7 *(Released Tue, 18 Oct 2011)*
9 Incompatible/important changes and bugfixes
10 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
12 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
13 resource's ``ip_check`` parameter changed from ``True`` to ``False``
14 to match the underlying LUXI interface.
15 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
16 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
17 not maintain backwards-compatibility as the underlying operation
18 changed in an incompatible way. The RAPI client can talk to old
19 servers, but it needs to be told so as the return value changed.
20 - When creating file-based instances via RAPI, the ``file_driver``
21 parameter no longer defaults to ``loop`` and must be specified.
22 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
24 - Support for the undocumented and deprecated RAPI instance creation
25 request format version 0 has been dropped. Use version 1, supported
26 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
27 - Pyparsing 1.4.6 or above is required, see :doc:`installation
28 documentation <install>`.
29 - The "cluster-verify" hooks are now executed per group by the
30 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
31 if you just run ``gnt-cluster verify``, which generates one opcode per
33 - The environment as passed to the OS scripts is cleared, and thus no
34 environment variables defined in the node daemon's environment will be
35 inherited by the scripts.
36 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
38 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
39 support operations involving multiple node groups.
40 - Offline nodes are ignored when failing over an instance.
45 - The ganeti-htools project has been merged into the ganeti-core source
46 tree and will be built as part of Ganeti (see :doc:`install-quick`).
47 - Implemented support for :doc:`shared storage <design-shared-storage>`.
48 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
49 GPT-style partition tables (requires `parted
50 <http://www.gnu.org/s/parted/>`_).
51 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
52 - Allowed adding tags on instance creation.
53 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
54 - Allow instances to be started in a paused state, enabling the user to
55 see the complete console output on boot using the console.
56 - Added new hypervisor flag to control default reboot behaviour
57 (``reboot_behavior``).
58 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
59 - Improved out-of-band management support:
61 - Added ``gnt-node health`` command reporting the health status of
63 - Added ``gnt-node power`` command to manage power status of nodes.
64 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
66 - Instance migration can fall back to failover if instance is not
68 - Filters can be used when listing nodes, instances, groups and locks;
69 see *ganeti(7)* manpage.
70 - Added post-execution status as variables to :doc:`hooks <hooks>`
72 - Instance tags are exported/imported together with the instance.
73 - When given an explicit job ID, ``gnt-job info`` will work for archived
75 - Jobs can define dependencies on other jobs (not yet supported via
76 RAPI or command line, but used by internal commands and usable via
79 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
82 - Instance failover is now available as a RAPI resource
83 (``/2/instances/[instance_name]/failover``).
84 - ``gnt-instance info`` defaults to static information if primary node
86 - Opcodes have a new ``comment`` attribute.
87 - Added basic SPICE support to KVM hypervisor.
88 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
90 Node group improvements
91 ~~~~~~~~~~~~~~~~~~~~~~~
93 - ``gnt-cluster verify`` has been modified to check groups separately,
94 thereby improving performance.
95 - Node group support has been added to ``gnt-cluster verify-disks``,
96 which now operates per node group.
97 - Watcher has been changed to work better with node groups.
99 - One process and state file per node group.
100 - Slow watcher in one group doesn't block other group's watcher.
102 - Added new command, ``gnt-group evacuate``, to move all instances in a
103 node group to other groups.
104 - Added ``gnt-instance change-group`` to move an instance to another
106 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
107 per-group operations.
108 - Node groups can be tagged.
109 - Some operations switch from an exclusive to a shared lock as soon as
111 - Instance's primary and secondary nodes' groups are now available as
112 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
113 and ``snodes.group.uuid``).
118 - Numerous updates to documentation and manpages.
120 - :doc:`RAPI <rapi>` documentation now has detailed parameter
123 - A lockset's internal lock is now also visible in lock monitor.
124 - Log messages from job queue workers now contain information about the
125 opcode they're processing.
126 - ``gnt-instance console`` no longer requires the instance lock.
127 - A short delay when waiting for job changes reduces the number of LUXI
128 requests significantly.
129 - DRBD metadata volumes are overwritten with zeros during disk creation.
135 *(Released Tue, 4 Oct 2011)*
137 This was the first release candidate of the 2.5 series.
143 *(Released Wed, 31 Aug 2011)*
145 This was the third beta release of the 2.5 series.
151 *(Released Mon, 22 Aug 2011)*
153 This was the second beta release of the 2.5 series.
159 *(Released Fri, 12 Aug 2011)*
161 This was the first beta release of the 2.5 series.
167 *(Released Thu, 27 Oct 2011)*
169 - Fixed bug when parsing command line parameter values ending in
171 - Fixed assertion error after unclean master shutdown
172 - Disable HTTP client pool for RPC, significantly reducing memory usage
174 - Fixed queue archive creation with wrong permissions
180 *(Released Tue, 23 Aug 2011)*
184 - Fixed documentation for importing with ``--src-dir`` option
185 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
186 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
192 *(Released Fri, 5 Aug 2011)*
194 Many bug-fixes and a few small features:
196 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
197 issues when you tried to add an instance with two MAC addresses in one
199 - KVM: fixed per-instance stored UID value
200 - KVM: configure bridged NICs at migration start
201 - KVM: Fix a bug where instance will not start with never KVM versions
203 - Added OS search path to ``gnt-cluster info``
204 - Fixed an issue with ``file_storage_dir`` where you were forced to
205 provide an absolute path, but the documentation states it is a
206 relative path, the documentation was right
207 - Added a new parameter to instance stop/start called ``--no-remember``
208 that will make the state change to not be remembered
209 - Implemented ``no_remember`` at RAPI level
210 - Improved the documentation
211 - Node evacuation: don't call IAllocator if node is already empty
212 - Fixed bug in DRBD8 replace disks on current nodes
213 - Fixed bug in recreate-disks for DRBD instances
214 - Moved assertion checking locks in ``gnt-instance replace-disks``
215 causing it to abort with not owning the right locks for some situation
216 - Job queue: Fixed potential race condition when cancelling queued jobs
217 - Fixed off-by-one bug in job serial generation
218 - ``gnt-node volumes``: Fix instance names
219 - Fixed aliases in bash completion
220 - Fixed a bug in reopening log files after being sent a SIGHUP
221 - Added a flag to burnin to allow specifying VCPU count
222 - Bugfixes to non-root Ganeti configuration
228 *(Released Thu, 12 May 2011)*
230 Many bug-fixes and a few new small features:
232 - Fixed a bug related to log opening failures
233 - Fixed a bug in instance listing with orphan instances
234 - Fixed a bug which prevented resetting the cluster-level node parameter
235 ``oob_program`` to the default
236 - Many fixes related to the ``cluster-merge`` tool
237 - Fixed a race condition in the lock monitor, which caused failures
238 during (at least) creation of many instances in parallel
239 - Improved output for gnt-job info
240 - Removed the quiet flag on some ssh calls which prevented debugging
242 - Improved the N+1 failure messages in cluster verify by actually
243 showing the memory values (needed and available)
244 - Increased lock attempt timeouts so that when executing long operations
245 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
246 too early and thus prevent the use of the 'fair' mechanism
247 - Changed instance query data (``gnt-instance info``) to not acquire
248 locks unless needed, thus allowing its use on locked instance if only
249 static information is asked for
250 - Improved behaviour with filesystems that do not support rename on an
252 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
253 kept locks on all nodes during the wipe, which is unneeded
254 - Fixed ``gnt-watcher`` handling of errors during hooks execution
255 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
256 10GiB) which caused the wipe to fail right at the end in some cases
257 - Fixed master IP activation when doing master failover with no-voting
258 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
259 the master node itself
260 - Fixed potential data-loss in under disk full conditions, where Ganeti
261 wouldn't check correctly the return code and would consider
262 partially-written files 'correct'
263 - Fixed bug related to multiple VGs and DRBD disk replacing
264 - Added new disk parameter ``metavg`` that allows placement of the meta
265 device for DRBD in a different volume group
266 - Fixed error handling in the node daemon when the system libc doesn't
267 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
268 - Fixed lock release during replace-disks, which kept cluster-wide locks
269 when doing disk replaces with an iallocator script
270 - Added check for missing bridges in cluster verify
271 - Handle EPIPE errors while writing to the terminal better, so that
272 piping the output to e.g. ``less`` doesn't cause a backtrace
273 - Fixed rare case where a ^C during Luxi calls could have been
274 interpreted as server errors, instead of simply terminating
275 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
277 - Added a few more parameters to the KVM hypervisor, allowing a second
278 CDROM, custom disk type for CDROMs and a floppy image
279 - Removed redundant message in instance rename when the name is given
281 - Added option to ``gnt-instance recreate-disks`` to allow creating the
282 disks on new nodes, allowing recreation when the original instance
283 nodes are completely gone
284 - Added option when converting disk templates to DRBD to skip waiting
285 for the resync, in order to make the instance available sooner
286 - Added two new variables to the OS scripts environment (containing the
288 - Made the root_path and optional parameter for the xen-pvm hypervisor,
289 to allow use of ``pvgrub`` as bootloader
290 - Changed the instance memory modifications to only check out-of-memory
291 conditions on memory increases, and turned the secondary node warnings
292 into errors (they can still be overridden via ``--force``)
293 - Fixed the handling of a corner case when the Python installation gets
294 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
295 to execute a command that doesn't exist
296 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
297 node of the instance returned failures during instance shutdown; this
298 adds the option ``--ignore-consistency`` to gnt-instance move
300 And as usual, various improvements to the error messages, documentation
306 *(Released Wed, 09 Mar 2011)*
308 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
309 the RAPI users file if run twice (even with ``--dry-run``).
311 The release fixes that bug (nothing else changed).
317 *(Released Mon, 07 Mar 2011)*
319 Final 2.4.0 release. Just a few small fixes:
321 - Fixed RAPI node evacuate
322 - Fixed the kvm-ifup script
323 - Fixed internal error handling for special job cases
324 - Updated man page to specify the escaping feature for options
330 *(Released Mon, 28 Feb 2011)*
332 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
333 that this feature wiped the disks of the wrong instance, leading to loss
338 - Fixed title of query field containing instance name
339 - Expanded the glossary in the documentation
340 - Fixed one unittest (internal issue)
346 *(Released Mon, 21 Feb 2011)*
348 A number of bug fixes plus just a couple functionality changes.
350 On the user-visible side, the ``gnt-* list`` command output has changed
351 with respect to "special" field states. The current rc1 style of display
352 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
353 the default output mode special fields states are displayed as follows:
355 - Offline resource: ``*``
356 - Unavailable/not applicable: ``-``
357 - Data missing (RPC failure): ``?``
358 - Unknown field: ``??``
360 Another user-visible change is the addition of ``--force-join`` to
365 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
366 - Fixed regression in RAPI/instance reinstall where all parameters were
367 required (instead of optional)
368 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
369 - Fixed iallocator usage (offline nodes were not considered offline)
370 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
371 - Fixed hypervisor and OS parameter validation with respect to
373 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
375 - Fixed ``tools/listrunner`` with respect to agent-based usage
381 *(Released Fri, 4 Feb 2011)*
383 Many changes and fixes since the beta1 release. While there were some
384 internal changes, the code has been mostly stabilised for the RC
387 Note: the dumb allocator was removed in this release, as it was not kept
388 up-to-date with the IAllocator protocol changes. It is recommended to
389 use the ``hail`` command from the ganeti-htools package.
391 Note: the 2.4 and up versions of Ganeti are not compatible with the
392 0.2.x branch of ganeti-htools. You need to upgrade to
393 ganeti-htools-0.3.0 (or later).
395 Regressions fixed from 2.3
396 ~~~~~~~~~~~~~~~~~~~~~~~~~~
398 - Fixed the ``gnt-cluster verify-disks`` command
399 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
401 - Fixed disk adoption breakage
402 - Fixed wrong headers in instance listing for field aliases
407 - Fixed corner case in KVM handling of NICs
408 - Fixed many cases of wrong handling of non-vm_capable nodes
409 - Fixed a bug where a missing instance symlink was not possible to
410 recreate with any ``gnt-*`` command (now ``gnt-instance
411 activate-disks`` does it)
412 - Fixed the volume group name as reported by ``gnt-cluster
414 - Increased timeouts for the import-export code, hopefully leading to
415 fewer aborts due network or instance timeouts
416 - Fixed bug in ``gnt-node list-storage``
417 - Fixed bug where not all daemons were started on cluster
418 initialisation, but only at the first watcher run
419 - Fixed many bugs in the OOB implementation
420 - Fixed watcher behaviour in presence of instances with offline
422 - Fixed instance list output for instances running on the wrong node
423 - a few fixes to the cluster-merge tool, but it still cannot merge
424 multi-node groups (currently it is not recommended to use this tool)
430 - Improved network configuration for the KVM hypervisor
431 - Added e1000 as a supported NIC for Xen-HVM
432 - Improved the lvmstrap tool to also be able to use partitions, as
433 opposed to full disks
434 - Improved speed of disk wiping (the cluster parameter
435 ``prealloc_wipe_disks``, so that it has a low impact on the total time
436 of instance creations
437 - Added documentation for the OS parameters
438 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
439 hypervisor is not responding
440 - Added display of blacklisted and hidden OS information in
442 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
443 NIC and node parameters, which might create problems with currently
444 invalid (but undetected) configuration files, but prevents validation
445 failures when unrelated parameters are modified
446 - Changed cluster initialisation to wait for the master daemon to become
448 - Expanded the RAPI interface:
450 - Added config redistribution resource
451 - Added activation/deactivation of instance disks
452 - Added export of console information
454 - Implemented log file reopening on SIGHUP, which allows using
455 logrotate(8) for the Ganeti log files
456 - Added a basic OOB helper script as an example
462 *(Released Fri, 14 Jan 2011)*
467 - Fixed timezone issues when formatting timestamps
468 - Added support for node groups, available via ``gnt-group`` and other
470 - Added out-of-band framework and management, see :doc:`design
471 document <design-oob>`
472 - Removed support for roman numbers from ``gnt-node list`` and
473 ``gnt-instance list``.
474 - Allowed modification of master network interface via ``gnt-cluster
475 modify --master-netdev``
476 - Accept offline secondaries while shutting down instance disks
477 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
478 - Added support for multiple LVM volume groups
479 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
481 - Added commands to list available fields:
483 - ``gnt-node list-fields``
484 - ``gnt-group list-fields``
485 - ``gnt-instance list-fields``
487 - Updated documentation and man pages
492 - Moved ``rapi_users`` file into separate directory, now named
493 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
495 - Added new tool for running commands on many machines,
496 ``tools/ganeti-listrunner``
497 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
498 improving the ``gnt-instance console`` output
499 - Allowed customisation of disk index separator at ``configure`` time
500 - Export node group allocation policy to :doc:`iallocator <iallocator>`
501 - Added support for non-partitioned md disks in ``lvmstrap``
502 - Added script to gracefully power off KVM instances
503 - Split ``utils`` module into smaller parts
504 - Changed query operations to return more detailed information, e.g.
505 whether an information is unavailable due to an offline node. To use
506 this new functionality, the LUXI call ``Query`` must be used. Field
507 information is now stored by the master daemon and can be retrieved
508 using ``QueryFields``. Instances, nodes and groups can also be queried
509 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
510 exposed via RAPI). The following commands make use of this
511 infrastructure change:
514 - ``gnt-group list-fields``
516 - ``gnt-node list-fields``
517 - ``gnt-instance list``
518 - ``gnt-instance list-fields``
519 - ``gnt-debug locks``
524 - New RAPI resources (see :doc:`rapi`):
528 - ``/2/groups/[group_name]``
529 - ``/2/groups/[group_name]/assign-nodes``
530 - ``/2/groups/[group_name]/modify``
531 - ``/2/groups/[group_name]/rename``
532 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
536 - Implemented ``no_install`` for instance creation
537 - Implemented OS parameters for instance reinstallation, allowing
538 use of special settings on reinstallation (e.g. for preserving data)
543 - Added IPv6 support in import/export
544 - Pause DRBD synchronization while wiping disks on instance creation
545 - Updated unittests and QA scripts
546 - Improved network parameters passed to KVM
547 - Converted man pages from docbook to reStructuredText
553 *(Released Mon, 20 Dec 2010)*
555 Released version 2.3.1~rc1 without any changes.
561 *(Released Wed, 1 Dec 2010)*
563 - impexpd: Disable OpenSSL compression in socat if possible (backport
564 from master, commit e90739d625b, see :doc:`installation guide
565 <install-quick>` for details)
566 - Changed unittest coverage report to exclude test scripts
567 - Added script to check version format
573 *(Released Wed, 1 Dec 2010)*
575 Released version 2.3.0~rc1 without any changes.
581 *(Released Fri, 19 Nov 2010)*
583 A number of bugfixes and documentation updates:
585 - Update ganeti-os-interface documentation
586 - Fixed a bug related to duplicate MACs or similar items which should be
588 - Fix breakage in OS state modify
589 - Reinstall instance: disallow offline secondaries (fixes bug related to
590 OS changing but reinstall failing)
591 - plus all the other fixes between 2.2.1 and 2.2.2
597 *(Released Tue, 2 Nov 2010)*
599 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
600 - Fixed master failover race with watcher
601 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
603 - Accept previously stopped instance for export with instance removal
604 - Simplify and extend the environment variables for instance OS scripts
605 - Added new node flags, ``master_capable`` and ``vm_capable``
606 - Added optional instance disk wiping prior during allocation. This is a
607 cluster-wide option and can be set/modified using
608 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
609 - Added IPv6 support, see :doc:`design document <design-2.3>` and
611 - Added a new watcher option (``--ignore-pause``)
612 - Added option to ignore offline node on instance start/stop
613 (``--ignore-offline``)
614 - Allow overriding OS parameters with ``gnt-instance reinstall``
615 - Added ability to change node's secondary IP address using ``gnt-node
617 - Implemented privilege separation for all daemons except
618 ``ganeti-noded``, see ``configure`` options
619 - Complain if an instance's disk is marked faulty in ``gnt-cluster
621 - Implemented job priorities (see ``ganeti(7)`` manpage)
622 - Ignore failures while shutting down instances during failover from
624 - Exit daemon's bootstrap process only once daemon is ready
625 - Export more information via ``LUInstanceQuery``/remote API
626 - Improved documentation, QA and unittests
627 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
628 restart if the file was created or changed
629 - Added LUXI protocol version sent with each request and response,
630 allowing detection of server/client mismatches
631 - Moved the Python scripts among gnt-* and ganeti-* into modules
632 - Moved all code related to setting up SSH to an external script,
634 - Infrastructure changes for node group support in future versions
640 *(Released Fri, 19 Nov 2010)*
642 A few small bugs fixed, and some improvements to the build system:
644 - Fix documentation regarding conversion to drbd
645 - Fix validation of parameters in cluster modify (``gnt-cluster modify
647 - Fix error handling in node modify with multiple changes
648 - Allow remote imports without checked names
654 *(Released Tue, 19 Oct 2010)*
656 - Disable SSL session ID cache in RPC client
662 *(Released Thu, 14 Oct 2010)*
664 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
665 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
667 - Fix problems with interaction between Curl and Python's HTTP server,
668 resulting in increased speed in many RPC calls
669 - Improve our release script to prevent breakage with older aclocal and
676 *(Released Thu, 7 Oct 2010)*
678 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
679 value retrieved from master
680 - Added support for blacklisted or hidden OS definitions
681 - Added simple lock monitor (accessible via (``gnt-debug locks``)
682 - Added support for -mem-path in KVM hypervisor abstraction layer
683 - Allow overriding instance parameters in tool for inter-cluster
684 instance moves (``tools/move-instance``)
685 - Improved opcode summaries (e.g. in ``gnt-job list``)
686 - Improve consistency of OS listing by sorting it
687 - Documentation updates
693 *(Released Fri, 8 Oct 2010)*
695 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
701 *(Released Mon, 4 Oct 2010)*
703 - Fixed regression in ``gnt-instance rename``
709 *(Released Wed, 22 Sep 2010)*
711 - Fixed OS_VARIANT variable for OS scripts
712 - Fixed cluster tag operations via RAPI
713 - Made ``setup-ssh`` exit with non-zero code if an error occurred
714 - Disabled RAPI CA checks in watcher
720 *(Released Mon, 23 Aug 2010)*
722 - Support DRBD versions of the format "a.b.c.d"
724 - Re-introduce support for usage from multiple threads in RAPI client
725 - Instance renames and modify via RAPI
726 - Work around race condition between processing and archival in job
728 - Mark opcodes following failed one as failed, too
729 - Job field ``lock_status`` was removed due to difficulties making it
730 work with the changed job queue in Ganeti 2.2; a better way to monitor
731 locks is expected for a later 2.2.x release
732 - Fixed dry-run behaviour with many commands
733 - Support ``ssh-agent`` again when adding nodes
734 - Many additional bugfixes
740 *(Released Fri, 30 Jul 2010)*
742 Important change: the internal RPC mechanism between Ganeti nodes has
743 changed from using a home-grown http library (based on the Python base
744 libraries) to use the PycURL library. This requires that PycURL is
745 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
746 against GnuTLS by default. cURL's support for GnuTLS had known issues
747 before cURL 7.21.0 and we recommend using the latest cURL release or
748 linking against OpenSSL. Most other distributions already link PycURL
749 and cURL against OpenSSL. The command::
751 python -c 'import pycurl; print pycurl.version'
753 can be used to determine the libraries PycURL and cURL are linked
756 Other significant changes:
758 - Rewrote much of the internals of the job queue, in order to achieve
759 better parallelism; this decouples job query operations from the job
760 processing, and it should allow much nicer behaviour of the master
761 daemon under load, and it also has uncovered some long-standing bugs
762 related to the job serialisation (now fixed)
763 - Added a default iallocator setting to the cluster parameters,
764 eliminating the need to always pass nodes or an iallocator for
765 operations that require selection of new node(s)
766 - Added experimental support for the LXC virtualization method
767 - Added support for OS parameters, which allows the installation of
768 instances to pass parameter to OS scripts in order to customise the
770 - Added a hypervisor parameter controlling the migration type (live or
771 non-live), since hypervisors have various levels of reliability; this
772 has renamed the 'live' parameter to 'mode'
773 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
774 logical volumes, meaning that cluster verify will ignore them and not
775 flag their presence as errors
776 - The watcher will now reset the error count for failed instances after
777 8 hours, thus allowing self-healing if the problem that caused the
778 instances to be down/fail to start has cleared in the meantime
779 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
780 check for, and warn, if the drbd module parameter ``usermode_helper``
781 is not consistent with the cluster-wide setting; this is needed to
782 make diagnose easier of failed drbd creations
783 - Started adding base IPv6 support, but this is not yet
784 enabled/available for use
785 - Rename operations (cluster, instance) will now return the new name,
786 which is especially useful if a short name was passed in
787 - Added support for instance migration in RAPI
788 - Added a tool to pre-configure nodes for the SSH setup, before joining
789 them to the cluster; this will allow in the future a simplified model
790 for node joining (but not yet fully enabled in 2.2); this needs the
791 paramiko python library
792 - Fixed handling of name-resolving errors
793 - Fixed consistency of job results on the error path
794 - Fixed master-failover race condition when executed multiple times in
796 - Fixed many bugs related to the job queue (mostly introduced during the
797 2.2 development cycle, so not all are impacting 2.1)
798 - Fixed instance migration with missing disk symlinks
799 - Fixed handling of unknown jobs in ``gnt-job archive``
800 - And many other small fixes/improvements
804 - Enhanced both the unittest and the QA coverage
805 - Switched the opcode validation to a generic model, and extended the
806 validation to all opcode parameters
807 - Changed more parts of the code that write shell scripts to use the
809 - Switched the master daemon to use the asyncore library for the Luxi
816 *(Released Thu, 17 Jun 2010)*
818 - Added tool (``move-instance``) and infrastructure to move instances
819 between separate clusters (see :doc:`separate documentation
820 <move-instance>` and :doc:`design document <design-2.2>`)
821 - Added per-request RPC timeout
822 - RAPI now requires a Content-Type header for requests with a body (e.g.
823 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
824 :rfc:`2616` (HTTP/1.1), section 7.2.1)
825 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
827 - Implemented initial support for running Ganeti daemons as separate
828 users, see configure-time flags ``--with-user-prefix`` and
829 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
831 - Instances can be removed after export (``gnt-backup export
833 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
834 key (instead of 1024 bit)
835 - Added new cluster configuration file for cluster domain secret
836 - Import/export now use SSL instead of SSH
837 - Added support for showing estimated time when exporting an instance,
838 see the ``ganeti-os-interface(7)`` manpage and look for
845 *(Released Tue, 16 Nov 2010)*
847 Some more bugfixes. Unless critical bugs occur, this will be the last
850 - Fix case of MAC special-values
851 - Fix mac checker regex
852 - backend: Fix typo causing "out of range" error
853 - Add missing --units in gnt-instance list man page
859 *(Released Tue, 24 Aug 2010)*
862 - Don't ignore secondary node silently on non-mirrored disk templates
864 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
865 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
866 - Properly document the usb_mouse parameter
867 - Fix path in ganeti-rapi(8) (issue 116)
868 - Adjust error message when the ganeti user's .ssh directory is
870 - Add same-node-check when changing the disk template to drbd
876 *(Released Fri, 16 Jul 2010)*
879 - Add an option to only select some reboot types during qa/burnin.
880 (on some hypervisors consequent reboots are not supported)
881 - Fix infrequent race condition in master failover. Sometimes the old
882 master ip address would be still detected as up for a short time
883 after it was removed, causing failover to fail.
884 - Decrease mlockall warnings when the ctypes module is missing. On
885 Python 2.4 we support running even if no ctypes module is installed,
886 but we were too verbose about this issue.
887 - Fix building on old distributions, on which man doesn't have a
889 - Fix RAPI not to ignore the MAC address on instance creation
890 - Implement the old instance creation format in the RAPI client.
896 *(Released Thu, 01 Jul 2010)*
898 A small bugfix release:
899 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
900 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
901 a non-existing option
902 - Fix instance create: the --force-variant option was ignored
903 - Improve pylint 0.21 compatibility and warnings with Python 2.6
904 - Fix modify node storage with non-FQDN arguments
905 - Fix RAPI client to authenticate under Python 2.6 when used
906 for more than 5 requests needing authentication
907 - Fix gnt-instance modify -t (storage) giving a wrong error message
908 when converting a non-shutdown drbd instance to plain
914 *(Released Fri, 18 Jun 2010)*
916 A small bugfix release:
918 - Fix live migration of KVM instances started with older Ganeti
919 versions which had fewer hypervisor parameters
920 - Fix gnt-instance grow-disk on down instances
921 - Fix an error-reporting bug during instance migration
922 - Better checking of the ``--net`` and ``--disk`` values, to avoid
923 silently ignoring broken ones
924 - Fix an RPC error reporting bug affecting, for example, RAPI client
926 - Fix bug triggered by different API version os-es on different nodes
927 - Fix a bug in instance startup with custom hvparams: OS level
928 parameters would fail to be applied.
929 - Fix the RAPI client under Python 2.6 (but more work is needed to
930 make it work completely well with OpenSSL)
931 - Fix handling of errors when resolving names from DNS
937 *(Released Thu, 3 Jun 2010)*
939 A medium sized development cycle. Some new features, and some
940 fixes/small improvements/cleanups.
945 The node deamon now tries to mlock itself into memory, unless the
946 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
947 its logs, and falls back to console logging. This allows emergency
948 features such as ``gnt-node powercycle`` to work even in the event of a
949 broken node disk (tested offlining the disk hosting the node's
950 filesystem and dropping its memory caches; don't try this at home)
952 KVM: add vhost-net acceleration support. It can be tested with a new
953 enough version of the kernel and of qemu-kvm.
955 KVM: Add instance chrooting feature. If you use privilege dropping for
956 your VMs you can also now force them to chroot to an empty directory,
957 before starting the emulated guest.
959 KVM: Add maximum migration bandwith and maximum downtime tweaking
960 support (requires a new-enough version of qemu-kvm).
962 Cluster verify will now warn if the master node doesn't have the master
965 Add a new (incompatible) instance creation request format to RAPI which
966 supports all parameters (previously only a subset was supported, and it
967 wasn't possible to extend the old format to accomodate all the new
968 features. The old format is still supported, and a client can check for
969 this feature, before using it, by checking for its presence in the
970 ``features`` RAPI resource.
972 Now with ancient latin support. Try it passing the ``--roman`` option to
973 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
974 (requires the python-roman module to be installed, in order to work).
979 As usual many internal code refactorings, documentation updates, and
982 - Lots of improvements and cleanups to the experimental Remote API
983 (RAPI) client library.
984 - A new unit test suite for the core daemon libraries.
985 - A fix to creating missing directories makes sure the umask is not
986 applied anymore. This enforces the same directory permissions
988 - Better handling terminating daemons with ctrl+c (used when running
989 them in debugging mode).
990 - Fix a race condition in live migrating a KVM instance, when stat()
991 on the old proc status file returned EINVAL, which is an unexpected
993 - Fixed manpage checking with newer man and utf-8 charachters. But now
994 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1000 *(Released Fri, 7 May 2010)*
1002 Fix a bug which prevented untagged KVM instances from starting.
1008 *(Released Fri, 7 May 2010)*
1010 Another release with a long development cycle, during which many
1011 different features were added.
1013 Significant features
1014 ~~~~~~~~~~~~~~~~~~~~
1016 The KVM hypervisor now can run the individual instances as non-root, to
1017 reduce the impact of a VM being hijacked due to bugs in the
1018 hypervisor. It is possible to run all instances as a single (non-root)
1019 user, to manually specify a user for each instance, or to dynamically
1020 allocate a user out of a cluster-wide pool to each instance, with the
1021 guarantee that no two instances will run under the same user ID on any
1024 An experimental RAPI client library, that can be used standalone
1025 (without the other Ganeti libraries), is provided in the source tree as
1026 ``lib/rapi/client.py``. Note this client might change its interface in
1027 the future, as we iterate on its capabilities.
1029 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1030 replace the cluster's certificates and crypto keys. This might help in
1031 case they have been compromised, or have simply expired.
1033 A new disk option for instance creation has been added that allows one
1034 to "adopt" currently existing logical volumes, with data
1035 preservation. This should allow easier migration to Ganeti from
1036 unmanaged (or managed via other software) instances.
1038 Another disk improvement is the possibility to convert between redundant
1039 (DRBD) and plain (LVM) disk configuration for an instance. This should
1040 allow better scalability (starting with one node and growing the
1041 cluster, or shrinking a two-node cluster to one node).
1043 A new feature that could help with automated node failovers has been
1044 implemented: if a node sees itself as offline (by querying the master
1045 candidates), it will try to shutdown (hard) all instances and any active
1046 DRBD devices. This reduces the risk of duplicate instances if an
1047 external script automatically failovers the instances on such nodes. To
1048 enable this, the cluster parameter ``maintain_node_health`` should be
1049 enabled; in the future this option (per the name) will enable other
1050 automatic maintenance features.
1052 Instance export/import now will reuse the original instance
1053 specifications for all parameters; that means exporting an instance,
1054 deleting it and the importing it back should give an almost identical
1055 instance. Note that the default import behaviour has changed from
1056 before, where it created only one NIC; now it recreates the original
1059 Cluster verify has added a few new checks: SSL certificates validity,
1060 /etc/hosts consistency across the cluster, etc.
1065 As usual, many internal changes were done, documentation fixes,
1068 - Fixed cluster initialization with disabled cluster storage (regression
1069 introduced in 2.1.1)
1070 - File-based storage supports growing the disks
1071 - Fixed behaviour of node role changes
1072 - Fixed cluster verify for some corner cases, plus a general rewrite of
1073 cluster verify to allow future extension with more checks
1074 - Fixed log spamming by watcher and node daemon (regression introduced
1076 - Fixed possible validation issues when changing the list of enabled
1078 - Fixed cleanup of /etc/hosts during node removal
1079 - Fixed RAPI response for invalid methods
1080 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1081 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1082 from ide disks, etc.)
1083 - Allow OS changes without re-installation (to record a changed OS
1084 outside of Ganeti, or to allow OS renames)
1085 - Allow instance creation without OS installation (useful for example if
1086 the OS will be installed manually, or restored from a backup not in
1088 - Implemented option to make cluster ``copyfile`` use the replication
1090 - Added list of enabled hypervisors to ssconf (possibly useful for
1092 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1094 - A partial form of node re-IP is possible via node readd, which now
1095 allows changed node primary IP
1096 - Command line utilities now show an informational message if the job is
1098 - The logs of the master daemon now show the PID/UID/GID of the
1105 *(Released Fri, 12 Mar 2010)*
1107 During the 2.1.0 long release candidate cycle, a lot of improvements and
1108 changes have accumulated with were released later as 2.1.1.
1113 The node evacuate command (``gnt-node evacuate``) was significantly
1114 rewritten, and as such the IAllocator protocol was changed - a new
1115 request type has been added. This unfortunate change during a stable
1116 series is designed to improve performance of node evacuations; on
1117 clusters with more than about five nodes and which are well-balanced,
1118 evacuation should proceed in parallel for all instances of the node
1119 being evacuated. As such, any existing IAllocator scripts need to be
1120 updated, otherwise the above command will fail due to the unknown
1121 request. The provided "dumb" allocator has not been updated; but the
1122 ganeti-htools package supports the new protocol since version 0.2.4.
1124 Another important change is increased validation of node and instance
1125 names. This might create problems in special cases, if invalid host
1126 names are being used.
1128 Also, a new layer of hypervisor parameters has been added, that sits at
1129 OS level between the cluster defaults and the instance ones. This allows
1130 customisation of virtualization parameters depending on the installed
1131 OS. For example instances with OS 'X' may have a different KVM kernel
1132 (or any other parameter) than the cluster defaults. This is intended to
1133 help managing a multiple OSes on the same cluster, without manual
1134 modification of each instance's parameters.
1136 A tool for merging clusters, ``cluster-merge``, has been added in the
1137 tools sub-directory.
1142 - Improved the int/float conversions that should make the code more
1143 robust in face of errors from the node daemons
1144 - Fixed the remove node code in case of internal configuration errors
1145 - Fixed the node daemon behaviour in face of inconsistent queue
1146 directory (e.g. read-only file-system where we can't open the files
1148 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1149 now it either aborts cleanly or, if given the new "auto_promote"
1150 parameter, will automatically promote other nodes as needed
1151 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1152 completely prevent Ganeti from working
1153 - Fixed bug for instance export when not all disks were successfully
1155 - Fixed behaviour of node add when the new node is slow in starting up
1157 - Fixed handling of signals in the LUXI client, which should improve
1158 behaviour of command-line scripts
1159 - Added checks for invalid node/instance names in the configuration (now
1160 flagged during cluster verify)
1161 - Fixed watcher behaviour for disk activation errors
1162 - Fixed two potentially endless loops in http library, which led to the
1163 RAPI daemon hanging and consuming 100% CPU in some cases
1164 - Fixed bug in RAPI daemon related to hashed passwords
1165 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1167 - Enhanced compatibility with non-Debian OSes, but not using absolute
1168 path in some commands and allowing customisation of the ssh
1169 configuration directory
1170 - Fixed possible future issue with new Python versions by abiding to the
1171 proper use of ``__slots__`` attribute on classes
1172 - Added checks that should prevent directory traversal attacks
1173 - Many documentation fixes based on feedback from users
1178 - Added an "early_release" more for instance replace disks and node
1179 evacuate, where we release locks earlier and thus allow higher
1180 parallelism within the cluster
1181 - Added watcher hooks, intended to allow the watcher to restart other
1182 daemons (e.g. from the ganeti-nbma project), but they can be used of
1183 course for any other purpose
1184 - Added a compile-time disable for DRBD barriers, to increase
1185 performance if the administrator trusts the power supply or the
1186 storage system to not lose writes
1187 - Added the option of using syslog for logging instead of, or in
1188 addition to, Ganeti's own log files
1189 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1190 can indeed boot from a paravirtual NIC
1191 - Added a generic debug level for many operations; while this is not
1192 used widely yet, it allows one to pass the debug value all the way to
1194 - Enhanced the hooks environment for instance moves (failovers,
1195 migrations) where the primary/secondary nodes changed during the
1196 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1197 - Enhanced data validations for many user-supplied values; one important
1198 item is the restrictions imposed on instance and node names, which
1199 might reject some (invalid) host names
1200 - Add a configure-time option to disable file-based storage, if it's not
1201 needed; this allows greater security separation between the master
1202 node and the other nodes from the point of view of the inter-node RPC
1204 - Added user notification in interactive tools if job is waiting in the
1205 job queue or trying to acquire locks
1206 - Added log messages when a job is waiting for locks
1207 - Added filtering by node tags in instance operations which admit
1208 multiple instances (start, stop, reboot, reinstall)
1209 - Added a new tool for cluster mergers, ``cluster-merge``
1210 - Parameters from command line which are of the form ``a=b,c=d`` can now
1211 use backslash escapes to pass in values which contain commas,
1212 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1214 - For KVM, the instance name is the first parameter passed to KVM, so
1215 that it's more visible in the process list
1221 *(Released Tue, 2 Mar 2010)*
1223 Ganeti 2.1 brings many improvements with it. Major changes:
1225 - Added infrastructure to ease automated disk repairs
1226 - Added new daemon to export configuration data in a cheaper way than
1227 using the remote API
1228 - Instance NICs can now be routed instead of being associated with a
1230 - Improved job locking logic to reduce impact of jobs acquiring multiple
1231 locks waiting for other long-running jobs
1233 In-depth implementation details can be found in the Ganeti 2.1 design
1239 - Added chroot hypervisor
1240 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1242 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1243 ``bootloader_path`` and ``bootloader_args``)
1244 - Added the ``use_localtime`` option for the xen-hvm and kvm
1245 hypervisors, and the default value for this has changed to false (in
1246 2.0 xen-hvm always enabled it)
1247 - Added luxi call to submit multiple jobs in one go
1248 - Added cluster initialization option to not modify ``/etc/hosts``
1250 - Added network interface parameters
1251 - Added dry run mode to some LUs
1252 - Added RAPI resources:
1254 - ``/2/instances/[instance_name]/info``
1255 - ``/2/instances/[instance_name]/replace-disks``
1256 - ``/2/nodes/[node_name]/evacuate``
1257 - ``/2/nodes/[node_name]/migrate``
1258 - ``/2/nodes/[node_name]/role``
1259 - ``/2/nodes/[node_name]/storage``
1260 - ``/2/nodes/[node_name]/storage/modify``
1261 - ``/2/nodes/[node_name]/storage/repair``
1263 - Added OpCodes to evacuate or migrate all instances on a node
1264 - Added new command to list storage elements on nodes (``gnt-node
1265 list-storage``) and modify them (``gnt-node modify-storage``)
1266 - Added new ssconf files with master candidate IP address
1267 (``ssconf_master_candidates_ips``), node primary IP address
1268 (``ssconf_node_primary_ips``) and node secondary IP address
1269 (``ssconf_node_secondary_ips``)
1270 - Added ``ganeti-confd`` and a client library to query the Ganeti
1271 configuration via UDP
1272 - Added ability to run hooks after cluster initialization and before
1274 - Added automatic mode for disk replace (``gnt-instance replace-disks
1276 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1277 after catastrophic data-loss
1278 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1280 - Added ``gnt-instance move`` command to move instances
1281 - Added ``gnt-cluster watcher`` command to control watcher
1282 - Added ``gnt-node powercycle`` command to powercycle nodes
1283 - Added new job status field ``lock_status``
1284 - Added parseable error codes to cluster verification (``gnt-cluster
1285 verify --error-codes``) and made output less verbose (use
1286 ``--verbose`` to restore previous behaviour)
1287 - Added UUIDs to the main config entities (cluster, nodes, instances)
1288 - Added support for OS variants
1289 - Added support for hashed passwords in the Ganeti remote API users file
1291 - Added option to specify maximum timeout on instance shutdown
1292 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1293 - Added new helper script to start and stop Ganeti daemons
1294 (``daemon-util``), with the intent to reduce the work necessary to
1295 adjust Ganeti for non-Debian distributions and to start/stop daemons
1297 - Added more unittests
1298 - Fixed critical bug in ganeti-masterd startup
1299 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1300 now customisable at the cluster level for both the KVM and Xen
1301 hypervisors using the new ``migration_port`` parameter
1302 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1303 reinstalling an instance
1304 - Allowed ``@`` in tag names
1305 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1306 - Many documentation updates
1307 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1308 - ``gnt-instance reinstall`` can now reinstall multiple instances
1309 - Updated many command line parameters
1310 - Introduced new OS API version 15
1311 - No longer support a default hypervisor
1312 - Treat virtual LVs as inexistent
1313 - Improved job locking logic to reduce lock contention
1314 - Match instance and node names case insensitively
1315 - Reimplemented bash completion script to be more complete
1322 *(Released Thu, 4 Feb 2010)*
1324 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1325 - Fix a string formatting bug
1326 - Improve safety of the code in some error paths
1327 - Improve data validation in the master of values returned from nodes
1333 *(Released Thu, 17 Dec 2009)*
1335 - Fix security issue due to missing validation of iallocator names; this
1336 allows local and remote execution of arbitrary executables
1337 - Fix failure of gnt-node list during instance removal
1338 - Ship the RAPI documentation in the archive
1344 *(Released Wed, 30 Sep 2009)*
1346 - Fixed many wrong messages
1347 - Fixed a few bugs related to the locking library
1348 - Fixed MAC checking at instance creation time
1349 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1350 - Fixed a few issues related to signal handling in both daemons and
1352 - Fixed the example startup script provided
1353 - Fixed insserv dependencies in the example startup script (patch from
1355 - Fixed handling of drained nodes in the iallocator framework
1356 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1358 - Fixed error related to invalid job IDs in job polling
1359 - Fixed job/opcode persistence on unclean master shutdown
1360 - Fixed handling of partial job processing after unclean master
1362 - Fixed error reporting from LUs, previously all errors were converted
1363 into execution errors
1364 - Fixed error reporting from burnin
1365 - Decreased significantly the memory usage of the job queue
1366 - Optimised slightly multi-job submission
1367 - Optimised slightly opcode loading
1368 - Backported the multi-job submit framework from the development
1369 branch; multi-instance start and stop should be faster
1370 - Added script to clean archived jobs after 21 days; this will reduce
1371 the size of the queue directory
1372 - Added some extra checks in disk size tracking
1373 - Added an example ethers hook script
1374 - Added a cluster parameter that prevents Ganeti from modifying of
1376 - Added more node information to RAPI responses
1377 - Added a ``gnt-job watch`` command that allows following the ouput of a
1379 - Added a bind-address option to ganeti-rapi
1380 - Added more checks to the configuration verify
1381 - Enhanced the burnin script such that some operations can be retried
1383 - Converted instance reinstall to multi-instance model
1389 *(Released Fri, 7 Aug 2009)*
1391 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1392 to allow using the pre-2.0.2 behaviour in activation, if any existing
1393 instances have mismatched disk sizes in the configuration
1394 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
1395 any configuration mismatches for disk sizes
1396 - Added ``gnt-master cluste-failover --no-voting`` to allow master
1397 failover to work on two-node clusters
1398 - Fixed the ``--net`` option of ``gnt-backup import``, which was
1400 - Fixed detection of OS script errors in ``gnt-backup export``
1401 - Fixed exit code of ``gnt-backup export``
1407 *(Released Fri, 17 Jul 2009)*
1409 - Added experimental support for stripped logical volumes; this should
1410 enhance performance but comes with a higher complexity in the block
1411 device handling; stripping is only enabled when passing
1412 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
1413 affected even in the non-stripped mode
1414 - Improved resiliency against transient failures at the end of DRBD
1415 resyncs, and in general of DRBD resync checks
1416 - Fixed a couple of issues with exports and snapshot errors
1417 - Fixed a couple of issues in instance listing
1418 - Added display of the disk size in ``gnt-instance info``
1419 - Fixed checking for valid OSes in instance creation
1420 - Fixed handling of the "vcpus" parameter in instance listing and in
1421 general of invalid parameters
1422 - Fixed http server library, and thus RAPI, to handle invalid
1423 username/password combinations correctly; this means that now they
1424 report unauthorized for queries too, not only for modifications,
1425 allowing earlier detect of configuration problems
1426 - Added a new "role" node list field, equivalent to the master/master
1427 candidate/drained/offline flags combinations
1428 - Fixed cluster modify and changes of candidate pool size
1429 - Fixed cluster verify error messages for wrong files on regular nodes
1430 - Fixed a couple of issues with node demotion from master candidate role
1431 - Fixed node readd issues
1432 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
1433 - Added a new ``--no-voting`` option for masterfailover to fix failover
1434 on two-nodes clusters when the former master node is unreachable
1435 - Added instance reinstall over RAPI
1441 *(Released Tue, 16 Jun 2009)*
1443 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
1444 allow re-adding the start in single-user option (regression from 1.2)
1445 - the watcher writes the instance status to a file, to allow monitoring
1446 to report the instance status (from the master) based on cached
1447 results of the watcher's queries; while this can get stale if the
1448 watcher is being locked due to other work on the cluster, this is
1449 still an improvement
1450 - the watcher now also restarts the node daemon and the rapi daemon if
1452 - fixed the watcher to handle full and drained queue cases
1453 - hooks export more instance data in the environment, which helps if
1454 hook scripts need to take action based on the instance's properties
1455 (no longer need to query back into ganeti)
1456 - instance failovers when the instance is stopped do not check for free
1457 RAM, so that failing over a stopped instance is possible in low memory
1459 - rapi uses queries for tags instead of jobs (for less job traffic), and
1460 for cluster tags it won't talk to masterd at all but read them from
1462 - a couple of error handling fixes in RAPI
1463 - drbd handling: improved the error handling of inconsistent disks after
1464 resync to reduce the frequency of "there are some degraded disks for
1465 this instance" messages
1466 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
1467 error handling path called a wrong function name)
1473 *(Released Wed, 27 May 2009)*
1475 - no changes from rc5
1478 Version 2.0 release candidate 5
1479 -------------------------------
1481 *(Released Wed, 20 May 2009)*
1483 - fix a couple of bugs (validation, argument checks)
1484 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
1485 - some small improvements to RAPI and IAllocator
1486 - make watcher automatically start the master daemon if down
1489 Version 2.0 release candidate 4
1490 -------------------------------
1492 *(Released Mon, 27 Apr 2009)*
1494 - change the OS list to not require locks; this helps with big clusters
1495 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
1496 volume group is broken
1497 - ``gnt-instance info``, without any arguments, doesn't run for all
1498 instances anymore; either pass ``--all`` or pass the desired
1499 instances; this helps against mistakes on big clusters where listing
1500 the information for all instances takes a long time
1501 - miscellaneous doc and man pages fixes
1504 Version 2.0 release candidate 3
1505 -------------------------------
1507 *(Released Wed, 8 Apr 2009)*
1509 - Change the internal locking model of some ``gnt-node`` commands, in
1510 order to reduce contention (and blocking of master daemon) when
1511 batching many creation/reinstall jobs
1512 - Fixes to Xen soft reboot
1513 - No longer build documentation at build time, instead distribute it in
1514 the archive, in order to reduce the need for the whole docbook/rst
1518 Version 2.0 release candidate 2
1519 -------------------------------
1521 *(Released Fri, 27 Mar 2009)*
1523 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
1524 - Fix watcher startup sequence, improves the behaviour of busy clusters
1525 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
1526 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
1527 bind address and other places
1528 - Some documentation fixes and updates
1531 Version 2.0 release candidate 1
1532 -------------------------------
1534 *(Released Mon, 2 Mar 2009)*
1536 - More documentation updates, now all docs should be more-or-less
1538 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
1540 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
1542 - Fix an issue related to $libdir/run/ganeti and cluster creation
1548 *(Released Thu, 19 Feb 2009)*
1550 - Xen PVM and KVM have switched the default value for the instance root
1551 disk to the first partition on the first drive, instead of the whole
1552 drive; this means that the OS installation scripts must be changed
1554 - Man pages have been updated
1555 - RAPI has been switched by default to HTTPS, and the exported functions
1556 should all work correctly
1557 - RAPI v1 has been removed
1558 - Many improvements to the KVM hypervisor
1559 - Block device errors are now better reported
1560 - Many other bugfixes and small improvements
1566 *(Released Mon, 26 Jan 2009)*
1568 - Version 2 is a general rewrite of the code and therefore the
1569 differences are too many to list, see the design document for 2.0 in
1570 the ``doc/`` subdirectory for more details
1571 - In this beta version there is not yet a migration path from 1.2 (there
1572 will be one in the final 2.0 release)
1573 - A few significant changes are:
1575 - all commands are executed by a daemon (``ganeti-masterd``) and the
1576 various ``gnt-*`` commands are just front-ends to it
1577 - all the commands are entered into, and executed from a job queue,
1578 see the ``gnt-job(8)`` manpage
1579 - the RAPI daemon supports read-write operations, secured by basic
1580 HTTP authentication on top of HTTPS
1581 - DRBD version 0.7 support has been removed, DRBD 8 is the only
1582 supported version (when migrating from Ganeti 1.2 to 2.0, you need
1583 to migrate to DRBD 8 first while still running Ganeti 1.2)
1584 - DRBD devices are using statically allocated minor numbers, which
1585 will be assigned to existing instances during the migration process
1586 - there is support for both Xen PVM and Xen HVM instances running on
1588 - KVM virtualization is supported too
1589 - file-based storage has been implemented, which means that it is
1590 possible to run the cluster without LVM and DRBD storage, for
1591 example using a shared filesystem exported from shared storage (and
1592 still have live migration)
1598 *(Released Tue, 13 Jan 2009)*
1600 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
1601 - Reuse the old instance mac address by default on instance import, if
1602 the instance name is the same.
1603 - Handle situations in which the node info rpc returns incomplete
1605 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
1606 - Improved version of batcher:
1608 - state file support
1609 - instance mac address support
1610 - support for HVM clusters/instances
1612 - Add an option to show the number of cpu sockets and nodes in
1614 - Support OSes that handle more than one version of the OS api (but do
1615 not change the current API in any other way)
1616 - Fix ``gnt-node migrate``
1617 - ``gnt-debug`` man page
1618 - Fixes various more typos and small issues
1619 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
1625 *(Released Wed, 24 Sep 2008)*
1627 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
1628 type of disk exported to fully virtualized instances.
1629 - provide access to the serial console of HVM instances
1630 - instance auto_balance flag, set by default. If turned off it will
1631 avoid warnings on cluster verify if there is not enough memory to fail
1632 over an instance. in the future it will prevent automatically failing
1633 it over when we will support that.
1634 - batcher tool for instance creation, see ``tools/README.batcher``
1635 - ``gnt-instance reinstall --select-os`` to interactively select a new
1636 operating system when reinstalling an instance.
1637 - when changing the memory amount on instance modify a check has been
1638 added that the instance will be able to start. also warnings are
1639 emitted if the instance will not be able to fail over, if auto_balance
1641 - documentation fixes
1642 - sync fields between ``gnt-instance list/modify/add/import``
1643 - fix a race condition in drbd when the sync speed was set after giving
1644 the device a remote peer.
1650 *(Released Tue, 22 Jul 2008)*
1652 - note: the allowed size and number of tags per object were reduced
1653 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1654 - fixed twisted 8.x compatibility
1655 - fixed ``gnt-instance replace-disks`` with iallocator
1656 - add TCP keepalives on twisted connections to detect restarted nodes
1657 - disk increase support, see ``gnt-instance grow-disk``
1658 - implement bulk node/instance query for RAPI
1659 - add tags in node/instance listing (optional)
1660 - experimental migration (and live migration) support, read the man page
1661 for ``gnt-instance migrate``
1662 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1663 has some small improvements in handling its state file
1669 *(Released Fri, 13 Jun 2008)*
1671 - Experimental readonly, REST-based remote API implementation;
1672 automatically started on master node, TCP port 5080, if enabled by
1673 ``--enable-rapi`` parameter to configure script.
1674 - Instance allocator support. Add and import instance accept a
1675 ``--iallocator`` parameter, and call that instance allocator to decide
1676 which node to use for the instance. The iallocator document describes
1677 what's expected from an allocator script.
1678 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1679 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1680 node is lost there is still enough memory to fail over the instances
1682 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1683 ``gnt-cluster verify``, to check for site-specific compliance. All the
1684 hooks will run, and their output, if any, will be displayed. Any
1685 failing hook will make the verification return an error value.
1686 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1687 primary and secondary interfaces
1688 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1689 that is still declared as part of the cluster and has failed.
1690 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1691 specifying output fields, that just adds the chosen fields to the
1693 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1694 export from the filesystem.
1695 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1696 have been added. Using them you can enable/disable acpi and pae
1697 support, and specify a path for a cd image to be exported to the
1698 instance. These parameters as the name suggest only work on HVM
1700 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1701 PAE support will be set to the previously hardcoded values, but the
1702 (previously hardcoded) path to the CDROM ISO image will be unset and
1703 if required, needs to be set manually with ``gnt-instance modify``
1705 - The address to which an instance's VNC console is bound is now
1706 selectable per-instance, rather than being cluster wide. Of course
1707 this only applies to instances controlled via VNC, so currently just
1708 applies to HVM clusters.
1714 *(Released Mon, 18 Feb 2008)*
1716 - more tweaks to the disk activation code (especially helpful for DRBD)
1717 - change the default ``gnt-instance list`` output format, now there is
1718 one combined status field (see the manpage for the exact values this
1720 - some more fixes for the mac export to hooks change
1721 - make Ganeti not break with DRBD 8.2.x (which changed the version
1722 format in ``/proc/drbd``) (issue 24)
1723 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1724 template, allowing migration from DRBD0.7+MD to DRBD8
1730 *(Released Wed, 30 Jan 2008)*
1732 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1734 - add command aliases infrastructure and a few aliases
1735 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1736 man pages and the ``--help`` option of ``gnt-node
1737 list``/``gnt-instance list``
1738 - fix ``gnt-backup list`` with down nodes (issue 21)
1739 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1740 - fix the dist archive and add a check for including svn/git files in
1742 - some developer-related changes: improve the burnin and the QA suite,
1743 add an upload script for testing during development
1749 *(Released Wed, 16 Jan 2008)*
1751 - experimental HVM support, read the install document, section
1752 "Initializing the cluster"
1753 - allow for the PVM hypervisor per-instance kernel and initrd paths
1754 - add a new command ``gnt-cluster verify-disks`` which uses a new
1755 algorithm to improve the reconnection of the DRBD pairs if the device
1756 on the secondary node has gone away
1757 - make logical volume code auto-activate LVs at disk activation time
1758 - slightly improve the speed of activating disks
1759 - allow specification of the MAC address at instance creation time, and
1760 changing it later via ``gnt-instance modify``
1761 - fix handling of external commands that generate lots of output on
1763 - update documentation with regard to minimum version of DRBD8 supported
1769 *(Released Tue, 4 Dec 2007)*
1771 - Log the ``xm create`` output to the node daemon log on failure (to
1772 help diagnosing the error)
1773 - In debug mode, log all external commands output if failed to the logs
1774 - Change parsing of lvm commands to ignore stderr
1780 *(Released Wed, 28 Nov 2007)*
1782 - Another round of updates to the DRBD 8 code to deal with more failures
1783 in the replace secondary node operation
1784 - Some more logging of failures in disk operations (lvm, drbd)
1785 - A few documentation updates
1792 *(Released Tue, 13 Nov 2007)*
1794 - Change configuration file format from Python's Pickle to JSON.
1795 Upgrading is possible using the cfgupgrade utility.
1796 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1797 faster replace disks and is more stable (DRBD 8 has many improvements
1798 compared to DRBD 0.7)
1799 - Added command line tags support (see man pages for ``gnt-instance``,
1800 ``gnt-node``, ``gnt-cluster``)
1801 - Added instance rename support
1802 - Added multi-instance startup/shutdown
1803 - Added cluster rename support
1804 - Added ``gnt-node evacuate`` to simplify some node operations
1805 - Added instance reboot operation that can speedup reboot as compared to
1807 - Soften the requirement that hostnames are in FQDN format
1808 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1810 - Removed dependency on debian's patched fping that uses the
1811 non-standard ``-S`` option
1812 - Now the OS definitions are searched for in multiple, configurable
1813 paths (easier for distros to package)
1814 - Some changes to the hooks infrastructure (especially the new
1815 post-configuration update hook)
1816 - Other small bugfixes
1818 .. vim: set textwidth=72 syntax=rst :