9 Many bug-fixes and a few new small features:
11 - Fixed a bug related to log opening failures
12 - Fixed a bug in instance listing with orphan instances
13 - Fixed a bug which prevented resetting the cluster-level node parameter
14 ``oob_program`` to the default
15 - Many fixes related to the ``cluster-merge`` tool
16 - Fixed a race condition in the lock monitor, which caused failures
17 during (at least) creation of many instances in parallel
18 - Improved output for gnt-job info
19 - Removed the quiet flag on some ssh calls which prevented debugging
21 - Improved the N+1 failure messages in cluster verify by actually
22 showing the memory values (needed and available)
23 - Increased lock attempt timeouts so that when executing long operations
24 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
25 too early and thus prevent the use of the 'fair' mechanism
26 - Changed instance query data (``gnt-instance info``) to not acquire
27 locks unless needed, thus allowing its use on locked instance if only
28 static information is asked for
29 - Improved behaviour with filesystems that do not support rename on an
31 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
32 kept locks on all nodes during the wipe, which is unneeded
33 - Fixed ``gnt-watcher`` handling of errors during hooks execution
34 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
35 10GiB) which caused the wipe to fail right at the end in some cases
36 - Fixed master IP activation when doing master failover with no-voting
37 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
38 the master node itself
39 - Fixed potential data-loss in under disk full conditions, where Ganeti
40 wouldn't check correctly the return code and would consider
41 partially-written files 'correct'
42 - Fixed bug related to multiple VGs and DRBD disk replacing
43 - Added new disk parameter ``metavg`` that allows placement of the meta
44 device for DRBD in a different volume group
45 - Fixed error handling in the node daemon when the system libc doesn't
46 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
47 - Fixed lock release during replace-disks, which kept cluster-wide locks
48 when doing disk replaces with an iallocator script
49 - Added check for missing bridges in cluster verify
50 - Handle EPIPE errors while writing to the terminal better, so that
51 piping the output to e.g. ``less`` doesn't cause a backtrace
52 - Fixed rare case where a ^C during Luxi calls could have been
53 interpreted as server errors, instead of simply terminating
54 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
56 - Added a few more parameters to the KVM hypervisor, allowing a second
57 CDROM, custom disk type for CDROMs and a floppy image
58 - Removed redundant message in instance rename when the name is given
60 - Added option to ``gnt-instance recreate-disks`` to allow creating the
61 disks on new nodes, allowing recreation when the original instance
62 nodes are completely gone
63 - Added option when converting disk templates to DRBD to skip waiting
64 for the resync, in order to make the instance available sooner
65 - Added two new variables to the OS scripts environment (containing the
67 - Made the root_path and optional parameter for the xen-pvm hypervisor,
68 to allow use of ``pvgrub`` as bootloader
70 And as usual, various improvements to the error messages, documentation
76 *(Released Wed, 09 Mar 2011)*
78 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
79 the RAPI users file if run twice (even with ``-dry-run``).
81 The release fixes that bug (nothing else changed).
87 *(Released Mon, 07 Mar 2011)*
89 Final 2.4.0 release. Just a few small fixes:
91 - Fixed RAPI node evacuate
92 - Fixed the kvm-ifup script
93 - Fixed internal error handling for special job cases
94 - Updated man page to specify the escaping feature for options
100 *(Released Mon, 28 Feb 2011)*
102 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
103 that this feature wiped the disks of the wrong instance, leading to loss
108 - Fixed title of query field containing instance name
109 - Expanded the glossary in the documentation
110 - Fixed one unittest (internal issue)
116 *(Released Mon, 21 Feb 2011)*
118 A number of bug fixes plus just a couple functionality changes.
120 On the user-visible side, the ``gnt-* list`` command output has changed
121 with respect to "special" field states. The current rc1 style of display
122 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
123 the default output mode special fields states are displayed as follows:
125 - Offline resource: ``*``
126 - Unavailable/not applicable: ``-``
127 - Data missing (RPC failure): ``?``
128 - Unknown field: ``??``
130 Another user-visible change is the addition of ``--force-join`` to
135 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
136 - Fixed regression in RAPI/instance reinstall where all parameters were
137 required (instead of optional)
138 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
139 - Fixed iallocator usage (offline nodes were not considered offline)
140 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
141 - Fixed hypervisor and OS parameter validation with respect to
143 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
145 - Fixed ``tools/listrunner`` with respect to agent-based usage
151 *(Released Fri, 4 Feb 2011)*
153 Many changes and fixes since the beta1 release. While there were some
154 internal changes, the code has been mostly stabilised for the RC
157 Note: the dumb allocator was removed in this release, as it was not kept
158 up-to-date with the IAllocator protocol changes. It is recommended to
159 use the ``hail`` command from the ganeti-htools package.
161 Note: the 2.4 and up versions of Ganeti are not compatible with the
162 0.2.x branch of ganeti-htools. You need to upgrade to
163 ganeti-htools-0.3.0 (or later).
165 Regressions fixed from 2.3
166 ~~~~~~~~~~~~~~~~~~~~~~~~~~
168 - Fixed the ``gnt-cluster verify-disks`` command
169 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
171 - Fixed disk adoption breakage
172 - Fixed wrong headers in instance listing for field aliases
177 - Fixed corner case in KVM handling of NICs
178 - Fixed many cases of wrong handling of non-vm_capable nodes
179 - Fixed a bug where a missing instance symlink was not possible to
180 recreate with any ``gnt-*`` command (now ``gnt-instance
181 activate-disks`` does it)
182 - Fixed the volume group name as reported by ``gnt-cluster
184 - Increased timeouts for the import-export code, hopefully leading to
185 fewer aborts due network or instance timeouts
186 - Fixed bug in ``gnt-node list-storage``
187 - Fixed bug where not all daemons were started on cluster
188 initialisation, but only at the first watcher run
189 - Fixed many bugs in the OOB implementation
190 - Fixed watcher behaviour in presence of instances with offline
192 - Fixed instance list output for instances running on the wrong node
193 - a few fixes to the cluster-merge tool, but it still cannot merge
194 multi-node groups (currently it is not recommended to use this tool)
200 - Improved network configuration for the KVM hypervisor
201 - Added e1000 as a supported NIC for Xen-HVM
202 - Improved the lvmstrap tool to also be able to use partitions, as
203 opposed to full disks
204 - Improved speed of disk wiping (the cluster parameter
205 ``prealloc_wipe_disks``, so that it has a low impact on the total time
206 of instance creations
207 - Added documentation for the OS parameters
208 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
209 hypervisor is not responding
210 - Added display of blacklisted and hidden OS information in
212 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
213 NIC and node parameters, which might create problems with currently
214 invalid (but undetected) configuration files, but prevents validation
215 failures when unrelated parameters are modified
216 - Changed cluster initialisation to wait for the master daemon to become
218 - Expanded the RAPI interface:
220 - Added config redistribution resource
221 - Added activation/deactivation of instance disks
222 - Added export of console information
224 - Implemented log file reopening on SIGHUP, which allows using
225 logrotate(8) for the Ganeti log files
226 - Added a basic OOB helper script as an example
232 *(Released Fri, 14 Jan 2011)*
237 - Fixed timezone issues when formatting timestamps
238 - Added support for node groups, available via ``gnt-group`` and other
240 - Added out-of-band framework and management, see :doc:`design
241 document <design-oob>`
242 - Removed support for roman numbers from ``gnt-node list`` and
243 ``gnt-instance list``.
244 - Allowed modification of master network interface via ``gnt-cluster
245 modify --master-netdev``
246 - Accept offline secondaries while shutting down instance disks
247 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
248 - Added support for multiple LVM volume groups
249 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
251 - Added commands to list available fields:
253 - ``gnt-node list-fields``
254 - ``gnt-group list-fields``
255 - ``gnt-instance list-fields``
257 - Updated documentation and man pages
262 - Moved ``rapi_users`` file into separate directory, now named
263 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
265 - Added new tool for running commands on many machines,
266 ``tools/ganeti-listrunner``
267 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
268 improving the ``gnt-instance console`` output
269 - Allowed customisation of disk index separator at ``configure`` time
270 - Export node group allocation policy to :doc:`iallocator <iallocator>`
271 - Added support for non-partitioned md disks in ``lvmstrap``
272 - Added script to gracefully power off KVM instances
273 - Split ``utils`` module into smaller parts
274 - Changed query operations to return more detailed information, e.g.
275 whether an information is unavailable due to an offline node. To use
276 this new functionality, the LUXI call ``Query`` must be used. Field
277 information is now stored by the master daemon and can be retrieved
278 using ``QueryFields``. Instances, nodes and groups can also be queried
279 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
280 exposed via RAPI). The following commands make use of this
281 infrastructure change:
284 - ``gnt-group list-fields``
286 - ``gnt-node list-fields``
287 - ``gnt-instance list``
288 - ``gnt-instance list-fields``
289 - ``gnt-debug locks``
294 - New RAPI resources (see :doc:`rapi`):
298 - ``/2/groups/[group_name]``
299 - ``/2/groups/[group_name]/assign-nodes``
300 - ``/2/groups/[group_name]/modify``
301 - ``/2/groups/[group_name]/rename``
302 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
306 - Implemented ``no_install`` for instance creation
307 - Implemented OS parameters for instance reinstallation, allowing
308 use of special settings on reinstallation (e.g. for preserving data)
313 - Added IPv6 support in import/export
314 - Pause DRBD synchronization while wiping disks on instance creation
315 - Updated unittests and QA scripts
316 - Improved network parameters passed to KVM
317 - Converted man pages from docbook to reStructuredText
323 *(Released Mon, 20 Dec 2010)*
325 Released version 2.3.1~rc1 without any changes.
331 *(Released Wed, 1 Dec 2010)*
333 - impexpd: Disable OpenSSL compression in socat if possible (backport
334 from master, commit e90739d625b, see :doc:`installation guide
335 <install-quick>` for details)
336 - Changed unittest coverage report to exclude test scripts
337 - Added script to check version format
343 *(Released Wed, 1 Dec 2010)*
345 Released version 2.3.0~rc1 without any changes.
351 *(Released Fri, 19 Nov 2010)*
353 A number of bugfixes and documentation updates:
355 - Update ganeti-os-interface documentation
356 - Fixed a bug related to duplicate MACs or similar items which should be
358 - Fix breakage in OS state modify
359 - Reinstall instance: disallow offline secondaries (fixes bug related to
360 OS changing but reinstall failing)
361 - plus all the other fixes between 2.2.1 and 2.2.2
367 *(Released Tue, 2 Nov 2010)*
369 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
370 - Fixed master failover race with watcher
371 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
373 - Accept previously stopped instance for export with instance removal
374 - Simplify and extend the environment variables for instance OS scripts
375 - Added new node flags, ``master_capable`` and ``vm_capable``
376 - Added optional instance disk wiping prior during allocation. This is a
377 cluster-wide option and can be set/modified using
378 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
379 - Added IPv6 support, see :doc:`design document <design-2.3>` and
381 - Added a new watcher option (``--ignore-pause``)
382 - Added option to ignore offline node on instance start/stop
383 (``--ignore-offline``)
384 - Allow overriding OS parameters with ``gnt-instance reinstall``
385 - Added ability to change node's secondary IP address using ``gnt-node
387 - Implemented privilege separation for all daemons except
388 ``ganeti-noded``, see ``configure`` options
389 - Complain if an instance's disk is marked faulty in ``gnt-cluster
391 - Implemented job priorities (see ``ganeti(7)`` manpage)
392 - Ignore failures while shutting down instances during failover from
394 - Exit daemon's bootstrap process only once daemon is ready
395 - Export more information via ``LUInstanceQuery``/remote API
396 - Improved documentation, QA and unittests
397 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
398 restart if the file was created or changed
399 - Added LUXI protocol version sent with each request and response,
400 allowing detection of server/client mismatches
401 - Moved the Python scripts among gnt-* and ganeti-* into modules
402 - Moved all code related to setting up SSH to an external script,
404 - Infrastructure changes for node group support in future versions
410 *(Released Fri, 19 Nov 2010)*
412 A few small bugs fixed, and some improvements to the build system:
414 - Fix documentation regarding conversion to drbd
415 - Fix validation of parameters in cluster modify (``gnt-cluster modify
417 - Fix error handling in node modify with multiple changes
418 - Allow remote imports without checked names
424 *(Released Tue, 19 Oct 2010)*
426 - Disable SSL session ID cache in RPC client
432 *(Released Thu, 14 Oct 2010)*
434 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
435 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
437 - Fix problems with interaction between Curl and Python's HTTP server,
438 resulting in increased speed in many RPC calls
439 - Improve our release script to prevent breakage with older aclocal and
446 *(Released Thu, 7 Oct 2010)*
448 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
449 value retrieved from master
450 - Added support for blacklisted or hidden OS definitions
451 - Added simple lock monitor (accessible via (``gnt-debug locks``)
452 - Added support for -mem-path in KVM hypervisor abstraction layer
453 - Allow overriding instance parameters in tool for inter-cluster
454 instance moves (``tools/move-instance``)
455 - Improved opcode summaries (e.g. in ``gnt-job list``)
456 - Improve consistency of OS listing by sorting it
457 - Documentation updates
463 *(Released Fri, 8 Oct 2010)*
465 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
471 *(Released Mon, 4 Oct 2010)*
473 - Fixed regression in ``gnt-instance rename``
479 *(Released Wed, 22 Sep 2010)*
481 - Fixed OS_VARIANT variable for OS scripts
482 - Fixed cluster tag operations via RAPI
483 - Made ``setup-ssh`` exit with non-zero code if an error occurred
484 - Disabled RAPI CA checks in watcher
490 *(Released Mon, 23 Aug 2010)*
492 - Support DRBD versions of the format "a.b.c.d"
494 - Re-introduce support for usage from multiple threads in RAPI client
495 - Instance renames and modify via RAPI
496 - Work around race condition between processing and archival in job
498 - Mark opcodes following failed one as failed, too
499 - Job field ``lock_status`` was removed due to difficulties making it
500 work with the changed job queue in Ganeti 2.2; a better way to monitor
501 locks is expected for a later 2.2.x release
502 - Fixed dry-run behaviour with many commands
503 - Support ``ssh-agent`` again when adding nodes
504 - Many additional bugfixes
510 *(Released Fri, 30 Jul 2010)*
512 Important change: the internal RPC mechanism between Ganeti nodes has
513 changed from using a home-grown http library (based on the Python base
514 libraries) to use the PycURL library. This requires that PycURL is
515 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
516 against GnuTLS by default. cURL's support for GnuTLS had known issues
517 before cURL 7.21.0 and we recommend using the latest cURL release or
518 linking against OpenSSL. Most other distributions already link PycURL
519 and cURL against OpenSSL. The command::
521 python -c 'import pycurl; print pycurl.version'
523 can be used to determine the libraries PycURL and cURL are linked
526 Other significant changes:
528 - Rewrote much of the internals of the job queue, in order to achieve
529 better parallelism; this decouples job query operations from the job
530 processing, and it should allow much nicer behaviour of the master
531 daemon under load, and it also has uncovered some long-standing bugs
532 related to the job serialisation (now fixed)
533 - Added a default iallocator setting to the cluster parameters,
534 eliminating the need to always pass nodes or an iallocator for
535 operations that require selection of new node(s)
536 - Added experimental support for the LXC virtualization method
537 - Added support for OS parameters, which allows the installation of
538 instances to pass parameter to OS scripts in order to customise the
540 - Added a hypervisor parameter controlling the migration type (live or
541 non-live), since hypervisors have various levels of reliability; this
542 has renamed the 'live' parameter to 'mode'
543 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
544 logical volumes, meaning that cluster verify will ignore them and not
545 flag their presence as errors
546 - The watcher will now reset the error count for failed instances after
547 8 hours, thus allowing self-healing if the problem that caused the
548 instances to be down/fail to start has cleared in the meantime
549 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
550 check for, and warn, if the drbd module parameter ``usermode_helper``
551 is not consistent with the cluster-wide setting; this is needed to
552 make diagnose easier of failed drbd creations
553 - Started adding base IPv6 support, but this is not yet
554 enabled/available for use
555 - Rename operations (cluster, instance) will now return the new name,
556 which is especially useful if a short name was passed in
557 - Added support for instance migration in RAPI
558 - Added a tool to pre-configure nodes for the SSH setup, before joining
559 them to the cluster; this will allow in the future a simplified model
560 for node joining (but not yet fully enabled in 2.2); this needs the
561 paramiko python library
562 - Fixed handling of name-resolving errors
563 - Fixed consistency of job results on the error path
564 - Fixed master-failover race condition when executed multiple times in
566 - Fixed many bugs related to the job queue (mostly introduced during the
567 2.2 development cycle, so not all are impacting 2.1)
568 - Fixed instance migration with missing disk symlinks
569 - Fixed handling of unknown jobs in ``gnt-job archive``
570 - And many other small fixes/improvements
574 - Enhanced both the unittest and the QA coverage
575 - Switched the opcode validation to a generic model, and extended the
576 validation to all opcode parameters
577 - Changed more parts of the code that write shell scripts to use the
579 - Switched the master daemon to use the asyncore library for the Luxi
586 *(Released Thu, 17 Jun 2010)*
588 - Added tool (``move-instance``) and infrastructure to move instances
589 between separate clusters (see :doc:`separate documentation
590 <move-instance>` and :doc:`design document <design-2.2>`)
591 - Added per-request RPC timeout
592 - RAPI now requires a Content-Type header for requests with a body (e.g.
593 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
594 :rfc:`2616` (HTTP/1.1), section 7.2.1)
595 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
597 - Implemented initial support for running Ganeti daemons as separate
598 users, see configure-time flags ``--with-user-prefix`` and
599 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
601 - Instances can be removed after export (``gnt-backup export
603 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
604 key (instead of 1024 bit)
605 - Added new cluster configuration file for cluster domain secret
606 - Import/export now use SSL instead of SSH
607 - Added support for showing estimated time when exporting an instance,
608 see the ``ganeti-os-interface(7)`` manpage and look for
615 *(Released Tue, 16 Nov 2010)*
617 Some more bugfixes. Unless critical bugs occur, this will be the last
620 - Fix case of MAC special-values
621 - Fix mac checker regex
622 - backend: Fix typo causing "out of range" error
623 - Add missing --units in gnt-instance list man page
629 *(Released Tue, 24 Aug 2010)*
632 - Don't ignore secondary node silently on non-mirrored disk templates
634 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
635 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
636 - Properly document the usb_mouse parameter
637 - Fix path in ganeti-rapi(8) (issue 116)
638 - Adjust error message when the ganeti user's .ssh directory is
640 - Add same-node-check when changing the disk template to drbd
646 *(Released Fri, 16 Jul 2010)*
649 - Add an option to only select some reboot types during qa/burnin.
650 (on some hypervisors consequent reboots are not supported)
651 - Fix infrequent race condition in master failover. Sometimes the old
652 master ip address would be still detected as up for a short time
653 after it was removed, causing failover to fail.
654 - Decrease mlockall warnings when the ctypes module is missing. On
655 Python 2.4 we support running even if no ctypes module is installed,
656 but we were too verbose about this issue.
657 - Fix building on old distributions, on which man doesn't have a
659 - Fix RAPI not to ignore the MAC address on instance creation
660 - Implement the old instance creation format in the RAPI client.
666 *(Released Thu, 01 Jul 2010)*
668 A small bugfix release:
669 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
670 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
671 a non-existing option
672 - Fix instance create: the --force-variant option was ignored
673 - Improve pylint 0.21 compatibility and warnings with Python 2.6
674 - Fix modify node storage with non-FQDN arguments
675 - Fix RAPI client to authenticate under Python 2.6 when used
676 for more than 5 requests needing authentication
677 - Fix gnt-instance modify -t (storage) giving a wrong error message
678 when converting a non-shutdown drbd instance to plain
684 *(Released Fri, 18 Jun 2010)*
686 A small bugfix release:
688 - Fix live migration of KVM instances started with older Ganeti
689 versions which had fewer hypervisor parameters
690 - Fix gnt-instance grow-disk on down instances
691 - Fix an error-reporting bug during instance migration
692 - Better checking of the ``--net`` and ``--disk`` values, to avoid
693 silently ignoring broken ones
694 - Fix an RPC error reporting bug affecting, for example, RAPI client
696 - Fix bug triggered by different API version os-es on different nodes
697 - Fix a bug in instance startup with custom hvparams: OS level
698 parameters would fail to be applied.
699 - Fix the RAPI client under Python 2.6 (but more work is needed to
700 make it work completely well with OpenSSL)
701 - Fix handling of errors when resolving names from DNS
707 *(Released Thu, 3 Jun 2010)*
709 A medium sized development cycle. Some new features, and some
710 fixes/small improvements/cleanups.
715 The node deamon now tries to mlock itself into memory, unless the
716 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
717 its logs, and falls back to console logging. This allows emergency
718 features such as ``gnt-node powercycle`` to work even in the event of a
719 broken node disk (tested offlining the disk hosting the node's
720 filesystem and dropping its memory caches; don't try this at home)
722 KVM: add vhost-net acceleration support. It can be tested with a new
723 enough version of the kernel and of qemu-kvm.
725 KVM: Add instance chrooting feature. If you use privilege dropping for
726 your VMs you can also now force them to chroot to an empty directory,
727 before starting the emulated guest.
729 KVM: Add maximum migration bandwith and maximum downtime tweaking
730 support (requires a new-enough version of qemu-kvm).
732 Cluster verify will now warn if the master node doesn't have the master
735 Add a new (incompatible) instance creation request format to RAPI which
736 supports all parameters (previously only a subset was supported, and it
737 wasn't possible to extend the old format to accomodate all the new
738 features. The old format is still supported, and a client can check for
739 this feature, before using it, by checking for its presence in the
740 ``features`` RAPI resource.
742 Now with ancient latin support. Try it passing the ``--roman`` option to
743 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
744 (requires the python-roman module to be installed, in order to work).
749 As usual many internal code refactorings, documentation updates, and
752 - Lots of improvements and cleanups to the experimental Remote API
753 (RAPI) client library.
754 - A new unit test suite for the core daemon libraries.
755 - A fix to creating missing directories makes sure the umask is not
756 applied anymore. This enforces the same directory permissions
758 - Better handling terminating daemons with ctrl+c (used when running
759 them in debugging mode).
760 - Fix a race condition in live migrating a KVM instance, when stat()
761 on the old proc status file returned EINVAL, which is an unexpected
763 - Fixed manpage checking with newer man and utf-8 charachters. But now
764 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
770 *(Released Fri, 7 May 2010)*
772 Fix a bug which prevented untagged KVM instances from starting.
778 *(Released Fri, 7 May 2010)*
780 Another release with a long development cycle, during which many
781 different features were added.
786 The KVM hypervisor now can run the individual instances as non-root, to
787 reduce the impact of a VM being hijacked due to bugs in the
788 hypervisor. It is possible to run all instances as a single (non-root)
789 user, to manually specify a user for each instance, or to dynamically
790 allocate a user out of a cluster-wide pool to each instance, with the
791 guarantee that no two instances will run under the same user ID on any
794 An experimental RAPI client library, that can be used standalone
795 (without the other Ganeti libraries), is provided in the source tree as
796 ``lib/rapi/client.py``. Note this client might change its interface in
797 the future, as we iterate on its capabilities.
799 A new command, ``gnt-cluster renew-crypto`` has been added to easily
800 replace the cluster's certificates and crypto keys. This might help in
801 case they have been compromised, or have simply expired.
803 A new disk option for instance creation has been added that allows one
804 to "adopt" currently existing logical volumes, with data
805 preservation. This should allow easier migration to Ganeti from
806 unmanaged (or managed via other software) instances.
808 Another disk improvement is the possibility to convert between redundant
809 (DRBD) and plain (LVM) disk configuration for an instance. This should
810 allow better scalability (starting with one node and growing the
811 cluster, or shrinking a two-node cluster to one node).
813 A new feature that could help with automated node failovers has been
814 implemented: if a node sees itself as offline (by querying the master
815 candidates), it will try to shutdown (hard) all instances and any active
816 DRBD devices. This reduces the risk of duplicate instances if an
817 external script automatically failovers the instances on such nodes. To
818 enable this, the cluster parameter ``maintain_node_health`` should be
819 enabled; in the future this option (per the name) will enable other
820 automatic maintenance features.
822 Instance export/import now will reuse the original instance
823 specifications for all parameters; that means exporting an instance,
824 deleting it and the importing it back should give an almost identical
825 instance. Note that the default import behaviour has changed from
826 before, where it created only one NIC; now it recreates the original
829 Cluster verify has added a few new checks: SSL certificates validity,
830 /etc/hosts consistency across the cluster, etc.
835 As usual, many internal changes were done, documentation fixes,
838 - Fixed cluster initialization with disabled cluster storage (regression
840 - File-based storage supports growing the disks
841 - Fixed behaviour of node role changes
842 - Fixed cluster verify for some corner cases, plus a general rewrite of
843 cluster verify to allow future extension with more checks
844 - Fixed log spamming by watcher and node daemon (regression introduced
846 - Fixed possible validation issues when changing the list of enabled
848 - Fixed cleanup of /etc/hosts during node removal
849 - Fixed RAPI response for invalid methods
850 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
851 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
852 from ide disks, etc.)
853 - Allow OS changes without re-installation (to record a changed OS
854 outside of Ganeti, or to allow OS renames)
855 - Allow instance creation without OS installation (useful for example if
856 the OS will be installed manually, or restored from a backup not in
858 - Implemented option to make cluster ``copyfile`` use the replication
860 - Added list of enabled hypervisors to ssconf (possibly useful for
862 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
864 - A partial form of node re-IP is possible via node readd, which now
865 allows changed node primary IP
866 - Command line utilities now show an informational message if the job is
868 - The logs of the master daemon now show the PID/UID/GID of the
875 *(Released Fri, 12 Mar 2010)*
877 During the 2.1.0 long release candidate cycle, a lot of improvements and
878 changes have accumulated with were released later as 2.1.1.
883 The node evacuate command (``gnt-node evacuate``) was significantly
884 rewritten, and as such the IAllocator protocol was changed - a new
885 request type has been added. This unfortunate change during a stable
886 series is designed to improve performance of node evacuations; on
887 clusters with more than about five nodes and which are well-balanced,
888 evacuation should proceed in parallel for all instances of the node
889 being evacuated. As such, any existing IAllocator scripts need to be
890 updated, otherwise the above command will fail due to the unknown
891 request. The provided "dumb" allocator has not been updated; but the
892 ganeti-htools package supports the new protocol since version 0.2.4.
894 Another important change is increased validation of node and instance
895 names. This might create problems in special cases, if invalid host
896 names are being used.
898 Also, a new layer of hypervisor parameters has been added, that sits at
899 OS level between the cluster defaults and the instance ones. This allows
900 customisation of virtualization parameters depending on the installed
901 OS. For example instances with OS 'X' may have a different KVM kernel
902 (or any other parameter) than the cluster defaults. This is intended to
903 help managing a multiple OSes on the same cluster, without manual
904 modification of each instance's parameters.
906 A tool for merging clusters, ``cluster-merge``, has been added in the
912 - Improved the int/float conversions that should make the code more
913 robust in face of errors from the node daemons
914 - Fixed the remove node code in case of internal configuration errors
915 - Fixed the node daemon behaviour in face of inconsistent queue
916 directory (e.g. read-only file-system where we can't open the files
918 - Fixed the behaviour of gnt-node modify for master candidate demotion;
919 now it either aborts cleanly or, if given the new "auto_promote"
920 parameter, will automatically promote other nodes as needed
921 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
922 completely prevent Ganeti from working
923 - Fixed bug for instance export when not all disks were successfully
925 - Fixed behaviour of node add when the new node is slow in starting up
927 - Fixed handling of signals in the LUXI client, which should improve
928 behaviour of command-line scripts
929 - Added checks for invalid node/instance names in the configuration (now
930 flagged during cluster verify)
931 - Fixed watcher behaviour for disk activation errors
932 - Fixed two potentially endless loops in http library, which led to the
933 RAPI daemon hanging and consuming 100% CPU in some cases
934 - Fixed bug in RAPI daemon related to hashed passwords
935 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
937 - Enhanced compatibility with non-Debian OSes, but not using absolute
938 path in some commands and allowing customisation of the ssh
939 configuration directory
940 - Fixed possible future issue with new Python versions by abiding to the
941 proper use of ``__slots__`` attribute on classes
942 - Added checks that should prevent directory traversal attacks
943 - Many documentation fixes based on feedback from users
948 - Added an "early_release" more for instance replace disks and node
949 evacuate, where we release locks earlier and thus allow higher
950 parallelism within the cluster
951 - Added watcher hooks, intended to allow the watcher to restart other
952 daemons (e.g. from the ganeti-nbma project), but they can be used of
953 course for any other purpose
954 - Added a compile-time disable for DRBD barriers, to increase
955 performance if the administrator trusts the power supply or the
956 storage system to not lose writes
957 - Added the option of using syslog for logging instead of, or in
958 addition to, Ganeti's own log files
959 - Removed boot restriction for paravirtual NICs for KVM, recent versions
960 can indeed boot from a paravirtual NIC
961 - Added a generic debug level for many operations; while this is not
962 used widely yet, it allows one to pass the debug value all the way to
964 - Enhanced the hooks environment for instance moves (failovers,
965 migrations) where the primary/secondary nodes changed during the
966 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
967 - Enhanced data validations for many user-supplied values; one important
968 item is the restrictions imposed on instance and node names, which
969 might reject some (invalid) host names
970 - Add a configure-time option to disable file-based storage, if it's not
971 needed; this allows greater security separation between the master
972 node and the other nodes from the point of view of the inter-node RPC
974 - Added user notification in interactive tools if job is waiting in the
975 job queue or trying to acquire locks
976 - Added log messages when a job is waiting for locks
977 - Added filtering by node tags in instance operations which admit
978 multiple instances (start, stop, reboot, reinstall)
979 - Added a new tool for cluster mergers, ``cluster-merge``
980 - Parameters from command line which are of the form ``a=b,c=d`` can now
981 use backslash escapes to pass in values which contain commas,
982 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
984 - For KVM, the instance name is the first parameter passed to KVM, so
985 that it's more visible in the process list
991 *(Released Tue, 2 Mar 2010)*
993 Ganeti 2.1 brings many improvements with it. Major changes:
995 - Added infrastructure to ease automated disk repairs
996 - Added new daemon to export configuration data in a cheaper way than
998 - Instance NICs can now be routed instead of being associated with a
1000 - Improved job locking logic to reduce impact of jobs acquiring multiple
1001 locks waiting for other long-running jobs
1003 In-depth implementation details can be found in the Ganeti 2.1 design
1009 - Added chroot hypervisor
1010 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1012 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1013 ``bootloader_path`` and ``bootloader_args``)
1014 - Added the ``use_localtime`` option for the xen-hvm and kvm
1015 hypervisors, and the default value for this has changed to false (in
1016 2.0 xen-hvm always enabled it)
1017 - Added luxi call to submit multiple jobs in one go
1018 - Added cluster initialization option to not modify ``/etc/hosts``
1020 - Added network interface parameters
1021 - Added dry run mode to some LUs
1022 - Added RAPI resources:
1024 - ``/2/instances/[instance_name]/info``
1025 - ``/2/instances/[instance_name]/replace-disks``
1026 - ``/2/nodes/[node_name]/evacuate``
1027 - ``/2/nodes/[node_name]/migrate``
1028 - ``/2/nodes/[node_name]/role``
1029 - ``/2/nodes/[node_name]/storage``
1030 - ``/2/nodes/[node_name]/storage/modify``
1031 - ``/2/nodes/[node_name]/storage/repair``
1033 - Added OpCodes to evacuate or migrate all instances on a node
1034 - Added new command to list storage elements on nodes (``gnt-node
1035 list-storage``) and modify them (``gnt-node modify-storage``)
1036 - Added new ssconf files with master candidate IP address
1037 (``ssconf_master_candidates_ips``), node primary IP address
1038 (``ssconf_node_primary_ips``) and node secondary IP address
1039 (``ssconf_node_secondary_ips``)
1040 - Added ``ganeti-confd`` and a client library to query the Ganeti
1041 configuration via UDP
1042 - Added ability to run hooks after cluster initialization and before
1044 - Added automatic mode for disk replace (``gnt-instance replace-disks
1046 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1047 after catastrophic data-loss
1048 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1050 - Added ``gnt-instance move`` command to move instances
1051 - Added ``gnt-cluster watcher`` command to control watcher
1052 - Added ``gnt-node powercycle`` command to powercycle nodes
1053 - Added new job status field ``lock_status``
1054 - Added parseable error codes to cluster verification (``gnt-cluster
1055 verify --error-codes``) and made output less verbose (use
1056 ``--verbose`` to restore previous behaviour)
1057 - Added UUIDs to the main config entities (cluster, nodes, instances)
1058 - Added support for OS variants
1059 - Added support for hashed passwords in the Ganeti remote API users file
1061 - Added option to specify maximum timeout on instance shutdown
1062 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1063 - Added new helper script to start and stop Ganeti daemons
1064 (``daemon-util``), with the intent to reduce the work necessary to
1065 adjust Ganeti for non-Debian distributions and to start/stop daemons
1067 - Added more unittests
1068 - Fixed critical bug in ganeti-masterd startup
1069 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1070 now customisable at the cluster level for both the KVM and Xen
1071 hypervisors using the new ``migration_port`` parameter
1072 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1073 reinstalling an instance
1074 - Allowed ``@`` in tag names
1075 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1076 - Many documentation updates
1077 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1078 - ``gnt-instance reinstall`` can now reinstall multiple instances
1079 - Updated many command line parameters
1080 - Introduced new OS API version 15
1081 - No longer support a default hypervisor
1082 - Treat virtual LVs as inexistent
1083 - Improved job locking logic to reduce lock contention
1084 - Match instance and node names case insensitively
1085 - Reimplemented bash completion script to be more complete
1092 *(Released Thu, 4 Feb 2010)*
1094 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1095 - Fix a string formatting bug
1096 - Improve safety of the code in some error paths
1097 - Improve data validation in the master of values returned from nodes
1103 *(Released Thu, 17 Dec 2009)*
1105 - Fix security issue due to missing validation of iallocator names; this
1106 allows local and remote execution of arbitrary executables
1107 - Fix failure of gnt-node list during instance removal
1108 - Ship the RAPI documentation in the archive
1114 *(Released Wed, 30 Sep 2009)*
1116 - Fixed many wrong messages
1117 - Fixed a few bugs related to the locking library
1118 - Fixed MAC checking at instance creation time
1119 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1120 - Fixed a few issues related to signal handling in both daemons and
1122 - Fixed the example startup script provided
1123 - Fixed insserv dependencies in the example startup script (patch from
1125 - Fixed handling of drained nodes in the iallocator framework
1126 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1128 - Fixed error related to invalid job IDs in job polling
1129 - Fixed job/opcode persistence on unclean master shutdown
1130 - Fixed handling of partial job processing after unclean master
1132 - Fixed error reporting from LUs, previously all errors were converted
1133 into execution errors
1134 - Fixed error reporting from burnin
1135 - Decreased significantly the memory usage of the job queue
1136 - Optimised slightly multi-job submission
1137 - Optimised slightly opcode loading
1138 - Backported the multi-job submit framework from the development
1139 branch; multi-instance start and stop should be faster
1140 - Added script to clean archived jobs after 21 days; this will reduce
1141 the size of the queue directory
1142 - Added some extra checks in disk size tracking
1143 - Added an example ethers hook script
1144 - Added a cluster parameter that prevents Ganeti from modifying of
1146 - Added more node information to RAPI responses
1147 - Added a ``gnt-job watch`` command that allows following the ouput of a
1149 - Added a bind-address option to ganeti-rapi
1150 - Added more checks to the configuration verify
1151 - Enhanced the burnin script such that some operations can be retried
1153 - Converted instance reinstall to multi-instance model
1159 *(Released Fri, 7 Aug 2009)*
1161 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1162 to allow using the pre-2.0.2 behaviour in activation, if any existing
1163 instances have mismatched disk sizes in the configuration
1164 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
1165 any configuration mismatches for disk sizes
1166 - Added ``gnt-master cluste-failover --no-voting`` to allow master
1167 failover to work on two-node clusters
1168 - Fixed the ``--net`` option of ``gnt-backup import``, which was
1170 - Fixed detection of OS script errors in ``gnt-backup export``
1171 - Fixed exit code of ``gnt-backup export``
1177 *(Released Fri, 17 Jul 2009)*
1179 - Added experimental support for stripped logical volumes; this should
1180 enhance performance but comes with a higher complexity in the block
1181 device handling; stripping is only enabled when passing
1182 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
1183 affected even in the non-stripped mode
1184 - Improved resiliency against transient failures at the end of DRBD
1185 resyncs, and in general of DRBD resync checks
1186 - Fixed a couple of issues with exports and snapshot errors
1187 - Fixed a couple of issues in instance listing
1188 - Added display of the disk size in ``gnt-instance info``
1189 - Fixed checking for valid OSes in instance creation
1190 - Fixed handling of the "vcpus" parameter in instance listing and in
1191 general of invalid parameters
1192 - Fixed http server library, and thus RAPI, to handle invalid
1193 username/password combinations correctly; this means that now they
1194 report unauthorized for queries too, not only for modifications,
1195 allowing earlier detect of configuration problems
1196 - Added a new "role" node list field, equivalent to the master/master
1197 candidate/drained/offline flags combinations
1198 - Fixed cluster modify and changes of candidate pool size
1199 - Fixed cluster verify error messages for wrong files on regular nodes
1200 - Fixed a couple of issues with node demotion from master candidate role
1201 - Fixed node readd issues
1202 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
1203 - Added a new ``--no-voting`` option for masterfailover to fix failover
1204 on two-nodes clusters when the former master node is unreachable
1205 - Added instance reinstall over RAPI
1211 *(Released Tue, 16 Jun 2009)*
1213 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
1214 allow re-adding the start in single-user option (regression from 1.2)
1215 - the watcher writes the instance status to a file, to allow monitoring
1216 to report the instance status (from the master) based on cached
1217 results of the watcher's queries; while this can get stale if the
1218 watcher is being locked due to other work on the cluster, this is
1219 still an improvement
1220 - the watcher now also restarts the node daemon and the rapi daemon if
1222 - fixed the watcher to handle full and drained queue cases
1223 - hooks export more instance data in the environment, which helps if
1224 hook scripts need to take action based on the instance's properties
1225 (no longer need to query back into ganeti)
1226 - instance failovers when the instance is stopped do not check for free
1227 RAM, so that failing over a stopped instance is possible in low memory
1229 - rapi uses queries for tags instead of jobs (for less job traffic), and
1230 for cluster tags it won't talk to masterd at all but read them from
1232 - a couple of error handling fixes in RAPI
1233 - drbd handling: improved the error handling of inconsistent disks after
1234 resync to reduce the frequency of "there are some degraded disks for
1235 this instance" messages
1236 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
1237 error handling path called a wrong function name)
1243 *(Released Wed, 27 May 2009)*
1245 - no changes from rc5
1248 Version 2.0 release candidate 5
1249 -------------------------------
1251 *(Released Wed, 20 May 2009)*
1253 - fix a couple of bugs (validation, argument checks)
1254 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
1255 - some small improvements to RAPI and IAllocator
1256 - make watcher automatically start the master daemon if down
1259 Version 2.0 release candidate 4
1260 -------------------------------
1262 *(Released Mon, 27 Apr 2009)*
1264 - change the OS list to not require locks; this helps with big clusters
1265 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
1266 volume group is broken
1267 - ``gnt-instance info``, without any arguments, doesn't run for all
1268 instances anymore; either pass ``--all`` or pass the desired
1269 instances; this helps against mistakes on big clusters where listing
1270 the information for all instances takes a long time
1271 - miscellaneous doc and man pages fixes
1274 Version 2.0 release candidate 3
1275 -------------------------------
1277 *(Released Wed, 8 Apr 2009)*
1279 - Change the internal locking model of some ``gnt-node`` commands, in
1280 order to reduce contention (and blocking of master daemon) when
1281 batching many creation/reinstall jobs
1282 - Fixes to Xen soft reboot
1283 - No longer build documentation at build time, instead distribute it in
1284 the archive, in order to reduce the need for the whole docbook/rst
1288 Version 2.0 release candidate 2
1289 -------------------------------
1291 *(Released Fri, 27 Mar 2009)*
1293 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
1294 - Fix watcher startup sequence, improves the behaviour of busy clusters
1295 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
1296 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
1297 bind address and other places
1298 - Some documentation fixes and updates
1301 Version 2.0 release candidate 1
1302 -------------------------------
1304 *(Released Mon, 2 Mar 2009)*
1306 - More documentation updates, now all docs should be more-or-less
1308 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
1310 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
1312 - Fix an issue related to $libdir/run/ganeti and cluster creation
1318 *(Released Thu, 19 Feb 2009)*
1320 - Xen PVM and KVM have switched the default value for the instance root
1321 disk to the first partition on the first drive, instead of the whole
1322 drive; this means that the OS installation scripts must be changed
1324 - Man pages have been updated
1325 - RAPI has been switched by default to HTTPS, and the exported functions
1326 should all work correctly
1327 - RAPI v1 has been removed
1328 - Many improvements to the KVM hypervisor
1329 - Block device errors are now better reported
1330 - Many other bugfixes and small improvements
1336 *(Released Mon, 26 Jan 2009)*
1338 - Version 2 is a general rewrite of the code and therefore the
1339 differences are too many to list, see the design document for 2.0 in
1340 the ``doc/`` subdirectory for more details
1341 - In this beta version there is not yet a migration path from 1.2 (there
1342 will be one in the final 2.0 release)
1343 - A few significant changes are:
1345 - all commands are executed by a daemon (``ganeti-masterd``) and the
1346 various ``gnt-*`` commands are just front-ends to it
1347 - all the commands are entered into, and executed from a job queue,
1348 see the ``gnt-job(8)`` manpage
1349 - the RAPI daemon supports read-write operations, secured by basic
1350 HTTP authentication on top of HTTPS
1351 - DRBD version 0.7 support has been removed, DRBD 8 is the only
1352 supported version (when migrating from Ganeti 1.2 to 2.0, you need
1353 to migrate to DRBD 8 first while still running Ganeti 1.2)
1354 - DRBD devices are using statically allocated minor numbers, which
1355 will be assigned to existing instances during the migration process
1356 - there is support for both Xen PVM and Xen HVM instances running on
1358 - KVM virtualization is supported too
1359 - file-based storage has been implemented, which means that it is
1360 possible to run the cluster without LVM and DRBD storage, for
1361 example using a shared filesystem exported from shared storage (and
1362 still have live migration)
1368 *(Released Tue, 13 Jan 2009)*
1370 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
1371 - Reuse the old instance mac address by default on instance import, if
1372 the instance name is the same.
1373 - Handle situations in which the node info rpc returns incomplete
1375 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
1376 - Improved version of batcher:
1378 - state file support
1379 - instance mac address support
1380 - support for HVM clusters/instances
1382 - Add an option to show the number of cpu sockets and nodes in
1384 - Support OSes that handle more than one version of the OS api (but do
1385 not change the current API in any other way)
1386 - Fix ``gnt-node migrate``
1387 - ``gnt-debug`` man page
1388 - Fixes various more typos and small issues
1389 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
1395 *(Released Wed, 24 Sep 2008)*
1397 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
1398 type of disk exported to fully virtualized instances.
1399 - provide access to the serial console of HVM instances
1400 - instance auto_balance flag, set by default. If turned off it will
1401 avoid warnings on cluster verify if there is not enough memory to fail
1402 over an instance. in the future it will prevent automatically failing
1403 it over when we will support that.
1404 - batcher tool for instance creation, see ``tools/README.batcher``
1405 - ``gnt-instance reinstall --select-os`` to interactively select a new
1406 operating system when reinstalling an instance.
1407 - when changing the memory amount on instance modify a check has been
1408 added that the instance will be able to start. also warnings are
1409 emitted if the instance will not be able to fail over, if auto_balance
1411 - documentation fixes
1412 - sync fields between ``gnt-instance list/modify/add/import``
1413 - fix a race condition in drbd when the sync speed was set after giving
1414 the device a remote peer.
1420 *(Released Tue, 22 Jul 2008)*
1422 - note: the allowed size and number of tags per object were reduced
1423 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1424 - fixed twisted 8.x compatibility
1425 - fixed ``gnt-instance replace-disks`` with iallocator
1426 - add TCP keepalives on twisted connections to detect restarted nodes
1427 - disk increase support, see ``gnt-instance grow-disk``
1428 - implement bulk node/instance query for RAPI
1429 - add tags in node/instance listing (optional)
1430 - experimental migration (and live migration) support, read the man page
1431 for ``gnt-instance migrate``
1432 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1433 has some small improvements in handling its state file
1439 *(Released Fri, 13 Jun 2008)*
1441 - Experimental readonly, REST-based remote API implementation;
1442 automatically started on master node, TCP port 5080, if enabled by
1443 ``--enable-rapi`` parameter to configure script.
1444 - Instance allocator support. Add and import instance accept a
1445 ``--iallocator`` parameter, and call that instance allocator to decide
1446 which node to use for the instance. The iallocator document describes
1447 what's expected from an allocator script.
1448 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1449 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1450 node is lost there is still enough memory to fail over the instances
1452 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1453 ``gnt-cluster verify``, to check for site-specific compliance. All the
1454 hooks will run, and their output, if any, will be displayed. Any
1455 failing hook will make the verification return an error value.
1456 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1457 primary and secondary interfaces
1458 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1459 that is still declared as part of the cluster and has failed.
1460 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1461 specifying output fields, that just adds the chosen fields to the
1463 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1464 export from the filesystem.
1465 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1466 have been added. Using them you can enable/disable acpi and pae
1467 support, and specify a path for a cd image to be exported to the
1468 instance. These parameters as the name suggest only work on HVM
1470 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1471 PAE support will be set to the previously hardcoded values, but the
1472 (previously hardcoded) path to the CDROM ISO image will be unset and
1473 if required, needs to be set manually with ``gnt-instance modify``
1475 - The address to which an instance's VNC console is bound is now
1476 selectable per-instance, rather than being cluster wide. Of course
1477 this only applies to instances controlled via VNC, so currently just
1478 applies to HVM clusters.
1484 *(Released Mon, 18 Feb 2008)*
1486 - more tweaks to the disk activation code (especially helpful for DRBD)
1487 - change the default ``gnt-instance list`` output format, now there is
1488 one combined status field (see the manpage for the exact values this
1490 - some more fixes for the mac export to hooks change
1491 - make Ganeti not break with DRBD 8.2.x (which changed the version
1492 format in ``/proc/drbd``) (issue 24)
1493 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1494 template, allowing migration from DRBD0.7+MD to DRBD8
1500 *(Released Wed, 30 Jan 2008)*
1502 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1504 - add command aliases infrastructure and a few aliases
1505 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1506 man pages and the ``--help`` option of ``gnt-node
1507 list``/``gnt-instance list``
1508 - fix ``gnt-backup list`` with down nodes (issue 21)
1509 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1510 - fix the dist archive and add a check for including svn/git files in
1512 - some developer-related changes: improve the burnin and the QA suite,
1513 add an upload script for testing during development
1519 *(Released Wed, 16 Jan 2008)*
1521 - experimental HVM support, read the install document, section
1522 "Initializing the cluster"
1523 - allow for the PVM hypervisor per-instance kernel and initrd paths
1524 - add a new command ``gnt-cluster verify-disks`` which uses a new
1525 algorithm to improve the reconnection of the DRBD pairs if the device
1526 on the secondary node has gone away
1527 - make logical volume code auto-activate LVs at disk activation time
1528 - slightly improve the speed of activating disks
1529 - allow specification of the MAC address at instance creation time, and
1530 changing it later via ``gnt-instance modify``
1531 - fix handling of external commands that generate lots of output on
1533 - update documentation with regard to minimum version of DRBD8 supported
1539 *(Released Tue, 4 Dec 2007)*
1541 - Log the ``xm create`` output to the node daemon log on failure (to
1542 help diagnosing the error)
1543 - In debug mode, log all external commands output if failed to the logs
1544 - Change parsing of lvm commands to ignore stderr
1550 *(Released Wed, 28 Nov 2007)*
1552 - Another round of updates to the DRBD 8 code to deal with more failures
1553 in the replace secondary node operation
1554 - Some more logging of failures in disk operations (lvm, drbd)
1555 - A few documentation updates
1562 *(Released Tue, 13 Nov 2007)*
1564 - Change configuration file format from Python's Pickle to JSON.
1565 Upgrading is possible using the cfgupgrade utility.
1566 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1567 faster replace disks and is more stable (DRBD 8 has many improvements
1568 compared to DRBD 0.7)
1569 - Added command line tags support (see man pages for ``gnt-instance``,
1570 ``gnt-node``, ``gnt-cluster``)
1571 - Added instance rename support
1572 - Added multi-instance startup/shutdown
1573 - Added cluster rename support
1574 - Added ``gnt-node evacuate`` to simplify some node operations
1575 - Added instance reboot operation that can speedup reboot as compared to
1577 - Soften the requirement that hostnames are in FQDN format
1578 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1580 - Removed dependency on debian's patched fping that uses the
1581 non-standard ``-S`` option
1582 - Now the OS definitions are searched for in multiple, configurable
1583 paths (easier for distros to package)
1584 - Some changes to the hooks infrastructure (especially the new
1585 post-configuration update hook)
1586 - Other small bugfixes
1588 .. vim: set textwidth=72 syntax=rst :