9 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
10 resource's ``ip_check`` parameter changed from ``True`` to ``False``
11 to match the underlying LUXI interface
17 *(Released Fri, 14 Jan 2011)*
22 - Fixed timezone issues when formatting timestamps
23 - Added support for node groups, available via ``gnt-group`` and other
25 - Added out-of-band framework and management, see :doc:`design
26 document <design-oob>`
27 - Removed support for roman numbers from ``gnt-node list`` and
28 ``gnt-instance list``.
29 - Allowed modification of master network interface via ``gnt-cluster
30 modify --master-netdev``
31 - Accept offline secondaries while shutting down instance disks
32 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
33 - Added support for multiple LVM volume groups
34 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
36 - Added commands to list available fields:
38 - ``gnt-node list-fields``
39 - ``gnt-group list-fields``
40 - ``gnt-instance list-fields``
42 - Updated documentation and man pages
47 - Moved ``rapi_users`` file into separate directory, now named
48 ``…/ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
50 - Added new tool for running commands on many machines,
51 ``tools/ganeti-listrunner``
52 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
53 improving the ``gnt-instance console`` output
54 - Allowed customisation of disk index separator at ``configure`` time
55 - Export node group allocation policy to :doc:`iallocator <iallocator>`
56 - Added support for non-partitioned md disks in ``lvmstrap``
57 - Added script to gracefully power off KVM instances
58 - Split ``utils`` module into smaller parts
59 - Changed query operations to return more detailed information, e.g.
60 whether an information is unavailable due to an offline node. To use
61 this new functionality, the LUXI call ``Query`` must be used. Field
62 information is now stored by the master daemon and can be retrieved
63 using ``QueryFields``. Instances, nodes and groups can also be queried
64 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
65 exposed via RAPI). The following commands make use of this
66 infrastructure change:
69 - ``gnt-group list-fields``
71 - ``gnt-node list-fields``
72 - ``gnt-instance list``
73 - ``gnt-instance list-fields``
79 - New RAPI resources (see :doc:`rapi`):
83 - ``/2/groups/[group_name]``
84 - ``/2/groups/[group_name]/assign-nodes``
85 - ``/2/groups/[group_name]/modify``
86 - ``/2/groups/[group_name]/rename``
87 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
91 - Implemented ``no_install`` for instance creation
92 - Implemented OS parameters for instance reinstallation, allowing
93 use of special settings on reinstallation (e.g. for preserving data)
98 - Added IPv6 support in import/export
99 - Pause DRBD synchronization while wiping disks on instance creation
100 - Updated unittests and QA scripts
101 - Improved network parameters passed to KVM
102 - Converted man pages from docbook to reStructuredText
108 *(Released Mon, 20 Dec 2010)*
110 Released version 2.3.1~rc1 without any changes.
116 *(Released Wed, 1 Dec 2010)*
118 - impexpd: Disable OpenSSL compression in socat if possible (backport
119 from master, commit e90739d625b, see :doc:`installation guide
120 <install-quick>` for details)
121 - Changed unittest coverage report to exclude test scripts
122 - Added script to check version format
128 *(Released Wed, 1 Dec 2010)*
130 Released version 2.3.0~rc1 without any changes.
136 *(Released Fri, 19 Nov 2010)*
138 A number of bugfixes and documentation updates:
140 - Update ganeti-os-interface documentation
141 - Fixed a bug related to duplicate MACs or similar items which should be
143 - Fix breakage in OS state modify
144 - Reinstall instance: disallow offline secondaries (fixes bug related to
145 OS changing but reinstall failing)
146 - plus all the other fixes between 2.2.1 and 2.2.2
152 *(Released Tue, 2 Nov 2010)*
154 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
155 - Fixed master failover race with watcher
156 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
158 - Accept previously stopped instance for export with instance removal
159 - Simplify and extend the environment variables for instance OS scripts
160 - Added new node flags, ``master_capable`` and ``vm_capable``
161 - Added optional instance disk wiping prior during allocation. This is a
162 cluster-wide option and can be set/modified using
163 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
164 - Added IPv6 support, see :doc:`design document <design-2.3>` and
166 - Added a new watcher option (``--ignore-pause``)
167 - Added option to ignore offline node on instance start/stop
168 (``--ignore-offline``)
169 - Allow overriding OS parameters with ``gnt-instance reinstall``
170 - Added ability to change node's secondary IP address using ``gnt-node
172 - Implemented privilege separation for all daemons except
173 ``ganeti-noded``, see ``configure`` options
174 - Complain if an instance's disk is marked faulty in ``gnt-cluster
176 - Implemented job priorities (see ``ganeti(7)`` manpage)
177 - Ignore failures while shutting down instances during failover from
179 - Exit daemon's bootstrap process only once daemon is ready
180 - Export more information via ``LUInstanceQuery``/remote API
181 - Improved documentation, QA and unittests
182 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
183 restart if the file was created or changed
184 - Added LUXI protocol version sent with each request and response,
185 allowing detection of server/client mismatches
186 - Moved the Python scripts among gnt-* and ganeti-* into modules
187 - Moved all code related to setting up SSH to an external script,
189 - Infrastructure changes for node group support in future versions
195 *(Released Fri, 19 Nov 2010)*
197 A few small bugs fixed, and some improvements to the build system:
199 - Fix documentation regarding conversion to drbd
200 - Fix validation of parameters in cluster modify (``gnt-cluster modify
202 - Fix error handling in node modify with multiple changes
203 - Allow remote imports without checked names
209 *(Released Tue, 19 Oct 2010)*
211 - Disable SSL session ID cache in RPC client
217 *(Released Thu, 14 Oct 2010)*
219 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
220 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
222 - Fix problems with interaction between Curl and Python's HTTP server,
223 resulting in increased speed in many RPC calls
224 - Improve our release script to prevent breakage with older aclocal and
231 *(Released Thu, 7 Oct 2010)*
233 - Fixed issue 125, replace hardcoded “xenvg” in ``gnt-cluster`` with
234 value retrieved from master
235 - Added support for blacklisted or hidden OS definitions
236 - Added simple lock monitor (accessible via (``gnt-debug locks``)
237 - Added support for -mem-path in KVM hypervisor abstraction layer
238 - Allow overriding instance parameters in tool for inter-cluster
239 instance moves (``tools/move-instance``)
240 - Improved opcode summaries (e.g. in ``gnt-job list``)
241 - Improve consistency of OS listing by sorting it
242 - Documentation updates
248 *(Released Fri, 8 Oct 2010)*
250 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
256 *(Released Mon, 4 Oct 2010)*
258 - Fixed regression in ``gnt-instance rename``
264 *(Released Wed, 22 Sep 2010)*
266 - Fixed OS_VARIANT variable for OS scripts
267 - Fixed cluster tag operations via RAPI
268 - Made ``setup-ssh`` exit with non-zero code if an error occurred
269 - Disabled RAPI CA checks in watcher
275 *(Released Mon, 23 Aug 2010)*
277 - Support DRBD versions of the format "a.b.c.d"
279 - Re-introduce support for usage from multiple threads in RAPI client
280 - Instance renames and modify via RAPI
281 - Work around race condition between processing and archival in job
283 - Mark opcodes following failed one as failed, too
284 - Job field ``lock_status`` was removed due to difficulties making it
285 work with the changed job queue in Ganeti 2.2; a better way to monitor
286 locks is expected for a later 2.2.x release
287 - Fixed dry-run behaviour with many commands
288 - Support ``ssh-agent`` again when adding nodes
289 - Many additional bugfixes
295 *(Released Fri, 30 Jul 2010)*
297 Important change: the internal RPC mechanism between Ganeti nodes has
298 changed from using a home-grown http library (based on the Python base
299 libraries) to use the PycURL library. This requires that PycURL is
300 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
301 against GnuTLS by default. cURL's support for GnuTLS had known issues
302 before cURL 7.21.0 and we recommend using the latest cURL release or
303 linking against OpenSSL. Most other distributions already link PycURL
304 and cURL against OpenSSL. The command::
306 python -c 'import pycurl; print pycurl.version'
308 can be used to determine the libraries PycURL and cURL are linked
311 Other significant changes:
313 - Rewrote much of the internals of the job queue, in order to achieve
314 better parallelism; this decouples job query operations from the job
315 processing, and it should allow much nicer behaviour of the master
316 daemon under load, and it also has uncovered some long-standing bugs
317 related to the job serialisation (now fixed)
318 - Added a default iallocator setting to the cluster parameters,
319 eliminating the need to always pass nodes or an iallocator for
320 operations that require selection of new node(s)
321 - Added experimental support for the LXC virtualization method
322 - Added support for OS parameters, which allows the installation of
323 instances to pass parameter to OS scripts in order to customise the
325 - Added a hypervisor parameter controlling the migration type (live or
326 non-live), since hypervisors have various levels of reliability; this
327 has renamed the 'live' parameter to 'mode'
328 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
329 logical volumes, meaning that cluster verify will ignore them and not
330 flag their presence as errors
331 - The watcher will now reset the error count for failed instances after
332 8 hours, thus allowing self-healing if the problem that caused the
333 instances to be down/fail to start has cleared in the meantime
334 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
335 check for, and warn, if the drbd module parameter ``usermode_helper``
336 is not consistent with the cluster-wide setting; this is needed to
337 make diagnose easier of failed drbd creations
338 - Started adding base IPv6 support, but this is not yet
339 enabled/available for use
340 - Rename operations (cluster, instance) will now return the new name,
341 which is especially useful if a short name was passed in
342 - Added support for instance migration in RAPI
343 - Added a tool to pre-configure nodes for the SSH setup, before joining
344 them to the cluster; this will allow in the future a simplified model
345 for node joining (but not yet fully enabled in 2.2); this needs the
346 paramiko python library
347 - Fixed handling of name-resolving errors
348 - Fixed consistency of job results on the error path
349 - Fixed master-failover race condition when executed multiple times in
351 - Fixed many bugs related to the job queue (mostly introduced during the
352 2.2 development cycle, so not all are impacting 2.1)
353 - Fixed instance migration with missing disk symlinks
354 - Fixed handling of unknown jobs in ``gnt-job archive``
355 - And many other small fixes/improvements
359 - Enhanced both the unittest and the QA coverage
360 - Switched the opcode validation to a generic model, and extended the
361 validation to all opcode parameters
362 - Changed more parts of the code that write shell scripts to use the
364 - Switched the master daemon to use the asyncore library for the Luxi
371 *(Released Thu, 17 Jun 2010)*
373 - Added tool (``move-instance``) and infrastructure to move instances
374 between separate clusters (see :doc:`separate documentation
375 <move-instance>` and :doc:`design document <design-2.2>`)
376 - Added per-request RPC timeout
377 - RAPI now requires a Content-Type header for requests with a body (e.g.
378 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
379 :rfc:`2616` (HTTP/1.1), section 7.2.1)
380 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
382 - Implemented initial support for running Ganeti daemons as separate
383 users, see configure-time flags ``--with-user-prefix`` and
384 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
386 - Instances can be removed after export (``gnt-backup export
388 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
389 key (instead of 1024 bit)
390 - Added new cluster configuration file for cluster domain secret
391 - Import/export now use SSL instead of SSH
392 - Added support for showing estimated time when exporting an instance,
393 see the ``ganeti-os-interface(7)`` manpage and look for
400 *(Released Tue, 16 Nov 2010)*
402 Some more bugfixes. Unless critical bugs occur, this will be the last
405 - Fix case of MAC special-values
406 - Fix mac checker regex
407 - backend: Fix typo causing “out of range” error
408 - Add missing --units in gnt-instance list man page
414 *(Released Tue, 24 Aug 2010)*
417 - Don't ignore secondary node silently on non-mirrored disk templates
419 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
420 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
421 - Properly document the usb_mouse parameter
422 - Fix path in ganeti-rapi(8) (issue 116)
423 - Adjust error message when the ganeti user's .ssh directory is
425 - Add same-node-check when changing the disk template to drbd
431 *(Released Fri, 16 Jul 2010)*
434 - Add an option to only select some reboot types during qa/burnin.
435 (on some hypervisors consequent reboots are not supported)
436 - Fix infrequent race condition in master failover. Sometimes the old
437 master ip address would be still detected as up for a short time
438 after it was removed, causing failover to fail.
439 - Decrease mlockall warnings when the ctypes module is missing. On
440 Python 2.4 we support running even if no ctypes module is installed,
441 but we were too verbose about this issue.
442 - Fix building on old distributions, on which man doesn't have a
444 - Fix RAPI not to ignore the MAC address on instance creation
445 - Implement the old instance creation format in the RAPI client.
451 *(Released Thu, 01 Jul 2010)*
453 A small bugfix release:
454 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
455 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
456 a non-existing option
457 - Fix instance create: the --force-variant option was ignored
458 - Improve pylint 0.21 compatibility and warnings with Python 2.6
459 - Fix modify node storage with non-FQDN arguments
460 - Fix RAPI client to authenticate under Python 2.6 when used
461 for more than 5 requests needing authentication
462 - Fix gnt-instance modify -t (storage) giving a wrong error message
463 when converting a non-shutdown drbd instance to plain
469 *(Released Fri, 18 Jun 2010)*
471 A small bugfix release:
473 - Fix live migration of KVM instances started with older Ganeti
474 versions which had fewer hypervisor parameters
475 - Fix gnt-instance grow-disk on down instances
476 - Fix an error-reporting bug during instance migration
477 - Better checking of the ``--net`` and ``--disk`` values, to avoid
478 silently ignoring broken ones
479 - Fix an RPC error reporting bug affecting, for example, RAPI client
481 - Fix bug triggered by different API version os-es on different nodes
482 - Fix a bug in instance startup with custom hvparams: OS level
483 parameters would fail to be applied.
484 - Fix the RAPI client under Python 2.6 (but more work is needed to
485 make it work completely well with OpenSSL)
486 - Fix handling of errors when resolving names from DNS
492 *(Released Thu, 3 Jun 2010)*
494 A medium sized development cycle. Some new features, and some
495 fixes/small improvements/cleanups.
500 The node deamon now tries to mlock itself into memory, unless the
501 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
502 its logs, and falls back to console logging. This allows emergency
503 features such as ``gnt-node powercycle`` to work even in the event of a
504 broken node disk (tested offlining the disk hosting the node's
505 filesystem and dropping its memory caches; don't try this at home)
507 KVM: add vhost-net acceleration support. It can be tested with a new
508 enough version of the kernel and of qemu-kvm.
510 KVM: Add instance chrooting feature. If you use privilege dropping for
511 your VMs you can also now force them to chroot to an empty directory,
512 before starting the emulated guest.
514 KVM: Add maximum migration bandwith and maximum downtime tweaking
515 support (requires a new-enough version of qemu-kvm).
517 Cluster verify will now warn if the master node doesn't have the master
520 Add a new (incompatible) instance creation request format to RAPI which
521 supports all parameters (previously only a subset was supported, and it
522 wasn't possible to extend the old format to accomodate all the new
523 features. The old format is still supported, and a client can check for
524 this feature, before using it, by checking for its presence in the
525 ``features`` RAPI resource.
527 Now with ancient latin support. Try it passing the ``--roman`` option to
528 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
529 (requires the python-roman module to be installed, in order to work).
534 As usual many internal code refactorings, documentation updates, and
537 - Lots of improvements and cleanups to the experimental Remote API
538 (RAPI) client library.
539 - A new unit test suite for the core daemon libraries.
540 - A fix to creating missing directories makes sure the umask is not
541 applied anymore. This enforces the same directory permissions
543 - Better handling terminating daemons with ctrl+c (used when running
544 them in debugging mode).
545 - Fix a race condition in live migrating a KVM instance, when stat()
546 on the old proc status file returned EINVAL, which is an unexpected
548 - Fixed manpage checking with newer man and utf-8 charachters. But now
549 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
555 *(Released Fri, 7 May 2010)*
557 Fix a bug which prevented untagged KVM instances from starting.
563 *(Released Fri, 7 May 2010)*
565 Another release with a long development cycle, during which many
566 different features were added.
571 The KVM hypervisor now can run the individual instances as non-root, to
572 reduce the impact of a VM being hijacked due to bugs in the
573 hypervisor. It is possible to run all instances as a single (non-root)
574 user, to manually specify a user for each instance, or to dynamically
575 allocate a user out of a cluster-wide pool to each instance, with the
576 guarantee that no two instances will run under the same user ID on any
579 An experimental RAPI client library, that can be used standalone
580 (without the other Ganeti libraries), is provided in the source tree as
581 ``lib/rapi/client.py``. Note this client might change its interface in
582 the future, as we iterate on its capabilities.
584 A new command, ``gnt-cluster renew-crypto`` has been added to easily
585 replace the cluster's certificates and crypto keys. This might help in
586 case they have been compromised, or have simply expired.
588 A new disk option for instance creation has been added that allows one
589 to "adopt" currently existing logical volumes, with data
590 preservation. This should allow easier migration to Ganeti from
591 unmanaged (or managed via other software) instances.
593 Another disk improvement is the possibility to convert between redundant
594 (DRBD) and plain (LVM) disk configuration for an instance. This should
595 allow better scalability (starting with one node and growing the
596 cluster, or shrinking a two-node cluster to one node).
598 A new feature that could help with automated node failovers has been
599 implemented: if a node sees itself as offline (by querying the master
600 candidates), it will try to shutdown (hard) all instances and any active
601 DRBD devices. This reduces the risk of duplicate instances if an
602 external script automatically failovers the instances on such nodes. To
603 enable this, the cluster parameter ``maintain_node_health`` should be
604 enabled; in the future this option (per the name) will enable other
605 automatic maintenance features.
607 Instance export/import now will reuse the original instance
608 specifications for all parameters; that means exporting an instance,
609 deleting it and the importing it back should give an almost identical
610 instance. Note that the default import behaviour has changed from
611 before, where it created only one NIC; now it recreates the original
614 Cluster verify has added a few new checks: SSL certificates validity,
615 /etc/hosts consistency across the cluster, etc.
620 As usual, many internal changes were done, documentation fixes,
623 - Fixed cluster initialization with disabled cluster storage (regression
625 - File-based storage supports growing the disks
626 - Fixed behaviour of node role changes
627 - Fixed cluster verify for some corner cases, plus a general rewrite of
628 cluster verify to allow future extension with more checks
629 - Fixed log spamming by watcher and node daemon (regression introduced
631 - Fixed possible validation issues when changing the list of enabled
633 - Fixed cleanup of /etc/hosts during node removal
634 - Fixed RAPI response for invalid methods
635 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
636 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
637 from ide disks, etc.)
638 - Allow OS changes without re-installation (to record a changed OS
639 outside of Ganeti, or to allow OS renames)
640 - Allow instance creation without OS installation (useful for example if
641 the OS will be installed manually, or restored from a backup not in
643 - Implemented option to make cluster ``copyfile`` use the replication
645 - Added list of enabled hypervisors to ssconf (possibly useful for
647 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
649 - A partial form of node re-IP is possible via node readd, which now
650 allows changed node primary IP
651 - Command line utilities now show an informational message if the job is
653 - The logs of the master daemon now show the PID/UID/GID of the
660 *(Released Fri, 12 Mar 2010)*
662 During the 2.1.0 long release candidate cycle, a lot of improvements and
663 changes have accumulated with were released later as 2.1.1.
668 The node evacuate command (``gnt-node evacuate``) was significantly
669 rewritten, and as such the IAllocator protocol was changed - a new
670 request type has been added. This unfortunate change during a stable
671 series is designed to improve performance of node evacuations; on
672 clusters with more than about five nodes and which are well-balanced,
673 evacuation should proceed in parallel for all instances of the node
674 being evacuated. As such, any existing IAllocator scripts need to be
675 updated, otherwise the above command will fail due to the unknown
676 request. The provided "dumb" allocator has not been updated; but the
677 ganeti-htools package supports the new protocol since version 0.2.4.
679 Another important change is increased validation of node and instance
680 names. This might create problems in special cases, if invalid host
681 names are being used.
683 Also, a new layer of hypervisor parameters has been added, that sits at
684 OS level between the cluster defaults and the instance ones. This allows
685 customisation of virtualization parameters depending on the installed
686 OS. For example instances with OS 'X' may have a different KVM kernel
687 (or any other parameter) than the cluster defaults. This is intended to
688 help managing a multiple OSes on the same cluster, without manual
689 modification of each instance's parameters.
691 A tool for merging clusters, ``cluster-merge``, has been added in the
697 - Improved the int/float conversions that should make the code more
698 robust in face of errors from the node daemons
699 - Fixed the remove node code in case of internal configuration errors
700 - Fixed the node daemon behaviour in face of inconsistent queue
701 directory (e.g. read-only file-system where we can't open the files
703 - Fixed the behaviour of gnt-node modify for master candidate demotion;
704 now it either aborts cleanly or, if given the new “auto_promote”
705 parameter, will automatically promote other nodes as needed
706 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
707 completely prevent Ganeti from working
708 - Fixed bug for instance export when not all disks were successfully
710 - Fixed behaviour of node add when the new node is slow in starting up
712 - Fixed handling of signals in the LUXI client, which should improve
713 behaviour of command-line scripts
714 - Added checks for invalid node/instance names in the configuration (now
715 flagged during cluster verify)
716 - Fixed watcher behaviour for disk activation errors
717 - Fixed two potentially endless loops in http library, which led to the
718 RAPI daemon hanging and consuming 100% CPU in some cases
719 - Fixed bug in RAPI daemon related to hashed passwords
720 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
722 - Enhanced compatibility with non-Debian OSes, but not using absolute
723 path in some commands and allowing customisation of the ssh
724 configuration directory
725 - Fixed possible future issue with new Python versions by abiding to the
726 proper use of ``__slots__`` attribute on classes
727 - Added checks that should prevent directory traversal attacks
728 - Many documentation fixes based on feedback from users
733 - Added an “early_release” more for instance replace disks and node
734 evacuate, where we release locks earlier and thus allow higher
735 parallelism within the cluster
736 - Added watcher hooks, intended to allow the watcher to restart other
737 daemons (e.g. from the ganeti-nbma project), but they can be used of
738 course for any other purpose
739 - Added a compile-time disable for DRBD barriers, to increase
740 performance if the administrator trusts the power supply or the
741 storage system to not lose writes
742 - Added the option of using syslog for logging instead of, or in
743 addition to, Ganeti's own log files
744 - Removed boot restriction for paravirtual NICs for KVM, recent versions
745 can indeed boot from a paravirtual NIC
746 - Added a generic debug level for many operations; while this is not
747 used widely yet, it allows one to pass the debug value all the way to
749 - Enhanced the hooks environment for instance moves (failovers,
750 migrations) where the primary/secondary nodes changed during the
751 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
752 - Enhanced data validations for many user-supplied values; one important
753 item is the restrictions imposed on instance and node names, which
754 might reject some (invalid) host names
755 - Add a configure-time option to disable file-based storage, if it's not
756 needed; this allows greater security separation between the master
757 node and the other nodes from the point of view of the inter-node RPC
759 - Added user notification in interactive tools if job is waiting in the
760 job queue or trying to acquire locks
761 - Added log messages when a job is waiting for locks
762 - Added filtering by node tags in instance operations which admit
763 multiple instances (start, stop, reboot, reinstall)
764 - Added a new tool for cluster mergers, ``cluster-merge``
765 - Parameters from command line which are of the form ``a=b,c=d`` can now
766 use backslash escapes to pass in values which contain commas,
767 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
769 - For KVM, the instance name is the first parameter passed to KVM, so
770 that it's more visible in the process list
776 *(Released Tue, 2 Mar 2010)*
778 Ganeti 2.1 brings many improvements with it. Major changes:
780 - Added infrastructure to ease automated disk repairs
781 - Added new daemon to export configuration data in a cheaper way than
783 - Instance NICs can now be routed instead of being associated with a
785 - Improved job locking logic to reduce impact of jobs acquiring multiple
786 locks waiting for other long-running jobs
788 In-depth implementation details can be found in the Ganeti 2.1 design
794 - Added chroot hypervisor
795 - Added more options to xen-hvm hypervisor (``kernel_path`` and
797 - Added more options to xen-pvm hypervisor (``use_bootloader``,
798 ``bootloader_path`` and ``bootloader_args``)
799 - Added the ``use_localtime`` option for the xen-hvm and kvm
800 hypervisors, and the default value for this has changed to false (in
801 2.0 xen-hvm always enabled it)
802 - Added luxi call to submit multiple jobs in one go
803 - Added cluster initialization option to not modify ``/etc/hosts``
805 - Added network interface parameters
806 - Added dry run mode to some LUs
807 - Added RAPI resources:
809 - ``/2/instances/[instance_name]/info``
810 - ``/2/instances/[instance_name]/replace-disks``
811 - ``/2/nodes/[node_name]/evacuate``
812 - ``/2/nodes/[node_name]/migrate``
813 - ``/2/nodes/[node_name]/role``
814 - ``/2/nodes/[node_name]/storage``
815 - ``/2/nodes/[node_name]/storage/modify``
816 - ``/2/nodes/[node_name]/storage/repair``
818 - Added OpCodes to evacuate or migrate all instances on a node
819 - Added new command to list storage elements on nodes (``gnt-node
820 list-storage``) and modify them (``gnt-node modify-storage``)
821 - Added new ssconf files with master candidate IP address
822 (``ssconf_master_candidates_ips``), node primary IP address
823 (``ssconf_node_primary_ips``) and node secondary IP address
824 (``ssconf_node_secondary_ips``)
825 - Added ``ganeti-confd`` and a client library to query the Ganeti
826 configuration via UDP
827 - Added ability to run hooks after cluster initialization and before
829 - Added automatic mode for disk replace (``gnt-instance replace-disks
831 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
832 after catastrophic data-loss
833 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
835 - Added ``gnt-instance move`` command to move instances
836 - Added ``gnt-cluster watcher`` command to control watcher
837 - Added ``gnt-node powercycle`` command to powercycle nodes
838 - Added new job status field ``lock_status``
839 - Added parseable error codes to cluster verification (``gnt-cluster
840 verify --error-codes``) and made output less verbose (use
841 ``--verbose`` to restore previous behaviour)
842 - Added UUIDs to the main config entities (cluster, nodes, instances)
843 - Added support for OS variants
844 - Added support for hashed passwords in the Ganeti remote API users file
846 - Added option to specify maximum timeout on instance shutdown
847 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
848 - Added new helper script to start and stop Ganeti daemons
849 (``daemon-util``), with the intent to reduce the work necessary to
850 adjust Ganeti for non-Debian distributions and to start/stop daemons
852 - Added more unittests
853 - Fixed critical bug in ganeti-masterd startup
854 - Removed the configure-time ``kvm-migration-port`` parameter, this is
855 now customisable at the cluster level for both the KVM and Xen
856 hypervisors using the new ``migration_port`` parameter
857 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
858 reinstalling an instance
859 - Allowed ``@`` in tag names
860 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
861 - Many documentation updates
862 - Distribute hypervisor files on ``gnt-cluster redist-conf``
863 - ``gnt-instance reinstall`` can now reinstall multiple instances
864 - Updated many command line parameters
865 - Introduced new OS API version 15
866 - No longer support a default hypervisor
867 - Treat virtual LVs as inexistent
868 - Improved job locking logic to reduce lock contention
869 - Match instance and node names case insensitively
870 - Reimplemented bash completion script to be more complete
877 *(Released Thu, 4 Feb 2010)*
879 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
880 - Fix a string formatting bug
881 - Improve safety of the code in some error paths
882 - Improve data validation in the master of values returned from nodes
888 *(Released Thu, 17 Dec 2009)*
890 - Fix security issue due to missing validation of iallocator names; this
891 allows local and remote execution of arbitrary executables
892 - Fix failure of gnt-node list during instance removal
893 - Ship the RAPI documentation in the archive
899 *(Released Wed, 30 Sep 2009)*
901 - Fixed many wrong messages
902 - Fixed a few bugs related to the locking library
903 - Fixed MAC checking at instance creation time
904 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
905 - Fixed a few issues related to signal handling in both daemons and
907 - Fixed the example startup script provided
908 - Fixed insserv dependencies in the example startup script (patch from
910 - Fixed handling of drained nodes in the iallocator framework
911 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
913 - Fixed error related to invalid job IDs in job polling
914 - Fixed job/opcode persistence on unclean master shutdown
915 - Fixed handling of partial job processing after unclean master
917 - Fixed error reporting from LUs, previously all errors were converted
918 into execution errors
919 - Fixed error reporting from burnin
920 - Decreased significantly the memory usage of the job queue
921 - Optimised slightly multi-job submission
922 - Optimised slightly opcode loading
923 - Backported the multi-job submit framework from the development
924 branch; multi-instance start and stop should be faster
925 - Added script to clean archived jobs after 21 days; this will reduce
926 the size of the queue directory
927 - Added some extra checks in disk size tracking
928 - Added an example ethers hook script
929 - Added a cluster parameter that prevents Ganeti from modifying of
931 - Added more node information to RAPI responses
932 - Added a ``gnt-job watch`` command that allows following the ouput of a
934 - Added a bind-address option to ganeti-rapi
935 - Added more checks to the configuration verify
936 - Enhanced the burnin script such that some operations can be retried
938 - Converted instance reinstall to multi-instance model
944 *(Released Fri, 7 Aug 2009)*
946 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
947 to allow using the pre-2.0.2 behaviour in activation, if any existing
948 instances have mismatched disk sizes in the configuration
949 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
950 any configuration mismatches for disk sizes
951 - Added ``gnt-master cluste-failover --no-voting`` to allow master
952 failover to work on two-node clusters
953 - Fixed the ``--net`` option of ``gnt-backup import``, which was
955 - Fixed detection of OS script errors in ``gnt-backup export``
956 - Fixed exit code of ``gnt-backup export``
962 *(Released Fri, 17 Jul 2009)*
964 - Added experimental support for stripped logical volumes; this should
965 enhance performance but comes with a higher complexity in the block
966 device handling; stripping is only enabled when passing
967 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
968 affected even in the non-stripped mode
969 - Improved resiliency against transient failures at the end of DRBD
970 resyncs, and in general of DRBD resync checks
971 - Fixed a couple of issues with exports and snapshot errors
972 - Fixed a couple of issues in instance listing
973 - Added display of the disk size in ``gnt-instance info``
974 - Fixed checking for valid OSes in instance creation
975 - Fixed handling of the "vcpus" parameter in instance listing and in
976 general of invalid parameters
977 - Fixed http server library, and thus RAPI, to handle invalid
978 username/password combinations correctly; this means that now they
979 report unauthorized for queries too, not only for modifications,
980 allowing earlier detect of configuration problems
981 - Added a new "role" node list field, equivalent to the master/master
982 candidate/drained/offline flags combinations
983 - Fixed cluster modify and changes of candidate pool size
984 - Fixed cluster verify error messages for wrong files on regular nodes
985 - Fixed a couple of issues with node demotion from master candidate role
986 - Fixed node readd issues
987 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
988 - Added a new ``--no-voting`` option for masterfailover to fix failover
989 on two-nodes clusters when the former master node is unreachable
990 - Added instance reinstall over RAPI
996 *(Released Tue, 16 Jun 2009)*
998 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
999 allow re-adding the start in single-user option (regression from 1.2)
1000 - the watcher writes the instance status to a file, to allow monitoring
1001 to report the instance status (from the master) based on cached
1002 results of the watcher's queries; while this can get stale if the
1003 watcher is being locked due to other work on the cluster, this is
1004 still an improvement
1005 - the watcher now also restarts the node daemon and the rapi daemon if
1007 - fixed the watcher to handle full and drained queue cases
1008 - hooks export more instance data in the environment, which helps if
1009 hook scripts need to take action based on the instance's properties
1010 (no longer need to query back into ganeti)
1011 - instance failovers when the instance is stopped do not check for free
1012 RAM, so that failing over a stopped instance is possible in low memory
1014 - rapi uses queries for tags instead of jobs (for less job traffic), and
1015 for cluster tags it won't talk to masterd at all but read them from
1017 - a couple of error handling fixes in RAPI
1018 - drbd handling: improved the error handling of inconsistent disks after
1019 resync to reduce the frequency of "there are some degraded disks for
1020 this instance" messages
1021 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
1022 error handling path called a wrong function name)
1028 *(Released Wed, 27 May 2009)*
1030 - no changes from rc5
1033 Version 2.0 release candidate 5
1034 -------------------------------
1036 *(Released Wed, 20 May 2009)*
1038 - fix a couple of bugs (validation, argument checks)
1039 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
1040 - some small improvements to RAPI and IAllocator
1041 - make watcher automatically start the master daemon if down
1044 Version 2.0 release candidate 4
1045 -------------------------------
1047 *(Released Mon, 27 Apr 2009)*
1049 - change the OS list to not require locks; this helps with big clusters
1050 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
1051 volume group is broken
1052 - ``gnt-instance info``, without any arguments, doesn't run for all
1053 instances anymore; either pass ``--all`` or pass the desired
1054 instances; this helps against mistakes on big clusters where listing
1055 the information for all instances takes a long time
1056 - miscellaneous doc and man pages fixes
1059 Version 2.0 release candidate 3
1060 -------------------------------
1062 *(Released Wed, 8 Apr 2009)*
1064 - Change the internal locking model of some ``gnt-node`` commands, in
1065 order to reduce contention (and blocking of master daemon) when
1066 batching many creation/reinstall jobs
1067 - Fixes to Xen soft reboot
1068 - No longer build documentation at build time, instead distribute it in
1069 the archive, in order to reduce the need for the whole docbook/rst
1073 Version 2.0 release candidate 2
1074 -------------------------------
1076 *(Released Fri, 27 Mar 2009)*
1078 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
1079 - Fix watcher startup sequence, improves the behaviour of busy clusters
1080 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
1081 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
1082 bind address and other places
1083 - Some documentation fixes and updates
1086 Version 2.0 release candidate 1
1087 -------------------------------
1089 *(Released Mon, 2 Mar 2009)*
1091 - More documentation updates, now all docs should be more-or-less
1093 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
1095 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
1097 - Fix an issue related to $libdir/run/ganeti and cluster creation
1103 *(Released Thu, 19 Feb 2009)*
1105 - Xen PVM and KVM have switched the default value for the instance root
1106 disk to the first partition on the first drive, instead of the whole
1107 drive; this means that the OS installation scripts must be changed
1109 - Man pages have been updated
1110 - RAPI has been switched by default to HTTPS, and the exported functions
1111 should all work correctly
1112 - RAPI v1 has been removed
1113 - Many improvements to the KVM hypervisor
1114 - Block device errors are now better reported
1115 - Many other bugfixes and small improvements
1121 *(Released Mon, 26 Jan 2009)*
1123 - Version 2 is a general rewrite of the code and therefore the
1124 differences are too many to list, see the design document for 2.0 in
1125 the ``doc/`` subdirectory for more details
1126 - In this beta version there is not yet a migration path from 1.2 (there
1127 will be one in the final 2.0 release)
1128 - A few significant changes are:
1130 - all commands are executed by a daemon (``ganeti-masterd``) and the
1131 various ``gnt-*`` commands are just front-ends to it
1132 - all the commands are entered into, and executed from a job queue,
1133 see the ``gnt-job(8)`` manpage
1134 - the RAPI daemon supports read-write operations, secured by basic
1135 HTTP authentication on top of HTTPS
1136 - DRBD version 0.7 support has been removed, DRBD 8 is the only
1137 supported version (when migrating from Ganeti 1.2 to 2.0, you need
1138 to migrate to DRBD 8 first while still running Ganeti 1.2)
1139 - DRBD devices are using statically allocated minor numbers, which
1140 will be assigned to existing instances during the migration process
1141 - there is support for both Xen PVM and Xen HVM instances running on
1143 - KVM virtualization is supported too
1144 - file-based storage has been implemented, which means that it is
1145 possible to run the cluster without LVM and DRBD storage, for
1146 example using a shared filesystem exported from shared storage (and
1147 still have live migration)
1153 *(Released Tue, 13 Jan 2009)*
1155 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
1156 - Reuse the old instance mac address by default on instance import, if
1157 the instance name is the same.
1158 - Handle situations in which the node info rpc returns incomplete
1160 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
1161 - Improved version of batcher:
1163 - state file support
1164 - instance mac address support
1165 - support for HVM clusters/instances
1167 - Add an option to show the number of cpu sockets and nodes in
1169 - Support OSes that handle more than one version of the OS api (but do
1170 not change the current API in any other way)
1171 - Fix ``gnt-node migrate``
1172 - ``gnt-debug`` man page
1173 - Fixes various more typos and small issues
1174 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
1180 *(Released Wed, 24 Sep 2008)*
1182 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
1183 type of disk exported to fully virtualized instances.
1184 - provide access to the serial console of HVM instances
1185 - instance auto_balance flag, set by default. If turned off it will
1186 avoid warnings on cluster verify if there is not enough memory to fail
1187 over an instance. in the future it will prevent automatically failing
1188 it over when we will support that.
1189 - batcher tool for instance creation, see ``tools/README.batcher``
1190 - ``gnt-instance reinstall --select-os`` to interactively select a new
1191 operating system when reinstalling an instance.
1192 - when changing the memory amount on instance modify a check has been
1193 added that the instance will be able to start. also warnings are
1194 emitted if the instance will not be able to fail over, if auto_balance
1196 - documentation fixes
1197 - sync fields between ``gnt-instance list/modify/add/import``
1198 - fix a race condition in drbd when the sync speed was set after giving
1199 the device a remote peer.
1205 *(Released Tue, 22 Jul 2008)*
1207 - note: the allowed size and number of tags per object were reduced
1208 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1209 - fixed twisted 8.x compatibility
1210 - fixed ``gnt-instance replace-disks`` with iallocator
1211 - add TCP keepalives on twisted connections to detect restarted nodes
1212 - disk increase support, see ``gnt-instance grow-disk``
1213 - implement bulk node/instance query for RAPI
1214 - add tags in node/instance listing (optional)
1215 - experimental migration (and live migration) support, read the man page
1216 for ``gnt-instance migrate``
1217 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1218 has some small improvements in handling its state file
1224 *(Released Fri, 13 Jun 2008)*
1226 - Experimental readonly, REST-based remote API implementation;
1227 automatically started on master node, TCP port 5080, if enabled by
1228 ``--enable-rapi`` parameter to configure script.
1229 - Instance allocator support. Add and import instance accept a
1230 ``--iallocator`` parameter, and call that instance allocator to decide
1231 which node to use for the instance. The iallocator document describes
1232 what's expected from an allocator script.
1233 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1234 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1235 node is lost there is still enough memory to fail over the instances
1237 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1238 ``gnt-cluster verify``, to check for site-specific compliance. All the
1239 hooks will run, and their output, if any, will be displayed. Any
1240 failing hook will make the verification return an error value.
1241 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1242 primary and secondary interfaces
1243 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1244 that is still declared as part of the cluster and has failed.
1245 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1246 specifying output fields, that just adds the chosen fields to the
1248 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1249 export from the filesystem.
1250 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1251 have been added. Using them you can enable/disable acpi and pae
1252 support, and specify a path for a cd image to be exported to the
1253 instance. These parameters as the name suggest only work on HVM
1255 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1256 PAE support will be set to the previously hardcoded values, but the
1257 (previously hardcoded) path to the CDROM ISO image will be unset and
1258 if required, needs to be set manually with ``gnt-instance modify``
1260 - The address to which an instance's VNC console is bound is now
1261 selectable per-instance, rather than being cluster wide. Of course
1262 this only applies to instances controlled via VNC, so currently just
1263 applies to HVM clusters.
1269 *(Released Mon, 18 Feb 2008)*
1271 - more tweaks to the disk activation code (especially helpful for DRBD)
1272 - change the default ``gnt-instance list`` output format, now there is
1273 one combined status field (see the manpage for the exact values this
1275 - some more fixes for the mac export to hooks change
1276 - make Ganeti not break with DRBD 8.2.x (which changed the version
1277 format in ``/proc/drbd``) (issue 24)
1278 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1279 template, allowing migration from DRBD0.7+MD to DRBD8
1285 *(Released Wed, 30 Jan 2008)*
1287 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1289 - add command aliases infrastructure and a few aliases
1290 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1291 man pages and the ``--help`` option of ``gnt-node
1292 list``/``gnt-instance list``
1293 - fix ``gnt-backup list`` with down nodes (issue 21)
1294 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1295 - fix the dist archive and add a check for including svn/git files in
1297 - some developer-related changes: improve the burnin and the QA suite,
1298 add an upload script for testing during development
1304 *(Released Wed, 16 Jan 2008)*
1306 - experimental HVM support, read the install document, section
1307 "Initializing the cluster"
1308 - allow for the PVM hypervisor per-instance kernel and initrd paths
1309 - add a new command ``gnt-cluster verify-disks`` which uses a new
1310 algorithm to improve the reconnection of the DRBD pairs if the device
1311 on the secondary node has gone away
1312 - make logical volume code auto-activate LVs at disk activation time
1313 - slightly improve the speed of activating disks
1314 - allow specification of the MAC address at instance creation time, and
1315 changing it later via ``gnt-instance modify``
1316 - fix handling of external commands that generate lots of output on
1318 - update documentation with regard to minimum version of DRBD8 supported
1324 *(Released Tue, 4 Dec 2007)*
1326 - Log the ``xm create`` output to the node daemon log on failure (to
1327 help diagnosing the error)
1328 - In debug mode, log all external commands output if failed to the logs
1329 - Change parsing of lvm commands to ignore stderr
1335 *(Released Wed, 28 Nov 2007)*
1337 - Another round of updates to the DRBD 8 code to deal with more failures
1338 in the replace secondary node operation
1339 - Some more logging of failures in disk operations (lvm, drbd)
1340 - A few documentation updates
1347 *(Released Tue, 13 Nov 2007)*
1349 - Change configuration file format from Python's Pickle to JSON.
1350 Upgrading is possible using the cfgupgrade utility.
1351 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1352 faster replace disks and is more stable (DRBD 8 has many improvements
1353 compared to DRBD 0.7)
1354 - Added command line tags support (see man pages for ``gnt-instance``,
1355 ``gnt-node``, ``gnt-cluster``)
1356 - Added instance rename support
1357 - Added multi-instance startup/shutdown
1358 - Added cluster rename support
1359 - Added ``gnt-node evacuate`` to simplify some node operations
1360 - Added instance reboot operation that can speedup reboot as compared to
1362 - Soften the requirement that hostnames are in FQDN format
1363 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1365 - Removed dependency on debian's patched fping that uses the
1366 non-standard ``-S`` option
1367 - Now the OS definitions are searched for in multiple, configurable
1368 paths (easier for distros to package)
1369 - Some changes to the hooks infrastructure (especially the new
1370 post-configuration update hook)
1371 - Other small bugfixes
1373 .. vim: set textwidth=72 syntax=rst :