8 *(Released Mon, 20 Dec 2010)*
10 Released version 2.3.1~rc1 without any changes.
16 *(Released Wed, 1 Dec 2010)*
18 - impexpd: Disable OpenSSL compression in socat if possible (backport
19 from master, commit e90739d625b, see :doc:`installation guide
20 <install-quick>` for details)
21 - Changed unittest coverage report to exclude test scripts
22 - Added script to check version format
28 *(Released Wed, 1 Dec 2010)*
30 Released version 2.3.0~rc1 without any changes.
36 *(Released Fri, 19 Nov 2010)*
38 A number of bugfixes and documentation updates:
40 - Update ganeti-os-interface documentation
41 - Fixed a bug related to duplicate MACs or similar items which should be
43 - Fix breakage in OS state modify
44 - Reinstall instance: disallow offline secondaries (fixes bug related to
45 OS changing but reinstall failing)
46 - plus all the other fixes between 2.2.1 and 2.2.2
52 *(Released Tue, 2 Nov 2010)*
54 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
55 - Fixed master failover race with watcher
56 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
58 - Accept previously stopped instance for export with instance removal
59 - Simplify and extend the environment variables for instance OS scripts
60 - Added new node flags, ``master_capable`` and ``vm_capable``
61 - Added optional instance disk wiping prior during allocation. This is a
62 cluster-wide option and can be set/modified using
63 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
64 - Added IPv6 support, see :doc:`design document <design-2.3>` and
66 - Added a new watcher option (``--ignore-pause``)
67 - Added option to ignore offline node on instance start/stop
68 (``--ignore-offline``)
69 - Allow overriding OS parameters with ``gnt-instance reinstall``
70 - Added ability to change node's secondary IP address using ``gnt-node
72 - Implemented privilege separation for all daemons except
73 ``ganeti-noded``, see ``configure`` options
74 - Complain if an instance's disk is marked faulty in ``gnt-cluster
76 - Implemented job priorities (see ``ganeti(7)`` manpage)
77 - Ignore failures while shutting down instances during failover from
79 - Exit daemon's bootstrap process only once daemon is ready
80 - Export more information via ``LUQueryInstances``/remote API
81 - Improved documentation, QA and unittests
82 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
83 restart if the file was created or changed
84 - Added LUXI protocol version sent with each request and response,
85 allowing detection of server/client mismatches
86 - Moved the Python scripts among gnt-* and ganeti-* into modules
87 - Moved all code related to setting up SSH to an external script,
89 - Infrastructure changes for node group support in future versions
95 *(Released Fri, 19 Nov 2010)*
97 A few small bugs fixed, and some improvements to the build system:
99 - Fix documentation regarding conversion to drbd
100 - Fix validation of parameters in cluster modify (``gnt-cluster modify
102 - Fix error handling in node modify with multiple changes
103 - Allow remote imports without checked names
109 *(Released Tue, 19 Oct 2010)*
111 - Disable SSL session ID cache in RPC client
117 *(Released Thu, 14 Oct 2010)*
119 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
120 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
122 - Fix problems with interaction between Curl and Python's HTTP server,
123 resulting in increased speed in many RPC calls
124 - Improve our release script to prevent breakage with older aclocal and
131 *(Released Thu, 7 Oct 2010)*
133 - Fixed issue 125, replace hardcoded “xenvg” in ``gnt-cluster`` with
134 value retrieved from master
135 - Added support for blacklisted or hidden OS definitions
136 - Added simple lock monitor (accessible via (``gnt-debug locks``)
137 - Added support for -mem-path in KVM hypervisor abstraction layer
138 - Allow overriding instance parameters in tool for inter-cluster
139 instance moves (``tools/move-instance``)
140 - Improved opcode summaries (e.g. in ``gnt-job list``)
141 - Improve consistency of OS listing by sorting it
142 - Documentation updates
148 *(Released Fri, 8 Oct 2010)*
150 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
156 *(Released Mon, 4 Oct 2010)*
158 - Fixed regression in ``gnt-instance rename``
164 *(Released Wed, 22 Sep 2010)*
166 - Fixed OS_VARIANT variable for OS scripts
167 - Fixed cluster tag operations via RAPI
168 - Made ``setup-ssh`` exit with non-zero code if an error occurred
169 - Disabled RAPI CA checks in watcher
175 *(Released Mon, 23 Aug 2010)*
177 - Support DRBD versions of the format "a.b.c.d"
179 - Re-introduce support for usage from multiple threads in RAPI client
180 - Instance renames and modify via RAPI
181 - Work around race condition between processing and archival in job
183 - Mark opcodes following failed one as failed, too
184 - Job field ``lock_status`` was removed due to difficulties making it
185 work with the changed job queue in Ganeti 2.2; a better way to monitor
186 locks is expected for a later 2.2.x release
187 - Fixed dry-run behaviour with many commands
188 - Support ``ssh-agent`` again when adding nodes
189 - Many additional bugfixes
195 *(Released Fri, 30 Jul 2010)*
197 Important change: the internal RPC mechanism between Ganeti nodes has
198 changed from using a home-grown http library (based on the Python base
199 libraries) to use the PycURL library. This requires that PycURL is
200 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
201 against GnuTLS by default. cURL's support for GnuTLS had known issues
202 before cURL 7.21.0 and we recommend using the latest cURL release or
203 linking against OpenSSL. Most other distributions already link PycURL
204 and cURL against OpenSSL. The command::
206 python -c 'import pycurl; print pycurl.version'
208 can be used to determine the libraries PycURL and cURL are linked
211 Other significant changes:
213 - Rewrote much of the internals of the job queue, in order to achieve
214 better parallelism; this decouples job query operations from the job
215 processing, and it should allow much nicer behaviour of the master
216 daemon under load, and it also has uncovered some long-standing bugs
217 related to the job serialisation (now fixed)
218 - Added a default iallocator setting to the cluster parameters,
219 eliminating the need to always pass nodes or an iallocator for
220 operations that require selection of new node(s)
221 - Added experimental support for the LXC virtualization method
222 - Added support for OS parameters, which allows the installation of
223 instances to pass parameter to OS scripts in order to customise the
225 - Added a hypervisor parameter controlling the migration type (live or
226 non-live), since hypervisors have various levels of reliability; this
227 has renamed the 'live' parameter to 'mode'
228 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
229 logical volumes, meaning that cluster verify will ignore them and not
230 flag their presence as errors
231 - The watcher will now reset the error count for failed instances after
232 8 hours, thus allowing self-healing if the problem that caused the
233 instances to be down/fail to start has cleared in the meantime
234 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
235 check for, and warn, if the drbd module parameter ``usermode_helper``
236 is not consistent with the cluster-wide setting; this is needed to
237 make diagnose easier of failed drbd creations
238 - Started adding base IPv6 support, but this is not yet
239 enabled/available for use
240 - Rename operations (cluster, instance) will now return the new name,
241 which is especially useful if a short name was passed in
242 - Added support for instance migration in RAPI
243 - Added a tool to pre-configure nodes for the SSH setup, before joining
244 them to the cluster; this will allow in the future a simplified model
245 for node joining (but not yet fully enabled in 2.2); this needs the
246 paramiko python library
247 - Fixed handling of name-resolving errors
248 - Fixed consistency of job results on the error path
249 - Fixed master-failover race condition when executed multiple times in
251 - Fixed many bugs related to the job queue (mostly introduced during the
252 2.2 development cycle, so not all are impacting 2.1)
253 - Fixed instance migration with missing disk symlinks
254 - Fixed handling of unknown jobs in ``gnt-job archive``
255 - And many other small fixes/improvements
259 - Enhanced both the unittest and the QA coverage
260 - Switched the opcode validation to a generic model, and extended the
261 validation to all opcode parameters
262 - Changed more parts of the code that write shell scripts to use the
264 - Switched the master daemon to use the asyncore library for the Luxi
271 *(Released Thu, 17 Jun 2010)*
273 - Added tool (``move-instance``) and infrastructure to move instances
274 between separate clusters (see :doc:`separate documentation
275 <move-instance>` and :doc:`design document <design-2.2>`)
276 - Added per-request RPC timeout
277 - RAPI now requires a Content-Type header for requests with a body (e.g.
278 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
279 :rfc:`2616` (HTTP/1.1), section 7.2.1)
280 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
282 - Implemented initial support for running Ganeti daemons as separate
283 users, see configure-time flags ``--with-user-prefix`` and
284 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
286 - Instances can be removed after export (``gnt-backup export
288 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
289 key (instead of 1024 bit)
290 - Added new cluster configuration file for cluster domain secret
291 - Import/export now use SSL instead of SSH
292 - Added support for showing estimated time when exporting an instance,
293 see the ``ganeti-os-interface(7)`` manpage and look for
300 *(Released Tue, 16 Nov 2010)*
302 Some more bugfixes. Unless critical bugs occur, this will be the last
305 - Fix case of MAC special-values
306 - Fix mac checker regex
307 - backend: Fix typo causing “out of range” error
308 - Add missing --units in gnt-instance list man page
314 *(Released Tue, 24 Aug 2010)*
317 - Don't ignore secondary node silently on non-mirrored disk templates
319 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
320 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
321 - Properly document the usb_mouse parameter
322 - Fix path in ganeti-rapi(8) (issue 116)
323 - Adjust error message when the ganeti user's .ssh directory is
325 - Add same-node-check when changing the disk template to drbd
331 *(Released Fri, 16 Jul 2010)*
334 - Add an option to only select some reboot types during qa/burnin.
335 (on some hypervisors consequent reboots are not supported)
336 - Fix infrequent race condition in master failover. Sometimes the old
337 master ip address would be still detected as up for a short time
338 after it was removed, causing failover to fail.
339 - Decrease mlockall warnings when the ctypes module is missing. On
340 Python 2.4 we support running even if no ctypes module is installed,
341 but we were too verbose about this issue.
342 - Fix building on old distributions, on which man doesn't have a
344 - Fix RAPI not to ignore the MAC address on instance creation
345 - Implement the old instance creation format in the RAPI client.
351 *(Released Thu, 01 Jul 2010)*
353 A small bugfix release:
354 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
355 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
356 a non-existing option
357 - Fix instance create: the --force-variant option was ignored
358 - Improve pylint 0.21 compatibility and warnings with Python 2.6
359 - Fix modify node storage with non-FQDN arguments
360 - Fix RAPI client to authenticate under Python 2.6 when used
361 for more than 5 requests needing authentication
362 - Fix gnt-instance modify -t (storage) giving a wrong error message
363 when converting a non-shutdown drbd instance to plain
369 *(Released Fri, 18 Jun 2010)*
371 A small bugfix release:
373 - Fix live migration of KVM instances started with older Ganeti
374 versions which had fewer hypervisor parameters
375 - Fix gnt-instance grow-disk on down instances
376 - Fix an error-reporting bug during instance migration
377 - Better checking of the ``--net`` and ``--disk`` values, to avoid
378 silently ignoring broken ones
379 - Fix an RPC error reporting bug affecting, for example, RAPI client
381 - Fix bug triggered by different API version os-es on different nodes
382 - Fix a bug in instance startup with custom hvparams: OS level
383 parameters would fail to be applied.
384 - Fix the RAPI client under Python 2.6 (but more work is needed to
385 make it work completely well with OpenSSL)
386 - Fix handling of errors when resolving names from DNS
392 *(Released Thu, 3 Jun 2010)*
394 A medium sized development cycle. Some new features, and some
395 fixes/small improvements/cleanups.
400 The node deamon now tries to mlock itself into memory, unless the
401 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
402 its logs, and falls back to console logging. This allows emergency
403 features such as ``gnt-node powercycle`` to work even in the event of a
404 broken node disk (tested offlining the disk hosting the node's
405 filesystem and dropping its memory caches; don't try this at home)
407 KVM: add vhost-net acceleration support. It can be tested with a new
408 enough version of the kernel and of qemu-kvm.
410 KVM: Add instance chrooting feature. If you use privilege dropping for
411 your VMs you can also now force them to chroot to an empty directory,
412 before starting the emulated guest.
414 KVM: Add maximum migration bandwith and maximum downtime tweaking
415 support (requires a new-enough version of qemu-kvm).
417 Cluster verify will now warn if the master node doesn't have the master
420 Add a new (incompatible) instance creation request format to RAPI which
421 supports all parameters (previously only a subset was supported, and it
422 wasn't possible to extend the old format to accomodate all the new
423 features. The old format is still supported, and a client can check for
424 this feature, before using it, by checking for its presence in the
425 ``features`` RAPI resource.
427 Now with ancient latin support. Try it passing the ``--roman`` option to
428 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
429 (requires the python-roman module to be installed, in order to work).
434 As usual many internal code refactorings, documentation updates, and
437 - Lots of improvements and cleanups to the experimental Remote API
438 (RAPI) client library.
439 - A new unit test suite for the core daemon libraries.
440 - A fix to creating missing directories makes sure the umask is not
441 applied anymore. This enforces the same directory permissions
443 - Better handling terminating daemons with ctrl+c (used when running
444 them in debugging mode).
445 - Fix a race condition in live migrating a KVM instance, when stat()
446 on the old proc status file returned EINVAL, which is an unexpected
448 - Fixed manpage checking with newer man and utf-8 charachters. But now
449 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
455 *(Released Fri, 7 May 2010)*
457 Fix a bug which prevented untagged KVM instances from starting.
463 *(Released Fri, 7 May 2010)*
465 Another release with a long development cycle, during which many
466 different features were added.
471 The KVM hypervisor now can run the individual instances as non-root, to
472 reduce the impact of a VM being hijacked due to bugs in the
473 hypervisor. It is possible to run all instances as a single (non-root)
474 user, to manually specify a user for each instance, or to dynamically
475 allocate a user out of a cluster-wide pool to each instance, with the
476 guarantee that no two instances will run under the same user ID on any
479 An experimental RAPI client library, that can be used standalone
480 (without the other Ganeti libraries), is provided in the source tree as
481 ``lib/rapi/client.py``. Note this client might change its interface in
482 the future, as we iterate on its capabilities.
484 A new command, ``gnt-cluster renew-crypto`` has been added to easily
485 replace the cluster's certificates and crypto keys. This might help in
486 case they have been compromised, or have simply expired.
488 A new disk option for instance creation has been added that allows one
489 to "adopt" currently existing logical volumes, with data
490 preservation. This should allow easier migration to Ganeti from
491 unmanaged (or managed via other software) instances.
493 Another disk improvement is the possibility to convert between redundant
494 (DRBD) and plain (LVM) disk configuration for an instance. This should
495 allow better scalability (starting with one node and growing the
496 cluster, or shrinking a two-node cluster to one node).
498 A new feature that could help with automated node failovers has been
499 implemented: if a node sees itself as offline (by querying the master
500 candidates), it will try to shutdown (hard) all instances and any active
501 DRBD devices. This reduces the risk of duplicate instances if an
502 external script automatically failovers the instances on such nodes. To
503 enable this, the cluster parameter ``maintain_node_health`` should be
504 enabled; in the future this option (per the name) will enable other
505 automatic maintenance features.
507 Instance export/import now will reuse the original instance
508 specifications for all parameters; that means exporting an instance,
509 deleting it and the importing it back should give an almost identical
510 instance. Note that the default import behaviour has changed from
511 before, where it created only one NIC; now it recreates the original
514 Cluster verify has added a few new checks: SSL certificates validity,
515 /etc/hosts consistency across the cluster, etc.
520 As usual, many internal changes were done, documentation fixes,
523 - Fixed cluster initialization with disabled cluster storage (regression
525 - File-based storage supports growing the disks
526 - Fixed behaviour of node role changes
527 - Fixed cluster verify for some corner cases, plus a general rewrite of
528 cluster verify to allow future extension with more checks
529 - Fixed log spamming by watcher and node daemon (regression introduced
531 - Fixed possible validation issues when changing the list of enabled
533 - Fixed cleanup of /etc/hosts during node removal
534 - Fixed RAPI response for invalid methods
535 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
536 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
537 from ide disks, etc.)
538 - Allow OS changes without re-installation (to record a changed OS
539 outside of Ganeti, or to allow OS renames)
540 - Allow instance creation without OS installation (useful for example if
541 the OS will be installed manually, or restored from a backup not in
543 - Implemented option to make cluster ``copyfile`` use the replication
545 - Added list of enabled hypervisors to ssconf (possibly useful for
547 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
549 - A partial form of node re-IP is possible via node readd, which now
550 allows changed node primary IP
551 - Command line utilities now show an informational message if the job is
553 - The logs of the master daemon now show the PID/UID/GID of the
560 *(Released Fri, 12 Mar 2010)*
562 During the 2.1.0 long release candidate cycle, a lot of improvements and
563 changes have accumulated with were released later as 2.1.1.
568 The node evacuate command (``gnt-node evacuate``) was significantly
569 rewritten, and as such the IAllocator protocol was changed - a new
570 request type has been added. This unfortunate change during a stable
571 series is designed to improve performance of node evacuations; on
572 clusters with more than about five nodes and which are well-balanced,
573 evacuation should proceed in parallel for all instances of the node
574 being evacuated. As such, any existing IAllocator scripts need to be
575 updated, otherwise the above command will fail due to the unknown
576 request. The provided "dumb" allocator has not been updated; but the
577 ganeti-htools package supports the new protocol since version 0.2.4.
579 Another important change is increased validation of node and instance
580 names. This might create problems in special cases, if invalid host
581 names are being used.
583 Also, a new layer of hypervisor parameters has been added, that sits at
584 OS level between the cluster defaults and the instance ones. This allows
585 customisation of virtualization parameters depending on the installed
586 OS. For example instances with OS 'X' may have a different KVM kernel
587 (or any other parameter) than the cluster defaults. This is intended to
588 help managing a multiple OSes on the same cluster, without manual
589 modification of each instance's parameters.
591 A tool for merging clusters, ``cluster-merge``, has been added in the
597 - Improved the int/float conversions that should make the code more
598 robust in face of errors from the node daemons
599 - Fixed the remove node code in case of internal configuration errors
600 - Fixed the node daemon behaviour in face of inconsistent queue
601 directory (e.g. read-only file-system where we can't open the files
603 - Fixed the behaviour of gnt-node modify for master candidate demotion;
604 now it either aborts cleanly or, if given the new “auto_promote”
605 parameter, will automatically promote other nodes as needed
606 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
607 completely prevent Ganeti from working
608 - Fixed bug for instance export when not all disks were successfully
610 - Fixed behaviour of node add when the new node is slow in starting up
612 - Fixed handling of signals in the LUXI client, which should improve
613 behaviour of command-line scripts
614 - Added checks for invalid node/instance names in the configuration (now
615 flagged during cluster verify)
616 - Fixed watcher behaviour for disk activation errors
617 - Fixed two potentially endless loops in http library, which led to the
618 RAPI daemon hanging and consuming 100% CPU in some cases
619 - Fixed bug in RAPI daemon related to hashed passwords
620 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
622 - Enhanced compatibility with non-Debian OSes, but not using absolute
623 path in some commands and allowing customisation of the ssh
624 configuration directory
625 - Fixed possible future issue with new Python versions by abiding to the
626 proper use of ``__slots__`` attribute on classes
627 - Added checks that should prevent directory traversal attacks
628 - Many documentation fixes based on feedback from users
633 - Added an “early_release” more for instance replace disks and node
634 evacuate, where we release locks earlier and thus allow higher
635 parallelism within the cluster
636 - Added watcher hooks, intended to allow the watcher to restart other
637 daemons (e.g. from the ganeti-nbma project), but they can be used of
638 course for any other purpose
639 - Added a compile-time disable for DRBD barriers, to increase
640 performance if the administrator trusts the power supply or the
641 storage system to not lose writes
642 - Added the option of using syslog for logging instead of, or in
643 addition to, Ganeti's own log files
644 - Removed boot restriction for paravirtual NICs for KVM, recent versions
645 can indeed boot from a paravirtual NIC
646 - Added a generic debug level for many operations; while this is not
647 used widely yet, it allows one to pass the debug value all the way to
649 - Enhanced the hooks environment for instance moves (failovers,
650 migrations) where the primary/secondary nodes changed during the
651 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
652 - Enhanced data validations for many user-supplied values; one important
653 item is the restrictions imposed on instance and node names, which
654 might reject some (invalid) host names
655 - Add a configure-time option to disable file-based storage, if it's not
656 needed; this allows greater security separation between the master
657 node and the other nodes from the point of view of the inter-node RPC
659 - Added user notification in interactive tools if job is waiting in the
660 job queue or trying to acquire locks
661 - Added log messages when a job is waiting for locks
662 - Added filtering by node tags in instance operations which admit
663 multiple instances (start, stop, reboot, reinstall)
664 - Added a new tool for cluster mergers, ``cluster-merge``
665 - Parameters from command line which are of the form ``a=b,c=d`` can now
666 use backslash escapes to pass in values which contain commas,
667 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
669 - For KVM, the instance name is the first parameter passed to KVM, so
670 that it's more visible in the process list
676 *(Released Tue, 2 Mar 2010)*
678 Ganeti 2.1 brings many improvements with it. Major changes:
680 - Added infrastructure to ease automated disk repairs
681 - Added new daemon to export configuration data in a cheaper way than
683 - Instance NICs can now be routed instead of being associated with a
685 - Improved job locking logic to reduce impact of jobs acquiring multiple
686 locks waiting for other long-running jobs
688 In-depth implementation details can be found in the Ganeti 2.1 design
694 - Added chroot hypervisor
695 - Added more options to xen-hvm hypervisor (``kernel_path`` and
697 - Added more options to xen-pvm hypervisor (``use_bootloader``,
698 ``bootloader_path`` and ``bootloader_args``)
699 - Added the ``use_localtime`` option for the xen-hvm and kvm
700 hypervisors, and the default value for this has changed to false (in
701 2.0 xen-hvm always enabled it)
702 - Added luxi call to submit multiple jobs in one go
703 - Added cluster initialization option to not modify ``/etc/hosts``
705 - Added network interface parameters
706 - Added dry run mode to some LUs
707 - Added RAPI resources:
709 - ``/2/instances/[instance_name]/info``
710 - ``/2/instances/[instance_name]/replace-disks``
711 - ``/2/nodes/[node_name]/evacuate``
712 - ``/2/nodes/[node_name]/migrate``
713 - ``/2/nodes/[node_name]/role``
714 - ``/2/nodes/[node_name]/storage``
715 - ``/2/nodes/[node_name]/storage/modify``
716 - ``/2/nodes/[node_name]/storage/repair``
718 - Added OpCodes to evacuate or migrate all instances on a node
719 - Added new command to list storage elements on nodes (``gnt-node
720 list-storage``) and modify them (``gnt-node modify-storage``)
721 - Added new ssconf files with master candidate IP address
722 (``ssconf_master_candidates_ips``), node primary IP address
723 (``ssconf_node_primary_ips``) and node secondary IP address
724 (``ssconf_node_secondary_ips``)
725 - Added ``ganeti-confd`` and a client library to query the Ganeti
726 configuration via UDP
727 - Added ability to run hooks after cluster initialization and before
729 - Added automatic mode for disk replace (``gnt-instance replace-disks
731 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
732 after catastrophic data-loss
733 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
735 - Added ``gnt-instance move`` command to move instances
736 - Added ``gnt-cluster watcher`` command to control watcher
737 - Added ``gnt-node powercycle`` command to powercycle nodes
738 - Added new job status field ``lock_status``
739 - Added parseable error codes to cluster verification (``gnt-cluster
740 verify --error-codes``) and made output less verbose (use
741 ``--verbose`` to restore previous behaviour)
742 - Added UUIDs to the main config entities (cluster, nodes, instances)
743 - Added support for OS variants
744 - Added support for hashed passwords in the Ganeti remote API users file
746 - Added option to specify maximum timeout on instance shutdown
747 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
748 - Added new helper script to start and stop Ganeti daemons
749 (``daemon-util``), with the intent to reduce the work necessary to
750 adjust Ganeti for non-Debian distributions and to start/stop daemons
752 - Added more unittests
753 - Fixed critical bug in ganeti-masterd startup
754 - Removed the configure-time ``kvm-migration-port`` parameter, this is
755 now customisable at the cluster level for both the KVM and Xen
756 hypervisors using the new ``migration_port`` parameter
757 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
758 reinstalling an instance
759 - Allowed ``@`` in tag names
760 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
761 - Many documentation updates
762 - Distribute hypervisor files on ``gnt-cluster redist-conf``
763 - ``gnt-instance reinstall`` can now reinstall multiple instances
764 - Updated many command line parameters
765 - Introduced new OS API version 15
766 - No longer support a default hypervisor
767 - Treat virtual LVs as inexistent
768 - Improved job locking logic to reduce lock contention
769 - Match instance and node names case insensitively
770 - Reimplemented bash completion script to be more complete
777 *(Released Thu, 4 Feb 2010)*
779 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
780 - Fix a string formatting bug
781 - Improve safety of the code in some error paths
782 - Improve data validation in the master of values returned from nodes
788 *(Released Thu, 17 Dec 2009)*
790 - Fix security issue due to missing validation of iallocator names; this
791 allows local and remote execution of arbitrary executables
792 - Fix failure of gnt-node list during instance removal
793 - Ship the RAPI documentation in the archive
799 *(Released Wed, 30 Sep 2009)*
801 - Fixed many wrong messages
802 - Fixed a few bugs related to the locking library
803 - Fixed MAC checking at instance creation time
804 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
805 - Fixed a few issues related to signal handling in both daemons and
807 - Fixed the example startup script provided
808 - Fixed insserv dependencies in the example startup script (patch from
810 - Fixed handling of drained nodes in the iallocator framework
811 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
813 - Fixed error related to invalid job IDs in job polling
814 - Fixed job/opcode persistence on unclean master shutdown
815 - Fixed handling of partial job processing after unclean master
817 - Fixed error reporting from LUs, previously all errors were converted
818 into execution errors
819 - Fixed error reporting from burnin
820 - Decreased significantly the memory usage of the job queue
821 - Optimised slightly multi-job submission
822 - Optimised slightly opcode loading
823 - Backported the multi-job submit framework from the development
824 branch; multi-instance start and stop should be faster
825 - Added script to clean archived jobs after 21 days; this will reduce
826 the size of the queue directory
827 - Added some extra checks in disk size tracking
828 - Added an example ethers hook script
829 - Added a cluster parameter that prevents Ganeti from modifying of
831 - Added more node information to RAPI responses
832 - Added a ``gnt-job watch`` command that allows following the ouput of a
834 - Added a bind-address option to ganeti-rapi
835 - Added more checks to the configuration verify
836 - Enhanced the burnin script such that some operations can be retried
838 - Converted instance reinstall to multi-instance model
844 *(Released Fri, 7 Aug 2009)*
846 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
847 to allow using the pre-2.0.2 behaviour in activation, if any existing
848 instances have mismatched disk sizes in the configuration
849 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
850 any configuration mismatches for disk sizes
851 - Added ``gnt-master cluste-failover --no-voting`` to allow master
852 failover to work on two-node clusters
853 - Fixed the ``--net`` option of ``gnt-backup import``, which was
855 - Fixed detection of OS script errors in ``gnt-backup export``
856 - Fixed exit code of ``gnt-backup export``
862 *(Released Fri, 17 Jul 2009)*
864 - Added experimental support for stripped logical volumes; this should
865 enhance performance but comes with a higher complexity in the block
866 device handling; stripping is only enabled when passing
867 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
868 affected even in the non-stripped mode
869 - Improved resiliency against transient failures at the end of DRBD
870 resyncs, and in general of DRBD resync checks
871 - Fixed a couple of issues with exports and snapshot errors
872 - Fixed a couple of issues in instance listing
873 - Added display of the disk size in ``gnt-instance info``
874 - Fixed checking for valid OSes in instance creation
875 - Fixed handling of the "vcpus" parameter in instance listing and in
876 general of invalid parameters
877 - Fixed http server library, and thus RAPI, to handle invalid
878 username/password combinations correctly; this means that now they
879 report unauthorized for queries too, not only for modifications,
880 allowing earlier detect of configuration problems
881 - Added a new "role" node list field, equivalent to the master/master
882 candidate/drained/offline flags combinations
883 - Fixed cluster modify and changes of candidate pool size
884 - Fixed cluster verify error messages for wrong files on regular nodes
885 - Fixed a couple of issues with node demotion from master candidate role
886 - Fixed node readd issues
887 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
888 - Added a new ``--no-voting`` option for masterfailover to fix failover
889 on two-nodes clusters when the former master node is unreachable
890 - Added instance reinstall over RAPI
896 *(Released Tue, 16 Jun 2009)*
898 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
899 allow re-adding the start in single-user option (regression from 1.2)
900 - the watcher writes the instance status to a file, to allow monitoring
901 to report the instance status (from the master) based on cached
902 results of the watcher's queries; while this can get stale if the
903 watcher is being locked due to other work on the cluster, this is
905 - the watcher now also restarts the node daemon and the rapi daemon if
907 - fixed the watcher to handle full and drained queue cases
908 - hooks export more instance data in the environment, which helps if
909 hook scripts need to take action based on the instance's properties
910 (no longer need to query back into ganeti)
911 - instance failovers when the instance is stopped do not check for free
912 RAM, so that failing over a stopped instance is possible in low memory
914 - rapi uses queries for tags instead of jobs (for less job traffic), and
915 for cluster tags it won't talk to masterd at all but read them from
917 - a couple of error handling fixes in RAPI
918 - drbd handling: improved the error handling of inconsistent disks after
919 resync to reduce the frequency of "there are some degraded disks for
920 this instance" messages
921 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
922 error handling path called a wrong function name)
928 *(Released Wed, 27 May 2009)*
930 - no changes from rc5
933 Version 2.0 release candidate 5
934 -------------------------------
936 *(Released Wed, 20 May 2009)*
938 - fix a couple of bugs (validation, argument checks)
939 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
940 - some small improvements to RAPI and IAllocator
941 - make watcher automatically start the master daemon if down
944 Version 2.0 release candidate 4
945 -------------------------------
947 *(Released Mon, 27 Apr 2009)*
949 - change the OS list to not require locks; this helps with big clusters
950 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
951 volume group is broken
952 - ``gnt-instance info``, without any arguments, doesn't run for all
953 instances anymore; either pass ``--all`` or pass the desired
954 instances; this helps against mistakes on big clusters where listing
955 the information for all instances takes a long time
956 - miscellaneous doc and man pages fixes
959 Version 2.0 release candidate 3
960 -------------------------------
962 *(Released Wed, 8 Apr 2009)*
964 - Change the internal locking model of some ``gnt-node`` commands, in
965 order to reduce contention (and blocking of master daemon) when
966 batching many creation/reinstall jobs
967 - Fixes to Xen soft reboot
968 - No longer build documentation at build time, instead distribute it in
969 the archive, in order to reduce the need for the whole docbook/rst
973 Version 2.0 release candidate 2
974 -------------------------------
976 *(Released Fri, 27 Mar 2009)*
978 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
979 - Fix watcher startup sequence, improves the behaviour of busy clusters
980 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
981 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
982 bind address and other places
983 - Some documentation fixes and updates
986 Version 2.0 release candidate 1
987 -------------------------------
989 *(Released Mon, 2 Mar 2009)*
991 - More documentation updates, now all docs should be more-or-less
993 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
995 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
997 - Fix an issue related to $libdir/run/ganeti and cluster creation
1003 *(Released Thu, 19 Feb 2009)*
1005 - Xen PVM and KVM have switched the default value for the instance root
1006 disk to the first partition on the first drive, instead of the whole
1007 drive; this means that the OS installation scripts must be changed
1009 - Man pages have been updated
1010 - RAPI has been switched by default to HTTPS, and the exported functions
1011 should all work correctly
1012 - RAPI v1 has been removed
1013 - Many improvements to the KVM hypervisor
1014 - Block device errors are now better reported
1015 - Many other bugfixes and small improvements
1021 *(Released Mon, 26 Jan 2009)*
1023 - Version 2 is a general rewrite of the code and therefore the
1024 differences are too many to list, see the design document for 2.0 in
1025 the ``doc/`` subdirectory for more details
1026 - In this beta version there is not yet a migration path from 1.2 (there
1027 will be one in the final 2.0 release)
1028 - A few significant changes are:
1030 - all commands are executed by a daemon (``ganeti-masterd``) and the
1031 various ``gnt-*`` commands are just front-ends to it
1032 - all the commands are entered into, and executed from a job queue,
1033 see the ``gnt-job(8)`` manpage
1034 - the RAPI daemon supports read-write operations, secured by basic
1035 HTTP authentication on top of HTTPS
1036 - DRBD version 0.7 support has been removed, DRBD 8 is the only
1037 supported version (when migrating from Ganeti 1.2 to 2.0, you need
1038 to migrate to DRBD 8 first while still running Ganeti 1.2)
1039 - DRBD devices are using statically allocated minor numbers, which
1040 will be assigned to existing instances during the migration process
1041 - there is support for both Xen PVM and Xen HVM instances running on
1043 - KVM virtualization is supported too
1044 - file-based storage has been implemented, which means that it is
1045 possible to run the cluster without LVM and DRBD storage, for
1046 example using a shared filesystem exported from shared storage (and
1047 still have live migration)
1053 *(Released Tue, 13 Jan 2009)*
1055 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
1056 - Reuse the old instance mac address by default on instance import, if
1057 the instance name is the same.
1058 - Handle situations in which the node info rpc returns incomplete
1060 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
1061 - Improved version of batcher:
1063 - state file support
1064 - instance mac address support
1065 - support for HVM clusters/instances
1067 - Add an option to show the number of cpu sockets and nodes in
1069 - Support OSes that handle more than one version of the OS api (but do
1070 not change the current API in any other way)
1071 - Fix ``gnt-node migrate``
1072 - ``gnt-debug`` man page
1073 - Fixes various more typos and small issues
1074 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
1080 *(Released Wed, 24 Sep 2008)*
1082 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
1083 type of disk exported to fully virtualized instances.
1084 - provide access to the serial console of HVM instances
1085 - instance auto_balance flag, set by default. If turned off it will
1086 avoid warnings on cluster verify if there is not enough memory to fail
1087 over an instance. in the future it will prevent automatically failing
1088 it over when we will support that.
1089 - batcher tool for instance creation, see ``tools/README.batcher``
1090 - ``gnt-instance reinstall --select-os`` to interactively select a new
1091 operating system when reinstalling an instance.
1092 - when changing the memory amount on instance modify a check has been
1093 added that the instance will be able to start. also warnings are
1094 emitted if the instance will not be able to fail over, if auto_balance
1096 - documentation fixes
1097 - sync fields between ``gnt-instance list/modify/add/import``
1098 - fix a race condition in drbd when the sync speed was set after giving
1099 the device a remote peer.
1105 *(Released Tue, 22 Jul 2008)*
1107 - note: the allowed size and number of tags per object were reduced
1108 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1109 - fixed twisted 8.x compatibility
1110 - fixed ``gnt-instance replace-disks`` with iallocator
1111 - add TCP keepalives on twisted connections to detect restarted nodes
1112 - disk increase support, see ``gnt-instance grow-disk``
1113 - implement bulk node/instance query for RAPI
1114 - add tags in node/instance listing (optional)
1115 - experimental migration (and live migration) support, read the man page
1116 for ``gnt-instance migrate``
1117 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1118 has some small improvements in handling its state file
1124 *(Released Fri, 13 Jun 2008)*
1126 - Experimental readonly, REST-based remote API implementation;
1127 automatically started on master node, TCP port 5080, if enabled by
1128 ``--enable-rapi`` parameter to configure script.
1129 - Instance allocator support. Add and import instance accept a
1130 ``--iallocator`` parameter, and call that instance allocator to decide
1131 which node to use for the instance. The iallocator document describes
1132 what's expected from an allocator script.
1133 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1134 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1135 node is lost there is still enough memory to fail over the instances
1137 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1138 ``gnt-cluster verify``, to check for site-specific compliance. All the
1139 hooks will run, and their output, if any, will be displayed. Any
1140 failing hook will make the verification return an error value.
1141 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1142 primary and secondary interfaces
1143 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1144 that is still declared as part of the cluster and has failed.
1145 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1146 specifying output fields, that just adds the chosen fields to the
1148 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1149 export from the filesystem.
1150 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1151 have been added. Using them you can enable/disable acpi and pae
1152 support, and specify a path for a cd image to be exported to the
1153 instance. These parameters as the name suggest only work on HVM
1155 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1156 PAE support will be set to the previously hardcoded values, but the
1157 (previously hardcoded) path to the CDROM ISO image will be unset and
1158 if required, needs to be set manually with ``gnt-instance modify``
1160 - The address to which an instance's VNC console is bound is now
1161 selectable per-instance, rather than being cluster wide. Of course
1162 this only applies to instances controlled via VNC, so currently just
1163 applies to HVM clusters.
1169 *(Released Mon, 18 Feb 2008)*
1171 - more tweaks to the disk activation code (especially helpful for DRBD)
1172 - change the default ``gnt-instance list`` output format, now there is
1173 one combined status field (see the manpage for the exact values this
1175 - some more fixes for the mac export to hooks change
1176 - make Ganeti not break with DRBD 8.2.x (which changed the version
1177 format in ``/proc/drbd``) (issue 24)
1178 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1179 template, allowing migration from DRBD0.7+MD to DRBD8
1185 *(Released Wed, 30 Jan 2008)*
1187 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1189 - add command aliases infrastructure and a few aliases
1190 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1191 man pages and the ``--help`` option of ``gnt-node
1192 list``/``gnt-instance list``
1193 - fix ``gnt-backup list`` with down nodes (issue 21)
1194 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1195 - fix the dist archive and add a check for including svn/git files in
1197 - some developer-related changes: improve the burnin and the QA suite,
1198 add an upload script for testing during development
1204 *(Released Wed, 16 Jan 2008)*
1206 - experimental HVM support, read the install document, section
1207 "Initializing the cluster"
1208 - allow for the PVM hypervisor per-instance kernel and initrd paths
1209 - add a new command ``gnt-cluster verify-disks`` which uses a new
1210 algorithm to improve the reconnection of the DRBD pairs if the device
1211 on the secondary node has gone away
1212 - make logical volume code auto-activate LVs at disk activation time
1213 - slightly improve the speed of activating disks
1214 - allow specification of the MAC address at instance creation time, and
1215 changing it later via ``gnt-instance modify``
1216 - fix handling of external commands that generate lots of output on
1218 - update documentation with regard to minimum version of DRBD8 supported
1224 *(Released Tue, 4 Dec 2007)*
1226 - Log the ``xm create`` output to the node daemon log on failure (to
1227 help diagnosing the error)
1228 - In debug mode, log all external commands output if failed to the logs
1229 - Change parsing of lvm commands to ignore stderr
1235 *(Released Wed, 28 Nov 2007)*
1237 - Another round of updates to the DRBD 8 code to deal with more failures
1238 in the replace secondary node operation
1239 - Some more logging of failures in disk operations (lvm, drbd)
1240 - A few documentation updates
1247 *(Released Tue, 13 Nov 2007)*
1249 - Change configuration file format from Python's Pickle to JSON.
1250 Upgrading is possible using the cfgupgrade utility.
1251 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1252 faster replace disks and is more stable (DRBD 8 has many improvements
1253 compared to DRBD 0.7)
1254 - Added command line tags support (see man pages for ``gnt-instance``,
1255 ``gnt-node``, ``gnt-cluster``)
1256 - Added instance rename support
1257 - Added multi-instance startup/shutdown
1258 - Added cluster rename support
1259 - Added ``gnt-node evacuate`` to simplify some node operations
1260 - Added instance reboot operation that can speedup reboot as compared to
1262 - Soften the requirement that hostnames are in FQDN format
1263 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1265 - Removed dependency on debian's patched fping that uses the
1266 non-standard ``-S`` option
1267 - Now the OS definitions are searched for in multiple, configurable
1268 paths (easier for distros to package)
1269 - Some changes to the hooks infrastructure (especially the new
1270 post-configuration update hook)
1271 - Other small bugfixes
1273 .. vim: set textwidth=72 syntax=rst :