8 *(Released Mon, 9 Jan 2012)*
10 Incompatible/important changes and bugfixes
11 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
13 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
14 resource's ``ip_check`` parameter changed from ``True`` to ``False``
15 to match the underlying LUXI interface.
16 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
17 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
18 not maintain backwards-compatibility as the underlying operation
19 changed in an incompatible way. The RAPI client can talk to old
20 servers, but it needs to be told so as the return value changed.
21 - When creating file-based instances via RAPI, the ``file_driver``
22 parameter no longer defaults to ``loop`` and must be specified.
23 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
25 - Support for the undocumented and deprecated RAPI instance creation
26 request format version 0 has been dropped. Use version 1, supported
27 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
28 - Pyparsing 1.4.6 or above is required, see :doc:`installation
29 documentation <install>`.
30 - The "cluster-verify" hooks are now executed per group by the
31 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
32 if you just run ``gnt-cluster verify``, which generates one opcode per
34 - The environment as passed to the OS scripts is cleared, and thus no
35 environment variables defined in the node daemon's environment will be
36 inherited by the scripts.
37 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
39 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
40 support operations involving multiple node groups.
41 - Offline nodes are ignored when failing over an instance.
42 - Support for KVM version 1.0, which changed the version reporting format
44 - Includes all bugfixes made in the 2.4 series
49 - The ganeti-htools project has been merged into the ganeti-core source
50 tree and will be built as part of Ganeti (see :doc:`install-quick`).
51 - Implemented support for :doc:`shared storage <design-shared-storage>`.
52 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
53 GPT-style partition tables (requires `parted
54 <http://www.gnu.org/s/parted/>`_).
55 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
56 - Allowed adding tags on instance creation.
57 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
58 - Allow instances to be started in a paused state, enabling the user to
59 see the complete console output on boot using the console.
60 - Added new hypervisor flag to control default reboot behaviour
61 (``reboot_behavior``).
62 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
63 - Improved out-of-band management support:
65 - Added ``gnt-node health`` command reporting the health status of
67 - Added ``gnt-node power`` command to manage power status of nodes.
68 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
70 - Instance migration can fall back to failover if instance is not
72 - Filters can be used when listing nodes, instances, groups and locks;
73 see *ganeti(7)* manpage.
74 - Added post-execution status as variables to :doc:`hooks <hooks>`
76 - Instance tags are exported/imported together with the instance.
77 - When given an explicit job ID, ``gnt-job info`` will work for archived
79 - Jobs can define dependencies on other jobs (not yet supported via
80 RAPI or command line, but used by internal commands and usable via
83 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
86 - Instance failover is now available as a RAPI resource
87 (``/2/instances/[instance_name]/failover``).
88 - ``gnt-instance info`` defaults to static information if primary node
90 - Opcodes have a new ``comment`` attribute.
91 - Added basic SPICE support to KVM hypervisor.
92 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
94 Node group improvements
95 ~~~~~~~~~~~~~~~~~~~~~~~
97 - ``gnt-cluster verify`` has been modified to check groups separately,
98 thereby improving performance.
99 - Node group support has been added to ``gnt-cluster verify-disks``,
100 which now operates per node group.
101 - Watcher has been changed to work better with node groups.
103 - One process and state file per node group.
104 - Slow watcher in one group doesn't block other group's watcher.
106 - Added new command, ``gnt-group evacuate``, to move all instances in a
107 node group to other groups.
108 - Added ``gnt-instance change-group`` to move an instance to another
110 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
111 per-group operations.
112 - Node groups can be tagged.
113 - Some operations switch from an exclusive to a shared lock as soon as
115 - Instance's primary and secondary nodes' groups are now available as
116 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
117 and ``snodes.group.uuid``).
122 - Numerous updates to documentation and manpages.
124 - :doc:`RAPI <rapi>` documentation now has detailed parameter
126 - Some opcode/job results are now also documented, see :doc:`RAPI
129 - A lockset's internal lock is now also visible in lock monitor.
130 - Log messages from job queue workers now contain information about the
131 opcode they're processing.
132 - ``gnt-instance console`` no longer requires the instance lock.
133 - A short delay when waiting for job changes reduces the number of LUXI
134 requests significantly.
135 - DRBD metadata volumes are overwritten with zeros during disk creation.
141 *(Released Thu, 27 Oct 2011)*
143 This was the fourth release candidate of the 2.5 series.
149 *(Released Wed, 26 Oct 2011)*
151 This was the third release candidate of the 2.5 series.
157 *(Released Tue, 18 Oct 2011)*
159 This was the second release candidate of the 2.5 series.
165 *(Released Tue, 4 Oct 2011)*
167 This was the first release candidate of the 2.5 series.
173 *(Released Wed, 31 Aug 2011)*
175 This was the third beta release of the 2.5 series.
181 *(Released Mon, 22 Aug 2011)*
183 This was the second beta release of the 2.5 series.
189 *(Released Fri, 12 Aug 2011)*
191 This was the first beta release of the 2.5 series.
197 *(Released Thu, 27 Oct 2011)*
199 - Fixed bug when parsing command line parameter values ending in
201 - Fixed assertion error after unclean master shutdown
202 - Disable HTTP client pool for RPC, significantly reducing memory usage
204 - Fixed queue archive creation with wrong permissions
210 *(Released Tue, 23 Aug 2011)*
214 - Fixed documentation for importing with ``--src-dir`` option
215 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
216 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
222 *(Released Fri, 5 Aug 2011)*
224 Many bug-fixes and a few small features:
226 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
227 issues when you tried to add an instance with two MAC addresses in one
229 - KVM: fixed per-instance stored UID value
230 - KVM: configure bridged NICs at migration start
231 - KVM: Fix a bug where instance will not start with never KVM versions
233 - Added OS search path to ``gnt-cluster info``
234 - Fixed an issue with ``file_storage_dir`` where you were forced to
235 provide an absolute path, but the documentation states it is a
236 relative path, the documentation was right
237 - Added a new parameter to instance stop/start called ``--no-remember``
238 that will make the state change to not be remembered
239 - Implemented ``no_remember`` at RAPI level
240 - Improved the documentation
241 - Node evacuation: don't call IAllocator if node is already empty
242 - Fixed bug in DRBD8 replace disks on current nodes
243 - Fixed bug in recreate-disks for DRBD instances
244 - Moved assertion checking locks in ``gnt-instance replace-disks``
245 causing it to abort with not owning the right locks for some situation
246 - Job queue: Fixed potential race condition when cancelling queued jobs
247 - Fixed off-by-one bug in job serial generation
248 - ``gnt-node volumes``: Fix instance names
249 - Fixed aliases in bash completion
250 - Fixed a bug in reopening log files after being sent a SIGHUP
251 - Added a flag to burnin to allow specifying VCPU count
252 - Bugfixes to non-root Ganeti configuration
258 *(Released Thu, 12 May 2011)*
260 Many bug-fixes and a few new small features:
262 - Fixed a bug related to log opening failures
263 - Fixed a bug in instance listing with orphan instances
264 - Fixed a bug which prevented resetting the cluster-level node parameter
265 ``oob_program`` to the default
266 - Many fixes related to the ``cluster-merge`` tool
267 - Fixed a race condition in the lock monitor, which caused failures
268 during (at least) creation of many instances in parallel
269 - Improved output for gnt-job info
270 - Removed the quiet flag on some ssh calls which prevented debugging
272 - Improved the N+1 failure messages in cluster verify by actually
273 showing the memory values (needed and available)
274 - Increased lock attempt timeouts so that when executing long operations
275 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
276 too early and thus prevent the use of the 'fair' mechanism
277 - Changed instance query data (``gnt-instance info``) to not acquire
278 locks unless needed, thus allowing its use on locked instance if only
279 static information is asked for
280 - Improved behaviour with filesystems that do not support rename on an
282 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
283 kept locks on all nodes during the wipe, which is unneeded
284 - Fixed ``gnt-watcher`` handling of errors during hooks execution
285 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
286 10GiB) which caused the wipe to fail right at the end in some cases
287 - Fixed master IP activation when doing master failover with no-voting
288 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
289 the master node itself
290 - Fixed potential data-loss in under disk full conditions, where Ganeti
291 wouldn't check correctly the return code and would consider
292 partially-written files 'correct'
293 - Fixed bug related to multiple VGs and DRBD disk replacing
294 - Added new disk parameter ``metavg`` that allows placement of the meta
295 device for DRBD in a different volume group
296 - Fixed error handling in the node daemon when the system libc doesn't
297 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
298 - Fixed lock release during replace-disks, which kept cluster-wide locks
299 when doing disk replaces with an iallocator script
300 - Added check for missing bridges in cluster verify
301 - Handle EPIPE errors while writing to the terminal better, so that
302 piping the output to e.g. ``less`` doesn't cause a backtrace
303 - Fixed rare case where a ^C during Luxi calls could have been
304 interpreted as server errors, instead of simply terminating
305 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
307 - Added a few more parameters to the KVM hypervisor, allowing a second
308 CDROM, custom disk type for CDROMs and a floppy image
309 - Removed redundant message in instance rename when the name is given
311 - Added option to ``gnt-instance recreate-disks`` to allow creating the
312 disks on new nodes, allowing recreation when the original instance
313 nodes are completely gone
314 - Added option when converting disk templates to DRBD to skip waiting
315 for the resync, in order to make the instance available sooner
316 - Added two new variables to the OS scripts environment (containing the
318 - Made the root_path and optional parameter for the xen-pvm hypervisor,
319 to allow use of ``pvgrub`` as bootloader
320 - Changed the instance memory modifications to only check out-of-memory
321 conditions on memory increases, and turned the secondary node warnings
322 into errors (they can still be overridden via ``--force``)
323 - Fixed the handling of a corner case when the Python installation gets
324 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
325 to execute a command that doesn't exist
326 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
327 node of the instance returned failures during instance shutdown; this
328 adds the option ``--ignore-consistency`` to gnt-instance move
330 And as usual, various improvements to the error messages, documentation
336 *(Released Wed, 09 Mar 2011)*
338 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
339 the RAPI users file if run twice (even with ``--dry-run``).
341 The release fixes that bug (nothing else changed).
347 *(Released Mon, 07 Mar 2011)*
349 Final 2.4.0 release. Just a few small fixes:
351 - Fixed RAPI node evacuate
352 - Fixed the kvm-ifup script
353 - Fixed internal error handling for special job cases
354 - Updated man page to specify the escaping feature for options
360 *(Released Mon, 28 Feb 2011)*
362 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
363 that this feature wiped the disks of the wrong instance, leading to loss
368 - Fixed title of query field containing instance name
369 - Expanded the glossary in the documentation
370 - Fixed one unittest (internal issue)
376 *(Released Mon, 21 Feb 2011)*
378 A number of bug fixes plus just a couple functionality changes.
380 On the user-visible side, the ``gnt-* list`` command output has changed
381 with respect to "special" field states. The current rc1 style of display
382 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
383 the default output mode special fields states are displayed as follows:
385 - Offline resource: ``*``
386 - Unavailable/not applicable: ``-``
387 - Data missing (RPC failure): ``?``
388 - Unknown field: ``??``
390 Another user-visible change is the addition of ``--force-join`` to
395 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
396 - Fixed regression in RAPI/instance reinstall where all parameters were
397 required (instead of optional)
398 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
399 - Fixed iallocator usage (offline nodes were not considered offline)
400 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
401 - Fixed hypervisor and OS parameter validation with respect to
403 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
405 - Fixed ``tools/listrunner`` with respect to agent-based usage
411 *(Released Fri, 4 Feb 2011)*
413 Many changes and fixes since the beta1 release. While there were some
414 internal changes, the code has been mostly stabilised for the RC
417 Note: the dumb allocator was removed in this release, as it was not kept
418 up-to-date with the IAllocator protocol changes. It is recommended to
419 use the ``hail`` command from the ganeti-htools package.
421 Note: the 2.4 and up versions of Ganeti are not compatible with the
422 0.2.x branch of ganeti-htools. You need to upgrade to
423 ganeti-htools-0.3.0 (or later).
425 Regressions fixed from 2.3
426 ~~~~~~~~~~~~~~~~~~~~~~~~~~
428 - Fixed the ``gnt-cluster verify-disks`` command
429 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
431 - Fixed disk adoption breakage
432 - Fixed wrong headers in instance listing for field aliases
437 - Fixed corner case in KVM handling of NICs
438 - Fixed many cases of wrong handling of non-vm_capable nodes
439 - Fixed a bug where a missing instance symlink was not possible to
440 recreate with any ``gnt-*`` command (now ``gnt-instance
441 activate-disks`` does it)
442 - Fixed the volume group name as reported by ``gnt-cluster
444 - Increased timeouts for the import-export code, hopefully leading to
445 fewer aborts due network or instance timeouts
446 - Fixed bug in ``gnt-node list-storage``
447 - Fixed bug where not all daemons were started on cluster
448 initialisation, but only at the first watcher run
449 - Fixed many bugs in the OOB implementation
450 - Fixed watcher behaviour in presence of instances with offline
452 - Fixed instance list output for instances running on the wrong node
453 - a few fixes to the cluster-merge tool, but it still cannot merge
454 multi-node groups (currently it is not recommended to use this tool)
460 - Improved network configuration for the KVM hypervisor
461 - Added e1000 as a supported NIC for Xen-HVM
462 - Improved the lvmstrap tool to also be able to use partitions, as
463 opposed to full disks
464 - Improved speed of disk wiping (the cluster parameter
465 ``prealloc_wipe_disks``, so that it has a low impact on the total time
466 of instance creations
467 - Added documentation for the OS parameters
468 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
469 hypervisor is not responding
470 - Added display of blacklisted and hidden OS information in
472 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
473 NIC and node parameters, which might create problems with currently
474 invalid (but undetected) configuration files, but prevents validation
475 failures when unrelated parameters are modified
476 - Changed cluster initialisation to wait for the master daemon to become
478 - Expanded the RAPI interface:
480 - Added config redistribution resource
481 - Added activation/deactivation of instance disks
482 - Added export of console information
484 - Implemented log file reopening on SIGHUP, which allows using
485 logrotate(8) for the Ganeti log files
486 - Added a basic OOB helper script as an example
492 *(Released Fri, 14 Jan 2011)*
497 - Fixed timezone issues when formatting timestamps
498 - Added support for node groups, available via ``gnt-group`` and other
500 - Added out-of-band framework and management, see :doc:`design
501 document <design-oob>`
502 - Removed support for roman numbers from ``gnt-node list`` and
503 ``gnt-instance list``.
504 - Allowed modification of master network interface via ``gnt-cluster
505 modify --master-netdev``
506 - Accept offline secondaries while shutting down instance disks
507 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
508 - Added support for multiple LVM volume groups
509 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
511 - Added commands to list available fields:
513 - ``gnt-node list-fields``
514 - ``gnt-group list-fields``
515 - ``gnt-instance list-fields``
517 - Updated documentation and man pages
522 - Moved ``rapi_users`` file into separate directory, now named
523 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
525 - Added new tool for running commands on many machines,
526 ``tools/ganeti-listrunner``
527 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
528 improving the ``gnt-instance console`` output
529 - Allowed customisation of disk index separator at ``configure`` time
530 - Export node group allocation policy to :doc:`iallocator <iallocator>`
531 - Added support for non-partitioned md disks in ``lvmstrap``
532 - Added script to gracefully power off KVM instances
533 - Split ``utils`` module into smaller parts
534 - Changed query operations to return more detailed information, e.g.
535 whether an information is unavailable due to an offline node. To use
536 this new functionality, the LUXI call ``Query`` must be used. Field
537 information is now stored by the master daemon and can be retrieved
538 using ``QueryFields``. Instances, nodes and groups can also be queried
539 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
540 exposed via RAPI). The following commands make use of this
541 infrastructure change:
544 - ``gnt-group list-fields``
546 - ``gnt-node list-fields``
547 - ``gnt-instance list``
548 - ``gnt-instance list-fields``
549 - ``gnt-debug locks``
554 - New RAPI resources (see :doc:`rapi`):
558 - ``/2/groups/[group_name]``
559 - ``/2/groups/[group_name]/assign-nodes``
560 - ``/2/groups/[group_name]/modify``
561 - ``/2/groups/[group_name]/rename``
562 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
566 - Implemented ``no_install`` for instance creation
567 - Implemented OS parameters for instance reinstallation, allowing
568 use of special settings on reinstallation (e.g. for preserving data)
573 - Added IPv6 support in import/export
574 - Pause DRBD synchronization while wiping disks on instance creation
575 - Updated unittests and QA scripts
576 - Improved network parameters passed to KVM
577 - Converted man pages from docbook to reStructuredText
583 *(Released Mon, 20 Dec 2010)*
585 Released version 2.3.1~rc1 without any changes.
591 *(Released Wed, 1 Dec 2010)*
593 - impexpd: Disable OpenSSL compression in socat if possible (backport
594 from master, commit e90739d625b, see :doc:`installation guide
595 <install-quick>` for details)
596 - Changed unittest coverage report to exclude test scripts
597 - Added script to check version format
603 *(Released Wed, 1 Dec 2010)*
605 Released version 2.3.0~rc1 without any changes.
611 *(Released Fri, 19 Nov 2010)*
613 A number of bugfixes and documentation updates:
615 - Update ganeti-os-interface documentation
616 - Fixed a bug related to duplicate MACs or similar items which should be
618 - Fix breakage in OS state modify
619 - Reinstall instance: disallow offline secondaries (fixes bug related to
620 OS changing but reinstall failing)
621 - plus all the other fixes between 2.2.1 and 2.2.2
627 *(Released Tue, 2 Nov 2010)*
629 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
630 - Fixed master failover race with watcher
631 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
633 - Accept previously stopped instance for export with instance removal
634 - Simplify and extend the environment variables for instance OS scripts
635 - Added new node flags, ``master_capable`` and ``vm_capable``
636 - Added optional instance disk wiping prior during allocation. This is a
637 cluster-wide option and can be set/modified using
638 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
639 - Added IPv6 support, see :doc:`design document <design-2.3>` and
641 - Added a new watcher option (``--ignore-pause``)
642 - Added option to ignore offline node on instance start/stop
643 (``--ignore-offline``)
644 - Allow overriding OS parameters with ``gnt-instance reinstall``
645 - Added ability to change node's secondary IP address using ``gnt-node
647 - Implemented privilege separation for all daemons except
648 ``ganeti-noded``, see ``configure`` options
649 - Complain if an instance's disk is marked faulty in ``gnt-cluster
651 - Implemented job priorities (see ``ganeti(7)`` manpage)
652 - Ignore failures while shutting down instances during failover from
654 - Exit daemon's bootstrap process only once daemon is ready
655 - Export more information via ``LUInstanceQuery``/remote API
656 - Improved documentation, QA and unittests
657 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
658 restart if the file was created or changed
659 - Added LUXI protocol version sent with each request and response,
660 allowing detection of server/client mismatches
661 - Moved the Python scripts among gnt-* and ganeti-* into modules
662 - Moved all code related to setting up SSH to an external script,
664 - Infrastructure changes for node group support in future versions
670 *(Released Fri, 19 Nov 2010)*
672 A few small bugs fixed, and some improvements to the build system:
674 - Fix documentation regarding conversion to drbd
675 - Fix validation of parameters in cluster modify (``gnt-cluster modify
677 - Fix error handling in node modify with multiple changes
678 - Allow remote imports without checked names
684 *(Released Tue, 19 Oct 2010)*
686 - Disable SSL session ID cache in RPC client
692 *(Released Thu, 14 Oct 2010)*
694 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
695 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
697 - Fix problems with interaction between Curl and Python's HTTP server,
698 resulting in increased speed in many RPC calls
699 - Improve our release script to prevent breakage with older aclocal and
706 *(Released Thu, 7 Oct 2010)*
708 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
709 value retrieved from master
710 - Added support for blacklisted or hidden OS definitions
711 - Added simple lock monitor (accessible via (``gnt-debug locks``)
712 - Added support for -mem-path in KVM hypervisor abstraction layer
713 - Allow overriding instance parameters in tool for inter-cluster
714 instance moves (``tools/move-instance``)
715 - Improved opcode summaries (e.g. in ``gnt-job list``)
716 - Improve consistency of OS listing by sorting it
717 - Documentation updates
723 *(Released Fri, 8 Oct 2010)*
725 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
731 *(Released Mon, 4 Oct 2010)*
733 - Fixed regression in ``gnt-instance rename``
739 *(Released Wed, 22 Sep 2010)*
741 - Fixed OS_VARIANT variable for OS scripts
742 - Fixed cluster tag operations via RAPI
743 - Made ``setup-ssh`` exit with non-zero code if an error occurred
744 - Disabled RAPI CA checks in watcher
750 *(Released Mon, 23 Aug 2010)*
752 - Support DRBD versions of the format "a.b.c.d"
754 - Re-introduce support for usage from multiple threads in RAPI client
755 - Instance renames and modify via RAPI
756 - Work around race condition between processing and archival in job
758 - Mark opcodes following failed one as failed, too
759 - Job field ``lock_status`` was removed due to difficulties making it
760 work with the changed job queue in Ganeti 2.2; a better way to monitor
761 locks is expected for a later 2.2.x release
762 - Fixed dry-run behaviour with many commands
763 - Support ``ssh-agent`` again when adding nodes
764 - Many additional bugfixes
770 *(Released Fri, 30 Jul 2010)*
772 Important change: the internal RPC mechanism between Ganeti nodes has
773 changed from using a home-grown http library (based on the Python base
774 libraries) to use the PycURL library. This requires that PycURL is
775 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
776 against GnuTLS by default. cURL's support for GnuTLS had known issues
777 before cURL 7.21.0 and we recommend using the latest cURL release or
778 linking against OpenSSL. Most other distributions already link PycURL
779 and cURL against OpenSSL. The command::
781 python -c 'import pycurl; print pycurl.version'
783 can be used to determine the libraries PycURL and cURL are linked
786 Other significant changes:
788 - Rewrote much of the internals of the job queue, in order to achieve
789 better parallelism; this decouples job query operations from the job
790 processing, and it should allow much nicer behaviour of the master
791 daemon under load, and it also has uncovered some long-standing bugs
792 related to the job serialisation (now fixed)
793 - Added a default iallocator setting to the cluster parameters,
794 eliminating the need to always pass nodes or an iallocator for
795 operations that require selection of new node(s)
796 - Added experimental support for the LXC virtualization method
797 - Added support for OS parameters, which allows the installation of
798 instances to pass parameter to OS scripts in order to customise the
800 - Added a hypervisor parameter controlling the migration type (live or
801 non-live), since hypervisors have various levels of reliability; this
802 has renamed the 'live' parameter to 'mode'
803 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
804 logical volumes, meaning that cluster verify will ignore them and not
805 flag their presence as errors
806 - The watcher will now reset the error count for failed instances after
807 8 hours, thus allowing self-healing if the problem that caused the
808 instances to be down/fail to start has cleared in the meantime
809 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
810 check for, and warn, if the drbd module parameter ``usermode_helper``
811 is not consistent with the cluster-wide setting; this is needed to
812 make diagnose easier of failed drbd creations
813 - Started adding base IPv6 support, but this is not yet
814 enabled/available for use
815 - Rename operations (cluster, instance) will now return the new name,
816 which is especially useful if a short name was passed in
817 - Added support for instance migration in RAPI
818 - Added a tool to pre-configure nodes for the SSH setup, before joining
819 them to the cluster; this will allow in the future a simplified model
820 for node joining (but not yet fully enabled in 2.2); this needs the
821 paramiko python library
822 - Fixed handling of name-resolving errors
823 - Fixed consistency of job results on the error path
824 - Fixed master-failover race condition when executed multiple times in
826 - Fixed many bugs related to the job queue (mostly introduced during the
827 2.2 development cycle, so not all are impacting 2.1)
828 - Fixed instance migration with missing disk symlinks
829 - Fixed handling of unknown jobs in ``gnt-job archive``
830 - And many other small fixes/improvements
834 - Enhanced both the unittest and the QA coverage
835 - Switched the opcode validation to a generic model, and extended the
836 validation to all opcode parameters
837 - Changed more parts of the code that write shell scripts to use the
839 - Switched the master daemon to use the asyncore library for the Luxi
846 *(Released Thu, 17 Jun 2010)*
848 - Added tool (``move-instance``) and infrastructure to move instances
849 between separate clusters (see :doc:`separate documentation
850 <move-instance>` and :doc:`design document <design-2.2>`)
851 - Added per-request RPC timeout
852 - RAPI now requires a Content-Type header for requests with a body (e.g.
853 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
854 :rfc:`2616` (HTTP/1.1), section 7.2.1)
855 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
857 - Implemented initial support for running Ganeti daemons as separate
858 users, see configure-time flags ``--with-user-prefix`` and
859 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
861 - Instances can be removed after export (``gnt-backup export
863 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
864 key (instead of 1024 bit)
865 - Added new cluster configuration file for cluster domain secret
866 - Import/export now use SSL instead of SSH
867 - Added support for showing estimated time when exporting an instance,
868 see the ``ganeti-os-interface(7)`` manpage and look for
875 *(Released Tue, 16 Nov 2010)*
877 Some more bugfixes. Unless critical bugs occur, this will be the last
880 - Fix case of MAC special-values
881 - Fix mac checker regex
882 - backend: Fix typo causing "out of range" error
883 - Add missing --units in gnt-instance list man page
889 *(Released Tue, 24 Aug 2010)*
892 - Don't ignore secondary node silently on non-mirrored disk templates
894 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
895 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
896 - Properly document the usb_mouse parameter
897 - Fix path in ganeti-rapi(8) (issue 116)
898 - Adjust error message when the ganeti user's .ssh directory is
900 - Add same-node-check when changing the disk template to drbd
906 *(Released Fri, 16 Jul 2010)*
909 - Add an option to only select some reboot types during qa/burnin.
910 (on some hypervisors consequent reboots are not supported)
911 - Fix infrequent race condition in master failover. Sometimes the old
912 master ip address would be still detected as up for a short time
913 after it was removed, causing failover to fail.
914 - Decrease mlockall warnings when the ctypes module is missing. On
915 Python 2.4 we support running even if no ctypes module is installed,
916 but we were too verbose about this issue.
917 - Fix building on old distributions, on which man doesn't have a
919 - Fix RAPI not to ignore the MAC address on instance creation
920 - Implement the old instance creation format in the RAPI client.
926 *(Released Thu, 01 Jul 2010)*
928 A small bugfix release:
929 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
930 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
931 a non-existing option
932 - Fix instance create: the --force-variant option was ignored
933 - Improve pylint 0.21 compatibility and warnings with Python 2.6
934 - Fix modify node storage with non-FQDN arguments
935 - Fix RAPI client to authenticate under Python 2.6 when used
936 for more than 5 requests needing authentication
937 - Fix gnt-instance modify -t (storage) giving a wrong error message
938 when converting a non-shutdown drbd instance to plain
944 *(Released Fri, 18 Jun 2010)*
946 A small bugfix release:
948 - Fix live migration of KVM instances started with older Ganeti
949 versions which had fewer hypervisor parameters
950 - Fix gnt-instance grow-disk on down instances
951 - Fix an error-reporting bug during instance migration
952 - Better checking of the ``--net`` and ``--disk`` values, to avoid
953 silently ignoring broken ones
954 - Fix an RPC error reporting bug affecting, for example, RAPI client
956 - Fix bug triggered by different API version os-es on different nodes
957 - Fix a bug in instance startup with custom hvparams: OS level
958 parameters would fail to be applied.
959 - Fix the RAPI client under Python 2.6 (but more work is needed to
960 make it work completely well with OpenSSL)
961 - Fix handling of errors when resolving names from DNS
967 *(Released Thu, 3 Jun 2010)*
969 A medium sized development cycle. Some new features, and some
970 fixes/small improvements/cleanups.
975 The node deamon now tries to mlock itself into memory, unless the
976 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
977 its logs, and falls back to console logging. This allows emergency
978 features such as ``gnt-node powercycle`` to work even in the event of a
979 broken node disk (tested offlining the disk hosting the node's
980 filesystem and dropping its memory caches; don't try this at home)
982 KVM: add vhost-net acceleration support. It can be tested with a new
983 enough version of the kernel and of qemu-kvm.
985 KVM: Add instance chrooting feature. If you use privilege dropping for
986 your VMs you can also now force them to chroot to an empty directory,
987 before starting the emulated guest.
989 KVM: Add maximum migration bandwith and maximum downtime tweaking
990 support (requires a new-enough version of qemu-kvm).
992 Cluster verify will now warn if the master node doesn't have the master
995 Add a new (incompatible) instance creation request format to RAPI which
996 supports all parameters (previously only a subset was supported, and it
997 wasn't possible to extend the old format to accomodate all the new
998 features. The old format is still supported, and a client can check for
999 this feature, before using it, by checking for its presence in the
1000 ``features`` RAPI resource.
1002 Now with ancient latin support. Try it passing the ``--roman`` option to
1003 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1004 (requires the python-roman module to be installed, in order to work).
1009 As usual many internal code refactorings, documentation updates, and
1012 - Lots of improvements and cleanups to the experimental Remote API
1013 (RAPI) client library.
1014 - A new unit test suite for the core daemon libraries.
1015 - A fix to creating missing directories makes sure the umask is not
1016 applied anymore. This enforces the same directory permissions
1018 - Better handling terminating daemons with ctrl+c (used when running
1019 them in debugging mode).
1020 - Fix a race condition in live migrating a KVM instance, when stat()
1021 on the old proc status file returned EINVAL, which is an unexpected
1023 - Fixed manpage checking with newer man and utf-8 charachters. But now
1024 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1030 *(Released Fri, 7 May 2010)*
1032 Fix a bug which prevented untagged KVM instances from starting.
1038 *(Released Fri, 7 May 2010)*
1040 Another release with a long development cycle, during which many
1041 different features were added.
1043 Significant features
1044 ~~~~~~~~~~~~~~~~~~~~
1046 The KVM hypervisor now can run the individual instances as non-root, to
1047 reduce the impact of a VM being hijacked due to bugs in the
1048 hypervisor. It is possible to run all instances as a single (non-root)
1049 user, to manually specify a user for each instance, or to dynamically
1050 allocate a user out of a cluster-wide pool to each instance, with the
1051 guarantee that no two instances will run under the same user ID on any
1054 An experimental RAPI client library, that can be used standalone
1055 (without the other Ganeti libraries), is provided in the source tree as
1056 ``lib/rapi/client.py``. Note this client might change its interface in
1057 the future, as we iterate on its capabilities.
1059 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1060 replace the cluster's certificates and crypto keys. This might help in
1061 case they have been compromised, or have simply expired.
1063 A new disk option for instance creation has been added that allows one
1064 to "adopt" currently existing logical volumes, with data
1065 preservation. This should allow easier migration to Ganeti from
1066 unmanaged (or managed via other software) instances.
1068 Another disk improvement is the possibility to convert between redundant
1069 (DRBD) and plain (LVM) disk configuration for an instance. This should
1070 allow better scalability (starting with one node and growing the
1071 cluster, or shrinking a two-node cluster to one node).
1073 A new feature that could help with automated node failovers has been
1074 implemented: if a node sees itself as offline (by querying the master
1075 candidates), it will try to shutdown (hard) all instances and any active
1076 DRBD devices. This reduces the risk of duplicate instances if an
1077 external script automatically failovers the instances on such nodes. To
1078 enable this, the cluster parameter ``maintain_node_health`` should be
1079 enabled; in the future this option (per the name) will enable other
1080 automatic maintenance features.
1082 Instance export/import now will reuse the original instance
1083 specifications for all parameters; that means exporting an instance,
1084 deleting it and the importing it back should give an almost identical
1085 instance. Note that the default import behaviour has changed from
1086 before, where it created only one NIC; now it recreates the original
1089 Cluster verify has added a few new checks: SSL certificates validity,
1090 /etc/hosts consistency across the cluster, etc.
1095 As usual, many internal changes were done, documentation fixes,
1098 - Fixed cluster initialization with disabled cluster storage (regression
1099 introduced in 2.1.1)
1100 - File-based storage supports growing the disks
1101 - Fixed behaviour of node role changes
1102 - Fixed cluster verify for some corner cases, plus a general rewrite of
1103 cluster verify to allow future extension with more checks
1104 - Fixed log spamming by watcher and node daemon (regression introduced
1106 - Fixed possible validation issues when changing the list of enabled
1108 - Fixed cleanup of /etc/hosts during node removal
1109 - Fixed RAPI response for invalid methods
1110 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1111 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1112 from ide disks, etc.)
1113 - Allow OS changes without re-installation (to record a changed OS
1114 outside of Ganeti, or to allow OS renames)
1115 - Allow instance creation without OS installation (useful for example if
1116 the OS will be installed manually, or restored from a backup not in
1118 - Implemented option to make cluster ``copyfile`` use the replication
1120 - Added list of enabled hypervisors to ssconf (possibly useful for
1122 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1124 - A partial form of node re-IP is possible via node readd, which now
1125 allows changed node primary IP
1126 - Command line utilities now show an informational message if the job is
1128 - The logs of the master daemon now show the PID/UID/GID of the
1135 *(Released Fri, 12 Mar 2010)*
1137 During the 2.1.0 long release candidate cycle, a lot of improvements and
1138 changes have accumulated with were released later as 2.1.1.
1143 The node evacuate command (``gnt-node evacuate``) was significantly
1144 rewritten, and as such the IAllocator protocol was changed - a new
1145 request type has been added. This unfortunate change during a stable
1146 series is designed to improve performance of node evacuations; on
1147 clusters with more than about five nodes and which are well-balanced,
1148 evacuation should proceed in parallel for all instances of the node
1149 being evacuated. As such, any existing IAllocator scripts need to be
1150 updated, otherwise the above command will fail due to the unknown
1151 request. The provided "dumb" allocator has not been updated; but the
1152 ganeti-htools package supports the new protocol since version 0.2.4.
1154 Another important change is increased validation of node and instance
1155 names. This might create problems in special cases, if invalid host
1156 names are being used.
1158 Also, a new layer of hypervisor parameters has been added, that sits at
1159 OS level between the cluster defaults and the instance ones. This allows
1160 customisation of virtualization parameters depending on the installed
1161 OS. For example instances with OS 'X' may have a different KVM kernel
1162 (or any other parameter) than the cluster defaults. This is intended to
1163 help managing a multiple OSes on the same cluster, without manual
1164 modification of each instance's parameters.
1166 A tool for merging clusters, ``cluster-merge``, has been added in the
1167 tools sub-directory.
1172 - Improved the int/float conversions that should make the code more
1173 robust in face of errors from the node daemons
1174 - Fixed the remove node code in case of internal configuration errors
1175 - Fixed the node daemon behaviour in face of inconsistent queue
1176 directory (e.g. read-only file-system where we can't open the files
1178 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1179 now it either aborts cleanly or, if given the new "auto_promote"
1180 parameter, will automatically promote other nodes as needed
1181 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1182 completely prevent Ganeti from working
1183 - Fixed bug for instance export when not all disks were successfully
1185 - Fixed behaviour of node add when the new node is slow in starting up
1187 - Fixed handling of signals in the LUXI client, which should improve
1188 behaviour of command-line scripts
1189 - Added checks for invalid node/instance names in the configuration (now
1190 flagged during cluster verify)
1191 - Fixed watcher behaviour for disk activation errors
1192 - Fixed two potentially endless loops in http library, which led to the
1193 RAPI daemon hanging and consuming 100% CPU in some cases
1194 - Fixed bug in RAPI daemon related to hashed passwords
1195 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1197 - Enhanced compatibility with non-Debian OSes, but not using absolute
1198 path in some commands and allowing customisation of the ssh
1199 configuration directory
1200 - Fixed possible future issue with new Python versions by abiding to the
1201 proper use of ``__slots__`` attribute on classes
1202 - Added checks that should prevent directory traversal attacks
1203 - Many documentation fixes based on feedback from users
1208 - Added an "early_release" more for instance replace disks and node
1209 evacuate, where we release locks earlier and thus allow higher
1210 parallelism within the cluster
1211 - Added watcher hooks, intended to allow the watcher to restart other
1212 daemons (e.g. from the ganeti-nbma project), but they can be used of
1213 course for any other purpose
1214 - Added a compile-time disable for DRBD barriers, to increase
1215 performance if the administrator trusts the power supply or the
1216 storage system to not lose writes
1217 - Added the option of using syslog for logging instead of, or in
1218 addition to, Ganeti's own log files
1219 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1220 can indeed boot from a paravirtual NIC
1221 - Added a generic debug level for many operations; while this is not
1222 used widely yet, it allows one to pass the debug value all the way to
1224 - Enhanced the hooks environment for instance moves (failovers,
1225 migrations) where the primary/secondary nodes changed during the
1226 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1227 - Enhanced data validations for many user-supplied values; one important
1228 item is the restrictions imposed on instance and node names, which
1229 might reject some (invalid) host names
1230 - Add a configure-time option to disable file-based storage, if it's not
1231 needed; this allows greater security separation between the master
1232 node and the other nodes from the point of view of the inter-node RPC
1234 - Added user notification in interactive tools if job is waiting in the
1235 job queue or trying to acquire locks
1236 - Added log messages when a job is waiting for locks
1237 - Added filtering by node tags in instance operations which admit
1238 multiple instances (start, stop, reboot, reinstall)
1239 - Added a new tool for cluster mergers, ``cluster-merge``
1240 - Parameters from command line which are of the form ``a=b,c=d`` can now
1241 use backslash escapes to pass in values which contain commas,
1242 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1244 - For KVM, the instance name is the first parameter passed to KVM, so
1245 that it's more visible in the process list
1251 *(Released Tue, 2 Mar 2010)*
1253 Ganeti 2.1 brings many improvements with it. Major changes:
1255 - Added infrastructure to ease automated disk repairs
1256 - Added new daemon to export configuration data in a cheaper way than
1257 using the remote API
1258 - Instance NICs can now be routed instead of being associated with a
1260 - Improved job locking logic to reduce impact of jobs acquiring multiple
1261 locks waiting for other long-running jobs
1263 In-depth implementation details can be found in the Ganeti 2.1 design
1269 - Added chroot hypervisor
1270 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1272 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1273 ``bootloader_path`` and ``bootloader_args``)
1274 - Added the ``use_localtime`` option for the xen-hvm and kvm
1275 hypervisors, and the default value for this has changed to false (in
1276 2.0 xen-hvm always enabled it)
1277 - Added luxi call to submit multiple jobs in one go
1278 - Added cluster initialization option to not modify ``/etc/hosts``
1280 - Added network interface parameters
1281 - Added dry run mode to some LUs
1282 - Added RAPI resources:
1284 - ``/2/instances/[instance_name]/info``
1285 - ``/2/instances/[instance_name]/replace-disks``
1286 - ``/2/nodes/[node_name]/evacuate``
1287 - ``/2/nodes/[node_name]/migrate``
1288 - ``/2/nodes/[node_name]/role``
1289 - ``/2/nodes/[node_name]/storage``
1290 - ``/2/nodes/[node_name]/storage/modify``
1291 - ``/2/nodes/[node_name]/storage/repair``
1293 - Added OpCodes to evacuate or migrate all instances on a node
1294 - Added new command to list storage elements on nodes (``gnt-node
1295 list-storage``) and modify them (``gnt-node modify-storage``)
1296 - Added new ssconf files with master candidate IP address
1297 (``ssconf_master_candidates_ips``), node primary IP address
1298 (``ssconf_node_primary_ips``) and node secondary IP address
1299 (``ssconf_node_secondary_ips``)
1300 - Added ``ganeti-confd`` and a client library to query the Ganeti
1301 configuration via UDP
1302 - Added ability to run hooks after cluster initialization and before
1304 - Added automatic mode for disk replace (``gnt-instance replace-disks
1306 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1307 after catastrophic data-loss
1308 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1310 - Added ``gnt-instance move`` command to move instances
1311 - Added ``gnt-cluster watcher`` command to control watcher
1312 - Added ``gnt-node powercycle`` command to powercycle nodes
1313 - Added new job status field ``lock_status``
1314 - Added parseable error codes to cluster verification (``gnt-cluster
1315 verify --error-codes``) and made output less verbose (use
1316 ``--verbose`` to restore previous behaviour)
1317 - Added UUIDs to the main config entities (cluster, nodes, instances)
1318 - Added support for OS variants
1319 - Added support for hashed passwords in the Ganeti remote API users file
1321 - Added option to specify maximum timeout on instance shutdown
1322 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1323 - Added new helper script to start and stop Ganeti daemons
1324 (``daemon-util``), with the intent to reduce the work necessary to
1325 adjust Ganeti for non-Debian distributions and to start/stop daemons
1327 - Added more unittests
1328 - Fixed critical bug in ganeti-masterd startup
1329 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1330 now customisable at the cluster level for both the KVM and Xen
1331 hypervisors using the new ``migration_port`` parameter
1332 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1333 reinstalling an instance
1334 - Allowed ``@`` in tag names
1335 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1336 - Many documentation updates
1337 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1338 - ``gnt-instance reinstall`` can now reinstall multiple instances
1339 - Updated many command line parameters
1340 - Introduced new OS API version 15
1341 - No longer support a default hypervisor
1342 - Treat virtual LVs as inexistent
1343 - Improved job locking logic to reduce lock contention
1344 - Match instance and node names case insensitively
1345 - Reimplemented bash completion script to be more complete
1352 *(Released Thu, 4 Feb 2010)*
1354 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1355 - Fix a string formatting bug
1356 - Improve safety of the code in some error paths
1357 - Improve data validation in the master of values returned from nodes
1363 *(Released Thu, 17 Dec 2009)*
1365 - Fix security issue due to missing validation of iallocator names; this
1366 allows local and remote execution of arbitrary executables
1367 - Fix failure of gnt-node list during instance removal
1368 - Ship the RAPI documentation in the archive
1374 *(Released Wed, 30 Sep 2009)*
1376 - Fixed many wrong messages
1377 - Fixed a few bugs related to the locking library
1378 - Fixed MAC checking at instance creation time
1379 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1380 - Fixed a few issues related to signal handling in both daemons and
1382 - Fixed the example startup script provided
1383 - Fixed insserv dependencies in the example startup script (patch from
1385 - Fixed handling of drained nodes in the iallocator framework
1386 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1388 - Fixed error related to invalid job IDs in job polling
1389 - Fixed job/opcode persistence on unclean master shutdown
1390 - Fixed handling of partial job processing after unclean master
1392 - Fixed error reporting from LUs, previously all errors were converted
1393 into execution errors
1394 - Fixed error reporting from burnin
1395 - Decreased significantly the memory usage of the job queue
1396 - Optimised slightly multi-job submission
1397 - Optimised slightly opcode loading
1398 - Backported the multi-job submit framework from the development
1399 branch; multi-instance start and stop should be faster
1400 - Added script to clean archived jobs after 21 days; this will reduce
1401 the size of the queue directory
1402 - Added some extra checks in disk size tracking
1403 - Added an example ethers hook script
1404 - Added a cluster parameter that prevents Ganeti from modifying of
1406 - Added more node information to RAPI responses
1407 - Added a ``gnt-job watch`` command that allows following the ouput of a
1409 - Added a bind-address option to ganeti-rapi
1410 - Added more checks to the configuration verify
1411 - Enhanced the burnin script such that some operations can be retried
1413 - Converted instance reinstall to multi-instance model
1419 *(Released Fri, 7 Aug 2009)*
1421 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1422 to allow using the pre-2.0.2 behaviour in activation, if any existing
1423 instances have mismatched disk sizes in the configuration
1424 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
1425 any configuration mismatches for disk sizes
1426 - Added ``gnt-master cluste-failover --no-voting`` to allow master
1427 failover to work on two-node clusters
1428 - Fixed the ``--net`` option of ``gnt-backup import``, which was
1430 - Fixed detection of OS script errors in ``gnt-backup export``
1431 - Fixed exit code of ``gnt-backup export``
1437 *(Released Fri, 17 Jul 2009)*
1439 - Added experimental support for stripped logical volumes; this should
1440 enhance performance but comes with a higher complexity in the block
1441 device handling; stripping is only enabled when passing
1442 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
1443 affected even in the non-stripped mode
1444 - Improved resiliency against transient failures at the end of DRBD
1445 resyncs, and in general of DRBD resync checks
1446 - Fixed a couple of issues with exports and snapshot errors
1447 - Fixed a couple of issues in instance listing
1448 - Added display of the disk size in ``gnt-instance info``
1449 - Fixed checking for valid OSes in instance creation
1450 - Fixed handling of the "vcpus" parameter in instance listing and in
1451 general of invalid parameters
1452 - Fixed http server library, and thus RAPI, to handle invalid
1453 username/password combinations correctly; this means that now they
1454 report unauthorized for queries too, not only for modifications,
1455 allowing earlier detect of configuration problems
1456 - Added a new "role" node list field, equivalent to the master/master
1457 candidate/drained/offline flags combinations
1458 - Fixed cluster modify and changes of candidate pool size
1459 - Fixed cluster verify error messages for wrong files on regular nodes
1460 - Fixed a couple of issues with node demotion from master candidate role
1461 - Fixed node readd issues
1462 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
1463 - Added a new ``--no-voting`` option for masterfailover to fix failover
1464 on two-nodes clusters when the former master node is unreachable
1465 - Added instance reinstall over RAPI
1471 *(Released Tue, 16 Jun 2009)*
1473 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
1474 allow re-adding the start in single-user option (regression from 1.2)
1475 - the watcher writes the instance status to a file, to allow monitoring
1476 to report the instance status (from the master) based on cached
1477 results of the watcher's queries; while this can get stale if the
1478 watcher is being locked due to other work on the cluster, this is
1479 still an improvement
1480 - the watcher now also restarts the node daemon and the rapi daemon if
1482 - fixed the watcher to handle full and drained queue cases
1483 - hooks export more instance data in the environment, which helps if
1484 hook scripts need to take action based on the instance's properties
1485 (no longer need to query back into ganeti)
1486 - instance failovers when the instance is stopped do not check for free
1487 RAM, so that failing over a stopped instance is possible in low memory
1489 - rapi uses queries for tags instead of jobs (for less job traffic), and
1490 for cluster tags it won't talk to masterd at all but read them from
1492 - a couple of error handling fixes in RAPI
1493 - drbd handling: improved the error handling of inconsistent disks after
1494 resync to reduce the frequency of "there are some degraded disks for
1495 this instance" messages
1496 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
1497 error handling path called a wrong function name)
1503 *(Released Wed, 27 May 2009)*
1505 - no changes from rc5
1508 Version 2.0 release candidate 5
1509 -------------------------------
1511 *(Released Wed, 20 May 2009)*
1513 - fix a couple of bugs (validation, argument checks)
1514 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
1515 - some small improvements to RAPI and IAllocator
1516 - make watcher automatically start the master daemon if down
1519 Version 2.0 release candidate 4
1520 -------------------------------
1522 *(Released Mon, 27 Apr 2009)*
1524 - change the OS list to not require locks; this helps with big clusters
1525 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
1526 volume group is broken
1527 - ``gnt-instance info``, without any arguments, doesn't run for all
1528 instances anymore; either pass ``--all`` or pass the desired
1529 instances; this helps against mistakes on big clusters where listing
1530 the information for all instances takes a long time
1531 - miscellaneous doc and man pages fixes
1534 Version 2.0 release candidate 3
1535 -------------------------------
1537 *(Released Wed, 8 Apr 2009)*
1539 - Change the internal locking model of some ``gnt-node`` commands, in
1540 order to reduce contention (and blocking of master daemon) when
1541 batching many creation/reinstall jobs
1542 - Fixes to Xen soft reboot
1543 - No longer build documentation at build time, instead distribute it in
1544 the archive, in order to reduce the need for the whole docbook/rst
1548 Version 2.0 release candidate 2
1549 -------------------------------
1551 *(Released Fri, 27 Mar 2009)*
1553 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
1554 - Fix watcher startup sequence, improves the behaviour of busy clusters
1555 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
1556 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
1557 bind address and other places
1558 - Some documentation fixes and updates
1561 Version 2.0 release candidate 1
1562 -------------------------------
1564 *(Released Mon, 2 Mar 2009)*
1566 - More documentation updates, now all docs should be more-or-less
1568 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
1570 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
1572 - Fix an issue related to $libdir/run/ganeti and cluster creation
1578 *(Released Thu, 19 Feb 2009)*
1580 - Xen PVM and KVM have switched the default value for the instance root
1581 disk to the first partition on the first drive, instead of the whole
1582 drive; this means that the OS installation scripts must be changed
1584 - Man pages have been updated
1585 - RAPI has been switched by default to HTTPS, and the exported functions
1586 should all work correctly
1587 - RAPI v1 has been removed
1588 - Many improvements to the KVM hypervisor
1589 - Block device errors are now better reported
1590 - Many other bugfixes and small improvements
1596 *(Released Mon, 26 Jan 2009)*
1598 - Version 2 is a general rewrite of the code and therefore the
1599 differences are too many to list, see the design document for 2.0 in
1600 the ``doc/`` subdirectory for more details
1601 - In this beta version there is not yet a migration path from 1.2 (there
1602 will be one in the final 2.0 release)
1603 - A few significant changes are:
1605 - all commands are executed by a daemon (``ganeti-masterd``) and the
1606 various ``gnt-*`` commands are just front-ends to it
1607 - all the commands are entered into, and executed from a job queue,
1608 see the ``gnt-job(8)`` manpage
1609 - the RAPI daemon supports read-write operations, secured by basic
1610 HTTP authentication on top of HTTPS
1611 - DRBD version 0.7 support has been removed, DRBD 8 is the only
1612 supported version (when migrating from Ganeti 1.2 to 2.0, you need
1613 to migrate to DRBD 8 first while still running Ganeti 1.2)
1614 - DRBD devices are using statically allocated minor numbers, which
1615 will be assigned to existing instances during the migration process
1616 - there is support for both Xen PVM and Xen HVM instances running on
1618 - KVM virtualization is supported too
1619 - file-based storage has been implemented, which means that it is
1620 possible to run the cluster without LVM and DRBD storage, for
1621 example using a shared filesystem exported from shared storage (and
1622 still have live migration)
1628 *(Released Tue, 13 Jan 2009)*
1630 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
1631 - Reuse the old instance mac address by default on instance import, if
1632 the instance name is the same.
1633 - Handle situations in which the node info rpc returns incomplete
1635 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
1636 - Improved version of batcher:
1638 - state file support
1639 - instance mac address support
1640 - support for HVM clusters/instances
1642 - Add an option to show the number of cpu sockets and nodes in
1644 - Support OSes that handle more than one version of the OS api (but do
1645 not change the current API in any other way)
1646 - Fix ``gnt-node migrate``
1647 - ``gnt-debug`` man page
1648 - Fixes various more typos and small issues
1649 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
1655 *(Released Wed, 24 Sep 2008)*
1657 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
1658 type of disk exported to fully virtualized instances.
1659 - provide access to the serial console of HVM instances
1660 - instance auto_balance flag, set by default. If turned off it will
1661 avoid warnings on cluster verify if there is not enough memory to fail
1662 over an instance. in the future it will prevent automatically failing
1663 it over when we will support that.
1664 - batcher tool for instance creation, see ``tools/README.batcher``
1665 - ``gnt-instance reinstall --select-os`` to interactively select a new
1666 operating system when reinstalling an instance.
1667 - when changing the memory amount on instance modify a check has been
1668 added that the instance will be able to start. also warnings are
1669 emitted if the instance will not be able to fail over, if auto_balance
1671 - documentation fixes
1672 - sync fields between ``gnt-instance list/modify/add/import``
1673 - fix a race condition in drbd when the sync speed was set after giving
1674 the device a remote peer.
1680 *(Released Tue, 22 Jul 2008)*
1682 - note: the allowed size and number of tags per object were reduced
1683 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1684 - fixed twisted 8.x compatibility
1685 - fixed ``gnt-instance replace-disks`` with iallocator
1686 - add TCP keepalives on twisted connections to detect restarted nodes
1687 - disk increase support, see ``gnt-instance grow-disk``
1688 - implement bulk node/instance query for RAPI
1689 - add tags in node/instance listing (optional)
1690 - experimental migration (and live migration) support, read the man page
1691 for ``gnt-instance migrate``
1692 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1693 has some small improvements in handling its state file
1699 *(Released Fri, 13 Jun 2008)*
1701 - Experimental readonly, REST-based remote API implementation;
1702 automatically started on master node, TCP port 5080, if enabled by
1703 ``--enable-rapi`` parameter to configure script.
1704 - Instance allocator support. Add and import instance accept a
1705 ``--iallocator`` parameter, and call that instance allocator to decide
1706 which node to use for the instance. The iallocator document describes
1707 what's expected from an allocator script.
1708 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1709 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1710 node is lost there is still enough memory to fail over the instances
1712 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1713 ``gnt-cluster verify``, to check for site-specific compliance. All the
1714 hooks will run, and their output, if any, will be displayed. Any
1715 failing hook will make the verification return an error value.
1716 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1717 primary and secondary interfaces
1718 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1719 that is still declared as part of the cluster and has failed.
1720 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1721 specifying output fields, that just adds the chosen fields to the
1723 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1724 export from the filesystem.
1725 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1726 have been added. Using them you can enable/disable acpi and pae
1727 support, and specify a path for a cd image to be exported to the
1728 instance. These parameters as the name suggest only work on HVM
1730 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1731 PAE support will be set to the previously hardcoded values, but the
1732 (previously hardcoded) path to the CDROM ISO image will be unset and
1733 if required, needs to be set manually with ``gnt-instance modify``
1735 - The address to which an instance's VNC console is bound is now
1736 selectable per-instance, rather than being cluster wide. Of course
1737 this only applies to instances controlled via VNC, so currently just
1738 applies to HVM clusters.
1744 *(Released Mon, 18 Feb 2008)*
1746 - more tweaks to the disk activation code (especially helpful for DRBD)
1747 - change the default ``gnt-instance list`` output format, now there is
1748 one combined status field (see the manpage for the exact values this
1750 - some more fixes for the mac export to hooks change
1751 - make Ganeti not break with DRBD 8.2.x (which changed the version
1752 format in ``/proc/drbd``) (issue 24)
1753 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1754 template, allowing migration from DRBD0.7+MD to DRBD8
1760 *(Released Wed, 30 Jan 2008)*
1762 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1764 - add command aliases infrastructure and a few aliases
1765 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1766 man pages and the ``--help`` option of ``gnt-node
1767 list``/``gnt-instance list``
1768 - fix ``gnt-backup list`` with down nodes (issue 21)
1769 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1770 - fix the dist archive and add a check for including svn/git files in
1772 - some developer-related changes: improve the burnin and the QA suite,
1773 add an upload script for testing during development
1779 *(Released Wed, 16 Jan 2008)*
1781 - experimental HVM support, read the install document, section
1782 "Initializing the cluster"
1783 - allow for the PVM hypervisor per-instance kernel and initrd paths
1784 - add a new command ``gnt-cluster verify-disks`` which uses a new
1785 algorithm to improve the reconnection of the DRBD pairs if the device
1786 on the secondary node has gone away
1787 - make logical volume code auto-activate LVs at disk activation time
1788 - slightly improve the speed of activating disks
1789 - allow specification of the MAC address at instance creation time, and
1790 changing it later via ``gnt-instance modify``
1791 - fix handling of external commands that generate lots of output on
1793 - update documentation with regard to minimum version of DRBD8 supported
1799 *(Released Tue, 4 Dec 2007)*
1801 - Log the ``xm create`` output to the node daemon log on failure (to
1802 help diagnosing the error)
1803 - In debug mode, log all external commands output if failed to the logs
1804 - Change parsing of lvm commands to ignore stderr
1810 *(Released Wed, 28 Nov 2007)*
1812 - Another round of updates to the DRBD 8 code to deal with more failures
1813 in the replace secondary node operation
1814 - Some more logging of failures in disk operations (lvm, drbd)
1815 - A few documentation updates
1822 *(Released Tue, 13 Nov 2007)*
1824 - Change configuration file format from Python's Pickle to JSON.
1825 Upgrading is possible using the cfgupgrade utility.
1826 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1827 faster replace disks and is more stable (DRBD 8 has many improvements
1828 compared to DRBD 0.7)
1829 - Added command line tags support (see man pages for ``gnt-instance``,
1830 ``gnt-node``, ``gnt-cluster``)
1831 - Added instance rename support
1832 - Added multi-instance startup/shutdown
1833 - Added cluster rename support
1834 - Added ``gnt-node evacuate`` to simplify some node operations
1835 - Added instance reboot operation that can speedup reboot as compared to
1837 - Soften the requirement that hostnames are in FQDN format
1838 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1840 - Removed dependency on debian's patched fping that uses the
1841 non-standard ``-S`` option
1842 - Now the OS definitions are searched for in multiple, configurable
1843 paths (easier for distros to package)
1844 - Some changes to the hooks infrastructure (especially the new
1845 post-configuration update hook)
1846 - Other small bugfixes
1848 .. vim: set textwidth=72 syntax=rst :