9 - Deprecated ``admin_up`` field. Instead, ``admin_state`` is introduced,
10 with 3 possible values -- ``up``,``down`` and ``offline``.
15 *(Released Tue, 18 Oct 2011)*
17 Incompatible/important changes and bugfixes
18 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
20 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
21 resource's ``ip_check`` parameter changed from ``True`` to ``False``
22 to match the underlying LUXI interface.
23 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
24 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
25 not maintain backwards-compatibility as the underlying operation
26 changed in an incompatible way. The RAPI client can talk to old
27 servers, but it needs to be told so as the return value changed.
28 - When creating file-based instances via RAPI, the ``file_driver``
29 parameter no longer defaults to ``loop`` and must be specified.
30 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
32 - Support for the undocumented and deprecated RAPI instance creation
33 request format version 0 has been dropped. Use version 1, supported
34 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
35 - Pyparsing 1.4.6 or above is required, see :doc:`installation
36 documentation <install>`.
37 - The "cluster-verify" hooks are now executed per group by the
38 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
39 if you just run ``gnt-cluster verify``, which generates one opcode per
41 - The environment as passed to the OS scripts is cleared, and thus no
42 environment variables defined in the node daemon's environment will be
43 inherited by the scripts.
44 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
46 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
47 support operations involving multiple node groups.
48 - Offline nodes are ignored when failing over an instance.
53 - The ganeti-htools project has been merged into the ganeti-core source
54 tree and will be built as part of Ganeti (see :doc:`install-quick`).
55 - Implemented support for :doc:`shared storage <design-shared-storage>`.
56 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
57 GPT-style partition tables (requires `parted
58 <http://www.gnu.org/s/parted/>`_).
59 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
60 - Allowed adding tags on instance creation.
61 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
62 - Allow instances to be started in a paused state, enabling the user to
63 see the complete console output on boot using the console.
64 - Added new hypervisor flag to control default reboot behaviour
65 (``reboot_behavior``).
66 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
67 - Improved out-of-band management support:
69 - Added ``gnt-node health`` command reporting the health status of
71 - Added ``gnt-node power`` command to manage power status of nodes.
72 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
74 - Instance migration can fall back to failover if instance is not
76 - Filters can be used when listing nodes, instances, groups and locks;
77 see *ganeti(7)* manpage.
78 - Added post-execution status as variables to :doc:`hooks <hooks>`
80 - Instance tags are exported/imported together with the instance.
81 - When given an explicit job ID, ``gnt-job info`` will work for archived
83 - Jobs can define dependencies on other jobs (not yet supported via
84 RAPI or command line, but used by internal commands and usable via
87 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
90 - Instance failover is now available as a RAPI resource
91 (``/2/instances/[instance_name]/failover``).
92 - ``gnt-instance info`` defaults to static information if primary node
94 - Opcodes have a new ``comment`` attribute.
95 - Added basic SPICE support to KVM hypervisor.
96 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
98 Node group improvements
99 ~~~~~~~~~~~~~~~~~~~~~~~
101 - ``gnt-cluster verify`` has been modified to check groups separately,
102 thereby improving performance.
103 - Node group support has been added to ``gnt-cluster verify-disks``,
104 which now operates per node group.
105 - Watcher has been changed to work better with node groups.
107 - One process and state file per node group.
108 - Slow watcher in one group doesn't block other group's watcher.
110 - Added new command, ``gnt-group evacuate``, to move all instances in a
111 node group to other groups.
112 - Added ``gnt-instance change-group`` to move an instance to another
114 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
115 per-group operations.
116 - Node groups can be tagged.
117 - Some operations switch from an exclusive to a shared lock as soon as
119 - Instance's primary and secondary nodes' groups are now available as
120 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
121 and ``snodes.group.uuid``).
126 - Numerous updates to documentation and manpages.
128 - :doc:`RAPI <rapi>` documentation now has detailed parameter
131 - A lockset's internal lock is now also visible in lock monitor.
132 - Log messages from job queue workers now contain information about the
133 opcode they're processing.
134 - ``gnt-instance console`` no longer requires the instance lock.
135 - A short delay when waiting for job changes reduces the number of LUXI
136 requests significantly.
137 - DRBD metadata volumes are overwritten with zeros during disk creation.
143 *(Released Tue, 4 Oct 2011)*
145 This was the first release candidate of the 2.5 series.
151 *(Released Wed, 31 Aug 2011)*
153 This was the third beta release of the 2.5 series.
159 *(Released Mon, 22 Aug 2011)*
161 This was the second beta release of the 2.5 series.
167 *(Released Fri, 12 Aug 2011)*
169 This was the first beta release of the 2.5 series.
175 *(Released Thu, 27 Oct 2011)*
177 - Fixed bug when parsing command line parameter values ending in
179 - Fixed assertion error after unclean master shutdown
180 - Disable HTTP client pool for RPC, significantly reducing memory usage
182 - Fixed queue archive creation with wrong permissions
188 *(Released Tue, 23 Aug 2011)*
192 - Fixed documentation for importing with ``--src-dir`` option
193 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
194 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
200 *(Released Fri, 5 Aug 2011)*
202 Many bug-fixes and a few small features:
204 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
205 issues when you tried to add an instance with two MAC addresses in one
207 - KVM: fixed per-instance stored UID value
208 - KVM: configure bridged NICs at migration start
209 - KVM: Fix a bug where instance will not start with never KVM versions
211 - Added OS search path to ``gnt-cluster info``
212 - Fixed an issue with ``file_storage_dir`` where you were forced to
213 provide an absolute path, but the documentation states it is a
214 relative path, the documentation was right
215 - Added a new parameter to instance stop/start called ``--no-remember``
216 that will make the state change to not be remembered
217 - Implemented ``no_remember`` at RAPI level
218 - Improved the documentation
219 - Node evacuation: don't call IAllocator if node is already empty
220 - Fixed bug in DRBD8 replace disks on current nodes
221 - Fixed bug in recreate-disks for DRBD instances
222 - Moved assertion checking locks in ``gnt-instance replace-disks``
223 causing it to abort with not owning the right locks for some situation
224 - Job queue: Fixed potential race condition when cancelling queued jobs
225 - Fixed off-by-one bug in job serial generation
226 - ``gnt-node volumes``: Fix instance names
227 - Fixed aliases in bash completion
228 - Fixed a bug in reopening log files after being sent a SIGHUP
229 - Added a flag to burnin to allow specifying VCPU count
230 - Bugfixes to non-root Ganeti configuration
236 *(Released Thu, 12 May 2011)*
238 Many bug-fixes and a few new small features:
240 - Fixed a bug related to log opening failures
241 - Fixed a bug in instance listing with orphan instances
242 - Fixed a bug which prevented resetting the cluster-level node parameter
243 ``oob_program`` to the default
244 - Many fixes related to the ``cluster-merge`` tool
245 - Fixed a race condition in the lock monitor, which caused failures
246 during (at least) creation of many instances in parallel
247 - Improved output for gnt-job info
248 - Removed the quiet flag on some ssh calls which prevented debugging
250 - Improved the N+1 failure messages in cluster verify by actually
251 showing the memory values (needed and available)
252 - Increased lock attempt timeouts so that when executing long operations
253 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
254 too early and thus prevent the use of the 'fair' mechanism
255 - Changed instance query data (``gnt-instance info``) to not acquire
256 locks unless needed, thus allowing its use on locked instance if only
257 static information is asked for
258 - Improved behaviour with filesystems that do not support rename on an
260 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
261 kept locks on all nodes during the wipe, which is unneeded
262 - Fixed ``gnt-watcher`` handling of errors during hooks execution
263 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
264 10GiB) which caused the wipe to fail right at the end in some cases
265 - Fixed master IP activation when doing master failover with no-voting
266 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
267 the master node itself
268 - Fixed potential data-loss in under disk full conditions, where Ganeti
269 wouldn't check correctly the return code and would consider
270 partially-written files 'correct'
271 - Fixed bug related to multiple VGs and DRBD disk replacing
272 - Added new disk parameter ``metavg`` that allows placement of the meta
273 device for DRBD in a different volume group
274 - Fixed error handling in the node daemon when the system libc doesn't
275 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
276 - Fixed lock release during replace-disks, which kept cluster-wide locks
277 when doing disk replaces with an iallocator script
278 - Added check for missing bridges in cluster verify
279 - Handle EPIPE errors while writing to the terminal better, so that
280 piping the output to e.g. ``less`` doesn't cause a backtrace
281 - Fixed rare case where a ^C during Luxi calls could have been
282 interpreted as server errors, instead of simply terminating
283 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
285 - Added a few more parameters to the KVM hypervisor, allowing a second
286 CDROM, custom disk type for CDROMs and a floppy image
287 - Removed redundant message in instance rename when the name is given
289 - Added option to ``gnt-instance recreate-disks`` to allow creating the
290 disks on new nodes, allowing recreation when the original instance
291 nodes are completely gone
292 - Added option when converting disk templates to DRBD to skip waiting
293 for the resync, in order to make the instance available sooner
294 - Added two new variables to the OS scripts environment (containing the
296 - Made the root_path and optional parameter for the xen-pvm hypervisor,
297 to allow use of ``pvgrub`` as bootloader
298 - Changed the instance memory modifications to only check out-of-memory
299 conditions on memory increases, and turned the secondary node warnings
300 into errors (they can still be overridden via ``--force``)
301 - Fixed the handling of a corner case when the Python installation gets
302 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
303 to execute a command that doesn't exist
304 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
305 node of the instance returned failures during instance shutdown; this
306 adds the option ``--ignore-consistency`` to gnt-instance move
308 And as usual, various improvements to the error messages, documentation
314 *(Released Wed, 09 Mar 2011)*
316 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
317 the RAPI users file if run twice (even with ``--dry-run``).
319 The release fixes that bug (nothing else changed).
325 *(Released Mon, 07 Mar 2011)*
327 Final 2.4.0 release. Just a few small fixes:
329 - Fixed RAPI node evacuate
330 - Fixed the kvm-ifup script
331 - Fixed internal error handling for special job cases
332 - Updated man page to specify the escaping feature for options
338 *(Released Mon, 28 Feb 2011)*
340 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
341 that this feature wiped the disks of the wrong instance, leading to loss
346 - Fixed title of query field containing instance name
347 - Expanded the glossary in the documentation
348 - Fixed one unittest (internal issue)
354 *(Released Mon, 21 Feb 2011)*
356 A number of bug fixes plus just a couple functionality changes.
358 On the user-visible side, the ``gnt-* list`` command output has changed
359 with respect to "special" field states. The current rc1 style of display
360 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
361 the default output mode special fields states are displayed as follows:
363 - Offline resource: ``*``
364 - Unavailable/not applicable: ``-``
365 - Data missing (RPC failure): ``?``
366 - Unknown field: ``??``
368 Another user-visible change is the addition of ``--force-join`` to
373 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
374 - Fixed regression in RAPI/instance reinstall where all parameters were
375 required (instead of optional)
376 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
377 - Fixed iallocator usage (offline nodes were not considered offline)
378 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
379 - Fixed hypervisor and OS parameter validation with respect to
381 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
383 - Fixed ``tools/listrunner`` with respect to agent-based usage
389 *(Released Fri, 4 Feb 2011)*
391 Many changes and fixes since the beta1 release. While there were some
392 internal changes, the code has been mostly stabilised for the RC
395 Note: the dumb allocator was removed in this release, as it was not kept
396 up-to-date with the IAllocator protocol changes. It is recommended to
397 use the ``hail`` command from the ganeti-htools package.
399 Note: the 2.4 and up versions of Ganeti are not compatible with the
400 0.2.x branch of ganeti-htools. You need to upgrade to
401 ganeti-htools-0.3.0 (or later).
403 Regressions fixed from 2.3
404 ~~~~~~~~~~~~~~~~~~~~~~~~~~
406 - Fixed the ``gnt-cluster verify-disks`` command
407 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
409 - Fixed disk adoption breakage
410 - Fixed wrong headers in instance listing for field aliases
415 - Fixed corner case in KVM handling of NICs
416 - Fixed many cases of wrong handling of non-vm_capable nodes
417 - Fixed a bug where a missing instance symlink was not possible to
418 recreate with any ``gnt-*`` command (now ``gnt-instance
419 activate-disks`` does it)
420 - Fixed the volume group name as reported by ``gnt-cluster
422 - Increased timeouts for the import-export code, hopefully leading to
423 fewer aborts due network or instance timeouts
424 - Fixed bug in ``gnt-node list-storage``
425 - Fixed bug where not all daemons were started on cluster
426 initialisation, but only at the first watcher run
427 - Fixed many bugs in the OOB implementation
428 - Fixed watcher behaviour in presence of instances with offline
430 - Fixed instance list output for instances running on the wrong node
431 - a few fixes to the cluster-merge tool, but it still cannot merge
432 multi-node groups (currently it is not recommended to use this tool)
438 - Improved network configuration for the KVM hypervisor
439 - Added e1000 as a supported NIC for Xen-HVM
440 - Improved the lvmstrap tool to also be able to use partitions, as
441 opposed to full disks
442 - Improved speed of disk wiping (the cluster parameter
443 ``prealloc_wipe_disks``, so that it has a low impact on the total time
444 of instance creations
445 - Added documentation for the OS parameters
446 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
447 hypervisor is not responding
448 - Added display of blacklisted and hidden OS information in
450 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
451 NIC and node parameters, which might create problems with currently
452 invalid (but undetected) configuration files, but prevents validation
453 failures when unrelated parameters are modified
454 - Changed cluster initialisation to wait for the master daemon to become
456 - Expanded the RAPI interface:
458 - Added config redistribution resource
459 - Added activation/deactivation of instance disks
460 - Added export of console information
462 - Implemented log file reopening on SIGHUP, which allows using
463 logrotate(8) for the Ganeti log files
464 - Added a basic OOB helper script as an example
470 *(Released Fri, 14 Jan 2011)*
475 - Fixed timezone issues when formatting timestamps
476 - Added support for node groups, available via ``gnt-group`` and other
478 - Added out-of-band framework and management, see :doc:`design
479 document <design-oob>`
480 - Removed support for roman numbers from ``gnt-node list`` and
481 ``gnt-instance list``.
482 - Allowed modification of master network interface via ``gnt-cluster
483 modify --master-netdev``
484 - Accept offline secondaries while shutting down instance disks
485 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
486 - Added support for multiple LVM volume groups
487 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
489 - Added commands to list available fields:
491 - ``gnt-node list-fields``
492 - ``gnt-group list-fields``
493 - ``gnt-instance list-fields``
495 - Updated documentation and man pages
500 - Moved ``rapi_users`` file into separate directory, now named
501 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
503 - Added new tool for running commands on many machines,
504 ``tools/ganeti-listrunner``
505 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
506 improving the ``gnt-instance console`` output
507 - Allowed customisation of disk index separator at ``configure`` time
508 - Export node group allocation policy to :doc:`iallocator <iallocator>`
509 - Added support for non-partitioned md disks in ``lvmstrap``
510 - Added script to gracefully power off KVM instances
511 - Split ``utils`` module into smaller parts
512 - Changed query operations to return more detailed information, e.g.
513 whether an information is unavailable due to an offline node. To use
514 this new functionality, the LUXI call ``Query`` must be used. Field
515 information is now stored by the master daemon and can be retrieved
516 using ``QueryFields``. Instances, nodes and groups can also be queried
517 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
518 exposed via RAPI). The following commands make use of this
519 infrastructure change:
522 - ``gnt-group list-fields``
524 - ``gnt-node list-fields``
525 - ``gnt-instance list``
526 - ``gnt-instance list-fields``
527 - ``gnt-debug locks``
532 - New RAPI resources (see :doc:`rapi`):
536 - ``/2/groups/[group_name]``
537 - ``/2/groups/[group_name]/assign-nodes``
538 - ``/2/groups/[group_name]/modify``
539 - ``/2/groups/[group_name]/rename``
540 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
544 - Implemented ``no_install`` for instance creation
545 - Implemented OS parameters for instance reinstallation, allowing
546 use of special settings on reinstallation (e.g. for preserving data)
551 - Added IPv6 support in import/export
552 - Pause DRBD synchronization while wiping disks on instance creation
553 - Updated unittests and QA scripts
554 - Improved network parameters passed to KVM
555 - Converted man pages from docbook to reStructuredText
561 *(Released Mon, 20 Dec 2010)*
563 Released version 2.3.1~rc1 without any changes.
569 *(Released Wed, 1 Dec 2010)*
571 - impexpd: Disable OpenSSL compression in socat if possible (backport
572 from master, commit e90739d625b, see :doc:`installation guide
573 <install-quick>` for details)
574 - Changed unittest coverage report to exclude test scripts
575 - Added script to check version format
581 *(Released Wed, 1 Dec 2010)*
583 Released version 2.3.0~rc1 without any changes.
589 *(Released Fri, 19 Nov 2010)*
591 A number of bugfixes and documentation updates:
593 - Update ganeti-os-interface documentation
594 - Fixed a bug related to duplicate MACs or similar items which should be
596 - Fix breakage in OS state modify
597 - Reinstall instance: disallow offline secondaries (fixes bug related to
598 OS changing but reinstall failing)
599 - plus all the other fixes between 2.2.1 and 2.2.2
605 *(Released Tue, 2 Nov 2010)*
607 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
608 - Fixed master failover race with watcher
609 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
611 - Accept previously stopped instance for export with instance removal
612 - Simplify and extend the environment variables for instance OS scripts
613 - Added new node flags, ``master_capable`` and ``vm_capable``
614 - Added optional instance disk wiping prior during allocation. This is a
615 cluster-wide option and can be set/modified using
616 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
617 - Added IPv6 support, see :doc:`design document <design-2.3>` and
619 - Added a new watcher option (``--ignore-pause``)
620 - Added option to ignore offline node on instance start/stop
621 (``--ignore-offline``)
622 - Allow overriding OS parameters with ``gnt-instance reinstall``
623 - Added ability to change node's secondary IP address using ``gnt-node
625 - Implemented privilege separation for all daemons except
626 ``ganeti-noded``, see ``configure`` options
627 - Complain if an instance's disk is marked faulty in ``gnt-cluster
629 - Implemented job priorities (see ``ganeti(7)`` manpage)
630 - Ignore failures while shutting down instances during failover from
632 - Exit daemon's bootstrap process only once daemon is ready
633 - Export more information via ``LUInstanceQuery``/remote API
634 - Improved documentation, QA and unittests
635 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
636 restart if the file was created or changed
637 - Added LUXI protocol version sent with each request and response,
638 allowing detection of server/client mismatches
639 - Moved the Python scripts among gnt-* and ganeti-* into modules
640 - Moved all code related to setting up SSH to an external script,
642 - Infrastructure changes for node group support in future versions
648 *(Released Fri, 19 Nov 2010)*
650 A few small bugs fixed, and some improvements to the build system:
652 - Fix documentation regarding conversion to drbd
653 - Fix validation of parameters in cluster modify (``gnt-cluster modify
655 - Fix error handling in node modify with multiple changes
656 - Allow remote imports without checked names
662 *(Released Tue, 19 Oct 2010)*
664 - Disable SSL session ID cache in RPC client
670 *(Released Thu, 14 Oct 2010)*
672 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
673 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
675 - Fix problems with interaction between Curl and Python's HTTP server,
676 resulting in increased speed in many RPC calls
677 - Improve our release script to prevent breakage with older aclocal and
684 *(Released Thu, 7 Oct 2010)*
686 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
687 value retrieved from master
688 - Added support for blacklisted or hidden OS definitions
689 - Added simple lock monitor (accessible via (``gnt-debug locks``)
690 - Added support for -mem-path in KVM hypervisor abstraction layer
691 - Allow overriding instance parameters in tool for inter-cluster
692 instance moves (``tools/move-instance``)
693 - Improved opcode summaries (e.g. in ``gnt-job list``)
694 - Improve consistency of OS listing by sorting it
695 - Documentation updates
701 *(Released Fri, 8 Oct 2010)*
703 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
709 *(Released Mon, 4 Oct 2010)*
711 - Fixed regression in ``gnt-instance rename``
717 *(Released Wed, 22 Sep 2010)*
719 - Fixed OS_VARIANT variable for OS scripts
720 - Fixed cluster tag operations via RAPI
721 - Made ``setup-ssh`` exit with non-zero code if an error occurred
722 - Disabled RAPI CA checks in watcher
728 *(Released Mon, 23 Aug 2010)*
730 - Support DRBD versions of the format "a.b.c.d"
732 - Re-introduce support for usage from multiple threads in RAPI client
733 - Instance renames and modify via RAPI
734 - Work around race condition between processing and archival in job
736 - Mark opcodes following failed one as failed, too
737 - Job field ``lock_status`` was removed due to difficulties making it
738 work with the changed job queue in Ganeti 2.2; a better way to monitor
739 locks is expected for a later 2.2.x release
740 - Fixed dry-run behaviour with many commands
741 - Support ``ssh-agent`` again when adding nodes
742 - Many additional bugfixes
748 *(Released Fri, 30 Jul 2010)*
750 Important change: the internal RPC mechanism between Ganeti nodes has
751 changed from using a home-grown http library (based on the Python base
752 libraries) to use the PycURL library. This requires that PycURL is
753 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
754 against GnuTLS by default. cURL's support for GnuTLS had known issues
755 before cURL 7.21.0 and we recommend using the latest cURL release or
756 linking against OpenSSL. Most other distributions already link PycURL
757 and cURL against OpenSSL. The command::
759 python -c 'import pycurl; print pycurl.version'
761 can be used to determine the libraries PycURL and cURL are linked
764 Other significant changes:
766 - Rewrote much of the internals of the job queue, in order to achieve
767 better parallelism; this decouples job query operations from the job
768 processing, and it should allow much nicer behaviour of the master
769 daemon under load, and it also has uncovered some long-standing bugs
770 related to the job serialisation (now fixed)
771 - Added a default iallocator setting to the cluster parameters,
772 eliminating the need to always pass nodes or an iallocator for
773 operations that require selection of new node(s)
774 - Added experimental support for the LXC virtualization method
775 - Added support for OS parameters, which allows the installation of
776 instances to pass parameter to OS scripts in order to customise the
778 - Added a hypervisor parameter controlling the migration type (live or
779 non-live), since hypervisors have various levels of reliability; this
780 has renamed the 'live' parameter to 'mode'
781 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
782 logical volumes, meaning that cluster verify will ignore them and not
783 flag their presence as errors
784 - The watcher will now reset the error count for failed instances after
785 8 hours, thus allowing self-healing if the problem that caused the
786 instances to be down/fail to start has cleared in the meantime
787 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
788 check for, and warn, if the drbd module parameter ``usermode_helper``
789 is not consistent with the cluster-wide setting; this is needed to
790 make diagnose easier of failed drbd creations
791 - Started adding base IPv6 support, but this is not yet
792 enabled/available for use
793 - Rename operations (cluster, instance) will now return the new name,
794 which is especially useful if a short name was passed in
795 - Added support for instance migration in RAPI
796 - Added a tool to pre-configure nodes for the SSH setup, before joining
797 them to the cluster; this will allow in the future a simplified model
798 for node joining (but not yet fully enabled in 2.2); this needs the
799 paramiko python library
800 - Fixed handling of name-resolving errors
801 - Fixed consistency of job results on the error path
802 - Fixed master-failover race condition when executed multiple times in
804 - Fixed many bugs related to the job queue (mostly introduced during the
805 2.2 development cycle, so not all are impacting 2.1)
806 - Fixed instance migration with missing disk symlinks
807 - Fixed handling of unknown jobs in ``gnt-job archive``
808 - And many other small fixes/improvements
812 - Enhanced both the unittest and the QA coverage
813 - Switched the opcode validation to a generic model, and extended the
814 validation to all opcode parameters
815 - Changed more parts of the code that write shell scripts to use the
817 - Switched the master daemon to use the asyncore library for the Luxi
824 *(Released Thu, 17 Jun 2010)*
826 - Added tool (``move-instance``) and infrastructure to move instances
827 between separate clusters (see :doc:`separate documentation
828 <move-instance>` and :doc:`design document <design-2.2>`)
829 - Added per-request RPC timeout
830 - RAPI now requires a Content-Type header for requests with a body (e.g.
831 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
832 :rfc:`2616` (HTTP/1.1), section 7.2.1)
833 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
835 - Implemented initial support for running Ganeti daemons as separate
836 users, see configure-time flags ``--with-user-prefix`` and
837 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
839 - Instances can be removed after export (``gnt-backup export
841 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
842 key (instead of 1024 bit)
843 - Added new cluster configuration file for cluster domain secret
844 - Import/export now use SSL instead of SSH
845 - Added support for showing estimated time when exporting an instance,
846 see the ``ganeti-os-interface(7)`` manpage and look for
853 *(Released Tue, 16 Nov 2010)*
855 Some more bugfixes. Unless critical bugs occur, this will be the last
858 - Fix case of MAC special-values
859 - Fix mac checker regex
860 - backend: Fix typo causing "out of range" error
861 - Add missing --units in gnt-instance list man page
867 *(Released Tue, 24 Aug 2010)*
870 - Don't ignore secondary node silently on non-mirrored disk templates
872 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
873 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
874 - Properly document the usb_mouse parameter
875 - Fix path in ganeti-rapi(8) (issue 116)
876 - Adjust error message when the ganeti user's .ssh directory is
878 - Add same-node-check when changing the disk template to drbd
884 *(Released Fri, 16 Jul 2010)*
887 - Add an option to only select some reboot types during qa/burnin.
888 (on some hypervisors consequent reboots are not supported)
889 - Fix infrequent race condition in master failover. Sometimes the old
890 master ip address would be still detected as up for a short time
891 after it was removed, causing failover to fail.
892 - Decrease mlockall warnings when the ctypes module is missing. On
893 Python 2.4 we support running even if no ctypes module is installed,
894 but we were too verbose about this issue.
895 - Fix building on old distributions, on which man doesn't have a
897 - Fix RAPI not to ignore the MAC address on instance creation
898 - Implement the old instance creation format in the RAPI client.
904 *(Released Thu, 01 Jul 2010)*
906 A small bugfix release:
907 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
908 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
909 a non-existing option
910 - Fix instance create: the --force-variant option was ignored
911 - Improve pylint 0.21 compatibility and warnings with Python 2.6
912 - Fix modify node storage with non-FQDN arguments
913 - Fix RAPI client to authenticate under Python 2.6 when used
914 for more than 5 requests needing authentication
915 - Fix gnt-instance modify -t (storage) giving a wrong error message
916 when converting a non-shutdown drbd instance to plain
922 *(Released Fri, 18 Jun 2010)*
924 A small bugfix release:
926 - Fix live migration of KVM instances started with older Ganeti
927 versions which had fewer hypervisor parameters
928 - Fix gnt-instance grow-disk on down instances
929 - Fix an error-reporting bug during instance migration
930 - Better checking of the ``--net`` and ``--disk`` values, to avoid
931 silently ignoring broken ones
932 - Fix an RPC error reporting bug affecting, for example, RAPI client
934 - Fix bug triggered by different API version os-es on different nodes
935 - Fix a bug in instance startup with custom hvparams: OS level
936 parameters would fail to be applied.
937 - Fix the RAPI client under Python 2.6 (but more work is needed to
938 make it work completely well with OpenSSL)
939 - Fix handling of errors when resolving names from DNS
945 *(Released Thu, 3 Jun 2010)*
947 A medium sized development cycle. Some new features, and some
948 fixes/small improvements/cleanups.
953 The node deamon now tries to mlock itself into memory, unless the
954 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
955 its logs, and falls back to console logging. This allows emergency
956 features such as ``gnt-node powercycle`` to work even in the event of a
957 broken node disk (tested offlining the disk hosting the node's
958 filesystem and dropping its memory caches; don't try this at home)
960 KVM: add vhost-net acceleration support. It can be tested with a new
961 enough version of the kernel and of qemu-kvm.
963 KVM: Add instance chrooting feature. If you use privilege dropping for
964 your VMs you can also now force them to chroot to an empty directory,
965 before starting the emulated guest.
967 KVM: Add maximum migration bandwith and maximum downtime tweaking
968 support (requires a new-enough version of qemu-kvm).
970 Cluster verify will now warn if the master node doesn't have the master
973 Add a new (incompatible) instance creation request format to RAPI which
974 supports all parameters (previously only a subset was supported, and it
975 wasn't possible to extend the old format to accomodate all the new
976 features. The old format is still supported, and a client can check for
977 this feature, before using it, by checking for its presence in the
978 ``features`` RAPI resource.
980 Now with ancient latin support. Try it passing the ``--roman`` option to
981 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
982 (requires the python-roman module to be installed, in order to work).
987 As usual many internal code refactorings, documentation updates, and
990 - Lots of improvements and cleanups to the experimental Remote API
991 (RAPI) client library.
992 - A new unit test suite for the core daemon libraries.
993 - A fix to creating missing directories makes sure the umask is not
994 applied anymore. This enforces the same directory permissions
996 - Better handling terminating daemons with ctrl+c (used when running
997 them in debugging mode).
998 - Fix a race condition in live migrating a KVM instance, when stat()
999 on the old proc status file returned EINVAL, which is an unexpected
1001 - Fixed manpage checking with newer man and utf-8 charachters. But now
1002 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1008 *(Released Fri, 7 May 2010)*
1010 Fix a bug which prevented untagged KVM instances from starting.
1016 *(Released Fri, 7 May 2010)*
1018 Another release with a long development cycle, during which many
1019 different features were added.
1021 Significant features
1022 ~~~~~~~~~~~~~~~~~~~~
1024 The KVM hypervisor now can run the individual instances as non-root, to
1025 reduce the impact of a VM being hijacked due to bugs in the
1026 hypervisor. It is possible to run all instances as a single (non-root)
1027 user, to manually specify a user for each instance, or to dynamically
1028 allocate a user out of a cluster-wide pool to each instance, with the
1029 guarantee that no two instances will run under the same user ID on any
1032 An experimental RAPI client library, that can be used standalone
1033 (without the other Ganeti libraries), is provided in the source tree as
1034 ``lib/rapi/client.py``. Note this client might change its interface in
1035 the future, as we iterate on its capabilities.
1037 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1038 replace the cluster's certificates and crypto keys. This might help in
1039 case they have been compromised, or have simply expired.
1041 A new disk option for instance creation has been added that allows one
1042 to "adopt" currently existing logical volumes, with data
1043 preservation. This should allow easier migration to Ganeti from
1044 unmanaged (or managed via other software) instances.
1046 Another disk improvement is the possibility to convert between redundant
1047 (DRBD) and plain (LVM) disk configuration for an instance. This should
1048 allow better scalability (starting with one node and growing the
1049 cluster, or shrinking a two-node cluster to one node).
1051 A new feature that could help with automated node failovers has been
1052 implemented: if a node sees itself as offline (by querying the master
1053 candidates), it will try to shutdown (hard) all instances and any active
1054 DRBD devices. This reduces the risk of duplicate instances if an
1055 external script automatically failovers the instances on such nodes. To
1056 enable this, the cluster parameter ``maintain_node_health`` should be
1057 enabled; in the future this option (per the name) will enable other
1058 automatic maintenance features.
1060 Instance export/import now will reuse the original instance
1061 specifications for all parameters; that means exporting an instance,
1062 deleting it and the importing it back should give an almost identical
1063 instance. Note that the default import behaviour has changed from
1064 before, where it created only one NIC; now it recreates the original
1067 Cluster verify has added a few new checks: SSL certificates validity,
1068 /etc/hosts consistency across the cluster, etc.
1073 As usual, many internal changes were done, documentation fixes,
1076 - Fixed cluster initialization with disabled cluster storage (regression
1077 introduced in 2.1.1)
1078 - File-based storage supports growing the disks
1079 - Fixed behaviour of node role changes
1080 - Fixed cluster verify for some corner cases, plus a general rewrite of
1081 cluster verify to allow future extension with more checks
1082 - Fixed log spamming by watcher and node daemon (regression introduced
1084 - Fixed possible validation issues when changing the list of enabled
1086 - Fixed cleanup of /etc/hosts during node removal
1087 - Fixed RAPI response for invalid methods
1088 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1089 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1090 from ide disks, etc.)
1091 - Allow OS changes without re-installation (to record a changed OS
1092 outside of Ganeti, or to allow OS renames)
1093 - Allow instance creation without OS installation (useful for example if
1094 the OS will be installed manually, or restored from a backup not in
1096 - Implemented option to make cluster ``copyfile`` use the replication
1098 - Added list of enabled hypervisors to ssconf (possibly useful for
1100 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1102 - A partial form of node re-IP is possible via node readd, which now
1103 allows changed node primary IP
1104 - Command line utilities now show an informational message if the job is
1106 - The logs of the master daemon now show the PID/UID/GID of the
1113 *(Released Fri, 12 Mar 2010)*
1115 During the 2.1.0 long release candidate cycle, a lot of improvements and
1116 changes have accumulated with were released later as 2.1.1.
1121 The node evacuate command (``gnt-node evacuate``) was significantly
1122 rewritten, and as such the IAllocator protocol was changed - a new
1123 request type has been added. This unfortunate change during a stable
1124 series is designed to improve performance of node evacuations; on
1125 clusters with more than about five nodes and which are well-balanced,
1126 evacuation should proceed in parallel for all instances of the node
1127 being evacuated. As such, any existing IAllocator scripts need to be
1128 updated, otherwise the above command will fail due to the unknown
1129 request. The provided "dumb" allocator has not been updated; but the
1130 ganeti-htools package supports the new protocol since version 0.2.4.
1132 Another important change is increased validation of node and instance
1133 names. This might create problems in special cases, if invalid host
1134 names are being used.
1136 Also, a new layer of hypervisor parameters has been added, that sits at
1137 OS level between the cluster defaults and the instance ones. This allows
1138 customisation of virtualization parameters depending on the installed
1139 OS. For example instances with OS 'X' may have a different KVM kernel
1140 (or any other parameter) than the cluster defaults. This is intended to
1141 help managing a multiple OSes on the same cluster, without manual
1142 modification of each instance's parameters.
1144 A tool for merging clusters, ``cluster-merge``, has been added in the
1145 tools sub-directory.
1150 - Improved the int/float conversions that should make the code more
1151 robust in face of errors from the node daemons
1152 - Fixed the remove node code in case of internal configuration errors
1153 - Fixed the node daemon behaviour in face of inconsistent queue
1154 directory (e.g. read-only file-system where we can't open the files
1156 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1157 now it either aborts cleanly or, if given the new "auto_promote"
1158 parameter, will automatically promote other nodes as needed
1159 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1160 completely prevent Ganeti from working
1161 - Fixed bug for instance export when not all disks were successfully
1163 - Fixed behaviour of node add when the new node is slow in starting up
1165 - Fixed handling of signals in the LUXI client, which should improve
1166 behaviour of command-line scripts
1167 - Added checks for invalid node/instance names in the configuration (now
1168 flagged during cluster verify)
1169 - Fixed watcher behaviour for disk activation errors
1170 - Fixed two potentially endless loops in http library, which led to the
1171 RAPI daemon hanging and consuming 100% CPU in some cases
1172 - Fixed bug in RAPI daemon related to hashed passwords
1173 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1175 - Enhanced compatibility with non-Debian OSes, but not using absolute
1176 path in some commands and allowing customisation of the ssh
1177 configuration directory
1178 - Fixed possible future issue with new Python versions by abiding to the
1179 proper use of ``__slots__`` attribute on classes
1180 - Added checks that should prevent directory traversal attacks
1181 - Many documentation fixes based on feedback from users
1186 - Added an "early_release" more for instance replace disks and node
1187 evacuate, where we release locks earlier and thus allow higher
1188 parallelism within the cluster
1189 - Added watcher hooks, intended to allow the watcher to restart other
1190 daemons (e.g. from the ganeti-nbma project), but they can be used of
1191 course for any other purpose
1192 - Added a compile-time disable for DRBD barriers, to increase
1193 performance if the administrator trusts the power supply or the
1194 storage system to not lose writes
1195 - Added the option of using syslog for logging instead of, or in
1196 addition to, Ganeti's own log files
1197 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1198 can indeed boot from a paravirtual NIC
1199 - Added a generic debug level for many operations; while this is not
1200 used widely yet, it allows one to pass the debug value all the way to
1202 - Enhanced the hooks environment for instance moves (failovers,
1203 migrations) where the primary/secondary nodes changed during the
1204 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1205 - Enhanced data validations for many user-supplied values; one important
1206 item is the restrictions imposed on instance and node names, which
1207 might reject some (invalid) host names
1208 - Add a configure-time option to disable file-based storage, if it's not
1209 needed; this allows greater security separation between the master
1210 node and the other nodes from the point of view of the inter-node RPC
1212 - Added user notification in interactive tools if job is waiting in the
1213 job queue or trying to acquire locks
1214 - Added log messages when a job is waiting for locks
1215 - Added filtering by node tags in instance operations which admit
1216 multiple instances (start, stop, reboot, reinstall)
1217 - Added a new tool for cluster mergers, ``cluster-merge``
1218 - Parameters from command line which are of the form ``a=b,c=d`` can now
1219 use backslash escapes to pass in values which contain commas,
1220 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1222 - For KVM, the instance name is the first parameter passed to KVM, so
1223 that it's more visible in the process list
1229 *(Released Tue, 2 Mar 2010)*
1231 Ganeti 2.1 brings many improvements with it. Major changes:
1233 - Added infrastructure to ease automated disk repairs
1234 - Added new daemon to export configuration data in a cheaper way than
1235 using the remote API
1236 - Instance NICs can now be routed instead of being associated with a
1238 - Improved job locking logic to reduce impact of jobs acquiring multiple
1239 locks waiting for other long-running jobs
1241 In-depth implementation details can be found in the Ganeti 2.1 design
1247 - Added chroot hypervisor
1248 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1250 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1251 ``bootloader_path`` and ``bootloader_args``)
1252 - Added the ``use_localtime`` option for the xen-hvm and kvm
1253 hypervisors, and the default value for this has changed to false (in
1254 2.0 xen-hvm always enabled it)
1255 - Added luxi call to submit multiple jobs in one go
1256 - Added cluster initialization option to not modify ``/etc/hosts``
1258 - Added network interface parameters
1259 - Added dry run mode to some LUs
1260 - Added RAPI resources:
1262 - ``/2/instances/[instance_name]/info``
1263 - ``/2/instances/[instance_name]/replace-disks``
1264 - ``/2/nodes/[node_name]/evacuate``
1265 - ``/2/nodes/[node_name]/migrate``
1266 - ``/2/nodes/[node_name]/role``
1267 - ``/2/nodes/[node_name]/storage``
1268 - ``/2/nodes/[node_name]/storage/modify``
1269 - ``/2/nodes/[node_name]/storage/repair``
1271 - Added OpCodes to evacuate or migrate all instances on a node
1272 - Added new command to list storage elements on nodes (``gnt-node
1273 list-storage``) and modify them (``gnt-node modify-storage``)
1274 - Added new ssconf files with master candidate IP address
1275 (``ssconf_master_candidates_ips``), node primary IP address
1276 (``ssconf_node_primary_ips``) and node secondary IP address
1277 (``ssconf_node_secondary_ips``)
1278 - Added ``ganeti-confd`` and a client library to query the Ganeti
1279 configuration via UDP
1280 - Added ability to run hooks after cluster initialization and before
1282 - Added automatic mode for disk replace (``gnt-instance replace-disks
1284 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1285 after catastrophic data-loss
1286 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1288 - Added ``gnt-instance move`` command to move instances
1289 - Added ``gnt-cluster watcher`` command to control watcher
1290 - Added ``gnt-node powercycle`` command to powercycle nodes
1291 - Added new job status field ``lock_status``
1292 - Added parseable error codes to cluster verification (``gnt-cluster
1293 verify --error-codes``) and made output less verbose (use
1294 ``--verbose`` to restore previous behaviour)
1295 - Added UUIDs to the main config entities (cluster, nodes, instances)
1296 - Added support for OS variants
1297 - Added support for hashed passwords in the Ganeti remote API users file
1299 - Added option to specify maximum timeout on instance shutdown
1300 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1301 - Added new helper script to start and stop Ganeti daemons
1302 (``daemon-util``), with the intent to reduce the work necessary to
1303 adjust Ganeti for non-Debian distributions and to start/stop daemons
1305 - Added more unittests
1306 - Fixed critical bug in ganeti-masterd startup
1307 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1308 now customisable at the cluster level for both the KVM and Xen
1309 hypervisors using the new ``migration_port`` parameter
1310 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1311 reinstalling an instance
1312 - Allowed ``@`` in tag names
1313 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1314 - Many documentation updates
1315 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1316 - ``gnt-instance reinstall`` can now reinstall multiple instances
1317 - Updated many command line parameters
1318 - Introduced new OS API version 15
1319 - No longer support a default hypervisor
1320 - Treat virtual LVs as inexistent
1321 - Improved job locking logic to reduce lock contention
1322 - Match instance and node names case insensitively
1323 - Reimplemented bash completion script to be more complete
1330 *(Released Thu, 4 Feb 2010)*
1332 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1333 - Fix a string formatting bug
1334 - Improve safety of the code in some error paths
1335 - Improve data validation in the master of values returned from nodes
1341 *(Released Thu, 17 Dec 2009)*
1343 - Fix security issue due to missing validation of iallocator names; this
1344 allows local and remote execution of arbitrary executables
1345 - Fix failure of gnt-node list during instance removal
1346 - Ship the RAPI documentation in the archive
1352 *(Released Wed, 30 Sep 2009)*
1354 - Fixed many wrong messages
1355 - Fixed a few bugs related to the locking library
1356 - Fixed MAC checking at instance creation time
1357 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1358 - Fixed a few issues related to signal handling in both daemons and
1360 - Fixed the example startup script provided
1361 - Fixed insserv dependencies in the example startup script (patch from
1363 - Fixed handling of drained nodes in the iallocator framework
1364 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1366 - Fixed error related to invalid job IDs in job polling
1367 - Fixed job/opcode persistence on unclean master shutdown
1368 - Fixed handling of partial job processing after unclean master
1370 - Fixed error reporting from LUs, previously all errors were converted
1371 into execution errors
1372 - Fixed error reporting from burnin
1373 - Decreased significantly the memory usage of the job queue
1374 - Optimised slightly multi-job submission
1375 - Optimised slightly opcode loading
1376 - Backported the multi-job submit framework from the development
1377 branch; multi-instance start and stop should be faster
1378 - Added script to clean archived jobs after 21 days; this will reduce
1379 the size of the queue directory
1380 - Added some extra checks in disk size tracking
1381 - Added an example ethers hook script
1382 - Added a cluster parameter that prevents Ganeti from modifying of
1384 - Added more node information to RAPI responses
1385 - Added a ``gnt-job watch`` command that allows following the ouput of a
1387 - Added a bind-address option to ganeti-rapi
1388 - Added more checks to the configuration verify
1389 - Enhanced the burnin script such that some operations can be retried
1391 - Converted instance reinstall to multi-instance model
1397 *(Released Fri, 7 Aug 2009)*
1399 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1400 to allow using the pre-2.0.2 behaviour in activation, if any existing
1401 instances have mismatched disk sizes in the configuration
1402 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
1403 any configuration mismatches for disk sizes
1404 - Added ``gnt-master cluste-failover --no-voting`` to allow master
1405 failover to work on two-node clusters
1406 - Fixed the ``--net`` option of ``gnt-backup import``, which was
1408 - Fixed detection of OS script errors in ``gnt-backup export``
1409 - Fixed exit code of ``gnt-backup export``
1415 *(Released Fri, 17 Jul 2009)*
1417 - Added experimental support for stripped logical volumes; this should
1418 enhance performance but comes with a higher complexity in the block
1419 device handling; stripping is only enabled when passing
1420 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
1421 affected even in the non-stripped mode
1422 - Improved resiliency against transient failures at the end of DRBD
1423 resyncs, and in general of DRBD resync checks
1424 - Fixed a couple of issues with exports and snapshot errors
1425 - Fixed a couple of issues in instance listing
1426 - Added display of the disk size in ``gnt-instance info``
1427 - Fixed checking for valid OSes in instance creation
1428 - Fixed handling of the "vcpus" parameter in instance listing and in
1429 general of invalid parameters
1430 - Fixed http server library, and thus RAPI, to handle invalid
1431 username/password combinations correctly; this means that now they
1432 report unauthorized for queries too, not only for modifications,
1433 allowing earlier detect of configuration problems
1434 - Added a new "role" node list field, equivalent to the master/master
1435 candidate/drained/offline flags combinations
1436 - Fixed cluster modify and changes of candidate pool size
1437 - Fixed cluster verify error messages for wrong files on regular nodes
1438 - Fixed a couple of issues with node demotion from master candidate role
1439 - Fixed node readd issues
1440 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
1441 - Added a new ``--no-voting`` option for masterfailover to fix failover
1442 on two-nodes clusters when the former master node is unreachable
1443 - Added instance reinstall over RAPI
1449 *(Released Tue, 16 Jun 2009)*
1451 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
1452 allow re-adding the start in single-user option (regression from 1.2)
1453 - the watcher writes the instance status to a file, to allow monitoring
1454 to report the instance status (from the master) based on cached
1455 results of the watcher's queries; while this can get stale if the
1456 watcher is being locked due to other work on the cluster, this is
1457 still an improvement
1458 - the watcher now also restarts the node daemon and the rapi daemon if
1460 - fixed the watcher to handle full and drained queue cases
1461 - hooks export more instance data in the environment, which helps if
1462 hook scripts need to take action based on the instance's properties
1463 (no longer need to query back into ganeti)
1464 - instance failovers when the instance is stopped do not check for free
1465 RAM, so that failing over a stopped instance is possible in low memory
1467 - rapi uses queries for tags instead of jobs (for less job traffic), and
1468 for cluster tags it won't talk to masterd at all but read them from
1470 - a couple of error handling fixes in RAPI
1471 - drbd handling: improved the error handling of inconsistent disks after
1472 resync to reduce the frequency of "there are some degraded disks for
1473 this instance" messages
1474 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
1475 error handling path called a wrong function name)
1481 *(Released Wed, 27 May 2009)*
1483 - no changes from rc5
1486 Version 2.0 release candidate 5
1487 -------------------------------
1489 *(Released Wed, 20 May 2009)*
1491 - fix a couple of bugs (validation, argument checks)
1492 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
1493 - some small improvements to RAPI and IAllocator
1494 - make watcher automatically start the master daemon if down
1497 Version 2.0 release candidate 4
1498 -------------------------------
1500 *(Released Mon, 27 Apr 2009)*
1502 - change the OS list to not require locks; this helps with big clusters
1503 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
1504 volume group is broken
1505 - ``gnt-instance info``, without any arguments, doesn't run for all
1506 instances anymore; either pass ``--all`` or pass the desired
1507 instances; this helps against mistakes on big clusters where listing
1508 the information for all instances takes a long time
1509 - miscellaneous doc and man pages fixes
1512 Version 2.0 release candidate 3
1513 -------------------------------
1515 *(Released Wed, 8 Apr 2009)*
1517 - Change the internal locking model of some ``gnt-node`` commands, in
1518 order to reduce contention (and blocking of master daemon) when
1519 batching many creation/reinstall jobs
1520 - Fixes to Xen soft reboot
1521 - No longer build documentation at build time, instead distribute it in
1522 the archive, in order to reduce the need for the whole docbook/rst
1526 Version 2.0 release candidate 2
1527 -------------------------------
1529 *(Released Fri, 27 Mar 2009)*
1531 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
1532 - Fix watcher startup sequence, improves the behaviour of busy clusters
1533 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
1534 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
1535 bind address and other places
1536 - Some documentation fixes and updates
1539 Version 2.0 release candidate 1
1540 -------------------------------
1542 *(Released Mon, 2 Mar 2009)*
1544 - More documentation updates, now all docs should be more-or-less
1546 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
1548 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
1550 - Fix an issue related to $libdir/run/ganeti and cluster creation
1556 *(Released Thu, 19 Feb 2009)*
1558 - Xen PVM and KVM have switched the default value for the instance root
1559 disk to the first partition on the first drive, instead of the whole
1560 drive; this means that the OS installation scripts must be changed
1562 - Man pages have been updated
1563 - RAPI has been switched by default to HTTPS, and the exported functions
1564 should all work correctly
1565 - RAPI v1 has been removed
1566 - Many improvements to the KVM hypervisor
1567 - Block device errors are now better reported
1568 - Many other bugfixes and small improvements
1574 *(Released Mon, 26 Jan 2009)*
1576 - Version 2 is a general rewrite of the code and therefore the
1577 differences are too many to list, see the design document for 2.0 in
1578 the ``doc/`` subdirectory for more details
1579 - In this beta version there is not yet a migration path from 1.2 (there
1580 will be one in the final 2.0 release)
1581 - A few significant changes are:
1583 - all commands are executed by a daemon (``ganeti-masterd``) and the
1584 various ``gnt-*`` commands are just front-ends to it
1585 - all the commands are entered into, and executed from a job queue,
1586 see the ``gnt-job(8)`` manpage
1587 - the RAPI daemon supports read-write operations, secured by basic
1588 HTTP authentication on top of HTTPS
1589 - DRBD version 0.7 support has been removed, DRBD 8 is the only
1590 supported version (when migrating from Ganeti 1.2 to 2.0, you need
1591 to migrate to DRBD 8 first while still running Ganeti 1.2)
1592 - DRBD devices are using statically allocated minor numbers, which
1593 will be assigned to existing instances during the migration process
1594 - there is support for both Xen PVM and Xen HVM instances running on
1596 - KVM virtualization is supported too
1597 - file-based storage has been implemented, which means that it is
1598 possible to run the cluster without LVM and DRBD storage, for
1599 example using a shared filesystem exported from shared storage (and
1600 still have live migration)
1606 *(Released Tue, 13 Jan 2009)*
1608 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
1609 - Reuse the old instance mac address by default on instance import, if
1610 the instance name is the same.
1611 - Handle situations in which the node info rpc returns incomplete
1613 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
1614 - Improved version of batcher:
1616 - state file support
1617 - instance mac address support
1618 - support for HVM clusters/instances
1620 - Add an option to show the number of cpu sockets and nodes in
1622 - Support OSes that handle more than one version of the OS api (but do
1623 not change the current API in any other way)
1624 - Fix ``gnt-node migrate``
1625 - ``gnt-debug`` man page
1626 - Fixes various more typos and small issues
1627 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
1633 *(Released Wed, 24 Sep 2008)*
1635 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
1636 type of disk exported to fully virtualized instances.
1637 - provide access to the serial console of HVM instances
1638 - instance auto_balance flag, set by default. If turned off it will
1639 avoid warnings on cluster verify if there is not enough memory to fail
1640 over an instance. in the future it will prevent automatically failing
1641 it over when we will support that.
1642 - batcher tool for instance creation, see ``tools/README.batcher``
1643 - ``gnt-instance reinstall --select-os`` to interactively select a new
1644 operating system when reinstalling an instance.
1645 - when changing the memory amount on instance modify a check has been
1646 added that the instance will be able to start. also warnings are
1647 emitted if the instance will not be able to fail over, if auto_balance
1649 - documentation fixes
1650 - sync fields between ``gnt-instance list/modify/add/import``
1651 - fix a race condition in drbd when the sync speed was set after giving
1652 the device a remote peer.
1658 *(Released Tue, 22 Jul 2008)*
1660 - note: the allowed size and number of tags per object were reduced
1661 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1662 - fixed twisted 8.x compatibility
1663 - fixed ``gnt-instance replace-disks`` with iallocator
1664 - add TCP keepalives on twisted connections to detect restarted nodes
1665 - disk increase support, see ``gnt-instance grow-disk``
1666 - implement bulk node/instance query for RAPI
1667 - add tags in node/instance listing (optional)
1668 - experimental migration (and live migration) support, read the man page
1669 for ``gnt-instance migrate``
1670 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1671 has some small improvements in handling its state file
1677 *(Released Fri, 13 Jun 2008)*
1679 - Experimental readonly, REST-based remote API implementation;
1680 automatically started on master node, TCP port 5080, if enabled by
1681 ``--enable-rapi`` parameter to configure script.
1682 - Instance allocator support. Add and import instance accept a
1683 ``--iallocator`` parameter, and call that instance allocator to decide
1684 which node to use for the instance. The iallocator document describes
1685 what's expected from an allocator script.
1686 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1687 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1688 node is lost there is still enough memory to fail over the instances
1690 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1691 ``gnt-cluster verify``, to check for site-specific compliance. All the
1692 hooks will run, and their output, if any, will be displayed. Any
1693 failing hook will make the verification return an error value.
1694 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1695 primary and secondary interfaces
1696 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1697 that is still declared as part of the cluster and has failed.
1698 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1699 specifying output fields, that just adds the chosen fields to the
1701 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1702 export from the filesystem.
1703 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1704 have been added. Using them you can enable/disable acpi and pae
1705 support, and specify a path for a cd image to be exported to the
1706 instance. These parameters as the name suggest only work on HVM
1708 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1709 PAE support will be set to the previously hardcoded values, but the
1710 (previously hardcoded) path to the CDROM ISO image will be unset and
1711 if required, needs to be set manually with ``gnt-instance modify``
1713 - The address to which an instance's VNC console is bound is now
1714 selectable per-instance, rather than being cluster wide. Of course
1715 this only applies to instances controlled via VNC, so currently just
1716 applies to HVM clusters.
1722 *(Released Mon, 18 Feb 2008)*
1724 - more tweaks to the disk activation code (especially helpful for DRBD)
1725 - change the default ``gnt-instance list`` output format, now there is
1726 one combined status field (see the manpage for the exact values this
1728 - some more fixes for the mac export to hooks change
1729 - make Ganeti not break with DRBD 8.2.x (which changed the version
1730 format in ``/proc/drbd``) (issue 24)
1731 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1732 template, allowing migration from DRBD0.7+MD to DRBD8
1738 *(Released Wed, 30 Jan 2008)*
1740 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1742 - add command aliases infrastructure and a few aliases
1743 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1744 man pages and the ``--help`` option of ``gnt-node
1745 list``/``gnt-instance list``
1746 - fix ``gnt-backup list`` with down nodes (issue 21)
1747 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1748 - fix the dist archive and add a check for including svn/git files in
1750 - some developer-related changes: improve the burnin and the QA suite,
1751 add an upload script for testing during development
1757 *(Released Wed, 16 Jan 2008)*
1759 - experimental HVM support, read the install document, section
1760 "Initializing the cluster"
1761 - allow for the PVM hypervisor per-instance kernel and initrd paths
1762 - add a new command ``gnt-cluster verify-disks`` which uses a new
1763 algorithm to improve the reconnection of the DRBD pairs if the device
1764 on the secondary node has gone away
1765 - make logical volume code auto-activate LVs at disk activation time
1766 - slightly improve the speed of activating disks
1767 - allow specification of the MAC address at instance creation time, and
1768 changing it later via ``gnt-instance modify``
1769 - fix handling of external commands that generate lots of output on
1771 - update documentation with regard to minimum version of DRBD8 supported
1777 *(Released Tue, 4 Dec 2007)*
1779 - Log the ``xm create`` output to the node daemon log on failure (to
1780 help diagnosing the error)
1781 - In debug mode, log all external commands output if failed to the logs
1782 - Change parsing of lvm commands to ignore stderr
1788 *(Released Wed, 28 Nov 2007)*
1790 - Another round of updates to the DRBD 8 code to deal with more failures
1791 in the replace secondary node operation
1792 - Some more logging of failures in disk operations (lvm, drbd)
1793 - A few documentation updates
1800 *(Released Tue, 13 Nov 2007)*
1802 - Change configuration file format from Python's Pickle to JSON.
1803 Upgrading is possible using the cfgupgrade utility.
1804 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1805 faster replace disks and is more stable (DRBD 8 has many improvements
1806 compared to DRBD 0.7)
1807 - Added command line tags support (see man pages for ``gnt-instance``,
1808 ``gnt-node``, ``gnt-cluster``)
1809 - Added instance rename support
1810 - Added multi-instance startup/shutdown
1811 - Added cluster rename support
1812 - Added ``gnt-node evacuate`` to simplify some node operations
1813 - Added instance reboot operation that can speedup reboot as compared to
1815 - Soften the requirement that hostnames are in FQDN format
1816 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1818 - Removed dependency on debian's patched fping that uses the
1819 non-standard ``-S`` option
1820 - Now the OS definitions are searched for in multiple, configurable
1821 paths (easier for distros to package)
1822 - Some changes to the hooks infrastructure (especially the new
1823 post-configuration update hook)
1824 - Other small bugfixes
1826 .. vim: set textwidth=72 syntax=rst :