9 - Moved ``rapi_users`` file into separate directory, now named
10 ``…/ganeti/rapi/users``.
11 - Changed query operations to return more detailed information, e.g.
12 whether an information is unavailable due to an offline node. To use
13 this new functionality, the LUXI call ``Query`` must be used. Field
14 information is now stored by the master daemon and can be retrieved
15 using ``QueryFields``. Instances and nodes can also be queried using
16 the new opcodes ``OpQuery`` and ``OpQueryFields``.
17 - Removed support for roman numbers from ``gnt-node list`` and
18 ``gnt-instance list``.
24 *(Released Mon, 20 Dec 2010)*
26 Released version 2.3.1~rc1 without any changes.
32 *(Released Wed, 1 Dec 2010)*
34 - impexpd: Disable OpenSSL compression in socat if possible (backport
35 from master, commit e90739d625b, see :doc:`installation guide
36 <install-quick>` for details)
37 - Changed unittest coverage report to exclude test scripts
38 - Added script to check version format
44 *(Released Wed, 1 Dec 2010)*
46 Released version 2.3.0~rc1 without any changes.
52 *(Released Fri, 19 Nov 2010)*
54 A number of bugfixes and documentation updates:
56 - Update ganeti-os-interface documentation
57 - Fixed a bug related to duplicate MACs or similar items which should be
59 - Fix breakage in OS state modify
60 - Reinstall instance: disallow offline secondaries (fixes bug related to
61 OS changing but reinstall failing)
62 - plus all the other fixes between 2.2.1 and 2.2.2
68 *(Released Tue, 2 Nov 2010)*
70 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
71 - Fixed master failover race with watcher
72 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
74 - Accept previously stopped instance for export with instance removal
75 - Simplify and extend the environment variables for instance OS scripts
76 - Added new node flags, ``master_capable`` and ``vm_capable``
77 - Added optional instance disk wiping prior during allocation. This is a
78 cluster-wide option and can be set/modified using
79 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
80 - Added IPv6 support, see :doc:`design document <design-2.3>` and
82 - Added a new watcher option (``--ignore-pause``)
83 - Added option to ignore offline node on instance start/stop
84 (``--ignore-offline``)
85 - Allow overriding OS parameters with ``gnt-instance reinstall``
86 - Added ability to change node's secondary IP address using ``gnt-node
88 - Implemented privilege separation for all daemons except
89 ``ganeti-noded``, see ``configure`` options
90 - Complain if an instance's disk is marked faulty in ``gnt-cluster
92 - Implemented job priorities (see ``ganeti(7)`` manpage)
93 - Ignore failures while shutting down instances during failover from
95 - Exit daemon's bootstrap process only once daemon is ready
96 - Export more information via ``LUQueryInstances``/remote API
97 - Improved documentation, QA and unittests
98 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
99 restart if the file was created or changed
100 - Added LUXI protocol version sent with each request and response,
101 allowing detection of server/client mismatches
102 - Moved the Python scripts among gnt-* and ganeti-* into modules
103 - Moved all code related to setting up SSH to an external script,
105 - Infrastructure changes for node group support in future versions
111 *(Released Fri, 19 Nov 2010)*
113 A few small bugs fixed, and some improvements to the build system:
115 - Fix documentation regarding conversion to drbd
116 - Fix validation of parameters in cluster modify (``gnt-cluster modify
118 - Fix error handling in node modify with multiple changes
119 - Allow remote imports without checked names
125 *(Released Tue, 19 Oct 2010)*
127 - Disable SSL session ID cache in RPC client
133 *(Released Thu, 14 Oct 2010)*
135 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
136 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
138 - Fix problems with interaction between Curl and Python's HTTP server,
139 resulting in increased speed in many RPC calls
140 - Improve our release script to prevent breakage with older aclocal and
147 *(Released Thu, 7 Oct 2010)*
149 - Fixed issue 125, replace hardcoded “xenvg” in ``gnt-cluster`` with
150 value retrieved from master
151 - Added support for blacklisted or hidden OS definitions
152 - Added simple lock monitor (accessible via (``gnt-debug locks``)
153 - Added support for -mem-path in KVM hypervisor abstraction layer
154 - Allow overriding instance parameters in tool for inter-cluster
155 instance moves (``tools/move-instance``)
156 - Improved opcode summaries (e.g. in ``gnt-job list``)
157 - Improve consistency of OS listing by sorting it
158 - Documentation updates
164 *(Released Fri, 8 Oct 2010)*
166 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
172 *(Released Mon, 4 Oct 2010)*
174 - Fixed regression in ``gnt-instance rename``
180 *(Released Wed, 22 Sep 2010)*
182 - Fixed OS_VARIANT variable for OS scripts
183 - Fixed cluster tag operations via RAPI
184 - Made ``setup-ssh`` exit with non-zero code if an error occurred
185 - Disabled RAPI CA checks in watcher
191 *(Released Mon, 23 Aug 2010)*
193 - Support DRBD versions of the format "a.b.c.d"
195 - Re-introduce support for usage from multiple threads in RAPI client
196 - Instance renames and modify via RAPI
197 - Work around race condition between processing and archival in job
199 - Mark opcodes following failed one as failed, too
200 - Job field ``lock_status`` was removed due to difficulties making it
201 work with the changed job queue in Ganeti 2.2; a better way to monitor
202 locks is expected for a later 2.2.x release
203 - Fixed dry-run behaviour with many commands
204 - Support ``ssh-agent`` again when adding nodes
205 - Many additional bugfixes
211 *(Released Fri, 30 Jul 2010)*
213 Important change: the internal RPC mechanism between Ganeti nodes has
214 changed from using a home-grown http library (based on the Python base
215 libraries) to use the PycURL library. This requires that PycURL is
216 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
217 against GnuTLS by default. cURL's support for GnuTLS had known issues
218 before cURL 7.21.0 and we recommend using the latest cURL release or
219 linking against OpenSSL. Most other distributions already link PycURL
220 and cURL against OpenSSL. The command::
222 python -c 'import pycurl; print pycurl.version'
224 can be used to determine the libraries PycURL and cURL are linked
227 Other significant changes:
229 - Rewrote much of the internals of the job queue, in order to achieve
230 better parallelism; this decouples job query operations from the job
231 processing, and it should allow much nicer behaviour of the master
232 daemon under load, and it also has uncovered some long-standing bugs
233 related to the job serialisation (now fixed)
234 - Added a default iallocator setting to the cluster parameters,
235 eliminating the need to always pass nodes or an iallocator for
236 operations that require selection of new node(s)
237 - Added experimental support for the LXC virtualization method
238 - Added support for OS parameters, which allows the installation of
239 instances to pass parameter to OS scripts in order to customise the
241 - Added a hypervisor parameter controlling the migration type (live or
242 non-live), since hypervisors have various levels of reliability; this
243 has renamed the 'live' parameter to 'mode'
244 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
245 logical volumes, meaning that cluster verify will ignore them and not
246 flag their presence as errors
247 - The watcher will now reset the error count for failed instances after
248 8 hours, thus allowing self-healing if the problem that caused the
249 instances to be down/fail to start has cleared in the meantime
250 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
251 check for, and warn, if the drbd module parameter ``usermode_helper``
252 is not consistent with the cluster-wide setting; this is needed to
253 make diagnose easier of failed drbd creations
254 - Started adding base IPv6 support, but this is not yet
255 enabled/available for use
256 - Rename operations (cluster, instance) will now return the new name,
257 which is especially useful if a short name was passed in
258 - Added support for instance migration in RAPI
259 - Added a tool to pre-configure nodes for the SSH setup, before joining
260 them to the cluster; this will allow in the future a simplified model
261 for node joining (but not yet fully enabled in 2.2); this needs the
262 paramiko python library
263 - Fixed handling of name-resolving errors
264 - Fixed consistency of job results on the error path
265 - Fixed master-failover race condition when executed multiple times in
267 - Fixed many bugs related to the job queue (mostly introduced during the
268 2.2 development cycle, so not all are impacting 2.1)
269 - Fixed instance migration with missing disk symlinks
270 - Fixed handling of unknown jobs in ``gnt-job archive``
271 - And many other small fixes/improvements
275 - Enhanced both the unittest and the QA coverage
276 - Switched the opcode validation to a generic model, and extended the
277 validation to all opcode parameters
278 - Changed more parts of the code that write shell scripts to use the
280 - Switched the master daemon to use the asyncore library for the Luxi
287 *(Released Thu, 17 Jun 2010)*
289 - Added tool (``move-instance``) and infrastructure to move instances
290 between separate clusters (see :doc:`separate documentation
291 <move-instance>` and :doc:`design document <design-2.2>`)
292 - Added per-request RPC timeout
293 - RAPI now requires a Content-Type header for requests with a body (e.g.
294 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
295 :rfc:`2616` (HTTP/1.1), section 7.2.1)
296 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
298 - Implemented initial support for running Ganeti daemons as separate
299 users, see configure-time flags ``--with-user-prefix`` and
300 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
302 - Instances can be removed after export (``gnt-backup export
304 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
305 key (instead of 1024 bit)
306 - Added new cluster configuration file for cluster domain secret
307 - Import/export now use SSL instead of SSH
308 - Added support for showing estimated time when exporting an instance,
309 see the ``ganeti-os-interface(7)`` manpage and look for
316 *(Released Tue, 16 Nov 2010)*
318 Some more bugfixes. Unless critical bugs occur, this will be the last
321 - Fix case of MAC special-values
322 - Fix mac checker regex
323 - backend: Fix typo causing “out of range” error
324 - Add missing --units in gnt-instance list man page
330 *(Released Tue, 24 Aug 2010)*
333 - Don't ignore secondary node silently on non-mirrored disk templates
335 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
336 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
337 - Properly document the usb_mouse parameter
338 - Fix path in ganeti-rapi(8) (issue 116)
339 - Adjust error message when the ganeti user's .ssh directory is
341 - Add same-node-check when changing the disk template to drbd
347 *(Released Fri, 16 Jul 2010)*
350 - Add an option to only select some reboot types during qa/burnin.
351 (on some hypervisors consequent reboots are not supported)
352 - Fix infrequent race condition in master failover. Sometimes the old
353 master ip address would be still detected as up for a short time
354 after it was removed, causing failover to fail.
355 - Decrease mlockall warnings when the ctypes module is missing. On
356 Python 2.4 we support running even if no ctypes module is installed,
357 but we were too verbose about this issue.
358 - Fix building on old distributions, on which man doesn't have a
360 - Fix RAPI not to ignore the MAC address on instance creation
361 - Implement the old instance creation format in the RAPI client.
367 *(Released Thu, 01 Jul 2010)*
369 A small bugfix release:
370 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
371 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
372 a non-existing option
373 - Fix instance create: the --force-variant option was ignored
374 - Improve pylint 0.21 compatibility and warnings with Python 2.6
375 - Fix modify node storage with non-FQDN arguments
376 - Fix RAPI client to authenticate under Python 2.6 when used
377 for more than 5 requests needing authentication
378 - Fix gnt-instance modify -t (storage) giving a wrong error message
379 when converting a non-shutdown drbd instance to plain
385 *(Released Fri, 18 Jun 2010)*
387 A small bugfix release:
389 - Fix live migration of KVM instances started with older Ganeti
390 versions which had fewer hypervisor parameters
391 - Fix gnt-instance grow-disk on down instances
392 - Fix an error-reporting bug during instance migration
393 - Better checking of the ``--net`` and ``--disk`` values, to avoid
394 silently ignoring broken ones
395 - Fix an RPC error reporting bug affecting, for example, RAPI client
397 - Fix bug triggered by different API version os-es on different nodes
398 - Fix a bug in instance startup with custom hvparams: OS level
399 parameters would fail to be applied.
400 - Fix the RAPI client under Python 2.6 (but more work is needed to
401 make it work completely well with OpenSSL)
402 - Fix handling of errors when resolving names from DNS
408 *(Released Thu, 3 Jun 2010)*
410 A medium sized development cycle. Some new features, and some
411 fixes/small improvements/cleanups.
416 The node deamon now tries to mlock itself into memory, unless the
417 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
418 its logs, and falls back to console logging. This allows emergency
419 features such as ``gnt-node powercycle`` to work even in the event of a
420 broken node disk (tested offlining the disk hosting the node's
421 filesystem and dropping its memory caches; don't try this at home)
423 KVM: add vhost-net acceleration support. It can be tested with a new
424 enough version of the kernel and of qemu-kvm.
426 KVM: Add instance chrooting feature. If you use privilege dropping for
427 your VMs you can also now force them to chroot to an empty directory,
428 before starting the emulated guest.
430 KVM: Add maximum migration bandwith and maximum downtime tweaking
431 support (requires a new-enough version of qemu-kvm).
433 Cluster verify will now warn if the master node doesn't have the master
436 Add a new (incompatible) instance creation request format to RAPI which
437 supports all parameters (previously only a subset was supported, and it
438 wasn't possible to extend the old format to accomodate all the new
439 features. The old format is still supported, and a client can check for
440 this feature, before using it, by checking for its presence in the
441 ``features`` RAPI resource.
443 Now with ancient latin support. Try it passing the ``--roman`` option to
444 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
445 (requires the python-roman module to be installed, in order to work).
450 As usual many internal code refactorings, documentation updates, and
453 - Lots of improvements and cleanups to the experimental Remote API
454 (RAPI) client library.
455 - A new unit test suite for the core daemon libraries.
456 - A fix to creating missing directories makes sure the umask is not
457 applied anymore. This enforces the same directory permissions
459 - Better handling terminating daemons with ctrl+c (used when running
460 them in debugging mode).
461 - Fix a race condition in live migrating a KVM instance, when stat()
462 on the old proc status file returned EINVAL, which is an unexpected
464 - Fixed manpage checking with newer man and utf-8 charachters. But now
465 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
471 *(Released Fri, 7 May 2010)*
473 Fix a bug which prevented untagged KVM instances from starting.
479 *(Released Fri, 7 May 2010)*
481 Another release with a long development cycle, during which many
482 different features were added.
487 The KVM hypervisor now can run the individual instances as non-root, to
488 reduce the impact of a VM being hijacked due to bugs in the
489 hypervisor. It is possible to run all instances as a single (non-root)
490 user, to manually specify a user for each instance, or to dynamically
491 allocate a user out of a cluster-wide pool to each instance, with the
492 guarantee that no two instances will run under the same user ID on any
495 An experimental RAPI client library, that can be used standalone
496 (without the other Ganeti libraries), is provided in the source tree as
497 ``lib/rapi/client.py``. Note this client might change its interface in
498 the future, as we iterate on its capabilities.
500 A new command, ``gnt-cluster renew-crypto`` has been added to easily
501 replace the cluster's certificates and crypto keys. This might help in
502 case they have been compromised, or have simply expired.
504 A new disk option for instance creation has been added that allows one
505 to "adopt" currently existing logical volumes, with data
506 preservation. This should allow easier migration to Ganeti from
507 unmanaged (or managed via other software) instances.
509 Another disk improvement is the possibility to convert between redundant
510 (DRBD) and plain (LVM) disk configuration for an instance. This should
511 allow better scalability (starting with one node and growing the
512 cluster, or shrinking a two-node cluster to one node).
514 A new feature that could help with automated node failovers has been
515 implemented: if a node sees itself as offline (by querying the master
516 candidates), it will try to shutdown (hard) all instances and any active
517 DRBD devices. This reduces the risk of duplicate instances if an
518 external script automatically failovers the instances on such nodes. To
519 enable this, the cluster parameter ``maintain_node_health`` should be
520 enabled; in the future this option (per the name) will enable other
521 automatic maintenance features.
523 Instance export/import now will reuse the original instance
524 specifications for all parameters; that means exporting an instance,
525 deleting it and the importing it back should give an almost identical
526 instance. Note that the default import behaviour has changed from
527 before, where it created only one NIC; now it recreates the original
530 Cluster verify has added a few new checks: SSL certificates validity,
531 /etc/hosts consistency across the cluster, etc.
536 As usual, many internal changes were done, documentation fixes,
539 - Fixed cluster initialization with disabled cluster storage (regression
541 - File-based storage supports growing the disks
542 - Fixed behaviour of node role changes
543 - Fixed cluster verify for some corner cases, plus a general rewrite of
544 cluster verify to allow future extension with more checks
545 - Fixed log spamming by watcher and node daemon (regression introduced
547 - Fixed possible validation issues when changing the list of enabled
549 - Fixed cleanup of /etc/hosts during node removal
550 - Fixed RAPI response for invalid methods
551 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
552 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
553 from ide disks, etc.)
554 - Allow OS changes without re-installation (to record a changed OS
555 outside of Ganeti, or to allow OS renames)
556 - Allow instance creation without OS installation (useful for example if
557 the OS will be installed manually, or restored from a backup not in
559 - Implemented option to make cluster ``copyfile`` use the replication
561 - Added list of enabled hypervisors to ssconf (possibly useful for
563 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
565 - A partial form of node re-IP is possible via node readd, which now
566 allows changed node primary IP
567 - Command line utilities now show an informational message if the job is
569 - The logs of the master daemon now show the PID/UID/GID of the
576 *(Released Fri, 12 Mar 2010)*
578 During the 2.1.0 long release candidate cycle, a lot of improvements and
579 changes have accumulated with were released later as 2.1.1.
584 The node evacuate command (``gnt-node evacuate``) was significantly
585 rewritten, and as such the IAllocator protocol was changed - a new
586 request type has been added. This unfortunate change during a stable
587 series is designed to improve performance of node evacuations; on
588 clusters with more than about five nodes and which are well-balanced,
589 evacuation should proceed in parallel for all instances of the node
590 being evacuated. As such, any existing IAllocator scripts need to be
591 updated, otherwise the above command will fail due to the unknown
592 request. The provided "dumb" allocator has not been updated; but the
593 ganeti-htools package supports the new protocol since version 0.2.4.
595 Another important change is increased validation of node and instance
596 names. This might create problems in special cases, if invalid host
597 names are being used.
599 Also, a new layer of hypervisor parameters has been added, that sits at
600 OS level between the cluster defaults and the instance ones. This allows
601 customisation of virtualization parameters depending on the installed
602 OS. For example instances with OS 'X' may have a different KVM kernel
603 (or any other parameter) than the cluster defaults. This is intended to
604 help managing a multiple OSes on the same cluster, without manual
605 modification of each instance's parameters.
607 A tool for merging clusters, ``cluster-merge``, has been added in the
613 - Improved the int/float conversions that should make the code more
614 robust in face of errors from the node daemons
615 - Fixed the remove node code in case of internal configuration errors
616 - Fixed the node daemon behaviour in face of inconsistent queue
617 directory (e.g. read-only file-system where we can't open the files
619 - Fixed the behaviour of gnt-node modify for master candidate demotion;
620 now it either aborts cleanly or, if given the new “auto_promote”
621 parameter, will automatically promote other nodes as needed
622 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
623 completely prevent Ganeti from working
624 - Fixed bug for instance export when not all disks were successfully
626 - Fixed behaviour of node add when the new node is slow in starting up
628 - Fixed handling of signals in the LUXI client, which should improve
629 behaviour of command-line scripts
630 - Added checks for invalid node/instance names in the configuration (now
631 flagged during cluster verify)
632 - Fixed watcher behaviour for disk activation errors
633 - Fixed two potentially endless loops in http library, which led to the
634 RAPI daemon hanging and consuming 100% CPU in some cases
635 - Fixed bug in RAPI daemon related to hashed passwords
636 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
638 - Enhanced compatibility with non-Debian OSes, but not using absolute
639 path in some commands and allowing customisation of the ssh
640 configuration directory
641 - Fixed possible future issue with new Python versions by abiding to the
642 proper use of ``__slots__`` attribute on classes
643 - Added checks that should prevent directory traversal attacks
644 - Many documentation fixes based on feedback from users
649 - Added an “early_release” more for instance replace disks and node
650 evacuate, where we release locks earlier and thus allow higher
651 parallelism within the cluster
652 - Added watcher hooks, intended to allow the watcher to restart other
653 daemons (e.g. from the ganeti-nbma project), but they can be used of
654 course for any other purpose
655 - Added a compile-time disable for DRBD barriers, to increase
656 performance if the administrator trusts the power supply or the
657 storage system to not lose writes
658 - Added the option of using syslog for logging instead of, or in
659 addition to, Ganeti's own log files
660 - Removed boot restriction for paravirtual NICs for KVM, recent versions
661 can indeed boot from a paravirtual NIC
662 - Added a generic debug level for many operations; while this is not
663 used widely yet, it allows one to pass the debug value all the way to
665 - Enhanced the hooks environment for instance moves (failovers,
666 migrations) where the primary/secondary nodes changed during the
667 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
668 - Enhanced data validations for many user-supplied values; one important
669 item is the restrictions imposed on instance and node names, which
670 might reject some (invalid) host names
671 - Add a configure-time option to disable file-based storage, if it's not
672 needed; this allows greater security separation between the master
673 node and the other nodes from the point of view of the inter-node RPC
675 - Added user notification in interactive tools if job is waiting in the
676 job queue or trying to acquire locks
677 - Added log messages when a job is waiting for locks
678 - Added filtering by node tags in instance operations which admit
679 multiple instances (start, stop, reboot, reinstall)
680 - Added a new tool for cluster mergers, ``cluster-merge``
681 - Parameters from command line which are of the form ``a=b,c=d`` can now
682 use backslash escapes to pass in values which contain commas,
683 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
685 - For KVM, the instance name is the first parameter passed to KVM, so
686 that it's more visible in the process list
692 *(Released Tue, 2 Mar 2010)*
694 Ganeti 2.1 brings many improvements with it. Major changes:
696 - Added infrastructure to ease automated disk repairs
697 - Added new daemon to export configuration data in a cheaper way than
699 - Instance NICs can now be routed instead of being associated with a
701 - Improved job locking logic to reduce impact of jobs acquiring multiple
702 locks waiting for other long-running jobs
704 In-depth implementation details can be found in the Ganeti 2.1 design
710 - Added chroot hypervisor
711 - Added more options to xen-hvm hypervisor (``kernel_path`` and
713 - Added more options to xen-pvm hypervisor (``use_bootloader``,
714 ``bootloader_path`` and ``bootloader_args``)
715 - Added the ``use_localtime`` option for the xen-hvm and kvm
716 hypervisors, and the default value for this has changed to false (in
717 2.0 xen-hvm always enabled it)
718 - Added luxi call to submit multiple jobs in one go
719 - Added cluster initialization option to not modify ``/etc/hosts``
721 - Added network interface parameters
722 - Added dry run mode to some LUs
723 - Added RAPI resources:
725 - ``/2/instances/[instance_name]/info``
726 - ``/2/instances/[instance_name]/replace-disks``
727 - ``/2/nodes/[node_name]/evacuate``
728 - ``/2/nodes/[node_name]/migrate``
729 - ``/2/nodes/[node_name]/role``
730 - ``/2/nodes/[node_name]/storage``
731 - ``/2/nodes/[node_name]/storage/modify``
732 - ``/2/nodes/[node_name]/storage/repair``
734 - Added OpCodes to evacuate or migrate all instances on a node
735 - Added new command to list storage elements on nodes (``gnt-node
736 list-storage``) and modify them (``gnt-node modify-storage``)
737 - Added new ssconf files with master candidate IP address
738 (``ssconf_master_candidates_ips``), node primary IP address
739 (``ssconf_node_primary_ips``) and node secondary IP address
740 (``ssconf_node_secondary_ips``)
741 - Added ``ganeti-confd`` and a client library to query the Ganeti
742 configuration via UDP
743 - Added ability to run hooks after cluster initialization and before
745 - Added automatic mode for disk replace (``gnt-instance replace-disks
747 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
748 after catastrophic data-loss
749 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
751 - Added ``gnt-instance move`` command to move instances
752 - Added ``gnt-cluster watcher`` command to control watcher
753 - Added ``gnt-node powercycle`` command to powercycle nodes
754 - Added new job status field ``lock_status``
755 - Added parseable error codes to cluster verification (``gnt-cluster
756 verify --error-codes``) and made output less verbose (use
757 ``--verbose`` to restore previous behaviour)
758 - Added UUIDs to the main config entities (cluster, nodes, instances)
759 - Added support for OS variants
760 - Added support for hashed passwords in the Ganeti remote API users file
762 - Added option to specify maximum timeout on instance shutdown
763 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
764 - Added new helper script to start and stop Ganeti daemons
765 (``daemon-util``), with the intent to reduce the work necessary to
766 adjust Ganeti for non-Debian distributions and to start/stop daemons
768 - Added more unittests
769 - Fixed critical bug in ganeti-masterd startup
770 - Removed the configure-time ``kvm-migration-port`` parameter, this is
771 now customisable at the cluster level for both the KVM and Xen
772 hypervisors using the new ``migration_port`` parameter
773 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
774 reinstalling an instance
775 - Allowed ``@`` in tag names
776 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
777 - Many documentation updates
778 - Distribute hypervisor files on ``gnt-cluster redist-conf``
779 - ``gnt-instance reinstall`` can now reinstall multiple instances
780 - Updated many command line parameters
781 - Introduced new OS API version 15
782 - No longer support a default hypervisor
783 - Treat virtual LVs as inexistent
784 - Improved job locking logic to reduce lock contention
785 - Match instance and node names case insensitively
786 - Reimplemented bash completion script to be more complete
793 *(Released Thu, 4 Feb 2010)*
795 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
796 - Fix a string formatting bug
797 - Improve safety of the code in some error paths
798 - Improve data validation in the master of values returned from nodes
804 *(Released Thu, 17 Dec 2009)*
806 - Fix security issue due to missing validation of iallocator names; this
807 allows local and remote execution of arbitrary executables
808 - Fix failure of gnt-node list during instance removal
809 - Ship the RAPI documentation in the archive
815 *(Released Wed, 30 Sep 2009)*
817 - Fixed many wrong messages
818 - Fixed a few bugs related to the locking library
819 - Fixed MAC checking at instance creation time
820 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
821 - Fixed a few issues related to signal handling in both daemons and
823 - Fixed the example startup script provided
824 - Fixed insserv dependencies in the example startup script (patch from
826 - Fixed handling of drained nodes in the iallocator framework
827 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
829 - Fixed error related to invalid job IDs in job polling
830 - Fixed job/opcode persistence on unclean master shutdown
831 - Fixed handling of partial job processing after unclean master
833 - Fixed error reporting from LUs, previously all errors were converted
834 into execution errors
835 - Fixed error reporting from burnin
836 - Decreased significantly the memory usage of the job queue
837 - Optimised slightly multi-job submission
838 - Optimised slightly opcode loading
839 - Backported the multi-job submit framework from the development
840 branch; multi-instance start and stop should be faster
841 - Added script to clean archived jobs after 21 days; this will reduce
842 the size of the queue directory
843 - Added some extra checks in disk size tracking
844 - Added an example ethers hook script
845 - Added a cluster parameter that prevents Ganeti from modifying of
847 - Added more node information to RAPI responses
848 - Added a ``gnt-job watch`` command that allows following the ouput of a
850 - Added a bind-address option to ganeti-rapi
851 - Added more checks to the configuration verify
852 - Enhanced the burnin script such that some operations can be retried
854 - Converted instance reinstall to multi-instance model
860 *(Released Fri, 7 Aug 2009)*
862 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
863 to allow using the pre-2.0.2 behaviour in activation, if any existing
864 instances have mismatched disk sizes in the configuration
865 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
866 any configuration mismatches for disk sizes
867 - Added ``gnt-master cluste-failover --no-voting`` to allow master
868 failover to work on two-node clusters
869 - Fixed the ``--net`` option of ``gnt-backup import``, which was
871 - Fixed detection of OS script errors in ``gnt-backup export``
872 - Fixed exit code of ``gnt-backup export``
878 *(Released Fri, 17 Jul 2009)*
880 - Added experimental support for stripped logical volumes; this should
881 enhance performance but comes with a higher complexity in the block
882 device handling; stripping is only enabled when passing
883 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
884 affected even in the non-stripped mode
885 - Improved resiliency against transient failures at the end of DRBD
886 resyncs, and in general of DRBD resync checks
887 - Fixed a couple of issues with exports and snapshot errors
888 - Fixed a couple of issues in instance listing
889 - Added display of the disk size in ``gnt-instance info``
890 - Fixed checking for valid OSes in instance creation
891 - Fixed handling of the "vcpus" parameter in instance listing and in
892 general of invalid parameters
893 - Fixed http server library, and thus RAPI, to handle invalid
894 username/password combinations correctly; this means that now they
895 report unauthorized for queries too, not only for modifications,
896 allowing earlier detect of configuration problems
897 - Added a new "role" node list field, equivalent to the master/master
898 candidate/drained/offline flags combinations
899 - Fixed cluster modify and changes of candidate pool size
900 - Fixed cluster verify error messages for wrong files on regular nodes
901 - Fixed a couple of issues with node demotion from master candidate role
902 - Fixed node readd issues
903 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
904 - Added a new ``--no-voting`` option for masterfailover to fix failover
905 on two-nodes clusters when the former master node is unreachable
906 - Added instance reinstall over RAPI
912 *(Released Tue, 16 Jun 2009)*
914 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
915 allow re-adding the start in single-user option (regression from 1.2)
916 - the watcher writes the instance status to a file, to allow monitoring
917 to report the instance status (from the master) based on cached
918 results of the watcher's queries; while this can get stale if the
919 watcher is being locked due to other work on the cluster, this is
921 - the watcher now also restarts the node daemon and the rapi daemon if
923 - fixed the watcher to handle full and drained queue cases
924 - hooks export more instance data in the environment, which helps if
925 hook scripts need to take action based on the instance's properties
926 (no longer need to query back into ganeti)
927 - instance failovers when the instance is stopped do not check for free
928 RAM, so that failing over a stopped instance is possible in low memory
930 - rapi uses queries for tags instead of jobs (for less job traffic), and
931 for cluster tags it won't talk to masterd at all but read them from
933 - a couple of error handling fixes in RAPI
934 - drbd handling: improved the error handling of inconsistent disks after
935 resync to reduce the frequency of "there are some degraded disks for
936 this instance" messages
937 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
938 error handling path called a wrong function name)
944 *(Released Wed, 27 May 2009)*
946 - no changes from rc5
949 Version 2.0 release candidate 5
950 -------------------------------
952 *(Released Wed, 20 May 2009)*
954 - fix a couple of bugs (validation, argument checks)
955 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
956 - some small improvements to RAPI and IAllocator
957 - make watcher automatically start the master daemon if down
960 Version 2.0 release candidate 4
961 -------------------------------
963 *(Released Mon, 27 Apr 2009)*
965 - change the OS list to not require locks; this helps with big clusters
966 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
967 volume group is broken
968 - ``gnt-instance info``, without any arguments, doesn't run for all
969 instances anymore; either pass ``--all`` or pass the desired
970 instances; this helps against mistakes on big clusters where listing
971 the information for all instances takes a long time
972 - miscellaneous doc and man pages fixes
975 Version 2.0 release candidate 3
976 -------------------------------
978 *(Released Wed, 8 Apr 2009)*
980 - Change the internal locking model of some ``gnt-node`` commands, in
981 order to reduce contention (and blocking of master daemon) when
982 batching many creation/reinstall jobs
983 - Fixes to Xen soft reboot
984 - No longer build documentation at build time, instead distribute it in
985 the archive, in order to reduce the need for the whole docbook/rst
989 Version 2.0 release candidate 2
990 -------------------------------
992 *(Released Fri, 27 Mar 2009)*
994 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
995 - Fix watcher startup sequence, improves the behaviour of busy clusters
996 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
997 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
998 bind address and other places
999 - Some documentation fixes and updates
1002 Version 2.0 release candidate 1
1003 -------------------------------
1005 *(Released Mon, 2 Mar 2009)*
1007 - More documentation updates, now all docs should be more-or-less
1009 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
1011 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
1013 - Fix an issue related to $libdir/run/ganeti and cluster creation
1019 *(Released Thu, 19 Feb 2009)*
1021 - Xen PVM and KVM have switched the default value for the instance root
1022 disk to the first partition on the first drive, instead of the whole
1023 drive; this means that the OS installation scripts must be changed
1025 - Man pages have been updated
1026 - RAPI has been switched by default to HTTPS, and the exported functions
1027 should all work correctly
1028 - RAPI v1 has been removed
1029 - Many improvements to the KVM hypervisor
1030 - Block device errors are now better reported
1031 - Many other bugfixes and small improvements
1037 *(Released Mon, 26 Jan 2009)*
1039 - Version 2 is a general rewrite of the code and therefore the
1040 differences are too many to list, see the design document for 2.0 in
1041 the ``doc/`` subdirectory for more details
1042 - In this beta version there is not yet a migration path from 1.2 (there
1043 will be one in the final 2.0 release)
1044 - A few significant changes are:
1046 - all commands are executed by a daemon (``ganeti-masterd``) and the
1047 various ``gnt-*`` commands are just front-ends to it
1048 - all the commands are entered into, and executed from a job queue,
1049 see the ``gnt-job(8)`` manpage
1050 - the RAPI daemon supports read-write operations, secured by basic
1051 HTTP authentication on top of HTTPS
1052 - DRBD version 0.7 support has been removed, DRBD 8 is the only
1053 supported version (when migrating from Ganeti 1.2 to 2.0, you need
1054 to migrate to DRBD 8 first while still running Ganeti 1.2)
1055 - DRBD devices are using statically allocated minor numbers, which
1056 will be assigned to existing instances during the migration process
1057 - there is support for both Xen PVM and Xen HVM instances running on
1059 - KVM virtualization is supported too
1060 - file-based storage has been implemented, which means that it is
1061 possible to run the cluster without LVM and DRBD storage, for
1062 example using a shared filesystem exported from shared storage (and
1063 still have live migration)
1069 *(Released Tue, 13 Jan 2009)*
1071 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
1072 - Reuse the old instance mac address by default on instance import, if
1073 the instance name is the same.
1074 - Handle situations in which the node info rpc returns incomplete
1076 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
1077 - Improved version of batcher:
1079 - state file support
1080 - instance mac address support
1081 - support for HVM clusters/instances
1083 - Add an option to show the number of cpu sockets and nodes in
1085 - Support OSes that handle more than one version of the OS api (but do
1086 not change the current API in any other way)
1087 - Fix ``gnt-node migrate``
1088 - ``gnt-debug`` man page
1089 - Fixes various more typos and small issues
1090 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
1096 *(Released Wed, 24 Sep 2008)*
1098 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
1099 type of disk exported to fully virtualized instances.
1100 - provide access to the serial console of HVM instances
1101 - instance auto_balance flag, set by default. If turned off it will
1102 avoid warnings on cluster verify if there is not enough memory to fail
1103 over an instance. in the future it will prevent automatically failing
1104 it over when we will support that.
1105 - batcher tool for instance creation, see ``tools/README.batcher``
1106 - ``gnt-instance reinstall --select-os`` to interactively select a new
1107 operating system when reinstalling an instance.
1108 - when changing the memory amount on instance modify a check has been
1109 added that the instance will be able to start. also warnings are
1110 emitted if the instance will not be able to fail over, if auto_balance
1112 - documentation fixes
1113 - sync fields between ``gnt-instance list/modify/add/import``
1114 - fix a race condition in drbd when the sync speed was set after giving
1115 the device a remote peer.
1121 *(Released Tue, 22 Jul 2008)*
1123 - note: the allowed size and number of tags per object were reduced
1124 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1125 - fixed twisted 8.x compatibility
1126 - fixed ``gnt-instance replace-disks`` with iallocator
1127 - add TCP keepalives on twisted connections to detect restarted nodes
1128 - disk increase support, see ``gnt-instance grow-disk``
1129 - implement bulk node/instance query for RAPI
1130 - add tags in node/instance listing (optional)
1131 - experimental migration (and live migration) support, read the man page
1132 for ``gnt-instance migrate``
1133 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1134 has some small improvements in handling its state file
1140 *(Released Fri, 13 Jun 2008)*
1142 - Experimental readonly, REST-based remote API implementation;
1143 automatically started on master node, TCP port 5080, if enabled by
1144 ``--enable-rapi`` parameter to configure script.
1145 - Instance allocator support. Add and import instance accept a
1146 ``--iallocator`` parameter, and call that instance allocator to decide
1147 which node to use for the instance. The iallocator document describes
1148 what's expected from an allocator script.
1149 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1150 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1151 node is lost there is still enough memory to fail over the instances
1153 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1154 ``gnt-cluster verify``, to check for site-specific compliance. All the
1155 hooks will run, and their output, if any, will be displayed. Any
1156 failing hook will make the verification return an error value.
1157 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1158 primary and secondary interfaces
1159 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1160 that is still declared as part of the cluster and has failed.
1161 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1162 specifying output fields, that just adds the chosen fields to the
1164 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1165 export from the filesystem.
1166 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1167 have been added. Using them you can enable/disable acpi and pae
1168 support, and specify a path for a cd image to be exported to the
1169 instance. These parameters as the name suggest only work on HVM
1171 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1172 PAE support will be set to the previously hardcoded values, but the
1173 (previously hardcoded) path to the CDROM ISO image will be unset and
1174 if required, needs to be set manually with ``gnt-instance modify``
1176 - The address to which an instance's VNC console is bound is now
1177 selectable per-instance, rather than being cluster wide. Of course
1178 this only applies to instances controlled via VNC, so currently just
1179 applies to HVM clusters.
1185 *(Released Mon, 18 Feb 2008)*
1187 - more tweaks to the disk activation code (especially helpful for DRBD)
1188 - change the default ``gnt-instance list`` output format, now there is
1189 one combined status field (see the manpage for the exact values this
1191 - some more fixes for the mac export to hooks change
1192 - make Ganeti not break with DRBD 8.2.x (which changed the version
1193 format in ``/proc/drbd``) (issue 24)
1194 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1195 template, allowing migration from DRBD0.7+MD to DRBD8
1201 *(Released Wed, 30 Jan 2008)*
1203 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1205 - add command aliases infrastructure and a few aliases
1206 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1207 man pages and the ``--help`` option of ``gnt-node
1208 list``/``gnt-instance list``
1209 - fix ``gnt-backup list`` with down nodes (issue 21)
1210 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1211 - fix the dist archive and add a check for including svn/git files in
1213 - some developer-related changes: improve the burnin and the QA suite,
1214 add an upload script for testing during development
1220 *(Released Wed, 16 Jan 2008)*
1222 - experimental HVM support, read the install document, section
1223 "Initializing the cluster"
1224 - allow for the PVM hypervisor per-instance kernel and initrd paths
1225 - add a new command ``gnt-cluster verify-disks`` which uses a new
1226 algorithm to improve the reconnection of the DRBD pairs if the device
1227 on the secondary node has gone away
1228 - make logical volume code auto-activate LVs at disk activation time
1229 - slightly improve the speed of activating disks
1230 - allow specification of the MAC address at instance creation time, and
1231 changing it later via ``gnt-instance modify``
1232 - fix handling of external commands that generate lots of output on
1234 - update documentation with regard to minimum version of DRBD8 supported
1240 *(Released Tue, 4 Dec 2007)*
1242 - Log the ``xm create`` output to the node daemon log on failure (to
1243 help diagnosing the error)
1244 - In debug mode, log all external commands output if failed to the logs
1245 - Change parsing of lvm commands to ignore stderr
1251 *(Released Wed, 28 Nov 2007)*
1253 - Another round of updates to the DRBD 8 code to deal with more failures
1254 in the replace secondary node operation
1255 - Some more logging of failures in disk operations (lvm, drbd)
1256 - A few documentation updates
1263 *(Released Tue, 13 Nov 2007)*
1265 - Change configuration file format from Python's Pickle to JSON.
1266 Upgrading is possible using the cfgupgrade utility.
1267 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1268 faster replace disks and is more stable (DRBD 8 has many improvements
1269 compared to DRBD 0.7)
1270 - Added command line tags support (see man pages for ``gnt-instance``,
1271 ``gnt-node``, ``gnt-cluster``)
1272 - Added instance rename support
1273 - Added multi-instance startup/shutdown
1274 - Added cluster rename support
1275 - Added ``gnt-node evacuate`` to simplify some node operations
1276 - Added instance reboot operation that can speedup reboot as compared to
1278 - Soften the requirement that hostnames are in FQDN format
1279 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1281 - Removed dependency on debian's patched fping that uses the
1282 non-standard ``-S`` option
1283 - Now the OS definitions are searched for in multiple, configurable
1284 paths (easier for distros to package)
1285 - Some changes to the hooks infrastructure (especially the new
1286 post-configuration update hook)
1287 - Other small bugfixes
1289 .. vim: set textwidth=72 syntax=rst :