7 *(Released Fri, 30 Jul 2010)*
9 Important change: the internal RPC mechanism between Ganeti nodes has
10 changed from using a home-grown http library (based on the Python base
11 libraries) to use the PycURL library. This requires that PycURL is
12 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
13 against GnuTLS by default. cURL's support for GnuTLS had known issues
14 before cURL 7.21.0 and we recommend using the latest cURL release or
15 linking against OpenSSL. Most other distributions already link PycURL
16 and cURL against OpenSSL. The command::
18 python -c 'import pycurl; print pycurl.version'
20 can be used to determine the libraries PycURL and cURL are linked
23 Other significant changes:
25 - Rewrote much of the internals of the job queue, in order to achieve
26 better parallelism; this decouples job query operations from the job
27 processing, and it should allow much nicer behaviour of the master
28 daemon under load, and it also has uncovered some long-standing bugs
29 related to the job serialisation (now fixed)
30 - Added a default iallocator setting to the cluster parameters,
31 eliminating the need to always pass nodes or an iallocator for
32 operations that require selection of new node(s)
33 - Added experimental support for the LXC virtualization method
34 - Added support for OS parameters, which allows the installation of
35 instances to pass parameter to OS scripts in order to customise the
37 - Added a hypervisor parameter controlling the migration type (live or
38 non-live), since hypervisors have various levels of reliability; this
39 has renamed the 'live' parameter to 'mode'
40 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
41 logical volumes, meaning that cluster verify will ignore them and not
42 flag their presence as errors
43 - The watcher will now reset the error count for failed instances after
44 8 hours, thus allowing self-healing if the problem that caused the
45 instances to be down/fail to start has cleared in the meantime
46 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
47 check for, and warn, if the drbd module parameter ``usermode_helper``
48 is not consistent with the cluster-wide setting; this is needed to
49 make diagnose easier of failed drbd creations
50 - Started adding base IPv6 support, but this is not yet
51 enabled/available for use
52 - Rename operations (cluster, instance) will now return the new name,
53 which is especially useful if a short name was passed in
54 - Added support for instance migration in RAPI
55 - Added a tool to pre-configure nodes for the SSH setup, before joining
56 them to the cluster; this will allow in the future a simplified model
57 for node joining (but not yet fully enabled in 2.2); this needs the
58 paramiko python library
59 - Fixed handling of name-resolving errors
60 - Fixed consistency of job results on the error path
61 - Fixed master-failover race condition when executed multiple times in
63 - Fixed many bugs related to the job queue (mostly introduced during the
64 2.2 development cycle, so not all are impacting 2.1)
65 - Fixed instance migration with missing disk symlinks
66 - Fixed handling of unknown jobs in ``gnt-job archive``
67 - And many other small fixes/improvements
71 - Enhanced both the unittest and the QA coverage
72 - Switched the opcode validation to a generic model, and extended the
73 validation to all opcode parameters
74 - Changed more parts of the code that write shell scripts to use the
76 - Switched the master daemon to use the asyncore library for the Luxi
83 *(Released Thu, 17 Jun 2010)*
85 - Added tool (``move-instance``) and infrastructure to move instances
86 between separate clusters (see :doc:`separate documentation
87 <move-instance>` and :doc:`design document <design-2.2>`)
88 - Added per-request RPC timeout
89 - RAPI now requires a Content-Type header for requests with a body (e.g.
90 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
91 RFC2616 (HTTP/1.1), section 7.2.1)
92 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
94 - Implemented initial support for running Ganeti daemons as separate
95 users, see configure-time flags ``--with-user-prefix`` and
96 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
98 - Instances can be removed after export (``gnt-backup export
100 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
101 key (instead of 1024 bit)
102 - Added new cluster configuration file for cluster domain secret
103 - Import/export now use SSL instead of SSH
104 - Added support for showing estimated time when exporting an instance,
105 see the ``ganeti-os-interface(7)`` manpage and look for
112 *(Released Fri, 16 Jul 2010)*
115 - Add an option to only select some reboot types during qa/burnin.
116 (on some hypervisors consequent reboots are not supported)
117 - Fix infrequent race condition in master failover. Sometimes the old
118 master ip address would be still detected as up for a short time
119 after it was removed, causing failover to fail.
120 - Decrease mlockall warnings when the ctypes module is missing. On
121 Python 2.4 we support running even if no ctypes module is installed,
122 but we were too verbose about this issue.
123 - Fix building on old distributions, on which man doesn't have a
125 - Fix RAPI not to ignore the MAC address on instance creation
126 - Implement the old instance creation format in the RAPI client.
132 *(Released Thu, 01 Jul 2010)*
134 A small bugfix release:
135 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
136 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
137 a non-existing option
138 - Fix instance create: the --force-variant option was ignored
139 - Improve pylint 0.21 compatibility and warnings with Python 2.6
140 - Fix modify node storage with non-FQDN arguments
141 - Fix RAPI client to authenticate under Python 2.6 when used
142 for more than 5 requests needing authentication
143 - Fix gnt-instance modify -t (storage) giving a wrong error message
144 when converting a non-shutdown drbd instance to plain
150 *(Released Fri, 18 Jun 2010)*
152 A small bugfix release:
154 - Fix live migration of KVM instances started with older Ganeti
155 versions which had fewer hypervisor parameters
156 - Fix gnt-instance grow-disk on down instances
157 - Fix an error-reporting bug during instance migration
158 - Better checking of the ``--net`` and ``--disk`` values, to avoid
159 silently ignoring broken ones
160 - Fix an RPC error reporting bug affecting, for example, RAPI client
162 - Fix bug triggered by different API version os-es on different nodes
163 - Fix a bug in instance startup with custom hvparams: OS level
164 parameters would fail to be applied.
165 - Fix the RAPI client under Python 2.6 (but more work is needed to
166 make it work completely well with OpenSSL)
167 - Fix handling of errors when resolving names from DNS
173 *(Released Thu, 3 Jun 2010)*
175 A medium sized development cycle. Some new features, and some
176 fixes/small improvements/cleanups.
181 The node deamon now tries to mlock itself into memory, unless the
182 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
183 its logs, and falls back to console logging. This allows emergency
184 features such as ``gnt-node powercycle`` to work even in the event of a
185 broken node disk (tested offlining the disk hosting the node's
186 filesystem and dropping its memory caches; don't try this at home)
188 KVM: add vhost-net acceleration support. It can be tested with a new
189 enough version of the kernel and of qemu-kvm.
191 KVM: Add instance chrooting feature. If you use privilege dropping for
192 your VMs you can also now force them to chroot to an empty directory,
193 before starting the emulated guest.
195 KVM: Add maximum migration bandwith and maximum downtime tweaking
196 support (requires a new-enough version of qemu-kvm).
198 Cluster verify will now warn if the master node doesn't have the master
201 Add a new (incompatible) instance creation request format to RAPI which
202 supports all parameters (previously only a subset was supported, and it
203 wasn't possible to extend the old format to accomodate all the new
204 features. The old format is still supported, and a client can check for
205 this feature, before using it, by checking for its presence in the
206 ``features`` RAPI resource.
208 Now with ancient latin support. Try it passing the ``--roman`` option to
209 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
210 (requires the python-roman module to be installed, in order to work).
215 As usual many internal code refactorings, documentation updates, and
218 - Lots of improvements and cleanups to the experimental Remote API
219 (RAPI) client library.
220 - A new unit test suite for the core daemon libraries.
221 - A fix to creating missing directories makes sure the umask is not
222 applied anymore. This enforces the same directory permissions
224 - Better handling terminating daemons with ctrl+c (used when running
225 them in debugging mode).
226 - Fix a race condition in live migrating a KVM instance, when stat()
227 on the old proc status file returned EINVAL, which is an unexpected
229 - Fixed manpage checking with newer man and utf-8 charachters. But now
230 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
236 *(Released Fri, 7 May 2010)*
238 Fix a bug which prevented untagged KVM instances from starting.
244 *(Released Fri, 7 May 2010)*
246 Another release with a long development cycle, during which many
247 different features were added.
252 The KVM hypervisor now can run the individual instances as non-root, to
253 reduce the impact of a VM being hijacked due to bugs in the
254 hypervisor. It is possible to run all instances as a single (non-root)
255 user, to manually specify a user for each instance, or to dynamically
256 allocate a user out of a cluster-wide pool to each instance, with the
257 guarantee that no two instances will run under the same user ID on any
260 An experimental RAPI client library, that can be used standalone
261 (without the other Ganeti libraries), is provided in the source tree as
262 ``lib/rapi/client.py``. Note this client might change its interface in
263 the future, as we iterate on its capabilities.
265 A new command, ``gnt-cluster renew-crypto`` has been added to easily
266 replace the cluster's certificates and crypto keys. This might help in
267 case they have been compromised, or have simply expired.
269 A new disk option for instance creation has been added that allows one
270 to "adopt" currently existing logical volumes, with data
271 preservation. This should allow easier migration to Ganeti from
272 unmanaged (or managed via other software) instances.
274 Another disk improvement is the possibility to convert between redundant
275 (DRBD) and plain (LVM) disk configuration for an instance. This should
276 allow better scalability (starting with one node and growing the
277 cluster, or shrinking a two-node cluster to one node).
279 A new feature that could help with automated node failovers has been
280 implemented: if a node sees itself as offline (by querying the master
281 candidates), it will try to shutdown (hard) all instances and any active
282 DRBD devices. This reduces the risk of duplicate instances if an
283 external script automatically failovers the instances on such nodes. To
284 enable this, the cluster parameter ``maintain_node_health`` should be
285 enabled; in the future this option (per the name) will enable other
286 automatic maintenance features.
288 Instance export/import now will reuse the original instance
289 specifications for all parameters; that means exporting an instance,
290 deleting it and the importing it back should give an almost identical
291 instance. Note that the default import behaviour has changed from
292 before, where it created only one NIC; now it recreates the original
295 Cluster verify has added a few new checks: SSL certificates validity,
296 /etc/hosts consistency across the cluster, etc.
301 As usual, many internal changes were done, documentation fixes,
304 - Fixed cluster initialization with disabled cluster storage (regression
306 - File-based storage supports growing the disks
307 - Fixed behaviour of node role changes
308 - Fixed cluster verify for some corner cases, plus a general rewrite of
309 cluster verify to allow future extension with more checks
310 - Fixed log spamming by watcher and node daemon (regression introduced
312 - Fixed possible validation issues when changing the list of enabled
314 - Fixed cleanup of /etc/hosts during node removal
315 - Fixed RAPI response for invalid methods
316 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
317 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
318 from ide disks, etc.)
319 - Allow OS changes without re-installation (to record a changed OS
320 outside of Ganeti, or to allow OS renames)
321 - Allow instance creation without OS installation (useful for example if
322 the OS will be installed manually, or restored from a backup not in
324 - Implemented option to make cluster ``copyfile`` use the replication
326 - Added list of enabled hypervisors to ssconf (possibly useful for
328 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
330 - A partial form of node re-IP is possible via node readd, which now
331 allows changed node primary IP
332 - Command line utilities now show an informational message if the job is
334 - The logs of the master daemon now show the PID/UID/GID of the
341 *(Released Fri, 12 Mar 2010)*
343 During the 2.1.0 long release candidate cycle, a lot of improvements and
344 changes have accumulated with were released later as 2.1.1.
349 The node evacuate command (``gnt-node evacuate``) was significantly
350 rewritten, and as such the IAllocator protocol was changed - a new
351 request type has been added. This unfortunate change during a stable
352 series is designed to improve performance of node evacuations; on
353 clusters with more than about five nodes and which are well-balanced,
354 evacuation should proceed in parallel for all instances of the node
355 being evacuated. As such, any existing IAllocator scripts need to be
356 updated, otherwise the above command will fail due to the unknown
357 request. The provided "dumb" allocator has not been updated; but the
358 ganeti-htools package supports the new protocol since version 0.2.4.
360 Another important change is increased validation of node and instance
361 names. This might create problems in special cases, if invalid host
362 names are being used.
364 Also, a new layer of hypervisor parameters has been added, that sits at
365 OS level between the cluster defaults and the instance ones. This allows
366 customisation of virtualization parameters depending on the installed
367 OS. For example instances with OS 'X' may have a different KVM kernel
368 (or any other parameter) than the cluster defaults. This is intended to
369 help managing a multiple OSes on the same cluster, without manual
370 modification of each instance's parameters.
372 A tool for merging clusters, ``cluster-merge``, has been added in the
378 - Improved the int/float conversions that should make the code more
379 robust in face of errors from the node daemons
380 - Fixed the remove node code in case of internal configuration errors
381 - Fixed the node daemon behaviour in face of inconsistent queue
382 directory (e.g. read-only file-system where we can't open the files
384 - Fixed the behaviour of gnt-node modify for master candidate demotion;
385 now it either aborts cleanly or, if given the new “auto_promote”
386 parameter, will automatically promote other nodes as needed
387 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
388 completely prevent Ganeti from working
389 - Fixed bug for instance export when not all disks were successfully
391 - Fixed behaviour of node add when the new node is slow in starting up
393 - Fixed handling of signals in the LUXI client, which should improve
394 behaviour of command-line scripts
395 - Added checks for invalid node/instance names in the configuration (now
396 flagged during cluster verify)
397 - Fixed watcher behaviour for disk activation errors
398 - Fixed two potentially endless loops in http library, which led to the
399 RAPI daemon hanging and consuming 100% CPU in some cases
400 - Fixed bug in RAPI daemon related to hashed passwords
401 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
403 - Enhanced compatibility with non-Debian OSes, but not using absolute
404 path in some commands and allowing customisation of the ssh
405 configuration directory
406 - Fixed possible future issue with new Python versions by abiding to the
407 proper use of ``__slots__`` attribute on classes
408 - Added checks that should prevent directory traversal attacks
409 - Many documentation fixes based on feedback from users
414 - Added an “early_release” more for instance replace disks and node
415 evacuate, where we release locks earlier and thus allow higher
416 parallelism within the cluster
417 - Added watcher hooks, intended to allow the watcher to restart other
418 daemons (e.g. from the ganeti-nbma project), but they can be used of
419 course for any other purpose
420 - Added a compile-time disable for DRBD barriers, to increase
421 performance if the administrator trusts the power supply or the
422 storage system to not lose writes
423 - Added the option of using syslog for logging instead of, or in
424 addition to, Ganeti's own log files
425 - Removed boot restriction for paravirtual NICs for KVM, recent versions
426 can indeed boot from a paravirtual NIC
427 - Added a generic debug level for many operations; while this is not
428 used widely yet, it allows one to pass the debug value all the way to
430 - Enhanced the hooks environment for instance moves (failovers,
431 migrations) where the primary/secondary nodes changed during the
432 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
433 - Enhanced data validations for many user-supplied values; one important
434 item is the restrictions imposed on instance and node names, which
435 might reject some (invalid) host names
436 - Add a configure-time option to disable file-based storage, if it's not
437 needed; this allows greater security separation between the master
438 node and the other nodes from the point of view of the inter-node RPC
440 - Added user notification in interactive tools if job is waiting in the
441 job queue or trying to acquire locks
442 - Added log messages when a job is waiting for locks
443 - Added filtering by node tags in instance operations which admit
444 multiple instances (start, stop, reboot, reinstall)
445 - Added a new tool for cluster mergers, ``cluster-merge``
446 - Parameters from command line which are of the form ``a=b,c=d`` can now
447 use backslash escapes to pass in values which contain commas,
448 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
450 - For KVM, the instance name is the first parameter passed to KVM, so
451 that it's more visible in the process list
457 *(Released Tue, 2 Mar 2010)*
459 Ganeti 2.1 brings many improvements with it. Major changes:
461 - Added infrastructure to ease automated disk repairs
462 - Added new daemon to export configuration data in a cheaper way than
464 - Instance NICs can now be routed instead of being associated with a
466 - Improved job locking logic to reduce impact of jobs acquiring multiple
467 locks waiting for other long-running jobs
469 In-depth implementation details can be found in the Ganeti 2.1 design
475 - Added chroot hypervisor
476 - Added more options to xen-hvm hypervisor (``kernel_path`` and
478 - Added more options to xen-pvm hypervisor (``use_bootloader``,
479 ``bootloader_path`` and ``bootloader_args``)
480 - Added the ``use_localtime`` option for the xen-hvm and kvm
481 hypervisors, and the default value for this has changed to false (in
482 2.0 xen-hvm always enabled it)
483 - Added luxi call to submit multiple jobs in one go
484 - Added cluster initialization option to not modify ``/etc/hosts``
486 - Added network interface parameters
487 - Added dry run mode to some LUs
488 - Added RAPI resources:
490 - ``/2/instances/[instance_name]/info``
491 - ``/2/instances/[instance_name]/replace-disks``
492 - ``/2/nodes/[node_name]/evacuate``
493 - ``/2/nodes/[node_name]/migrate``
494 - ``/2/nodes/[node_name]/role``
495 - ``/2/nodes/[node_name]/storage``
496 - ``/2/nodes/[node_name]/storage/modify``
497 - ``/2/nodes/[node_name]/storage/repair``
499 - Added OpCodes to evacuate or migrate all instances on a node
500 - Added new command to list storage elements on nodes (``gnt-node
501 list-storage``) and modify them (``gnt-node modify-storage``)
502 - Added new ssconf files with master candidate IP address
503 (``ssconf_master_candidates_ips``), node primary IP address
504 (``ssconf_node_primary_ips``) and node secondary IP address
505 (``ssconf_node_secondary_ips``)
506 - Added ``ganeti-confd`` and a client library to query the Ganeti
507 configuration via UDP
508 - Added ability to run hooks after cluster initialization and before
510 - Added automatic mode for disk replace (``gnt-instance replace-disks
512 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
513 after catastrophic data-loss
514 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
516 - Added ``gnt-instance move`` command to move instances
517 - Added ``gnt-cluster watcher`` command to control watcher
518 - Added ``gnt-node powercycle`` command to powercycle nodes
519 - Added new job status field ``lock_status``
520 - Added parseable error codes to cluster verification (``gnt-cluster
521 verify --error-codes``) and made output less verbose (use
522 ``--verbose`` to restore previous behaviour)
523 - Added UUIDs to the main config entities (cluster, nodes, instances)
524 - Added support for OS variants
525 - Added support for hashed passwords in the Ganeti remote API users file
527 - Added option to specify maximum timeout on instance shutdown
528 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
529 - Added new helper script to start and stop Ganeti daemons
530 (``daemon-util``), with the intent to reduce the work necessary to
531 adjust Ganeti for non-Debian distributions and to start/stop daemons
533 - Added more unittests
534 - Fixed critical bug in ganeti-masterd startup
535 - Removed the configure-time ``kvm-migration-port`` parameter, this is
536 now customisable at the cluster level for both the KVM and Xen
537 hypervisors using the new ``migration_port`` parameter
538 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
539 reinstalling an instance
540 - Allowed ``@`` in tag names
541 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
542 - Many documentation updates
543 - Distribute hypervisor files on ``gnt-cluster redist-conf``
544 - ``gnt-instance reinstall`` can now reinstall multiple instances
545 - Updated many command line parameters
546 - Introduced new OS API version 15
547 - No longer support a default hypervisor
548 - Treat virtual LVs as inexistent
549 - Improved job locking logic to reduce lock contention
550 - Match instance and node names case insensitively
551 - Reimplemented bash completion script to be more complete
558 *(Released Thu, 4 Feb 2010)*
560 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
561 - Fix a string formatting bug
562 - Improve safety of the code in some error paths
563 - Improve data validation in the master of values returned from nodes
569 *(Released Thu, 17 Dec 2009)*
571 - Fix security issue due to missing validation of iallocator names; this
572 allows local and remote execution of arbitrary executables
573 - Fix failure of gnt-node list during instance removal
574 - Ship the RAPI documentation in the archive
580 *(Released Wed, 30 Sep 2009)*
582 - Fixed many wrong messages
583 - Fixed a few bugs related to the locking library
584 - Fixed MAC checking at instance creation time
585 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
586 - Fixed a few issues related to signal handling in both daemons and
588 - Fixed the example startup script provided
589 - Fixed insserv dependencies in the example startup script (patch from
591 - Fixed handling of drained nodes in the iallocator framework
592 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
594 - Fixed error related to invalid job IDs in job polling
595 - Fixed job/opcode persistence on unclean master shutdown
596 - Fixed handling of partial job processing after unclean master
598 - Fixed error reporting from LUs, previously all errors were converted
599 into execution errors
600 - Fixed error reporting from burnin
601 - Decreased significantly the memory usage of the job queue
602 - Optimised slightly multi-job submission
603 - Optimised slightly opcode loading
604 - Backported the multi-job submit framework from the development
605 branch; multi-instance start and stop should be faster
606 - Added script to clean archived jobs after 21 days; this will reduce
607 the size of the queue directory
608 - Added some extra checks in disk size tracking
609 - Added an example ethers hook script
610 - Added a cluster parameter that prevents Ganeti from modifying of
612 - Added more node information to RAPI responses
613 - Added a ``gnt-job watch`` command that allows following the ouput of a
615 - Added a bind-address option to ganeti-rapi
616 - Added more checks to the configuration verify
617 - Enhanced the burnin script such that some operations can be retried
619 - Converted instance reinstall to multi-instance model
625 *(Released Fri, 7 Aug 2009)*
627 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
628 to allow using the pre-2.0.2 behaviour in activation, if any existing
629 instances have mismatched disk sizes in the configuration
630 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
631 any configuration mismatches for disk sizes
632 - Added ``gnt-master cluste-failover --no-voting`` to allow master
633 failover to work on two-node clusters
634 - Fixed the ``--net`` option of ``gnt-backup import``, which was
636 - Fixed detection of OS script errors in ``gnt-backup export``
637 - Fixed exit code of ``gnt-backup export``
643 *(Released Fri, 17 Jul 2009)*
645 - Added experimental support for stripped logical volumes; this should
646 enhance performance but comes with a higher complexity in the block
647 device handling; stripping is only enabled when passing
648 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
649 affected even in the non-stripped mode
650 - Improved resiliency against transient failures at the end of DRBD
651 resyncs, and in general of DRBD resync checks
652 - Fixed a couple of issues with exports and snapshot errors
653 - Fixed a couple of issues in instance listing
654 - Added display of the disk size in ``gnt-instance info``
655 - Fixed checking for valid OSes in instance creation
656 - Fixed handling of the "vcpus" parameter in instance listing and in
657 general of invalid parameters
658 - Fixed http server library, and thus RAPI, to handle invalid
659 username/password combinations correctly; this means that now they
660 report unauthorized for queries too, not only for modifications,
661 allowing earlier detect of configuration problems
662 - Added a new "role" node list field, equivalent to the master/master
663 candidate/drained/offline flags combinations
664 - Fixed cluster modify and changes of candidate pool size
665 - Fixed cluster verify error messages for wrong files on regular nodes
666 - Fixed a couple of issues with node demotion from master candidate role
667 - Fixed node readd issues
668 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
669 - Added a new ``--no-voting`` option for masterfailover to fix failover
670 on two-nodes clusters when the former master node is unreachable
671 - Added instance reinstall over RAPI
677 *(Released Tue, 16 Jun 2009)*
679 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
680 allow re-adding the start in single-user option (regression from 1.2)
681 - the watcher writes the instance status to a file, to allow monitoring
682 to report the instance status (from the master) based on cached
683 results of the watcher's queries; while this can get stale if the
684 watcher is being locked due to other work on the cluster, this is
686 - the watcher now also restarts the node daemon and the rapi daemon if
688 - fixed the watcher to handle full and drained queue cases
689 - hooks export more instance data in the environment, which helps if
690 hook scripts need to take action based on the instance's properties
691 (no longer need to query back into ganeti)
692 - instance failovers when the instance is stopped do not check for free
693 RAM, so that failing over a stopped instance is possible in low memory
695 - rapi uses queries for tags instead of jobs (for less job traffic), and
696 for cluster tags it won't talk to masterd at all but read them from
698 - a couple of error handling fixes in RAPI
699 - drbd handling: improved the error handling of inconsistent disks after
700 resync to reduce the frequency of "there are some degraded disks for
701 this instance" messages
702 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
703 error handling path called a wrong function name)
709 *(Released Wed, 27 May 2009)*
711 - no changes from rc5
714 Version 2.0 release candidate 5
715 -------------------------------
717 *(Released Wed, 20 May 2009)*
719 - fix a couple of bugs (validation, argument checks)
720 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
721 - some small improvements to RAPI and IAllocator
722 - make watcher automatically start the master daemon if down
725 Version 2.0 release candidate 4
726 -------------------------------
728 *(Released Mon, 27 Apr 2009)*
730 - change the OS list to not require locks; this helps with big clusters
731 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
732 volume group is broken
733 - ``gnt-instance info``, without any arguments, doesn't run for all
734 instances anymore; either pass ``--all`` or pass the desired
735 instances; this helps against mistakes on big clusters where listing
736 the information for all instances takes a long time
737 - miscellaneous doc and man pages fixes
740 Version 2.0 release candidate 3
741 -------------------------------
743 *(Released Wed, 8 Apr 2009)*
745 - Change the internal locking model of some ``gnt-node`` commands, in
746 order to reduce contention (and blocking of master daemon) when
747 batching many creation/reinstall jobs
748 - Fixes to Xen soft reboot
749 - No longer build documentation at build time, instead distribute it in
750 the archive, in order to reduce the need for the whole docbook/rst
754 Version 2.0 release candidate 2
755 -------------------------------
757 *(Released Fri, 27 Mar 2009)*
759 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
760 - Fix watcher startup sequence, improves the behaviour of busy clusters
761 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
762 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
763 bind address and other places
764 - Some documentation fixes and updates
767 Version 2.0 release candidate 1
768 -------------------------------
770 *(Released Mon, 2 Mar 2009)*
772 - More documentation updates, now all docs should be more-or-less
774 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
776 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
778 - Fix an issue related to $libdir/run/ganeti and cluster creation
784 *(Released Thu, 19 Feb 2009)*
786 - Xen PVM and KVM have switched the default value for the instance root
787 disk to the first partition on the first drive, instead of the whole
788 drive; this means that the OS installation scripts must be changed
790 - Man pages have been updated
791 - RAPI has been switched by default to HTTPS, and the exported functions
792 should all work correctly
793 - RAPI v1 has been removed
794 - Many improvements to the KVM hypervisor
795 - Block device errors are now better reported
796 - Many other bugfixes and small improvements
802 *(Released Mon, 26 Jan 2009)*
804 - Version 2 is a general rewrite of the code and therefore the
805 differences are too many to list, see the design document for 2.0 in
806 the ``doc/`` subdirectory for more details
807 - In this beta version there is not yet a migration path from 1.2 (there
808 will be one in the final 2.0 release)
809 - A few significant changes are:
811 - all commands are executed by a daemon (``ganeti-masterd``) and the
812 various ``gnt-*`` commands are just front-ends to it
813 - all the commands are entered into, and executed from a job queue,
814 see the ``gnt-job(8)`` manpage
815 - the RAPI daemon supports read-write operations, secured by basic
816 HTTP authentication on top of HTTPS
817 - DRBD version 0.7 support has been removed, DRBD 8 is the only
818 supported version (when migrating from Ganeti 1.2 to 2.0, you need
819 to migrate to DRBD 8 first while still running Ganeti 1.2)
820 - DRBD devices are using statically allocated minor numbers, which
821 will be assigned to existing instances during the migration process
822 - there is support for both Xen PVM and Xen HVM instances running on
824 - KVM virtualization is supported too
825 - file-based storage has been implemented, which means that it is
826 possible to run the cluster without LVM and DRBD storage, for
827 example using a shared filesystem exported from shared storage (and
828 still have live migration)
834 *(Released Tue, 13 Jan 2009)*
836 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
837 - Reuse the old instance mac address by default on instance import, if
838 the instance name is the same.
839 - Handle situations in which the node info rpc returns incomplete
841 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
842 - Improved version of batcher:
845 - instance mac address support
846 - support for HVM clusters/instances
848 - Add an option to show the number of cpu sockets and nodes in
850 - Support OSes that handle more than one version of the OS api (but do
851 not change the current API in any other way)
852 - Fix ``gnt-node migrate``
853 - ``gnt-debug`` man page
854 - Fixes various more typos and small issues
855 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
861 *(Released Wed, 24 Sep 2008)*
863 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
864 type of disk exported to fully virtualized instances.
865 - provide access to the serial console of HVM instances
866 - instance auto_balance flag, set by default. If turned off it will
867 avoid warnings on cluster verify if there is not enough memory to fail
868 over an instance. in the future it will prevent automatically failing
869 it over when we will support that.
870 - batcher tool for instance creation, see ``tools/README.batcher``
871 - ``gnt-instance reinstall --select-os`` to interactively select a new
872 operating system when reinstalling an instance.
873 - when changing the memory amount on instance modify a check has been
874 added that the instance will be able to start. also warnings are
875 emitted if the instance will not be able to fail over, if auto_balance
877 - documentation fixes
878 - sync fields between ``gnt-instance list/modify/add/import``
879 - fix a race condition in drbd when the sync speed was set after giving
880 the device a remote peer.
886 *(Released Tue, 22 Jul 2008)*
888 - note: the allowed size and number of tags per object were reduced
889 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
890 - fixed twisted 8.x compatibility
891 - fixed ``gnt-instance replace-disks`` with iallocator
892 - add TCP keepalives on twisted connections to detect restarted nodes
893 - disk increase support, see ``gnt-instance grow-disk``
894 - implement bulk node/instance query for RAPI
895 - add tags in node/instance listing (optional)
896 - experimental migration (and live migration) support, read the man page
897 for ``gnt-instance migrate``
898 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
899 has some small improvements in handling its state file
905 *(Released Fri, 13 Jun 2008)*
907 - Experimental readonly, REST-based remote API implementation;
908 automatically started on master node, TCP port 5080, if enabled by
909 ``--enable-rapi`` parameter to configure script.
910 - Instance allocator support. Add and import instance accept a
911 ``--iallocator`` parameter, and call that instance allocator to decide
912 which node to use for the instance. The iallocator document describes
913 what's expected from an allocator script.
914 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
915 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
916 node is lost there is still enough memory to fail over the instances
918 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
919 ``gnt-cluster verify``, to check for site-specific compliance. All the
920 hooks will run, and their output, if any, will be displayed. Any
921 failing hook will make the verification return an error value.
922 - ``gnt-cluster verify`` now checks that its peers are reachable on the
923 primary and secondary interfaces
924 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
925 that is still declared as part of the cluster and has failed.
926 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
927 specifying output fields, that just adds the chosen fields to the
929 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
930 export from the filesystem.
931 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
932 have been added. Using them you can enable/disable acpi and pae
933 support, and specify a path for a cd image to be exported to the
934 instance. These parameters as the name suggest only work on HVM
936 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
937 PAE support will be set to the previously hardcoded values, but the
938 (previously hardcoded) path to the CDROM ISO image will be unset and
939 if required, needs to be set manually with ``gnt-instance modify``
941 - The address to which an instance's VNC console is bound is now
942 selectable per-instance, rather than being cluster wide. Of course
943 this only applies to instances controlled via VNC, so currently just
944 applies to HVM clusters.
950 *(Released Mon, 18 Feb 2008)*
952 - more tweaks to the disk activation code (especially helpful for DRBD)
953 - change the default ``gnt-instance list`` output format, now there is
954 one combined status field (see the manpage for the exact values this
956 - some more fixes for the mac export to hooks change
957 - make Ganeti not break with DRBD 8.2.x (which changed the version
958 format in ``/proc/drbd``) (issue 24)
959 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
960 template, allowing migration from DRBD0.7+MD to DRBD8
966 *(Released Wed, 30 Jan 2008)*
968 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
970 - add command aliases infrastructure and a few aliases
971 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
972 man pages and the ``--help`` option of ``gnt-node
973 list``/``gnt-instance list``
974 - fix ``gnt-backup list`` with down nodes (issue 21)
975 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
976 - fix the dist archive and add a check for including svn/git files in
978 - some developer-related changes: improve the burnin and the QA suite,
979 add an upload script for testing during development
985 *(Released Wed, 16 Jan 2008)*
987 - experimental HVM support, read the install document, section
988 "Initializing the cluster"
989 - allow for the PVM hypervisor per-instance kernel and initrd paths
990 - add a new command ``gnt-cluster verify-disks`` which uses a new
991 algorithm to improve the reconnection of the DRBD pairs if the device
992 on the secondary node has gone away
993 - make logical volume code auto-activate LVs at disk activation time
994 - slightly improve the speed of activating disks
995 - allow specification of the MAC address at instance creation time, and
996 changing it later via ``gnt-instance modify``
997 - fix handling of external commands that generate lots of output on
999 - update documentation with regard to minimum version of DRBD8 supported
1005 *(Released Tue, 4 Dec 2007)*
1007 - Log the ``xm create`` output to the node daemon log on failure (to
1008 help diagnosing the error)
1009 - In debug mode, log all external commands output if failed to the logs
1010 - Change parsing of lvm commands to ignore stderr
1016 *(Released Wed, 28 Nov 2007)*
1018 - Another round of updates to the DRBD 8 code to deal with more failures
1019 in the replace secondary node operation
1020 - Some more logging of failures in disk operations (lvm, drbd)
1021 - A few documentation updates
1028 *(Released Tue, 13 Nov 2007)*
1030 - Change configuration file format from Python's Pickle to JSON.
1031 Upgrading is possible using the cfgupgrade utility.
1032 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1033 faster replace disks and is more stable (DRBD 8 has many improvements
1034 compared to DRBD 0.7)
1035 - Added command line tags support (see man pages for ``gnt-instance``,
1036 ``gnt-node``, ``gnt-cluster``)
1037 - Added instance rename support
1038 - Added multi-instance startup/shutdown
1039 - Added cluster rename support
1040 - Added ``gnt-node evacuate`` to simplify some node operations
1041 - Added instance reboot operation that can speedup reboot as compared to
1043 - Soften the requirement that hostnames are in FQDN format
1044 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1046 - Removed dependency on debian's patched fping that uses the
1047 non-standard ``-S`` option
1048 - Now the OS definitions are searched for in multiple, configurable
1049 paths (easier for distros to package)
1050 - Some changes to the hooks infrastructure (especially the new
1051 post-configuration update hook)
1052 - Other small bugfixes
1054 .. vim: set textwidth=72 :