8 *(Released Fri, 12 Oct 2012)*
10 A small bugfix release. Among the bugs fixed:
12 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
13 made the command unusable.
14 - Commands that issue many jobs don't fail anymore just because some jobs
15 take so long that other jobs are archived.
16 - Failures during ``gnt-instance reinstall`` are reflected by the exit
18 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
20 - When ``always_failover`` is set, ``--allow-failover`` is not required
21 in migrate commands anymore.
22 - ``bash_completion`` works even if extglob is disabled.
23 - Fixed bug with locks that made failover for RDB-based instances fail.
24 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
25 a random node instead of one based on the allocator metric.
26 - Support for newer versions of pylint and pep8.
27 - Hail doesn't fail anymore when trying to add an instance of type
28 ``file``, ``sharedfile`` or ``rbd``.
29 - Added new Makefile target to rebuild the whole distribution, so that
30 all files are included.
36 *(Released Fri, 27 Jul 2012)*
39 .. attention:: The ``LUXI`` protocol has been made more consistent
40 regarding its handling of command arguments. This, however, leads to
41 incompatibility issues with previous versions. Please ensure that you
42 restart Ganeti daemons soon after the upgrade, otherwise most
43 ``LUXI`` calls (job submission, setting/resetting the drain flag,
44 pausing/resuming the watcher, cancelling and archiving jobs, querying
45 the cluster configuration) will fail.
54 The current ``admin_up`` field, which used to denote whether an instance
55 should be running or not, has been removed. Instead, ``admin_state`` is
56 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
58 The rational behind this is that an instance being “down” can have
61 - it could be down during a reboot
62 - it could be temporarily be down for a reinstall
63 - or it could be down because it is deprecated and kept just for its
66 The previous Boolean state was making it difficult to do capacity
67 calculations: should Ganeti reserve memory for a down instance? Now, the
68 tri-state field makes it clear:
70 - in ``up`` and ``down`` state, all resources are reserved for the
71 instance, and it can be at any time brought up if it is down
72 - in ``offline`` state, only disk space is reserved for it, but not
75 The field can have an extra use: since the transition between ``up`` and
76 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
77 transition between ``offline`` and ``down`` is done via ``gnt-instance
78 modify``, it is possible to given different rights to users. For
79 example, owners of an instance could be allowed to start/stop it, but
80 not transition it out of the offline state.
82 Instance policies and specs
83 +++++++++++++++++++++++++++
85 In previous Ganeti versions, an instance creation request was not
86 limited on the minimum size and on the maximum size just by the cluster
87 resources. As such, any policy could be implemented only in third-party
88 clients (RAPI clients, or shell wrappers over ``gnt-*``
89 tools). Furthermore, calculating cluster capacity via ``hspace`` again
90 required external input with regards to instance sizes.
92 In order to improve these workflows and to allow for example better
93 per-node group differentiation, we introduced instance specs, which
96 - minimum instance disk size, disk count, memory size, cpu count
97 - maximum values for the above metrics
98 - and “standard” values (used in ``hspace`` to calculate the standard
101 The minimum/maximum values can be also customised at node-group level,
102 for example allowing more powerful hardware to support bigger instance
105 Beside the instance specs, there are a few other settings belonging to
106 the instance policy framework. It is possible now to customise, per
107 cluster and node-group:
109 - the list of allowed disk templates
110 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
111 - the maximum ratio of instance to spindles (see below for more
112 information) for local storage
114 All these together should allow all tools that talk to Ganeti to know
115 what are the ranges of allowed values for instances and the
116 over-subscription that is allowed.
118 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
119 instance configuration, and the physical CPU configuration from the
120 node. For the spindle ratios however, we didn't track before these
121 values, so new parameters have been added:
123 - a new node parameter ``spindle_count``, defaults to 1, customisable at
124 node group or node level
125 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
127 Note that spindles in this context doesn't need to mean actual
128 mechanical hard-drives; it's just a relative number for both the node
129 I/O capacity and instance I/O consumption.
131 Instance migration behaviour
132 ++++++++++++++++++++++++++++
134 While live-migration is in general desirable over failover, it is
135 possible that for some workloads it is actually worse, due to the
136 variable time of the “suspend” phase during live migration.
138 To allow the tools to work consistently over such instances (without
139 having to hard-code instance names), a new backend parameter
140 ``always_failover`` has been added to control the migration/failover
141 behaviour. When set to True, all migration requests for an instance will
142 instead fall-back to failover.
144 Instance memory ballooning
145 ++++++++++++++++++++++++++
147 Initial support for memory ballooning has been added. The memory for an
148 instance is no longer fixed (backend parameter ``memory``), but instead
149 can vary between minimum and maximum values (backend parameters
150 ``minmem`` and ``maxmem``). Currently we only change an instance's
153 - live migrating or failing over and instance and the target node
154 doesn't have enough memory
155 - user requests changing the memory via ``gnt-instance modify
161 In order to control the use of specific CPUs by instance, support for
162 controlling CPU pinning has been added for the Xen, HVM and LXC
163 hypervisors. This is controlled by a new hypervisor parameter
164 ``cpu_mask``; details about possible values for this are in the
165 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
166 VCPU-to-CPU mapping) form will work well only when all nodes in your
167 cluster have the same amount of CPUs.
172 Another area in which Ganeti was not customisable were the parameters
173 used for storage configuration, e.g. how many stripes to use for LVM,
174 DRBD resync configuration, etc.
176 To improve this area, we've added disks parameters, which are
177 customisable at cluster and node group level, and which allow to
178 specify various parameters for disks (DRBD has the most parameters
179 currently), for example:
181 - DRBD resync algorithm and parameters (e.g. speed)
182 - the default VG for meta-data volumes for DRBD
183 - number of stripes for LVM (plain disk template)
186 These parameters can be modified via ``gnt-cluster modify -D …`` and
187 ``gnt-group modify -D …``, and are used at either instance creation (in
188 case of LVM stripes, for example) or at disk “activation” time
191 Rados block device support
192 ++++++++++++++++++++++++++
194 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
195 denoted by the ``rbd`` disk template type. This is considered
196 experimental, feedback is welcome. For details on configuring it, see
197 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
202 The existing master IP functionality works well only in simple setups (a
203 single network shared by all nodes); however, if nodes belong to
204 different networks, then the ``/32`` setup and lack of routing
205 information is not enough.
207 To allow the master IP to function well in more complex cases, the
208 system was reworked as follows:
210 - a master IP netmask setting has been added
211 - the master IP activation/turn-down code was moved from the node daemon
213 - whether to run the Ganeti-supplied master IP script or a user-supplied
214 on is a ``gnt-cluster init`` setting
216 Details about the location of the standard and custom setup scripts are
217 in the man page :manpage:`gnt-cluster(8)`; for information about the
218 setup script protocol, look at the Ganeti-supplied script.
223 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
226 It is now possible to use TLS-protected connections, and when renewing
227 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
228 it is now possible to specify spice or spice CA certificates. Also, it
229 is possible to configure a password for SPICE sessions via the
230 hypervisor parameter ``spice_password_file``.
232 There are also new parameters to control the compression and streaming
233 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
234 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
235 for the spice parameters.
237 Lastly, it is now possible to see the SPICE connection information via
238 ``gnt-instance console``.
243 A new tool (``tools/ovfconverter``) has been added that supports
244 conversion between Ganeti and the `Open Virtualization Format
245 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
248 This relies on the ``qemu-img`` tool to convert the disk formats, so the
249 actual compatibility with other virtualization solutions depends on it.
254 The configuration query daemon (``ganeti-confd``) is now optional, and
255 has been rewritten in Haskell; whether to use the daemon at all, use the
256 Python (default) or the Haskell version is selectable at configure time
257 via the ``--enable-confd`` parameter, which can take one of the
258 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
259 daemon will result in a smaller footprint; for larger systems, we
260 welcome feedback on the Haskell version which might become the default
263 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
264 daemon running. The Python version doesn't implement the new call.
267 User interface changes
268 ~~~~~~~~~~~~~~~~~~~~~~
270 We have replaced the ``--disks`` option of ``gnt-instance
271 replace-disks`` with a more flexible ``--disk`` option, which allows
272 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
273 disk size and mode can be changed upon recreation (via ``gnt-instance
274 recreate-disks``, which accepts the same ``--disk`` option).
276 As many people are used to a ``show`` command, we have added that as an
277 alias to ``info`` on all ``gnt-*`` commands.
279 The ``gnt-instance grow-disk`` command has a new mode in which it can
280 accept the target size of the disk, instead of the delta; this can be
281 more safe since two runs in absolute mode will be idempotent, and
282 sometimes it's also easier to specify the desired size directly.
284 Also the handling of instances with regard to offline secondaries has
285 been improved. Instance operations should not fail because one of it's
286 secondary nodes is offline, even though it's safe to proceed.
288 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
289 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
290 minors to instance name.
295 RAPI coverage has improved, with (for example) new resources for
296 recreate-disks, node power-cycle, etc.
301 There is partial support for ``xl`` in the Xen hypervisor; feedback is
304 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
305 whether to still support Python 2.4 or move to Python 2.6 as minimum
308 Support for Fedora has been slightly improved; the provided example
309 init.d script should work better on it and the INSTALL file should
310 document the needed dependencies.
315 The deprecated ``QueryLocks`` LUXI request has been removed. Use
316 ``Query(what=QR_LOCK, ...)`` instead.
318 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
319 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
320 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
321 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
322 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
324 RAPI client: ``CertificateError`` now derives from
325 ``GanetiApiError``. This should make it more easy to handle Ganeti
328 Deprecation warnings due to PyCrypto/paramiko import in
329 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
330 make sure to run an up-to-date paramiko version, if you use this tool.
332 The QA scripts now depend on Python 2.5 or above (the main code base
333 still works with Python 2.4).
335 The configuration file (``config.data``) is now written without
336 indentation for performance reasons; if you want to edit it, it can be
337 re-formatted via ``tools/fmtjson``.
339 A number of bugs has been fixed in the cluster merge tool.
341 ``x509`` certification verification (used in import-export) has been
342 changed to allow the same clock skew as permitted by the cluster
343 verification. This will remove some rare but hard to diagnose errors in
350 *(Released Thu, 19 Jul 2012)*
352 Very few changes from rc4 to the final release, only bugfixes:
354 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
355 instance, fix CDROM booting for KVM instances)
356 - fixed node group modification of node parameters
357 - fixed issue in LUClusterVerifyGroup with multi-group clusters
358 - fixed generation of bash completion to ensure a stable ordering
365 *(Released Fri, 13 Jul 2012)*
367 Third release candidate for 2.6. The following changes were done from
370 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
371 - Fixed an inconsistency in the LUXI protocol with the provided
372 arguments (NOT backwards compatible)
373 - Fixed a bug with node groups ipolicy where ``min`` was greater than
374 the cluster ``std`` value
375 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
376 easier instance debugging on nodes (requires ``hconfd`` to work)
382 *(Released Tue, 03 Jul 2012)*
384 Second release candidate for 2.6. The following changes were done from
387 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
389 - Fixed a RAPI regression on missing beparams/memory
390 - Fixed redistribution of files on offline nodes
391 - Added possibility to run activate-disks even though secondaries are
392 offline. With this change it relaxes also the strictness on some other
393 commands which use activate disks internally:
394 * ``gnt-instance start|reboot|rename|backup|export``
395 - Made it possible to remove safely an instance if its secondaries are
397 - Made it possible to reinstall even though secondaries are offline
403 *(Released Mon, 25 Jun 2012)*
405 First release candidate for 2.6. The following changes were done from
408 - Fixed bugs with disk parameters and ``rbd`` templates as well as
410 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
411 behaviour. It supports now the modify operation
412 - ``hcheck`` implemented to analyze cluster health and possibility of
413 improving health by rebalance
414 - ``hbal`` has been improved in dealing with split instances
420 *(Released Mon, 11 Jun 2012)*
422 Second beta release of 2.6. The following changes were done from beta2
425 - Fixed ``daemon-util`` with non-root user models
426 - Fixed creation of plain instances with ``--no-wait-for-sync``
427 - Fix wrong iv_names when running ``cfgupgrade``
428 - Export more information in RAPI group queries
429 - Fixed bug when changing instance network interfaces
430 - Extended burnin to do NIC changes
431 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
432 - Changed default for DRBD barriers
433 - Fixed DRBD error reporting for syncer rate
434 - Verify the options on disk parameters
436 And of course various fixes to documentation and improved unittests and
443 *(Released Wed, 23 May 2012)*
445 First beta release of 2.6. The following changes were done from beta1 to
448 - integrated patch for distributions without ``start-stop-daemon``
449 - adapted example init.d script to work on Fedora
450 - fixed log handling in Haskell daemons
451 - adapted checks in the watcher for pycurl linked against libnss
452 - add partial support for ``xl`` instead of ``xm`` for Xen
453 - fixed a type issue in cluster verification
454 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
457 Plus integrated fixes from the 2.5 branch:
459 - fixed ``kvm-ifup`` to use ``/bin/bash``
460 - fixed parallel build failures
461 - KVM live migration when using a custom keymap
467 *(Released Tue, 24 Jul 2012)*
469 A small bugfix release, with no new features:
471 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
472 different default shell (e.g. Debian, Ubuntu)
473 - fixed KVM startup and live migration with a custom keymap (fixes Issue
474 243 and Debian bug #650664)
475 - fixed compatibility with KVM versions that don't support multiple boot
476 devices (fixes Issue 230 and Debian bug #624256)
478 Additionally, a few fixes were done to the build system (fixed parallel
479 build failures) and to the unittests (fixed race condition in test for
480 FileID functions, and the default enable/disable mode for QA test is now
487 *(Released Fri, 11 May 2012)*
489 A small bugfix release.
491 The main issues solved are on the topic of compatibility with newer LVM
494 - fixed parsing of ``lv_attr`` field
495 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
496 the ``--force`` flag is needed
498 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
499 to accept kernel 3.x too (was hardcoded to 2.6.*).
501 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
502 of hook results and that also made display of other errors suboptimal
503 was fixed; the code behaves now like 2.4 and earlier.
505 Another change in 2.5, the cleanup of the OS scripts environment, is too
506 aggressive: it removed even the ``PATH`` variable, which requires the OS
507 scripts to *always* need to export it. Since this is a bit too strict,
508 we now export a minimal PATH, the same that we export for hooks.
510 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
511 integrated into this release.
513 Finally, a few other miscellaneous changes were done (no new features,
514 just small improvements):
516 - Fix ``gnt-group --help`` display
517 - Fix hardcoded Xen kernel path
518 - Fix grow-disk handling of invalid units
519 - Update synopsis for ``gnt-cluster repair-disk-sizes``
520 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
526 *(Released Thu, 12 Apr 2012)*
528 Incompatible/important changes and bugfixes
529 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
531 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
532 resource's ``ip_check`` parameter changed from ``True`` to ``False``
533 to match the underlying LUXI interface.
534 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
535 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
536 not maintain backwards-compatibility as the underlying operation
537 changed in an incompatible way. The RAPI client can talk to old
538 servers, but it needs to be told so as the return value changed.
539 - When creating file-based instances via RAPI, the ``file_driver``
540 parameter no longer defaults to ``loop`` and must be specified.
541 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
543 - Support for the undocumented and deprecated RAPI instance creation
544 request format version 0 has been dropped. Use version 1, supported
545 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
546 - Pyparsing 1.4.6 or above is required, see :doc:`installation
547 documentation <install>`.
548 - The "cluster-verify" hooks are now executed per group by the
549 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
550 if you just run ``gnt-cluster verify``, which generates one opcode per
552 - The environment as passed to the OS scripts is cleared, and thus no
553 environment variables defined in the node daemon's environment will be
554 inherited by the scripts.
555 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
557 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
558 support operations involving multiple node groups.
559 - Offline nodes are ignored when failing over an instance.
560 - Support for KVM version 1.0, which changed the version reporting format
562 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
564 - ``Makefile`` is now compatible with Automake 1.11.2
565 - Includes all bugfixes made in the 2.4 series
570 - The ganeti-htools project has been merged into the ganeti-core source
571 tree and will be built as part of Ganeti (see :doc:`install-quick`).
572 - Implemented support for :doc:`shared storage <design-shared-storage>`.
573 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
574 GPT-style partition tables (requires `parted
575 <http://www.gnu.org/s/parted/>`_).
576 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
577 - Allowed adding tags on instance creation.
578 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
579 - Allow instances to be started in a paused state, enabling the user to
580 see the complete console output on boot using the console.
581 - Added new hypervisor flag to control default reboot behaviour
582 (``reboot_behavior``).
583 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
584 - Improved out-of-band management support:
586 - Added ``gnt-node health`` command reporting the health status of
588 - Added ``gnt-node power`` command to manage power status of nodes.
589 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
591 - Instance migration can fall back to failover if instance is not
593 - Filters can be used when listing nodes, instances, groups and locks;
594 see *ganeti(7)* manpage.
595 - Added post-execution status as variables to :doc:`hooks <hooks>`
597 - Instance tags are exported/imported together with the instance.
598 - When given an explicit job ID, ``gnt-job info`` will work for archived
600 - Jobs can define dependencies on other jobs (not yet supported via
601 RAPI or command line, but used by internal commands and usable via
604 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
607 - Instance failover is now available as a RAPI resource
608 (``/2/instances/[instance_name]/failover``).
609 - ``gnt-instance info`` defaults to static information if primary node
611 - Opcodes have a new ``comment`` attribute.
612 - Added basic SPICE support to KVM hypervisor.
613 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
615 Node group improvements
616 ~~~~~~~~~~~~~~~~~~~~~~~
618 - ``gnt-cluster verify`` has been modified to check groups separately,
619 thereby improving performance.
620 - Node group support has been added to ``gnt-cluster verify-disks``,
621 which now operates per node group.
622 - Watcher has been changed to work better with node groups.
624 - One process and state file per node group.
625 - Slow watcher in one group doesn't block other group's watcher.
627 - Added new command, ``gnt-group evacuate``, to move all instances in a
628 node group to other groups.
629 - Added ``gnt-instance change-group`` to move an instance to another
631 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
632 per-group operations.
633 - Node groups can be tagged.
634 - Some operations switch from an exclusive to a shared lock as soon as
636 - Instance's primary and secondary nodes' groups are now available as
637 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
638 and ``snodes.group.uuid``).
643 - Numerous updates to documentation and manpages.
645 - :doc:`RAPI <rapi>` documentation now has detailed parameter
647 - Some opcode/job results are now also documented, see :doc:`RAPI
650 - A lockset's internal lock is now also visible in lock monitor.
651 - Log messages from job queue workers now contain information about the
652 opcode they're processing.
653 - ``gnt-instance console`` no longer requires the instance lock.
654 - A short delay when waiting for job changes reduces the number of LUXI
655 requests significantly.
656 - DRBD metadata volumes are overwritten with zeros during disk creation.
657 - Out-of-band commands no longer acquire the cluster lock in exclusive
659 - ``devel/upload`` now uses correct permissions for directories.
665 *(Released Fri, 23 Mar 2012)*
667 This was the sixth release candidate of the 2.5 series.
673 *(Released Mon, 9 Jan 2012)*
675 This was the fifth release candidate of the 2.5 series.
681 *(Released Thu, 27 Oct 2011)*
683 This was the fourth release candidate of the 2.5 series.
689 *(Released Wed, 26 Oct 2011)*
691 This was the third release candidate of the 2.5 series.
697 *(Released Tue, 18 Oct 2011)*
699 This was the second release candidate of the 2.5 series.
705 *(Released Tue, 4 Oct 2011)*
707 This was the first release candidate of the 2.5 series.
713 *(Released Wed, 31 Aug 2011)*
715 This was the third beta release of the 2.5 series.
721 *(Released Mon, 22 Aug 2011)*
723 This was the second beta release of the 2.5 series.
729 *(Released Fri, 12 Aug 2011)*
731 This was the first beta release of the 2.5 series.
737 *(Released Thu, 27 Oct 2011)*
739 - Fixed bug when parsing command line parameter values ending in
741 - Fixed assertion error after unclean master shutdown
742 - Disable HTTP client pool for RPC, significantly reducing memory usage
744 - Fixed queue archive creation with wrong permissions
750 *(Released Tue, 23 Aug 2011)*
754 - Fixed documentation for importing with ``--src-dir`` option
755 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
756 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
762 *(Released Fri, 5 Aug 2011)*
764 Many bug-fixes and a few small features:
766 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
767 issues when you tried to add an instance with two MAC addresses in one
769 - KVM: fixed per-instance stored UID value
770 - KVM: configure bridged NICs at migration start
771 - KVM: Fix a bug where instance will not start with never KVM versions
773 - Added OS search path to ``gnt-cluster info``
774 - Fixed an issue with ``file_storage_dir`` where you were forced to
775 provide an absolute path, but the documentation states it is a
776 relative path, the documentation was right
777 - Added a new parameter to instance stop/start called ``--no-remember``
778 that will make the state change to not be remembered
779 - Implemented ``no_remember`` at RAPI level
780 - Improved the documentation
781 - Node evacuation: don't call IAllocator if node is already empty
782 - Fixed bug in DRBD8 replace disks on current nodes
783 - Fixed bug in recreate-disks for DRBD instances
784 - Moved assertion checking locks in ``gnt-instance replace-disks``
785 causing it to abort with not owning the right locks for some situation
786 - Job queue: Fixed potential race condition when cancelling queued jobs
787 - Fixed off-by-one bug in job serial generation
788 - ``gnt-node volumes``: Fix instance names
789 - Fixed aliases in bash completion
790 - Fixed a bug in reopening log files after being sent a SIGHUP
791 - Added a flag to burnin to allow specifying VCPU count
792 - Bugfixes to non-root Ganeti configuration
798 *(Released Thu, 12 May 2011)*
800 Many bug-fixes and a few new small features:
802 - Fixed a bug related to log opening failures
803 - Fixed a bug in instance listing with orphan instances
804 - Fixed a bug which prevented resetting the cluster-level node parameter
805 ``oob_program`` to the default
806 - Many fixes related to the ``cluster-merge`` tool
807 - Fixed a race condition in the lock monitor, which caused failures
808 during (at least) creation of many instances in parallel
809 - Improved output for gnt-job info
810 - Removed the quiet flag on some ssh calls which prevented debugging
812 - Improved the N+1 failure messages in cluster verify by actually
813 showing the memory values (needed and available)
814 - Increased lock attempt timeouts so that when executing long operations
815 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
816 too early and thus prevent the use of the 'fair' mechanism
817 - Changed instance query data (``gnt-instance info``) to not acquire
818 locks unless needed, thus allowing its use on locked instance if only
819 static information is asked for
820 - Improved behaviour with filesystems that do not support rename on an
822 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
823 kept locks on all nodes during the wipe, which is unneeded
824 - Fixed ``gnt-watcher`` handling of errors during hooks execution
825 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
826 10GiB) which caused the wipe to fail right at the end in some cases
827 - Fixed master IP activation when doing master failover with no-voting
828 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
829 the master node itself
830 - Fixed potential data-loss in under disk full conditions, where Ganeti
831 wouldn't check correctly the return code and would consider
832 partially-written files 'correct'
833 - Fixed bug related to multiple VGs and DRBD disk replacing
834 - Added new disk parameter ``metavg`` that allows placement of the meta
835 device for DRBD in a different volume group
836 - Fixed error handling in the node daemon when the system libc doesn't
837 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
838 - Fixed lock release during replace-disks, which kept cluster-wide locks
839 when doing disk replaces with an iallocator script
840 - Added check for missing bridges in cluster verify
841 - Handle EPIPE errors while writing to the terminal better, so that
842 piping the output to e.g. ``less`` doesn't cause a backtrace
843 - Fixed rare case where a ^C during Luxi calls could have been
844 interpreted as server errors, instead of simply terminating
845 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
847 - Added a few more parameters to the KVM hypervisor, allowing a second
848 CDROM, custom disk type for CDROMs and a floppy image
849 - Removed redundant message in instance rename when the name is given
851 - Added option to ``gnt-instance recreate-disks`` to allow creating the
852 disks on new nodes, allowing recreation when the original instance
853 nodes are completely gone
854 - Added option when converting disk templates to DRBD to skip waiting
855 for the resync, in order to make the instance available sooner
856 - Added two new variables to the OS scripts environment (containing the
858 - Made the root_path and optional parameter for the xen-pvm hypervisor,
859 to allow use of ``pvgrub`` as bootloader
860 - Changed the instance memory modifications to only check out-of-memory
861 conditions on memory increases, and turned the secondary node warnings
862 into errors (they can still be overridden via ``--force``)
863 - Fixed the handling of a corner case when the Python installation gets
864 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
865 to execute a command that doesn't exist
866 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
867 node of the instance returned failures during instance shutdown; this
868 adds the option ``--ignore-consistency`` to gnt-instance move
870 And as usual, various improvements to the error messages, documentation
877 *(Released Wed, 09 Mar 2011)*
879 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
880 the RAPI users file if run twice (even with ``--dry-run``).
882 The release fixes that bug (nothing else changed).
888 *(Released Mon, 07 Mar 2011)*
890 Final 2.4.0 release. Just a few small fixes:
892 - Fixed RAPI node evacuate
893 - Fixed the kvm-ifup script
894 - Fixed internal error handling for special job cases
895 - Updated man page to specify the escaping feature for options
901 *(Released Mon, 28 Feb 2011)*
903 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
904 that this feature wiped the disks of the wrong instance, leading to loss
909 - Fixed title of query field containing instance name
910 - Expanded the glossary in the documentation
911 - Fixed one unittest (internal issue)
917 *(Released Mon, 21 Feb 2011)*
919 A number of bug fixes plus just a couple functionality changes.
921 On the user-visible side, the ``gnt-* list`` command output has changed
922 with respect to "special" field states. The current rc1 style of display
923 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
924 the default output mode special fields states are displayed as follows:
926 - Offline resource: ``*``
927 - Unavailable/not applicable: ``-``
928 - Data missing (RPC failure): ``?``
929 - Unknown field: ``??``
931 Another user-visible change is the addition of ``--force-join`` to
936 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
937 - Fixed regression in RAPI/instance reinstall where all parameters were
938 required (instead of optional)
939 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
940 - Fixed iallocator usage (offline nodes were not considered offline)
941 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
942 - Fixed hypervisor and OS parameter validation with respect to
944 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
946 - Fixed ``tools/listrunner`` with respect to agent-based usage
952 *(Released Fri, 4 Feb 2011)*
954 Many changes and fixes since the beta1 release. While there were some
955 internal changes, the code has been mostly stabilised for the RC
958 Note: the dumb allocator was removed in this release, as it was not kept
959 up-to-date with the IAllocator protocol changes. It is recommended to
960 use the ``hail`` command from the ganeti-htools package.
962 Note: the 2.4 and up versions of Ganeti are not compatible with the
963 0.2.x branch of ganeti-htools. You need to upgrade to
964 ganeti-htools-0.3.0 (or later).
966 Regressions fixed from 2.3
967 ~~~~~~~~~~~~~~~~~~~~~~~~~~
969 - Fixed the ``gnt-cluster verify-disks`` command
970 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
972 - Fixed disk adoption breakage
973 - Fixed wrong headers in instance listing for field aliases
978 - Fixed corner case in KVM handling of NICs
979 - Fixed many cases of wrong handling of non-vm_capable nodes
980 - Fixed a bug where a missing instance symlink was not possible to
981 recreate with any ``gnt-*`` command (now ``gnt-instance
982 activate-disks`` does it)
983 - Fixed the volume group name as reported by ``gnt-cluster
985 - Increased timeouts for the import-export code, hopefully leading to
986 fewer aborts due network or instance timeouts
987 - Fixed bug in ``gnt-node list-storage``
988 - Fixed bug where not all daemons were started on cluster
989 initialisation, but only at the first watcher run
990 - Fixed many bugs in the OOB implementation
991 - Fixed watcher behaviour in presence of instances with offline
993 - Fixed instance list output for instances running on the wrong node
994 - a few fixes to the cluster-merge tool, but it still cannot merge
995 multi-node groups (currently it is not recommended to use this tool)
1001 - Improved network configuration for the KVM hypervisor
1002 - Added e1000 as a supported NIC for Xen-HVM
1003 - Improved the lvmstrap tool to also be able to use partitions, as
1004 opposed to full disks
1005 - Improved speed of disk wiping (the cluster parameter
1006 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1007 of instance creations
1008 - Added documentation for the OS parameters
1009 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1010 hypervisor is not responding
1011 - Added display of blacklisted and hidden OS information in
1012 ``gnt-cluster info``
1013 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1014 NIC and node parameters, which might create problems with currently
1015 invalid (but undetected) configuration files, but prevents validation
1016 failures when unrelated parameters are modified
1017 - Changed cluster initialisation to wait for the master daemon to become
1019 - Expanded the RAPI interface:
1021 - Added config redistribution resource
1022 - Added activation/deactivation of instance disks
1023 - Added export of console information
1025 - Implemented log file reopening on SIGHUP, which allows using
1026 logrotate(8) for the Ganeti log files
1027 - Added a basic OOB helper script as an example
1033 *(Released Fri, 14 Jan 2011)*
1038 - Fixed timezone issues when formatting timestamps
1039 - Added support for node groups, available via ``gnt-group`` and other
1041 - Added out-of-band framework and management, see :doc:`design
1042 document <design-oob>`
1043 - Removed support for roman numbers from ``gnt-node list`` and
1044 ``gnt-instance list``.
1045 - Allowed modification of master network interface via ``gnt-cluster
1046 modify --master-netdev``
1047 - Accept offline secondaries while shutting down instance disks
1048 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1049 - Added support for multiple LVM volume groups
1050 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1052 - Added commands to list available fields:
1054 - ``gnt-node list-fields``
1055 - ``gnt-group list-fields``
1056 - ``gnt-instance list-fields``
1058 - Updated documentation and man pages
1063 - Moved ``rapi_users`` file into separate directory, now named
1064 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1066 - Added new tool for running commands on many machines,
1067 ``tools/ganeti-listrunner``
1068 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1069 improving the ``gnt-instance console`` output
1070 - Allowed customisation of disk index separator at ``configure`` time
1071 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1072 - Added support for non-partitioned md disks in ``lvmstrap``
1073 - Added script to gracefully power off KVM instances
1074 - Split ``utils`` module into smaller parts
1075 - Changed query operations to return more detailed information, e.g.
1076 whether an information is unavailable due to an offline node. To use
1077 this new functionality, the LUXI call ``Query`` must be used. Field
1078 information is now stored by the master daemon and can be retrieved
1079 using ``QueryFields``. Instances, nodes and groups can also be queried
1080 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1081 exposed via RAPI). The following commands make use of this
1082 infrastructure change:
1084 - ``gnt-group list``
1085 - ``gnt-group list-fields``
1087 - ``gnt-node list-fields``
1088 - ``gnt-instance list``
1089 - ``gnt-instance list-fields``
1090 - ``gnt-debug locks``
1095 - New RAPI resources (see :doc:`rapi`):
1099 - ``/2/groups/[group_name]``
1100 - ``/2/groups/[group_name]/assign-nodes``
1101 - ``/2/groups/[group_name]/modify``
1102 - ``/2/groups/[group_name]/rename``
1103 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1107 - Implemented ``no_install`` for instance creation
1108 - Implemented OS parameters for instance reinstallation, allowing
1109 use of special settings on reinstallation (e.g. for preserving data)
1114 - Added IPv6 support in import/export
1115 - Pause DRBD synchronization while wiping disks on instance creation
1116 - Updated unittests and QA scripts
1117 - Improved network parameters passed to KVM
1118 - Converted man pages from docbook to reStructuredText
1124 *(Released Mon, 20 Dec 2010)*
1126 Released version 2.3.1~rc1 without any changes.
1132 *(Released Wed, 1 Dec 2010)*
1134 - impexpd: Disable OpenSSL compression in socat if possible (backport
1135 from master, commit e90739d625b, see :doc:`installation guide
1136 <install-quick>` for details)
1137 - Changed unittest coverage report to exclude test scripts
1138 - Added script to check version format
1144 *(Released Wed, 1 Dec 2010)*
1146 Released version 2.3.0~rc1 without any changes.
1152 *(Released Fri, 19 Nov 2010)*
1154 A number of bugfixes and documentation updates:
1156 - Update ganeti-os-interface documentation
1157 - Fixed a bug related to duplicate MACs or similar items which should be
1159 - Fix breakage in OS state modify
1160 - Reinstall instance: disallow offline secondaries (fixes bug related to
1161 OS changing but reinstall failing)
1162 - plus all the other fixes between 2.2.1 and 2.2.2
1168 *(Released Tue, 2 Nov 2010)*
1170 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1171 - Fixed master failover race with watcher
1172 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1174 - Accept previously stopped instance for export with instance removal
1175 - Simplify and extend the environment variables for instance OS scripts
1176 - Added new node flags, ``master_capable`` and ``vm_capable``
1177 - Added optional instance disk wiping prior during allocation. This is a
1178 cluster-wide option and can be set/modified using
1179 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1180 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1181 :doc:`install-quick`
1182 - Added a new watcher option (``--ignore-pause``)
1183 - Added option to ignore offline node on instance start/stop
1184 (``--ignore-offline``)
1185 - Allow overriding OS parameters with ``gnt-instance reinstall``
1186 - Added ability to change node's secondary IP address using ``gnt-node
1188 - Implemented privilege separation for all daemons except
1189 ``ganeti-noded``, see ``configure`` options
1190 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1192 - Implemented job priorities (see ``ganeti(7)`` manpage)
1193 - Ignore failures while shutting down instances during failover from
1195 - Exit daemon's bootstrap process only once daemon is ready
1196 - Export more information via ``LUInstanceQuery``/remote API
1197 - Improved documentation, QA and unittests
1198 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1199 restart if the file was created or changed
1200 - Added LUXI protocol version sent with each request and response,
1201 allowing detection of server/client mismatches
1202 - Moved the Python scripts among gnt-* and ganeti-* into modules
1203 - Moved all code related to setting up SSH to an external script,
1205 - Infrastructure changes for node group support in future versions
1211 *(Released Fri, 19 Nov 2010)*
1213 A few small bugs fixed, and some improvements to the build system:
1215 - Fix documentation regarding conversion to drbd
1216 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1218 - Fix error handling in node modify with multiple changes
1219 - Allow remote imports without checked names
1225 *(Released Tue, 19 Oct 2010)*
1227 - Disable SSL session ID cache in RPC client
1233 *(Released Thu, 14 Oct 2010)*
1235 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1236 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1238 - Fix problems with interaction between Curl and Python's HTTP server,
1239 resulting in increased speed in many RPC calls
1240 - Improve our release script to prevent breakage with older aclocal and
1247 *(Released Thu, 7 Oct 2010)*
1249 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1250 value retrieved from master
1251 - Added support for blacklisted or hidden OS definitions
1252 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1253 - Added support for -mem-path in KVM hypervisor abstraction layer
1254 - Allow overriding instance parameters in tool for inter-cluster
1255 instance moves (``tools/move-instance``)
1256 - Improved opcode summaries (e.g. in ``gnt-job list``)
1257 - Improve consistency of OS listing by sorting it
1258 - Documentation updates
1264 *(Released Fri, 8 Oct 2010)*
1266 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1272 *(Released Mon, 4 Oct 2010)*
1274 - Fixed regression in ``gnt-instance rename``
1280 *(Released Wed, 22 Sep 2010)*
1282 - Fixed OS_VARIANT variable for OS scripts
1283 - Fixed cluster tag operations via RAPI
1284 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1285 - Disabled RAPI CA checks in watcher
1291 *(Released Mon, 23 Aug 2010)*
1293 - Support DRBD versions of the format "a.b.c.d"
1295 - Re-introduce support for usage from multiple threads in RAPI client
1296 - Instance renames and modify via RAPI
1297 - Work around race condition between processing and archival in job
1299 - Mark opcodes following failed one as failed, too
1300 - Job field ``lock_status`` was removed due to difficulties making it
1301 work with the changed job queue in Ganeti 2.2; a better way to monitor
1302 locks is expected for a later 2.2.x release
1303 - Fixed dry-run behaviour with many commands
1304 - Support ``ssh-agent`` again when adding nodes
1305 - Many additional bugfixes
1311 *(Released Fri, 30 Jul 2010)*
1313 Important change: the internal RPC mechanism between Ganeti nodes has
1314 changed from using a home-grown http library (based on the Python base
1315 libraries) to use the PycURL library. This requires that PycURL is
1316 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1317 against GnuTLS by default. cURL's support for GnuTLS had known issues
1318 before cURL 7.21.0 and we recommend using the latest cURL release or
1319 linking against OpenSSL. Most other distributions already link PycURL
1320 and cURL against OpenSSL. The command::
1322 python -c 'import pycurl; print pycurl.version'
1324 can be used to determine the libraries PycURL and cURL are linked
1327 Other significant changes:
1329 - Rewrote much of the internals of the job queue, in order to achieve
1330 better parallelism; this decouples job query operations from the job
1331 processing, and it should allow much nicer behaviour of the master
1332 daemon under load, and it also has uncovered some long-standing bugs
1333 related to the job serialisation (now fixed)
1334 - Added a default iallocator setting to the cluster parameters,
1335 eliminating the need to always pass nodes or an iallocator for
1336 operations that require selection of new node(s)
1337 - Added experimental support for the LXC virtualization method
1338 - Added support for OS parameters, which allows the installation of
1339 instances to pass parameter to OS scripts in order to customise the
1341 - Added a hypervisor parameter controlling the migration type (live or
1342 non-live), since hypervisors have various levels of reliability; this
1343 has renamed the 'live' parameter to 'mode'
1344 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1345 logical volumes, meaning that cluster verify will ignore them and not
1346 flag their presence as errors
1347 - The watcher will now reset the error count for failed instances after
1348 8 hours, thus allowing self-healing if the problem that caused the
1349 instances to be down/fail to start has cleared in the meantime
1350 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1351 check for, and warn, if the drbd module parameter ``usermode_helper``
1352 is not consistent with the cluster-wide setting; this is needed to
1353 make diagnose easier of failed drbd creations
1354 - Started adding base IPv6 support, but this is not yet
1355 enabled/available for use
1356 - Rename operations (cluster, instance) will now return the new name,
1357 which is especially useful if a short name was passed in
1358 - Added support for instance migration in RAPI
1359 - Added a tool to pre-configure nodes for the SSH setup, before joining
1360 them to the cluster; this will allow in the future a simplified model
1361 for node joining (but not yet fully enabled in 2.2); this needs the
1362 paramiko python library
1363 - Fixed handling of name-resolving errors
1364 - Fixed consistency of job results on the error path
1365 - Fixed master-failover race condition when executed multiple times in
1367 - Fixed many bugs related to the job queue (mostly introduced during the
1368 2.2 development cycle, so not all are impacting 2.1)
1369 - Fixed instance migration with missing disk symlinks
1370 - Fixed handling of unknown jobs in ``gnt-job archive``
1371 - And many other small fixes/improvements
1375 - Enhanced both the unittest and the QA coverage
1376 - Switched the opcode validation to a generic model, and extended the
1377 validation to all opcode parameters
1378 - Changed more parts of the code that write shell scripts to use the
1380 - Switched the master daemon to use the asyncore library for the Luxi
1387 *(Released Thu, 17 Jun 2010)*
1389 - Added tool (``move-instance``) and infrastructure to move instances
1390 between separate clusters (see :doc:`separate documentation
1391 <move-instance>` and :doc:`design document <design-2.2>`)
1392 - Added per-request RPC timeout
1393 - RAPI now requires a Content-Type header for requests with a body (e.g.
1394 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1395 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1396 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1398 - Implemented initial support for running Ganeti daemons as separate
1399 users, see configure-time flags ``--with-user-prefix`` and
1400 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1402 - Instances can be removed after export (``gnt-backup export
1403 --remove-instance``)
1404 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1405 key (instead of 1024 bit)
1406 - Added new cluster configuration file for cluster domain secret
1407 - Import/export now use SSL instead of SSH
1408 - Added support for showing estimated time when exporting an instance,
1409 see the ``ganeti-os-interface(7)`` manpage and look for
1416 *(Released Tue, 16 Nov 2010)*
1418 Some more bugfixes. Unless critical bugs occur, this will be the last
1421 - Fix case of MAC special-values
1422 - Fix mac checker regex
1423 - backend: Fix typo causing "out of range" error
1424 - Add missing --units in gnt-instance list man page
1430 *(Released Tue, 24 Aug 2010)*
1433 - Don't ignore secondary node silently on non-mirrored disk templates
1435 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1436 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1437 - Properly document the usb_mouse parameter
1438 - Fix path in ganeti-rapi(8) (issue 116)
1439 - Adjust error message when the ganeti user's .ssh directory is
1441 - Add same-node-check when changing the disk template to drbd
1447 *(Released Fri, 16 Jul 2010)*
1450 - Add an option to only select some reboot types during qa/burnin.
1451 (on some hypervisors consequent reboots are not supported)
1452 - Fix infrequent race condition in master failover. Sometimes the old
1453 master ip address would be still detected as up for a short time
1454 after it was removed, causing failover to fail.
1455 - Decrease mlockall warnings when the ctypes module is missing. On
1456 Python 2.4 we support running even if no ctypes module is installed,
1457 but we were too verbose about this issue.
1458 - Fix building on old distributions, on which man doesn't have a
1460 - Fix RAPI not to ignore the MAC address on instance creation
1461 - Implement the old instance creation format in the RAPI client.
1467 *(Released Thu, 01 Jul 2010)*
1469 A small bugfix release:
1470 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1471 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1472 a non-existing option
1473 - Fix instance create: the --force-variant option was ignored
1474 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1475 - Fix modify node storage with non-FQDN arguments
1476 - Fix RAPI client to authenticate under Python 2.6 when used
1477 for more than 5 requests needing authentication
1478 - Fix gnt-instance modify -t (storage) giving a wrong error message
1479 when converting a non-shutdown drbd instance to plain
1485 *(Released Fri, 18 Jun 2010)*
1487 A small bugfix release:
1489 - Fix live migration of KVM instances started with older Ganeti
1490 versions which had fewer hypervisor parameters
1491 - Fix gnt-instance grow-disk on down instances
1492 - Fix an error-reporting bug during instance migration
1493 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1494 silently ignoring broken ones
1495 - Fix an RPC error reporting bug affecting, for example, RAPI client
1497 - Fix bug triggered by different API version os-es on different nodes
1498 - Fix a bug in instance startup with custom hvparams: OS level
1499 parameters would fail to be applied.
1500 - Fix the RAPI client under Python 2.6 (but more work is needed to
1501 make it work completely well with OpenSSL)
1502 - Fix handling of errors when resolving names from DNS
1508 *(Released Thu, 3 Jun 2010)*
1510 A medium sized development cycle. Some new features, and some
1511 fixes/small improvements/cleanups.
1513 Significant features
1514 ~~~~~~~~~~~~~~~~~~~~
1516 The node deamon now tries to mlock itself into memory, unless the
1517 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1518 its logs, and falls back to console logging. This allows emergency
1519 features such as ``gnt-node powercycle`` to work even in the event of a
1520 broken node disk (tested offlining the disk hosting the node's
1521 filesystem and dropping its memory caches; don't try this at home)
1523 KVM: add vhost-net acceleration support. It can be tested with a new
1524 enough version of the kernel and of qemu-kvm.
1526 KVM: Add instance chrooting feature. If you use privilege dropping for
1527 your VMs you can also now force them to chroot to an empty directory,
1528 before starting the emulated guest.
1530 KVM: Add maximum migration bandwith and maximum downtime tweaking
1531 support (requires a new-enough version of qemu-kvm).
1533 Cluster verify will now warn if the master node doesn't have the master
1534 ip configured on it.
1536 Add a new (incompatible) instance creation request format to RAPI which
1537 supports all parameters (previously only a subset was supported, and it
1538 wasn't possible to extend the old format to accomodate all the new
1539 features. The old format is still supported, and a client can check for
1540 this feature, before using it, by checking for its presence in the
1541 ``features`` RAPI resource.
1543 Now with ancient latin support. Try it passing the ``--roman`` option to
1544 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1545 (requires the python-roman module to be installed, in order to work).
1550 As usual many internal code refactorings, documentation updates, and
1553 - Lots of improvements and cleanups to the experimental Remote API
1554 (RAPI) client library.
1555 - A new unit test suite for the core daemon libraries.
1556 - A fix to creating missing directories makes sure the umask is not
1557 applied anymore. This enforces the same directory permissions
1559 - Better handling terminating daemons with ctrl+c (used when running
1560 them in debugging mode).
1561 - Fix a race condition in live migrating a KVM instance, when stat()
1562 on the old proc status file returned EINVAL, which is an unexpected
1564 - Fixed manpage checking with newer man and utf-8 charachters. But now
1565 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1571 *(Released Fri, 7 May 2010)*
1573 Fix a bug which prevented untagged KVM instances from starting.
1579 *(Released Fri, 7 May 2010)*
1581 Another release with a long development cycle, during which many
1582 different features were added.
1584 Significant features
1585 ~~~~~~~~~~~~~~~~~~~~
1587 The KVM hypervisor now can run the individual instances as non-root, to
1588 reduce the impact of a VM being hijacked due to bugs in the
1589 hypervisor. It is possible to run all instances as a single (non-root)
1590 user, to manually specify a user for each instance, or to dynamically
1591 allocate a user out of a cluster-wide pool to each instance, with the
1592 guarantee that no two instances will run under the same user ID on any
1595 An experimental RAPI client library, that can be used standalone
1596 (without the other Ganeti libraries), is provided in the source tree as
1597 ``lib/rapi/client.py``. Note this client might change its interface in
1598 the future, as we iterate on its capabilities.
1600 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1601 replace the cluster's certificates and crypto keys. This might help in
1602 case they have been compromised, or have simply expired.
1604 A new disk option for instance creation has been added that allows one
1605 to "adopt" currently existing logical volumes, with data
1606 preservation. This should allow easier migration to Ganeti from
1607 unmanaged (or managed via other software) instances.
1609 Another disk improvement is the possibility to convert between redundant
1610 (DRBD) and plain (LVM) disk configuration for an instance. This should
1611 allow better scalability (starting with one node and growing the
1612 cluster, or shrinking a two-node cluster to one node).
1614 A new feature that could help with automated node failovers has been
1615 implemented: if a node sees itself as offline (by querying the master
1616 candidates), it will try to shutdown (hard) all instances and any active
1617 DRBD devices. This reduces the risk of duplicate instances if an
1618 external script automatically failovers the instances on such nodes. To
1619 enable this, the cluster parameter ``maintain_node_health`` should be
1620 enabled; in the future this option (per the name) will enable other
1621 automatic maintenance features.
1623 Instance export/import now will reuse the original instance
1624 specifications for all parameters; that means exporting an instance,
1625 deleting it and the importing it back should give an almost identical
1626 instance. Note that the default import behaviour has changed from
1627 before, where it created only one NIC; now it recreates the original
1630 Cluster verify has added a few new checks: SSL certificates validity,
1631 /etc/hosts consistency across the cluster, etc.
1636 As usual, many internal changes were done, documentation fixes,
1639 - Fixed cluster initialization with disabled cluster storage (regression
1640 introduced in 2.1.1)
1641 - File-based storage supports growing the disks
1642 - Fixed behaviour of node role changes
1643 - Fixed cluster verify for some corner cases, plus a general rewrite of
1644 cluster verify to allow future extension with more checks
1645 - Fixed log spamming by watcher and node daemon (regression introduced
1647 - Fixed possible validation issues when changing the list of enabled
1649 - Fixed cleanup of /etc/hosts during node removal
1650 - Fixed RAPI response for invalid methods
1651 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1652 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1653 from ide disks, etc.)
1654 - Allow OS changes without re-installation (to record a changed OS
1655 outside of Ganeti, or to allow OS renames)
1656 - Allow instance creation without OS installation (useful for example if
1657 the OS will be installed manually, or restored from a backup not in
1659 - Implemented option to make cluster ``copyfile`` use the replication
1661 - Added list of enabled hypervisors to ssconf (possibly useful for
1663 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1665 - A partial form of node re-IP is possible via node readd, which now
1666 allows changed node primary IP
1667 - Command line utilities now show an informational message if the job is
1669 - The logs of the master daemon now show the PID/UID/GID of the
1676 *(Released Fri, 12 Mar 2010)*
1678 During the 2.1.0 long release candidate cycle, a lot of improvements and
1679 changes have accumulated with were released later as 2.1.1.
1684 The node evacuate command (``gnt-node evacuate``) was significantly
1685 rewritten, and as such the IAllocator protocol was changed - a new
1686 request type has been added. This unfortunate change during a stable
1687 series is designed to improve performance of node evacuations; on
1688 clusters with more than about five nodes and which are well-balanced,
1689 evacuation should proceed in parallel for all instances of the node
1690 being evacuated. As such, any existing IAllocator scripts need to be
1691 updated, otherwise the above command will fail due to the unknown
1692 request. The provided "dumb" allocator has not been updated; but the
1693 ganeti-htools package supports the new protocol since version 0.2.4.
1695 Another important change is increased validation of node and instance
1696 names. This might create problems in special cases, if invalid host
1697 names are being used.
1699 Also, a new layer of hypervisor parameters has been added, that sits at
1700 OS level between the cluster defaults and the instance ones. This allows
1701 customisation of virtualization parameters depending on the installed
1702 OS. For example instances with OS 'X' may have a different KVM kernel
1703 (or any other parameter) than the cluster defaults. This is intended to
1704 help managing a multiple OSes on the same cluster, without manual
1705 modification of each instance's parameters.
1707 A tool for merging clusters, ``cluster-merge``, has been added in the
1708 tools sub-directory.
1713 - Improved the int/float conversions that should make the code more
1714 robust in face of errors from the node daemons
1715 - Fixed the remove node code in case of internal configuration errors
1716 - Fixed the node daemon behaviour in face of inconsistent queue
1717 directory (e.g. read-only file-system where we can't open the files
1719 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1720 now it either aborts cleanly or, if given the new "auto_promote"
1721 parameter, will automatically promote other nodes as needed
1722 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1723 completely prevent Ganeti from working
1724 - Fixed bug for instance export when not all disks were successfully
1726 - Fixed behaviour of node add when the new node is slow in starting up
1728 - Fixed handling of signals in the LUXI client, which should improve
1729 behaviour of command-line scripts
1730 - Added checks for invalid node/instance names in the configuration (now
1731 flagged during cluster verify)
1732 - Fixed watcher behaviour for disk activation errors
1733 - Fixed two potentially endless loops in http library, which led to the
1734 RAPI daemon hanging and consuming 100% CPU in some cases
1735 - Fixed bug in RAPI daemon related to hashed passwords
1736 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1738 - Enhanced compatibility with non-Debian OSes, but not using absolute
1739 path in some commands and allowing customisation of the ssh
1740 configuration directory
1741 - Fixed possible future issue with new Python versions by abiding to the
1742 proper use of ``__slots__`` attribute on classes
1743 - Added checks that should prevent directory traversal attacks
1744 - Many documentation fixes based on feedback from users
1749 - Added an "early_release" more for instance replace disks and node
1750 evacuate, where we release locks earlier and thus allow higher
1751 parallelism within the cluster
1752 - Added watcher hooks, intended to allow the watcher to restart other
1753 daemons (e.g. from the ganeti-nbma project), but they can be used of
1754 course for any other purpose
1755 - Added a compile-time disable for DRBD barriers, to increase
1756 performance if the administrator trusts the power supply or the
1757 storage system to not lose writes
1758 - Added the option of using syslog for logging instead of, or in
1759 addition to, Ganeti's own log files
1760 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1761 can indeed boot from a paravirtual NIC
1762 - Added a generic debug level for many operations; while this is not
1763 used widely yet, it allows one to pass the debug value all the way to
1765 - Enhanced the hooks environment for instance moves (failovers,
1766 migrations) where the primary/secondary nodes changed during the
1767 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1768 - Enhanced data validations for many user-supplied values; one important
1769 item is the restrictions imposed on instance and node names, which
1770 might reject some (invalid) host names
1771 - Add a configure-time option to disable file-based storage, if it's not
1772 needed; this allows greater security separation between the master
1773 node and the other nodes from the point of view of the inter-node RPC
1775 - Added user notification in interactive tools if job is waiting in the
1776 job queue or trying to acquire locks
1777 - Added log messages when a job is waiting for locks
1778 - Added filtering by node tags in instance operations which admit
1779 multiple instances (start, stop, reboot, reinstall)
1780 - Added a new tool for cluster mergers, ``cluster-merge``
1781 - Parameters from command line which are of the form ``a=b,c=d`` can now
1782 use backslash escapes to pass in values which contain commas,
1783 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1785 - For KVM, the instance name is the first parameter passed to KVM, so
1786 that it's more visible in the process list
1792 *(Released Tue, 2 Mar 2010)*
1794 Ganeti 2.1 brings many improvements with it. Major changes:
1796 - Added infrastructure to ease automated disk repairs
1797 - Added new daemon to export configuration data in a cheaper way than
1798 using the remote API
1799 - Instance NICs can now be routed instead of being associated with a
1801 - Improved job locking logic to reduce impact of jobs acquiring multiple
1802 locks waiting for other long-running jobs
1804 In-depth implementation details can be found in the Ganeti 2.1 design
1810 - Added chroot hypervisor
1811 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1813 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1814 ``bootloader_path`` and ``bootloader_args``)
1815 - Added the ``use_localtime`` option for the xen-hvm and kvm
1816 hypervisors, and the default value for this has changed to false (in
1817 2.0 xen-hvm always enabled it)
1818 - Added luxi call to submit multiple jobs in one go
1819 - Added cluster initialization option to not modify ``/etc/hosts``
1821 - Added network interface parameters
1822 - Added dry run mode to some LUs
1823 - Added RAPI resources:
1825 - ``/2/instances/[instance_name]/info``
1826 - ``/2/instances/[instance_name]/replace-disks``
1827 - ``/2/nodes/[node_name]/evacuate``
1828 - ``/2/nodes/[node_name]/migrate``
1829 - ``/2/nodes/[node_name]/role``
1830 - ``/2/nodes/[node_name]/storage``
1831 - ``/2/nodes/[node_name]/storage/modify``
1832 - ``/2/nodes/[node_name]/storage/repair``
1834 - Added OpCodes to evacuate or migrate all instances on a node
1835 - Added new command to list storage elements on nodes (``gnt-node
1836 list-storage``) and modify them (``gnt-node modify-storage``)
1837 - Added new ssconf files with master candidate IP address
1838 (``ssconf_master_candidates_ips``), node primary IP address
1839 (``ssconf_node_primary_ips``) and node secondary IP address
1840 (``ssconf_node_secondary_ips``)
1841 - Added ``ganeti-confd`` and a client library to query the Ganeti
1842 configuration via UDP
1843 - Added ability to run hooks after cluster initialization and before
1845 - Added automatic mode for disk replace (``gnt-instance replace-disks
1847 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1848 after catastrophic data-loss
1849 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1851 - Added ``gnt-instance move`` command to move instances
1852 - Added ``gnt-cluster watcher`` command to control watcher
1853 - Added ``gnt-node powercycle`` command to powercycle nodes
1854 - Added new job status field ``lock_status``
1855 - Added parseable error codes to cluster verification (``gnt-cluster
1856 verify --error-codes``) and made output less verbose (use
1857 ``--verbose`` to restore previous behaviour)
1858 - Added UUIDs to the main config entities (cluster, nodes, instances)
1859 - Added support for OS variants
1860 - Added support for hashed passwords in the Ganeti remote API users file
1862 - Added option to specify maximum timeout on instance shutdown
1863 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1864 - Added new helper script to start and stop Ganeti daemons
1865 (``daemon-util``), with the intent to reduce the work necessary to
1866 adjust Ganeti for non-Debian distributions and to start/stop daemons
1868 - Added more unittests
1869 - Fixed critical bug in ganeti-masterd startup
1870 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1871 now customisable at the cluster level for both the KVM and Xen
1872 hypervisors using the new ``migration_port`` parameter
1873 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1874 reinstalling an instance
1875 - Allowed ``@`` in tag names
1876 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1877 - Many documentation updates
1878 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1879 - ``gnt-instance reinstall`` can now reinstall multiple instances
1880 - Updated many command line parameters
1881 - Introduced new OS API version 15
1882 - No longer support a default hypervisor
1883 - Treat virtual LVs as inexistent
1884 - Improved job locking logic to reduce lock contention
1885 - Match instance and node names case insensitively
1886 - Reimplemented bash completion script to be more complete
1893 *(Released Thu, 4 Feb 2010)*
1895 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1896 - Fix a string formatting bug
1897 - Improve safety of the code in some error paths
1898 - Improve data validation in the master of values returned from nodes
1904 *(Released Thu, 17 Dec 2009)*
1906 - Fix security issue due to missing validation of iallocator names; this
1907 allows local and remote execution of arbitrary executables
1908 - Fix failure of gnt-node list during instance removal
1909 - Ship the RAPI documentation in the archive
1915 *(Released Wed, 30 Sep 2009)*
1917 - Fixed many wrong messages
1918 - Fixed a few bugs related to the locking library
1919 - Fixed MAC checking at instance creation time
1920 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1921 - Fixed a few issues related to signal handling in both daemons and
1923 - Fixed the example startup script provided
1924 - Fixed insserv dependencies in the example startup script (patch from
1926 - Fixed handling of drained nodes in the iallocator framework
1927 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1929 - Fixed error related to invalid job IDs in job polling
1930 - Fixed job/opcode persistence on unclean master shutdown
1931 - Fixed handling of partial job processing after unclean master
1933 - Fixed error reporting from LUs, previously all errors were converted
1934 into execution errors
1935 - Fixed error reporting from burnin
1936 - Decreased significantly the memory usage of the job queue
1937 - Optimised slightly multi-job submission
1938 - Optimised slightly opcode loading
1939 - Backported the multi-job submit framework from the development
1940 branch; multi-instance start and stop should be faster
1941 - Added script to clean archived jobs after 21 days; this will reduce
1942 the size of the queue directory
1943 - Added some extra checks in disk size tracking
1944 - Added an example ethers hook script
1945 - Added a cluster parameter that prevents Ganeti from modifying of
1947 - Added more node information to RAPI responses
1948 - Added a ``gnt-job watch`` command that allows following the ouput of a
1950 - Added a bind-address option to ganeti-rapi
1951 - Added more checks to the configuration verify
1952 - Enhanced the burnin script such that some operations can be retried
1954 - Converted instance reinstall to multi-instance model
1960 *(Released Fri, 7 Aug 2009)*
1962 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1963 to allow using the pre-2.0.2 behaviour in activation, if any existing
1964 instances have mismatched disk sizes in the configuration
1965 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
1966 any configuration mismatches for disk sizes
1967 - Added ``gnt-master cluste-failover --no-voting`` to allow master
1968 failover to work on two-node clusters
1969 - Fixed the ``--net`` option of ``gnt-backup import``, which was
1971 - Fixed detection of OS script errors in ``gnt-backup export``
1972 - Fixed exit code of ``gnt-backup export``
1978 *(Released Fri, 17 Jul 2009)*
1980 - Added experimental support for stripped logical volumes; this should
1981 enhance performance but comes with a higher complexity in the block
1982 device handling; stripping is only enabled when passing
1983 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
1984 affected even in the non-stripped mode
1985 - Improved resiliency against transient failures at the end of DRBD
1986 resyncs, and in general of DRBD resync checks
1987 - Fixed a couple of issues with exports and snapshot errors
1988 - Fixed a couple of issues in instance listing
1989 - Added display of the disk size in ``gnt-instance info``
1990 - Fixed checking for valid OSes in instance creation
1991 - Fixed handling of the "vcpus" parameter in instance listing and in
1992 general of invalid parameters
1993 - Fixed http server library, and thus RAPI, to handle invalid
1994 username/password combinations correctly; this means that now they
1995 report unauthorized for queries too, not only for modifications,
1996 allowing earlier detect of configuration problems
1997 - Added a new "role" node list field, equivalent to the master/master
1998 candidate/drained/offline flags combinations
1999 - Fixed cluster modify and changes of candidate pool size
2000 - Fixed cluster verify error messages for wrong files on regular nodes
2001 - Fixed a couple of issues with node demotion from master candidate role
2002 - Fixed node readd issues
2003 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2004 - Added a new ``--no-voting`` option for masterfailover to fix failover
2005 on two-nodes clusters when the former master node is unreachable
2006 - Added instance reinstall over RAPI
2012 *(Released Tue, 16 Jun 2009)*
2014 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2015 allow re-adding the start in single-user option (regression from 1.2)
2016 - the watcher writes the instance status to a file, to allow monitoring
2017 to report the instance status (from the master) based on cached
2018 results of the watcher's queries; while this can get stale if the
2019 watcher is being locked due to other work on the cluster, this is
2020 still an improvement
2021 - the watcher now also restarts the node daemon and the rapi daemon if
2023 - fixed the watcher to handle full and drained queue cases
2024 - hooks export more instance data in the environment, which helps if
2025 hook scripts need to take action based on the instance's properties
2026 (no longer need to query back into ganeti)
2027 - instance failovers when the instance is stopped do not check for free
2028 RAM, so that failing over a stopped instance is possible in low memory
2030 - rapi uses queries for tags instead of jobs (for less job traffic), and
2031 for cluster tags it won't talk to masterd at all but read them from
2033 - a couple of error handling fixes in RAPI
2034 - drbd handling: improved the error handling of inconsistent disks after
2035 resync to reduce the frequency of "there are some degraded disks for
2036 this instance" messages
2037 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2038 error handling path called a wrong function name)
2044 *(Released Wed, 27 May 2009)*
2046 - no changes from rc5
2052 *(Released Wed, 20 May 2009)*
2054 - fix a couple of bugs (validation, argument checks)
2055 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2056 - some small improvements to RAPI and IAllocator
2057 - make watcher automatically start the master daemon if down
2063 *(Released Mon, 27 Apr 2009)*
2065 - change the OS list to not require locks; this helps with big clusters
2066 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2067 volume group is broken
2068 - ``gnt-instance info``, without any arguments, doesn't run for all
2069 instances anymore; either pass ``--all`` or pass the desired
2070 instances; this helps against mistakes on big clusters where listing
2071 the information for all instances takes a long time
2072 - miscellaneous doc and man pages fixes
2078 *(Released Wed, 8 Apr 2009)*
2080 - Change the internal locking model of some ``gnt-node`` commands, in
2081 order to reduce contention (and blocking of master daemon) when
2082 batching many creation/reinstall jobs
2083 - Fixes to Xen soft reboot
2084 - No longer build documentation at build time, instead distribute it in
2085 the archive, in order to reduce the need for the whole docbook/rst
2092 *(Released Fri, 27 Mar 2009)*
2094 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2095 - Fix watcher startup sequence, improves the behaviour of busy clusters
2096 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2097 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2098 bind address and other places
2099 - Some documentation fixes and updates
2105 *(Released Mon, 2 Mar 2009)*
2107 - More documentation updates, now all docs should be more-or-less
2109 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2111 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2113 - Fix an issue related to $libdir/run/ganeti and cluster creation
2119 *(Released Thu, 19 Feb 2009)*
2121 - Xen PVM and KVM have switched the default value for the instance root
2122 disk to the first partition on the first drive, instead of the whole
2123 drive; this means that the OS installation scripts must be changed
2125 - Man pages have been updated
2126 - RAPI has been switched by default to HTTPS, and the exported functions
2127 should all work correctly
2128 - RAPI v1 has been removed
2129 - Many improvements to the KVM hypervisor
2130 - Block device errors are now better reported
2131 - Many other bugfixes and small improvements
2137 *(Released Mon, 26 Jan 2009)*
2139 - Version 2 is a general rewrite of the code and therefore the
2140 differences are too many to list, see the design document for 2.0 in
2141 the ``doc/`` subdirectory for more details
2142 - In this beta version there is not yet a migration path from 1.2 (there
2143 will be one in the final 2.0 release)
2144 - A few significant changes are:
2146 - all commands are executed by a daemon (``ganeti-masterd``) and the
2147 various ``gnt-*`` commands are just front-ends to it
2148 - all the commands are entered into, and executed from a job queue,
2149 see the ``gnt-job(8)`` manpage
2150 - the RAPI daemon supports read-write operations, secured by basic
2151 HTTP authentication on top of HTTPS
2152 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2153 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2154 to migrate to DRBD 8 first while still running Ganeti 1.2)
2155 - DRBD devices are using statically allocated minor numbers, which
2156 will be assigned to existing instances during the migration process
2157 - there is support for both Xen PVM and Xen HVM instances running on
2159 - KVM virtualization is supported too
2160 - file-based storage has been implemented, which means that it is
2161 possible to run the cluster without LVM and DRBD storage, for
2162 example using a shared filesystem exported from shared storage (and
2163 still have live migration)
2169 *(Released Tue, 13 Jan 2009)*
2171 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2172 - Reuse the old instance mac address by default on instance import, if
2173 the instance name is the same.
2174 - Handle situations in which the node info rpc returns incomplete
2176 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2177 - Improved version of batcher:
2179 - state file support
2180 - instance mac address support
2181 - support for HVM clusters/instances
2183 - Add an option to show the number of cpu sockets and nodes in
2185 - Support OSes that handle more than one version of the OS api (but do
2186 not change the current API in any other way)
2187 - Fix ``gnt-node migrate``
2188 - ``gnt-debug`` man page
2189 - Fixes various more typos and small issues
2190 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2196 *(Released Wed, 24 Sep 2008)*
2198 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2199 type of disk exported to fully virtualized instances.
2200 - provide access to the serial console of HVM instances
2201 - instance auto_balance flag, set by default. If turned off it will
2202 avoid warnings on cluster verify if there is not enough memory to fail
2203 over an instance. in the future it will prevent automatically failing
2204 it over when we will support that.
2205 - batcher tool for instance creation, see ``tools/README.batcher``
2206 - ``gnt-instance reinstall --select-os`` to interactively select a new
2207 operating system when reinstalling an instance.
2208 - when changing the memory amount on instance modify a check has been
2209 added that the instance will be able to start. also warnings are
2210 emitted if the instance will not be able to fail over, if auto_balance
2212 - documentation fixes
2213 - sync fields between ``gnt-instance list/modify/add/import``
2214 - fix a race condition in drbd when the sync speed was set after giving
2215 the device a remote peer.
2221 *(Released Tue, 22 Jul 2008)*
2223 - note: the allowed size and number of tags per object were reduced
2224 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2225 - fixed twisted 8.x compatibility
2226 - fixed ``gnt-instance replace-disks`` with iallocator
2227 - add TCP keepalives on twisted connections to detect restarted nodes
2228 - disk increase support, see ``gnt-instance grow-disk``
2229 - implement bulk node/instance query for RAPI
2230 - add tags in node/instance listing (optional)
2231 - experimental migration (and live migration) support, read the man page
2232 for ``gnt-instance migrate``
2233 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2234 has some small improvements in handling its state file
2240 *(Released Fri, 13 Jun 2008)*
2242 - Experimental readonly, REST-based remote API implementation;
2243 automatically started on master node, TCP port 5080, if enabled by
2244 ``--enable-rapi`` parameter to configure script.
2245 - Instance allocator support. Add and import instance accept a
2246 ``--iallocator`` parameter, and call that instance allocator to decide
2247 which node to use for the instance. The iallocator document describes
2248 what's expected from an allocator script.
2249 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2250 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2251 node is lost there is still enough memory to fail over the instances
2253 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2254 ``gnt-cluster verify``, to check for site-specific compliance. All the
2255 hooks will run, and their output, if any, will be displayed. Any
2256 failing hook will make the verification return an error value.
2257 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2258 primary and secondary interfaces
2259 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2260 that is still declared as part of the cluster and has failed.
2261 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2262 specifying output fields, that just adds the chosen fields to the
2264 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2265 export from the filesystem.
2266 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2267 have been added. Using them you can enable/disable acpi and pae
2268 support, and specify a path for a cd image to be exported to the
2269 instance. These parameters as the name suggest only work on HVM
2271 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2272 PAE support will be set to the previously hardcoded values, but the
2273 (previously hardcoded) path to the CDROM ISO image will be unset and
2274 if required, needs to be set manually with ``gnt-instance modify``
2276 - The address to which an instance's VNC console is bound is now
2277 selectable per-instance, rather than being cluster wide. Of course
2278 this only applies to instances controlled via VNC, so currently just
2279 applies to HVM clusters.
2285 *(Released Mon, 18 Feb 2008)*
2287 - more tweaks to the disk activation code (especially helpful for DRBD)
2288 - change the default ``gnt-instance list`` output format, now there is
2289 one combined status field (see the manpage for the exact values this
2291 - some more fixes for the mac export to hooks change
2292 - make Ganeti not break with DRBD 8.2.x (which changed the version
2293 format in ``/proc/drbd``) (issue 24)
2294 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2295 template, allowing migration from DRBD0.7+MD to DRBD8
2301 *(Released Wed, 30 Jan 2008)*
2303 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2305 - add command aliases infrastructure and a few aliases
2306 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2307 man pages and the ``--help`` option of ``gnt-node
2308 list``/``gnt-instance list``
2309 - fix ``gnt-backup list`` with down nodes (issue 21)
2310 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2311 - fix the dist archive and add a check for including svn/git files in
2313 - some developer-related changes: improve the burnin and the QA suite,
2314 add an upload script for testing during development
2320 *(Released Wed, 16 Jan 2008)*
2322 - experimental HVM support, read the install document, section
2323 "Initializing the cluster"
2324 - allow for the PVM hypervisor per-instance kernel and initrd paths
2325 - add a new command ``gnt-cluster verify-disks`` which uses a new
2326 algorithm to improve the reconnection of the DRBD pairs if the device
2327 on the secondary node has gone away
2328 - make logical volume code auto-activate LVs at disk activation time
2329 - slightly improve the speed of activating disks
2330 - allow specification of the MAC address at instance creation time, and
2331 changing it later via ``gnt-instance modify``
2332 - fix handling of external commands that generate lots of output on
2334 - update documentation with regard to minimum version of DRBD8 supported
2340 *(Released Tue, 4 Dec 2007)*
2342 - Log the ``xm create`` output to the node daemon log on failure (to
2343 help diagnosing the error)
2344 - In debug mode, log all external commands output if failed to the logs
2345 - Change parsing of lvm commands to ignore stderr
2351 *(Released Wed, 28 Nov 2007)*
2353 - Another round of updates to the DRBD 8 code to deal with more failures
2354 in the replace secondary node operation
2355 - Some more logging of failures in disk operations (lvm, drbd)
2356 - A few documentation updates
2363 *(Released Tue, 13 Nov 2007)*
2365 - Change configuration file format from Python's Pickle to JSON.
2366 Upgrading is possible using the cfgupgrade utility.
2367 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2368 faster replace disks and is more stable (DRBD 8 has many improvements
2369 compared to DRBD 0.7)
2370 - Added command line tags support (see man pages for ``gnt-instance``,
2371 ``gnt-node``, ``gnt-cluster``)
2372 - Added instance rename support
2373 - Added multi-instance startup/shutdown
2374 - Added cluster rename support
2375 - Added ``gnt-node evacuate`` to simplify some node operations
2376 - Added instance reboot operation that can speedup reboot as compared to
2378 - Soften the requirement that hostnames are in FQDN format
2379 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2381 - Removed dependency on debian's patched fping that uses the
2382 non-standard ``-S`` option
2383 - Now the OS definitions are searched for in multiple, configurable
2384 paths (easier for distros to package)
2385 - Some changes to the hooks infrastructure (especially the new
2386 post-configuration update hook)
2387 - Other small bugfixes
2389 .. vim: set textwidth=72 syntax=rst :