record a changed OS
1638 outside of Ganeti, or to allow OS renames)
1639 - Allow instance creation without OS installation (useful for example if
1640 the OS will be installed manually, or restored from a backup not in
1642 - Implemented option to make cluster ``copyfile`` use the replication
1644 - Added list of enabled hypervisors to ssconf (possibly useful for
1646 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1648 - A partial form of node re-IP is possible via node readd, which now
1649 allows changed node primary IP
1650 - Command line utilities now show an informational message if the job is
1652 - The logs of the master daemon now show the PID/UID/GID of the
1659 *(Released Fri, 12 Mar 2010)*
1661 During the 2.1.0 long release candidate cycle, a lot of improvements and
1662 changes have accumulated with were released later as 2.1.1.
1667 The node evacuate command (``gnt-node evacuate``) was significantly
1668 rewritten, and as such the IAllocator protocol was changed - a new
1669 request type has been added. This unfortunate change during a stable
1670 series is designed to improve performance of node evacuations; on
1671 clusters with more than about five nodes and which are well-balanced,
1672 evacuation should proceed in parallel for all instances of the node
1673 being evacuated. As such, any existing IAllocator scripts need to be
1674 updated, otherwise the above command will fail due to the unknown
1675 request. The provided "dumb" allocator has not been updated; but the
1676 ganeti-htools package supports the new protocol since version 0.2.4.
1678 Another important change is increased validation of node and instance
1679 names. This might create problems in special cases, if invalid host
1680 names are being used.
1682 Also, a new layer of hypervisor parameters has been added, that sits at
1683 OS level between the cluster defaults and the instance ones. This allows
1684 customisation of virtualization parameters depending on the installed
1685 OS. For example instances with OS 'X' may have a different KVM kernel
1686 (or any other parameter) than the cluster defaults. This is intended to
1687 help managing a multiple OSes on the same cluster, without manual
1688 modification of each instance's parameters.
1690 A tool for merging clusters, ``cluster-merge``, has been added in the
1691 tools sub-directory.
1696 - Improved the int/float conversions that should make the code more
1697 robust in face of errors from the node daemons
1698 - Fixed the remove node code in case of internal configuration errors
1699 - Fixed the node daemon behaviour in face of inconsistent queue
1700 directory (e.g. read-only file-system where we can't open the files
1702 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1703 now it either aborts cleanly or, if given the new "auto_promote"
1704 parameter, will automatically promote other nodes as needed
1705 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1706 completely prevent Ganeti from working
1707 - Fixed bug for instance export when not all disks were successfully
1709 - Fixed behaviour of node add when the new node is slow in starting up
1711 - Fixed handling of signals in the LUXI client, which should improve
1712 behaviour of command-line scripts
1713 - Added checks for invalid node/instance names in the configuration (now
1714 flagged during cluster verify)
1715 - Fixed watcher behaviour for disk activation errors
1716 - Fixed two potentially endless loops in http library, which led to the
1717 RAPI daemon hanging and consuming 100% CPU in some cases
1718 - Fixed bug in RAPI daemon related to hashed passwords
1719 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1721 - Enhanced compatibility with non-Debian OSes, but not using absolute
1722 path in some commands and allowing customisation of the ssh
1723 configuration directory
1724 - Fixed possible future issue with new Python versions by abiding to the
1725 proper use of ``__slots__`` attribute on classes
1726 - Added checks that should prevent directory traversal attacks
1727 - Many documentation fixes based on feedback from users
1732 - Added an "early_release" more for instance replace disks and node
1733 evacuate, where we release locks earlier and thus allow higher
1734 parallelism within the cluster
1735 - Added watcher hooks, intended to allow the watcher to restart other
1736 daemons (e.g. from the ganeti-nbma project), but they can be used of
1737 course for any other purpose
1738 - Added a compile-time disable for DRBD barriers, to increase
1739 performance if the administrator trusts the power supply or the
1740 storage system to not lose writes
1741 - Added the option of using syslog for logging instead of, or in
1742 addition to, Ganeti's own log files
1743 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1744 can indeed boot from a paravirtual NIC
1745 - Added a generic debug level for many operations; while this is not
1746 used widely yet, it allows one to pass the debug value all the way to
1748 - Enhanced the hooks environment for instance moves (failovers,
1749 migrations) where the primary/secondary nodes changed during the
1750 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1751 - Enhanced data validations for many user-supplied values; one important
1752 item is the restrictions imposed on instance and node names, which
1753 might reject some (invalid) host names
1754 - Add a configure-time option to disable file-based storage, if it's not
1755 needed; this allows greater security separation between the master
1756 node and the other nodes from the point of view of the inter-node RPC
1758 - Added user notification in interactive tools if job is waiting in the
1759 job queue or trying to acquire locks
1760 - Added log messages when a job is waiting for locks
1761 - Added filtering by node tags in instance operations which admit
1762 multiple instances (start, stop, reboot, reinstall)
1763 - Added a new tool for cluster mergers, ``cluster-merge``
1764 - Parameters from command line which are of the form ``a=b,c=d`` can now
1765 use backslash escapes to pass in values which contain commas,
1766 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1768 - For KVM, the instance name is the first parameter passed to KVM, so
1769 that it's more visible in the process list
1775 *(Released Tue, 2 Mar 2010)*
1777 Ganeti 2.1 brings many improvements with it. Major changes:
1779 - Added infrastructure to ease automated disk repairs
1780 - Added new daemon to export configuration data in a cheaper way than
1781 using the remote API
1782 - Instance NICs can now be routed instead of being associated with a
1784 - Improved job locking logic to reduce impact of jobs acquiring multiple
1785 locks waiting for other long-running jobs
1787 In-depth implementation details can be found in the Ganeti 2.1 design
1793 - Added chroot hypervisor
1794 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1796 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1797 ``bootloader_path`` and ``bootloader_args``)
1798 - Added the ``use_localtime`` option for the xen-hvm and kvm
1799 hypervisors, and the default value for this has changed to false (in
1800 2.0 xen-hvm always enabled it)
1801 - Added luxi call to submit multiple jobs in one go
1802 - Added cluster initialization option to not modify ``/etc/hosts``
1804 - Added network interface parameters
1805 - Added dry run mode to some LUs
1806 - Added RAPI resources:
1808 - ``/2/instances/[instance_name]/info``
1809 - ``/2/instances/[instance_name]/replace-disks``
1810 - ``/2/nodes/[node_name]/evacuate``
1811 - ``/2/nodes/[node_name]/migrate``
1812 - ``/2/nodes/[node_name]/role``
1813 - ``/2/nodes/[node_name]/storage``
1814 - ``/2/nodes/[node_name]/storage/modify``
1815 - ``/2/nodes/[node_name]/storage/repair``
1817 - Added OpCodes to evacuate or migrate all instances on a node
1818 - Added new command to list storage elements on nodes (``gnt-node
1819 list-storage``) and modify them (``gnt-node modify-storage``)
1820 - Added new ssconf files with master candidate IP address
1821 (``ssconf_master_candidates_ips``), node primary IP address
1822 (``ssconf_node_primary_ips``) and node secondary IP address
1823 (``ssconf_node_secondary_ips``)
1824 - Added ``ganeti-confd`` and a client library to query the Ganeti
1825 configuration via UDP
1826 - Added ability to run hooks after cluster initialization and before
1828 - Added automatic mode for disk replace (``gnt-instance replace-disks
1830 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1831 after catastrophic data-loss
1832 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1834 - Added ``gnt-instance move`` command to move instances
1835 - Added ``gnt-cluster watcher`` command to control watcher
1836 - Added ``gnt-node powercycle`` command to powercycle nodes
1837 - Added new job status field ``lock_status``
1838 - Added parseable error codes to cluster verification (``gnt-cluster
1839 verify --error-codes``) and made output less verbose (use
1840 ``--verbose`` to restore previous behaviour)
1841 - Added UUIDs to the main config entities (cluster, nodes, instances)
1842 - Added support for OS variants
1843 - Added support for hashed passwords in the Ganeti remote API users file
1845 - Added option to specify maximum timeout on instance shutdown
1846 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1847 - Added new helper script to start and stop Ganeti daemons
1848 (``daemon-util``), with the intent to reduce the work necessary to
1849 adjust Ganeti for non-Debian distributions and to start/stop daemons
1851 - Added more unittests
1852 - Fixed critical bug in ganeti-masterd startup
1853 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1854 now customisable at the cluster level for both the KVM and Xen
1855 hypervisors using the new ``migration_port`` parameter
1856 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1857 reinstalling an instance
1858 - Allowed ``@`` in tag names
1859 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1860 - Many documentation updates
1861 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1862 - ``gnt-instance reinstall`` can now reinstall multiple instances
1863 - Updated many command line parameters
1864 - Introduced new OS API version 15
1865 - No longer support a default hypervisor
1866 - Treat virtual LVs as inexistent
1867 - Improved job locking logic to reduce lock contention
1868 - Match instance and node names case insensitively
1869 - Reimplemented bash completion script to be more complete
1876 *(Released Thu, 4 Feb 2010)*
1878 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1879 - Fix a string formatting bug
1880 - Improve safety of the code in some error paths
1881 - Improve data validation in the master of values returned from nodes
1887 *(Released Thu, 17 Dec 2009)*
1889 - Fix security issue due to missing validation of iallocator names; this
1890 allows local and remote execution of arbitrary executables
1891 - Fix failure of gnt-node list during instance removal
1892 - Ship the RAPI documentation in the archive
1898 *(Released Wed, 30 Sep 2009)*
1900 - Fixed many wrong messages
1901 - Fixed a few bugs related to the locking library
1902 - Fixed MAC checking at instance creation time
1903 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1904 - Fixed a few issues related to signal handling in both daemons and
1906 - Fixed the example startup script provided
1907 - Fixed insserv dependencies in the example startup script (patch from
1909 - Fixed handling of drained nodes in the iallocator framework
1910 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1912 - Fixed error related to invalid job IDs in job polling
1913 - Fixed job/opcode persistence on unclean master shutdown
1914 - Fixed handling of partial job processing after unclean master
1916 - Fixed error reporting from LUs, previously all errors were converted
1917 into execution errors
1918 - Fixed error reporting from burnin
1919 - Decreased significantly the memory usage of the job queue
1920 - Optimised slightly multi-job submission
1921 - Optimised slightly opcode loading
1922 - Backported the multi-job submit framework from the development
1923 branch; multi-instance start and stop should be faster
1924 - Added script to clean archived jobs after 21 days; this will reduce
1925 the size of the queue directory
1926 - Added some extra checks in disk size tracking
1927 - Added an example ethers hook script
1928 - Added a cluster parameter that prevents Ganeti from modifying of
1930 - Added more node information to RAPI responses
1931 - Added a ``gnt-job watch`` command that allows following the ouput of a
1933 - Added a bind-address option to ganeti-rapi
1934 - Added more checks to the configuration verify
1935 - Enhanced the burnin script such that some operations can be retried
1937 - Converted instance reinstall to multi-instance model
1943 *(Released Fri, 7 Aug 2009)*
1945 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1946 to allow using the pre-2.0.2 behaviour in activation, if any existing
1947 instances have mismatched disk sizes in the configuration
1948 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
1949 any configuration mismatches for disk sizes
1950 - Added ``gnt-master cluste-failover --no-voting`` to allow master
1951 failover to work on two-node clusters
1952 - Fixed the ``--net`` option of ``gnt-backup import``, which was
1954 - Fixed detection of OS script errors in ``gnt-backup export``
1955 - Fixed exit code of ``gnt-backup export``
1961 *(Released Fri, 17 Jul 2009)*
1963 - Added experimental support for stripped logical volumes; this should
1964 enhance performance but comes with a higher complexity in the block
1965 device handling; stripping is only enabled when passing
1966 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
1967 affected even in the non-stripped mode
1968 - Improved resiliency against transient failures at the end of DRBD
1969 resyncs, and in general of DRBD resync checks
1970 - Fixed a couple of issues with exports and snapshot errors
1971 - Fixed a couple of issues in instance listing
1972 - Added display of the disk size in ``gnt-instance info``
1973 - Fixed checking for valid OSes in instance creation
1974 - Fixed handling of the "vcpus" parameter in instance listing and in
1975 general of invalid parameters
1976 - Fixed http server library, and thus RAPI, to handle invalid
1977 username/password combinations correctly; this means that now they
1978 report unauthorized for queries too, not only for modifications,
1979 allowing earlier detect of configuration problems
1980 - Added a new "role" node list field, equivalent to the master/master
1981 candidate/drained/offline flags combinations
1982 - Fixed cluster modify and changes of candidate pool size
1983 - Fixed cluster verify error messages for wrong files on regular nodes
1984 - Fixed a couple of issues with node demotion from master candidate role
1985 - Fixed node readd issues
1986 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
1987 - Added a new ``--no-voting`` option for masterfailover to fix failover
1988 on two-nodes clusters when the former master node is unreachable
1989 - Added instance reinstall over RAPI
1995 *(Released Tue, 16 Jun 2009)*
1997 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
1998 allow re-adding the start in single-user option (regression from 1.2)
1999 - the watcher writes the instance status to a file, to allow monitoring
2000 to report the instance status (from the master) based on cached
2001 results of the watcher's queries; while this can get stale if the
2002 watcher is being locked due to other work on the cluster, this is
2003 still an improvement
2004 - the watcher now also restarts the node daemon and the rapi daemon if
2006 - fixed the watcher to handle full and drained queue cases
2007 - hooks export more instance data in the environment, which helps if
2008 hook scripts need to take action based on the instance's properties
2009 (no longer need to query back into ganeti)
2010 - instance failovers when the instance is stopped do not check for free
2011 RAM, so that failing over a stopped instance is possible in low memory
2013 - rapi uses queries for tags instead of jobs (for less job traffic), and
2014 for cluster tags it won't talk to masterd at all but read them from
2016 - a couple of error handling fixes in RAPI
2017 - drbd handling: improved the error handling of inconsistent disks after
2018 resync to reduce the frequency of "there are some degraded disks for
2019 this instance" messages
2020 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2021 error handling path called a wrong function name)
2027 *(Released Wed, 27 May 2009)*
2029 - no changes from rc5
2035 *(Released Wed, 20 May 2009)*
2037 - fix a couple of bugs (validation, argument checks)
2038 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2039 - some small improvements to RAPI and IAllocator
2040 - make watcher automatically start the master daemon if down
2046 *(Released Mon, 27 Apr 2009)*
2048 - change the OS list to not require locks; this helps with big clusters
2049 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2050 volume group is broken
2051 - ``gnt-instance info``, without any arguments, doesn't run for all
2052 instances anymore; either pass ``--all`` or pass the desired
2053 instances; this helps against mistakes on big clusters where listing
2054 the information for all instances takes a long time
2055 - miscellaneous doc and man pages fixes
2061 *(Released Wed, 8 Apr 2009)*
2063 - Change the internal locking model of some ``gnt-node`` commands, in
2064 order to reduce contention (and blocking of master daemon) when
2065 batching many creation/reinstall jobs
2066 - Fixes to Xen soft reboot
2067 - No longer build documentation at build time, instead distribute it in
2068 the archive, in order to reduce the need for the whole docbook/rst
2075 *(Released Fri, 27 Mar 2009)*
2077 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2078 - Fix watcher startup sequence, improves the behaviour of busy clusters
2079 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2080 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2081 bind address and other places
2082 - Some documentation fixes and updates
2088 *(Released Mon, 2 Mar 2009)*
2090 - More documentation updates, now all docs should be more-or-less
2092 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2094 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2096 - Fix an issue related to $libdir/run/ganeti and cluster creation
2102 *(Released Thu, 19 Feb 2009)*
2104 - Xen PVM and KVM have switched the default value for the instance root
2105 disk to the first partition on the first drive, instead of the whole
2106 drive; this means that the OS installation scripts must be changed
2108 - Man pages have been updated
2109 - RAPI has been switched by default to HTTPS, and the exported functions
2110 should all work correctly
2111 - RAPI v1 has been removed
2112 - Many improvements to the KVM hypervisor
2113 - Block device errors are now better reported
2114 - Many other bugfixes and small improvements
2120 *(Released Mon, 26 Jan 2009)*
2122 - Version 2 is a general rewrite of the code and therefore the
2123 differences are too many to list, see the design document for 2.0 in
2124 the ``doc/`` subdirectory for more details
2125 - In this beta version there is not yet a migration path from 1.2 (there
2126 will be one in the final 2.0 release)
2127 - A few significant changes are:
2129 - all commands are executed by a daemon (``ganeti-masterd``) and the
2130 various ``gnt-*`` commands are just front-ends to it
2131 - all the commands are entered into, and executed from a job queue,
2132 see the ``gnt-job(8)`` manpage
2133 - the RAPI daemon supports read-write operations, secured by basic
2134 HTTP authentication on top of HTTPS
2135 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2136 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2137 to migrate to DRBD 8 first while still running Ganeti 1.2)
2138 - DRBD devices are using statically allocated minor numbers, which
2139 will be assigned to existing instances during the migration process
2140 - there is support for both Xen PVM and Xen HVM instances running on
2142 - KVM virtualization is supported too
2143 - file-based storage has been implemented, which means that it is
2144 possible to run the cluster without LVM and DRBD storage, for
2145 example using a shared filesystem exported from shared storage (and
2146 still have live migration)
2152 *(Released Tue, 13 Jan 2009)*
2154 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2155 - Reuse the old instance mac address by default on instance import, if
2156 the instance name is the same.
2157 - Handle situations in which the node info rpc returns incomplete
2159 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2160 - Improved version of batcher:
2162 - state file support
2163 - instance mac address support
2164 - support for HVM clusters/instances
2166 - Add an option to show the number of cpu sockets and nodes in
2168 - Support OSes that handle more than one version of the OS api (but do
2169 not change the current API in any other way)
2170 - Fix ``gnt-node migrate``
2171 - ``gnt-debug`` man page
2172 - Fixes various more typos and small issues
2173 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2179 *(Released Wed, 24 Sep 2008)*
2181 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2182 type of disk exported to fully virtualized instances.
2183 - provide access to the serial console of HVM instances
2184 - instance auto_balance flag, set by default. If turned off it will
2185 avoid warnings on cluster verify if there is not enough memory to fail
2186 over an instance. in the future it will prevent automatically failing
2187 it over when we will support that.
2188 - batcher tool for instance creation, see ``tools/README.batcher``
2189 - ``gnt-instance reinstall --select-os`` to interactively select a new
2190 operating system when reinstalling an instance.
2191 - when changing the memory amount on instance modify a check has been
2192 added that the instance will be able to start. also warnings are
2193 emitted if the instance will not be able to fail over, if auto_balance
2195 - documentation fixes
2196 - sync fields between ``gnt-instance list/modify/add/import``
2197 - fix a race condition in drbd when the sync speed was set after giving
2198 the device a remote peer.
2204 *(Released Tue, 22 Jul 2008)*
2206 - note: the allowed size and number of tags per object were reduced
2207 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2208 - fixed twisted 8.x compatibility
2209 - fixed ``gnt-instance replace-disks`` with iallocator
2210 - add TCP keepalives on twisted connections to detect restarted nodes
2211 - disk increase support, see ``gnt-instance grow-disk``
2212 - implement bulk node/instance query for RAPI
2213 - add tags in node/instance listing (optional)
2214 - experimental migration (and live migration) support, read the man page
2215 for ``gnt-instance migrate``
2216 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2217 has some small improvements in handling its state file
2223 *(Released Fri, 13 Jun 2008)*
2225 - Experimental readonly, REST-based remote API implementation;
2226 automatically started on master node, TCP port 5080, if enabled by
2227 ``--enable-rapi`` parameter to configure script.
2228 - Instance allocator support. Add and import instance accept a
2229 ``--iallocator`` parameter, and call that instance allocator to decide
2230 which node to use for the instance. The iallocator document describes
2231 what's expected from an allocator script.
2232 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2233 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2234 node is lost there is still enough memory to fail over the instances
2236 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2237 ``gnt-cluster verify``, to check for site-specific compliance. All the
2238 hooks will run, and their output, if any, will be displayed. Any
2239 failing hook will make the verification return an error value.
2240 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2241 primary and secondary interfaces
2242 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2243 that is still declared as part of the cluster and has failed.
2244 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2245 specifying output fields, that just adds the chosen fields to the
2247 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2248 export from the filesystem.
2249 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2250 have been added. Using them you can enable/disable acpi and pae
2251 support, and specify a path for a cd image to be exported to the
2252 instance. These parameters as the name suggest only work on HVM
2254 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2255 PAE support will be set to the previously hardcoded values, but the
2256 (previously hardcoded) path to the CDROM ISO image will be unset and
2257 if required, needs to be set manually with ``gnt-instance modify``
2259 - The address to which an instance's VNC console is bound is now
2260 selectable per-instance, rather than being cluster wide. Of course
2261 this only applies to instances controlled via VNC, so currently just
2262 applies to HVM clusters.
2268 *(Released Mon, 18 Feb 2008)*
2270 - more tweaks to the disk activation code (especially helpful for DRBD)
2271 - change the default ``gnt-instance list`` output format, now there is
2272 one combined status field (see the manpage for the exact values this
2274 - some more fixes for the mac export to hooks change
2275 - make Ganeti not break with DRBD 8.2.x (which changed the version
2276 format in ``/proc/drbd``) (issue 24)
2277 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2278 template, allowing migration from DRBD0.7+MD to DRBD8
2284 *(Released Wed, 30 Jan 2008)*
2286 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2288 - add command aliases infrastructure and a few aliases
2289 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2290 man pages and the ``--help`` option of ``gnt-node
2291 list``/``gnt-instance list``
2292 - fix ``gnt-backup list`` with down nodes (issue 21)
2293 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2294 - fix the dist archive and add a check for including svn/git files in
2296 - some developer-related changes: improve the burnin and the QA suite,
2297 add an upload script for testing during development
2303 *(Released Wed, 16 Jan 2008)*
2305 - experimental HVM support, read the install document, section
2306 "Initializing the cluster"
2307 - allow for the PVM hypervisor per-instance kernel and initrd paths
2308 - add a new command ``gnt-cluster verify-disks`` which uses a new
2309 algorithm to improve the reconnection of the DRBD pairs if the device
2310 on the secondary node has gone away
2311 - make logical volume code auto-activate LVs at disk activation time
2312 - slightly improve the speed of activating disks
2313 - allow specification of the MAC address at instance creation time, and
2314 changing it later via ``gnt-instance modify``
2315 - fix handling of external commands that generate lots of output on
2317 - update documentation with regard to minimum version of DRBD8 supported
2323 *(Released Tue, 4 Dec 2007)*
2325 - Log the ``xm create`` output to the node daemon log on failure (to
2326 help diagnosing the error)
2327 - In debug mode, log all external commands output if failed to the logs
2328 - Change parsing of lvm commands to ignore stderr
2334 *(Released Wed, 28 Nov 2007)*
2336 - Another round of updates to the DRBD 8 code to deal with more failures
2337 in the replace secondary node operation
2338 - Some more logging of failures in disk operations (lvm, drbd)
2339 - A few documentation updates
2346 *(Released Tue, 13 Nov 2007)*
2348 - Change configuration file format from Python's Pickle to JSON.
2349 Upgrading is possible using the cfgupgrade utility.
2350 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2351 faster replace disks and is more stable (DRBD 8 has many improvements
2352 compared to DRBD 0.7)
2353 - Added command line tags support (see man pages for ``gnt-instance``,
2354 ``gnt-node``, ``gnt-cluster``)
2355 - Added instance rename support
2356 - Added multi-instance startup/shutdown
2357 - Added cluster rename support
2358 - Added ``gnt-node evacuate`` to simplify some node operations
2359 - Added instance reboot operation that can speedup reboot as compared to
2361 - Soften the requirement that hostnames are in FQDN format
2362 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2364 - Removed dependency on debian's patched fping that uses the
2365 non-standard ``-S`` option
2366 - Now the OS definitions are searched for in multiple, configurable
2367 paths (easier for distros to package)
2368 - Some changes to the hooks infrastructure (especially the new
2369 post-configuration update hook)
2370 - Other small bugfixes
2372 .. vim: set textwidth=72 syntax=rst :