7 *(Released Mon, 4 Oct 2010)*
9 - Fixed regression in ``gnt-instance rename``
15 *(Released Wed, 22 Sep 2010)*
17 - Fixed OS_VARIANT variable for OS scripts
18 - Fixed cluster tag operations via RAPI
19 - Made ``setup-ssh`` exit with non-zero code if an error occurred
20 - Disabled RAPI CA checks in watcher
26 *(Released Mon, 23 Aug 2010)*
28 - Support DRBD versions of the format "a.b.c.d"
30 - Re-introduce support for usage from multiple threads in RAPI client
31 - Instance renames and modify via RAPI
32 - Work around race condition between processing and archival in job
34 - Mark opcodes following failed one as failed, too
35 - Job field ``lock_status`` was removed due to difficulties making it
36 work with the changed job queue in Ganeti 2.2; a better way to monitor
37 locks is expected for a later 2.2.x release
38 - Fixed dry-run behaviour with many commands
39 - Support ``ssh-agent`` again when adding nodes
40 - Many additional bugfixes
46 *(Released Fri, 30 Jul 2010)*
48 Important change: the internal RPC mechanism between Ganeti nodes has
49 changed from using a home-grown http library (based on the Python base
50 libraries) to use the PycURL library. This requires that PycURL is
51 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
52 against GnuTLS by default. cURL's support for GnuTLS had known issues
53 before cURL 7.21.0 and we recommend using the latest cURL release or
54 linking against OpenSSL. Most other distributions already link PycURL
55 and cURL against OpenSSL. The command::
57 python -c 'import pycurl; print pycurl.version'
59 can be used to determine the libraries PycURL and cURL are linked
62 Other significant changes:
64 - Rewrote much of the internals of the job queue, in order to achieve
65 better parallelism; this decouples job query operations from the job
66 processing, and it should allow much nicer behaviour of the master
67 daemon under load, and it also has uncovered some long-standing bugs
68 related to the job serialisation (now fixed)
69 - Added a default iallocator setting to the cluster parameters,
70 eliminating the need to always pass nodes or an iallocator for
71 operations that require selection of new node(s)
72 - Added experimental support for the LXC virtualization method
73 - Added support for OS parameters, which allows the installation of
74 instances to pass parameter to OS scripts in order to customise the
76 - Added a hypervisor parameter controlling the migration type (live or
77 non-live), since hypervisors have various levels of reliability; this
78 has renamed the 'live' parameter to 'mode'
79 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
80 logical volumes, meaning that cluster verify will ignore them and not
81 flag their presence as errors
82 - The watcher will now reset the error count for failed instances after
83 8 hours, thus allowing self-healing if the problem that caused the
84 instances to be down/fail to start has cleared in the meantime
85 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
86 check for, and warn, if the drbd module parameter ``usermode_helper``
87 is not consistent with the cluster-wide setting; this is needed to
88 make diagnose easier of failed drbd creations
89 - Started adding base IPv6 support, but this is not yet
90 enabled/available for use
91 - Rename operations (cluster, instance) will now return the new name,
92 which is especially useful if a short name was passed in
93 - Added support for instance migration in RAPI
94 - Added a tool to pre-configure nodes for the SSH setup, before joining
95 them to the cluster; this will allow in the future a simplified model
96 for node joining (but not yet fully enabled in 2.2); this needs the
97 paramiko python library
98 - Fixed handling of name-resolving errors
99 - Fixed consistency of job results on the error path
100 - Fixed master-failover race condition when executed multiple times in
102 - Fixed many bugs related to the job queue (mostly introduced during the
103 2.2 development cycle, so not all are impacting 2.1)
104 - Fixed instance migration with missing disk symlinks
105 - Fixed handling of unknown jobs in ``gnt-job archive``
106 - And many other small fixes/improvements
110 - Enhanced both the unittest and the QA coverage
111 - Switched the opcode validation to a generic model, and extended the
112 validation to all opcode parameters
113 - Changed more parts of the code that write shell scripts to use the
115 - Switched the master daemon to use the asyncore library for the Luxi
122 *(Released Thu, 17 Jun 2010)*
124 - Added tool (``move-instance``) and infrastructure to move instances
125 between separate clusters (see :doc:`separate documentation
126 <move-instance>` and :doc:`design document <design-2.2>`)
127 - Added per-request RPC timeout
128 - RAPI now requires a Content-Type header for requests with a body (e.g.
129 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
130 :rfc:`2616` (HTTP/1.1), section 7.2.1)
131 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
133 - Implemented initial support for running Ganeti daemons as separate
134 users, see configure-time flags ``--with-user-prefix`` and
135 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
137 - Instances can be removed after export (``gnt-backup export
139 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
140 key (instead of 1024 bit)
141 - Added new cluster configuration file for cluster domain secret
142 - Import/export now use SSL instead of SSH
143 - Added support for showing estimated time when exporting an instance,
144 see the ``ganeti-os-interface(7)`` manpage and look for
151 *(Released Fri, 16 Jul 2010)*
154 - Add an option to only select some reboot types during qa/burnin.
155 (on some hypervisors consequent reboots are not supported)
156 - Fix infrequent race condition in master failover. Sometimes the old
157 master ip address would be still detected as up for a short time
158 after it was removed, causing failover to fail.
159 - Decrease mlockall warnings when the ctypes module is missing. On
160 Python 2.4 we support running even if no ctypes module is installed,
161 but we were too verbose about this issue.
162 - Fix building on old distributions, on which man doesn't have a
164 - Fix RAPI not to ignore the MAC address on instance creation
165 - Implement the old instance creation format in the RAPI client.
171 *(Released Thu, 01 Jul 2010)*
173 A small bugfix release:
174 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
175 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
176 a non-existing option
177 - Fix instance create: the --force-variant option was ignored
178 - Improve pylint 0.21 compatibility and warnings with Python 2.6
179 - Fix modify node storage with non-FQDN arguments
180 - Fix RAPI client to authenticate under Python 2.6 when used
181 for more than 5 requests needing authentication
182 - Fix gnt-instance modify -t (storage) giving a wrong error message
183 when converting a non-shutdown drbd instance to plain
189 *(Released Fri, 18 Jun 2010)*
191 A small bugfix release:
193 - Fix live migration of KVM instances started with older Ganeti
194 versions which had fewer hypervisor parameters
195 - Fix gnt-instance grow-disk on down instances
196 - Fix an error-reporting bug during instance migration
197 - Better checking of the ``--net`` and ``--disk`` values, to avoid
198 silently ignoring broken ones
199 - Fix an RPC error reporting bug affecting, for example, RAPI client
201 - Fix bug triggered by different API version os-es on different nodes
202 - Fix a bug in instance startup with custom hvparams: OS level
203 parameters would fail to be applied.
204 - Fix the RAPI client under Python 2.6 (but more work is needed to
205 make it work completely well with OpenSSL)
206 - Fix handling of errors when resolving names from DNS
212 *(Released Thu, 3 Jun 2010)*
214 A medium sized development cycle. Some new features, and some
215 fixes/small improvements/cleanups.
220 The node deamon now tries to mlock itself into memory, unless the
221 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
222 its logs, and falls back to console logging. This allows emergency
223 features such as ``gnt-node powercycle`` to work even in the event of a
224 broken node disk (tested offlining the disk hosting the node's
225 filesystem and dropping its memory caches; don't try this at home)
227 KVM: add vhost-net acceleration support. It can be tested with a new
228 enough version of the kernel and of qemu-kvm.
230 KVM: Add instance chrooting feature. If you use privilege dropping for
231 your VMs you can also now force them to chroot to an empty directory,
232 before starting the emulated guest.
234 KVM: Add maximum migration bandwith and maximum downtime tweaking
235 support (requires a new-enough version of qemu-kvm).
237 Cluster verify will now warn if the master node doesn't have the master
240 Add a new (incompatible) instance creation request format to RAPI which
241 supports all parameters (previously only a subset was supported, and it
242 wasn't possible to extend the old format to accomodate all the new
243 features. The old format is still supported, and a client can check for
244 this feature, before using it, by checking for its presence in the
245 ``features`` RAPI resource.
247 Now with ancient latin support. Try it passing the ``--roman`` option to
248 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
249 (requires the python-roman module to be installed, in order to work).
254 As usual many internal code refactorings, documentation updates, and
257 - Lots of improvements and cleanups to the experimental Remote API
258 (RAPI) client library.
259 - A new unit test suite for the core daemon libraries.
260 - A fix to creating missing directories makes sure the umask is not
261 applied anymore. This enforces the same directory permissions
263 - Better handling terminating daemons with ctrl+c (used when running
264 them in debugging mode).
265 - Fix a race condition in live migrating a KVM instance, when stat()
266 on the old proc status file returned EINVAL, which is an unexpected
268 - Fixed manpage checking with newer man and utf-8 charachters. But now
269 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
275 *(Released Fri, 7 May 2010)*
277 Fix a bug which prevented untagged KVM instances from starting.
283 *(Released Fri, 7 May 2010)*
285 Another release with a long development cycle, during which many
286 different features were added.
291 The KVM hypervisor now can run the individual instances as non-root, to
292 reduce the impact of a VM being hijacked due to bugs in the
293 hypervisor. It is possible to run all instances as a single (non-root)
294 user, to manually specify a user for each instance, or to dynamically
295 allocate a user out of a cluster-wide pool to each instance, with the
296 guarantee that no two instances will run under the same user ID on any
299 An experimental RAPI client library, that can be used standalone
300 (without the other Ganeti libraries), is provided in the source tree as
301 ``lib/rapi/client.py``. Note this client might change its interface in
302 the future, as we iterate on its capabilities.
304 A new command, ``gnt-cluster renew-crypto`` has been added to easily
305 replace the cluster's certificates and crypto keys. This might help in
306 case they have been compromised, or have simply expired.
308 A new disk option for instance creation has been added that allows one
309 to "adopt" currently existing logical volumes, with data
310 preservation. This should allow easier migration to Ganeti from
311 unmanaged (or managed via other software) instances.
313 Another disk improvement is the possibility to convert between redundant
314 (DRBD) and plain (LVM) disk configuration for an instance. This should
315 allow better scalability (starting with one node and growing the
316 cluster, or shrinking a two-node cluster to one node).
318 A new feature that could help with automated node failovers has been
319 implemented: if a node sees itself as offline (by querying the master
320 candidates), it will try to shutdown (hard) all instances and any active
321 DRBD devices. This reduces the risk of duplicate instances if an
322 external script automatically failovers the instances on such nodes. To
323 enable this, the cluster parameter ``maintain_node_health`` should be
324 enabled; in the future this option (per the name) will enable other
325 automatic maintenance features.
327 Instance export/import now will reuse the original instance
328 specifications for all parameters; that means exporting an instance,
329 deleting it and the importing it back should give an almost identical
330 instance. Note that the default import behaviour has changed from
331 before, where it created only one NIC; now it recreates the original
334 Cluster verify has added a few new checks: SSL certificates validity,
335 /etc/hosts consistency across the cluster, etc.
340 As usual, many internal changes were done, documentation fixes,
343 - Fixed cluster initialization with disabled cluster storage (regression
345 - File-based storage supports growing the disks
346 - Fixed behaviour of node role changes
347 - Fixed cluster verify for some corner cases, plus a general rewrite of
348 cluster verify to allow future extension with more checks
349 - Fixed log spamming by watcher and node daemon (regression introduced
351 - Fixed possible validation issues when changing the list of enabled
353 - Fixed cleanup of /etc/hosts during node removal
354 - Fixed RAPI response for invalid methods
355 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
356 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
357 from ide disks, etc.)
358 - Allow OS changes without re-installation (to record a changed OS
359 outside of Ganeti, or to allow OS renames)
360 - Allow instance creation without OS installation (useful for example if
361 the OS will be installed manually, or restored from a backup not in
363 - Implemented option to make cluster ``copyfile`` use the replication
365 - Added list of enabled hypervisors to ssconf (possibly useful for
367 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
369 - A partial form of node re-IP is possible via node readd, which now
370 allows changed node primary IP
371 - Command line utilities now show an informational message if the job is
373 - The logs of the master daemon now show the PID/UID/GID of the
380 *(Released Fri, 12 Mar 2010)*
382 During the 2.1.0 long release candidate cycle, a lot of improvements and
383 changes have accumulated with were released later as 2.1.1.
388 The node evacuate command (``gnt-node evacuate``) was significantly
389 rewritten, and as such the IAllocator protocol was changed - a new
390 request type has been added. This unfortunate change during a stable
391 series is designed to improve performance of node evacuations; on
392 clusters with more than about five nodes and which are well-balanced,
393 evacuation should proceed in parallel for all instances of the node
394 being evacuated. As such, any existing IAllocator scripts need to be
395 updated, otherwise the above command will fail due to the unknown
396 request. The provided "dumb" allocator has not been updated; but the
397 ganeti-htools package supports the new protocol since version 0.2.4.
399 Another important change is increased validation of node and instance
400 names. This might create problems in special cases, if invalid host
401 names are being used.
403 Also, a new layer of hypervisor parameters has been added, that sits at
404 OS level between the cluster defaults and the instance ones. This allows
405 customisation of virtualization parameters depending on the installed
406 OS. For example instances with OS 'X' may have a different KVM kernel
407 (or any other parameter) than the cluster defaults. This is intended to
408 help managing a multiple OSes on the same cluster, without manual
409 modification of each instance's parameters.
411 A tool for merging clusters, ``cluster-merge``, has been added in the
417 - Improved the int/float conversions that should make the code more
418 robust in face of errors from the node daemons
419 - Fixed the remove node code in case of internal configuration errors
420 - Fixed the node daemon behaviour in face of inconsistent queue
421 directory (e.g. read-only file-system where we can't open the files
423 - Fixed the behaviour of gnt-node modify for master candidate demotion;
424 now it either aborts cleanly or, if given the new “auto_promote”
425 parameter, will automatically promote other nodes as needed
426 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
427 completely prevent Ganeti from working
428 - Fixed bug for instance export when not all disks were successfully
430 - Fixed behaviour of node add when the new node is slow in starting up
432 - Fixed handling of signals in the LUXI client, which should improve
433 behaviour of command-line scripts
434 - Added checks for invalid node/instance names in the configuration (now
435 flagged during cluster verify)
436 - Fixed watcher behaviour for disk activation errors
437 - Fixed two potentially endless loops in http library, which led to the
438 RAPI daemon hanging and consuming 100% CPU in some cases
439 - Fixed bug in RAPI daemon related to hashed passwords
440 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
442 - Enhanced compatibility with non-Debian OSes, but not using absolute
443 path in some commands and allowing customisation of the ssh
444 configuration directory
445 - Fixed possible future issue with new Python versions by abiding to the
446 proper use of ``__slots__`` attribute on classes
447 - Added checks that should prevent directory traversal attacks
448 - Many documentation fixes based on feedback from users
453 - Added an “early_release” more for instance replace disks and node
454 evacuate, where we release locks earlier and thus allow higher
455 parallelism within the cluster
456 - Added watcher hooks, intended to allow the watcher to restart other
457 daemons (e.g. from the ganeti-nbma project), but they can be used of
458 course for any other purpose
459 - Added a compile-time disable for DRBD barriers, to increase
460 performance if the administrator trusts the power supply or the
461 storage system to not lose writes
462 - Added the option of using syslog for logging instead of, or in
463 addition to, Ganeti's own log files
464 - Removed boot restriction for paravirtual NICs for KVM, recent versions
465 can indeed boot from a paravirtual NIC
466 - Added a generic debug level for many operations; while this is not
467 used widely yet, it allows one to pass the debug value all the way to
469 - Enhanced the hooks environment for instance moves (failovers,
470 migrations) where the primary/secondary nodes changed during the
471 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
472 - Enhanced data validations for many user-supplied values; one important
473 item is the restrictions imposed on instance and node names, which
474 might reject some (invalid) host names
475 - Add a configure-time option to disable file-based storage, if it's not
476 needed; this allows greater security separation between the master
477 node and the other nodes from the point of view of the inter-node RPC
479 - Added user notification in interactive tools if job is waiting in the
480 job queue or trying to acquire locks
481 - Added log messages when a job is waiting for locks
482 - Added filtering by node tags in instance operations which admit
483 multiple instances (start, stop, reboot, reinstall)
484 - Added a new tool for cluster mergers, ``cluster-merge``
485 - Parameters from command line which are of the form ``a=b,c=d`` can now
486 use backslash escapes to pass in values which contain commas,
487 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
489 - For KVM, the instance name is the first parameter passed to KVM, so
490 that it's more visible in the process list
496 *(Released Tue, 2 Mar 2010)*
498 Ganeti 2.1 brings many improvements with it. Major changes:
500 - Added infrastructure to ease automated disk repairs
501 - Added new daemon to export configuration data in a cheaper way than
503 - Instance NICs can now be routed instead of being associated with a
505 - Improved job locking logic to reduce impact of jobs acquiring multiple
506 locks waiting for other long-running jobs
508 In-depth implementation details can be found in the Ganeti 2.1 design
514 - Added chroot hypervisor
515 - Added more options to xen-hvm hypervisor (``kernel_path`` and
517 - Added more options to xen-pvm hypervisor (``use_bootloader``,
518 ``bootloader_path`` and ``bootloader_args``)
519 - Added the ``use_localtime`` option for the xen-hvm and kvm
520 hypervisors, and the default value for this has changed to false (in
521 2.0 xen-hvm always enabled it)
522 - Added luxi call to submit multiple jobs in one go
523 - Added cluster initialization option to not modify ``/etc/hosts``
525 - Added network interface parameters
526 - Added dry run mode to some LUs
527 - Added RAPI resources:
529 - ``/2/instances/[instance_name]/info``
530 - ``/2/instances/[instance_name]/replace-disks``
531 - ``/2/nodes/[node_name]/evacuate``
532 - ``/2/nodes/[node_name]/migrate``
533 - ``/2/nodes/[node_name]/role``
534 - ``/2/nodes/[node_name]/storage``
535 - ``/2/nodes/[node_name]/storage/modify``
536 - ``/2/nodes/[node_name]/storage/repair``
538 - Added OpCodes to evacuate or migrate all instances on a node
539 - Added new command to list storage elements on nodes (``gnt-node
540 list-storage``) and modify them (``gnt-node modify-storage``)
541 - Added new ssconf files with master candidate IP address
542 (``ssconf_master_candidates_ips``), node primary IP address
543 (``ssconf_node_primary_ips``) and node secondary IP address
544 (``ssconf_node_secondary_ips``)
545 - Added ``ganeti-confd`` and a client library to query the Ganeti
546 configuration via UDP
547 - Added ability to run hooks after cluster initialization and before
549 - Added automatic mode for disk replace (``gnt-instance replace-disks
551 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
552 after catastrophic data-loss
553 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
555 - Added ``gnt-instance move`` command to move instances
556 - Added ``gnt-cluster watcher`` command to control watcher
557 - Added ``gnt-node powercycle`` command to powercycle nodes
558 - Added new job status field ``lock_status``
559 - Added parseable error codes to cluster verification (``gnt-cluster
560 verify --error-codes``) and made output less verbose (use
561 ``--verbose`` to restore previous behaviour)
562 - Added UUIDs to the main config entities (cluster, nodes, instances)
563 - Added support for OS variants
564 - Added support for hashed passwords in the Ganeti remote API users file
566 - Added option to specify maximum timeout on instance shutdown
567 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
568 - Added new helper script to start and stop Ganeti daemons
569 (``daemon-util``), with the intent to reduce the work necessary to
570 adjust Ganeti for non-Debian distributions and to start/stop daemons
572 - Added more unittests
573 - Fixed critical bug in ganeti-masterd startup
574 - Removed the configure-time ``kvm-migration-port`` parameter, this is
575 now customisable at the cluster level for both the KVM and Xen
576 hypervisors using the new ``migration_port`` parameter
577 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
578 reinstalling an instance
579 - Allowed ``@`` in tag names
580 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
581 - Many documentation updates
582 - Distribute hypervisor files on ``gnt-cluster redist-conf``
583 - ``gnt-instance reinstall`` can now reinstall multiple instances
584 - Updated many command line parameters
585 - Introduced new OS API version 15
586 - No longer support a default hypervisor
587 - Treat virtual LVs as inexistent
588 - Improved job locking logic to reduce lock contention
589 - Match instance and node names case insensitively
590 - Reimplemented bash completion script to be more complete
597 *(Released Thu, 4 Feb 2010)*
599 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
600 - Fix a string formatting bug
601 - Improve safety of the code in some error paths
602 - Improve data validation in the master of values returned from nodes
608 *(Released Thu, 17 Dec 2009)*
610 - Fix security issue due to missing validation of iallocator names; this
611 allows local and remote execution of arbitrary executables
612 - Fix failure of gnt-node list during instance removal
613 - Ship the RAPI documentation in the archive
619 *(Released Wed, 30 Sep 2009)*
621 - Fixed many wrong messages
622 - Fixed a few bugs related to the locking library
623 - Fixed MAC checking at instance creation time
624 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
625 - Fixed a few issues related to signal handling in both daemons and
627 - Fixed the example startup script provided
628 - Fixed insserv dependencies in the example startup script (patch from
630 - Fixed handling of drained nodes in the iallocator framework
631 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
633 - Fixed error related to invalid job IDs in job polling
634 - Fixed job/opcode persistence on unclean master shutdown
635 - Fixed handling of partial job processing after unclean master
637 - Fixed error reporting from LUs, previously all errors were converted
638 into execution errors
639 - Fixed error reporting from burnin
640 - Decreased significantly the memory usage of the job queue
641 - Optimised slightly multi-job submission
642 - Optimised slightly opcode loading
643 - Backported the multi-job submit framework from the development
644 branch; multi-instance start and stop should be faster
645 - Added script to clean archived jobs after 21 days; this will reduce
646 the size of the queue directory
647 - Added some extra checks in disk size tracking
648 - Added an example ethers hook script
649 - Added a cluster parameter that prevents Ganeti from modifying of
651 - Added more node information to RAPI responses
652 - Added a ``gnt-job watch`` command that allows following the ouput of a
654 - Added a bind-address option to ganeti-rapi
655 - Added more checks to the configuration verify
656 - Enhanced the burnin script such that some operations can be retried
658 - Converted instance reinstall to multi-instance model
664 *(Released Fri, 7 Aug 2009)*
666 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
667 to allow using the pre-2.0.2 behaviour in activation, if any existing
668 instances have mismatched disk sizes in the configuration
669 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
670 any configuration mismatches for disk sizes
671 - Added ``gnt-master cluste-failover --no-voting`` to allow master
672 failover to work on two-node clusters
673 - Fixed the ``--net`` option of ``gnt-backup import``, which was
675 - Fixed detection of OS script errors in ``gnt-backup export``
676 - Fixed exit code of ``gnt-backup export``
682 *(Released Fri, 17 Jul 2009)*
684 - Added experimental support for stripped logical volumes; this should
685 enhance performance but comes with a higher complexity in the block
686 device handling; stripping is only enabled when passing
687 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
688 affected even in the non-stripped mode
689 - Improved resiliency against transient failures at the end of DRBD
690 resyncs, and in general of DRBD resync checks
691 - Fixed a couple of issues with exports and snapshot errors
692 - Fixed a couple of issues in instance listing
693 - Added display of the disk size in ``gnt-instance info``
694 - Fixed checking for valid OSes in instance creation
695 - Fixed handling of the "vcpus" parameter in instance listing and in
696 general of invalid parameters
697 - Fixed http server library, and thus RAPI, to handle invalid
698 username/password combinations correctly; this means that now they
699 report unauthorized for queries too, not only for modifications,
700 allowing earlier detect of configuration problems
701 - Added a new "role" node list field, equivalent to the master/master
702 candidate/drained/offline flags combinations
703 - Fixed cluster modify and changes of candidate pool size
704 - Fixed cluster verify error messages for wrong files on regular nodes
705 - Fixed a couple of issues with node demotion from master candidate role
706 - Fixed node readd issues
707 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
708 - Added a new ``--no-voting`` option for masterfailover to fix failover
709 on two-nodes clusters when the former master node is unreachable
710 - Added instance reinstall over RAPI
716 *(Released Tue, 16 Jun 2009)*
718 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
719 allow re-adding the start in single-user option (regression from 1.2)
720 - the watcher writes the instance status to a file, to allow monitoring
721 to report the instance status (from the master) based on cached
722 results of the watcher's queries; while this can get stale if the
723 watcher is being locked due to other work on the cluster, this is
725 - the watcher now also restarts the node daemon and the rapi daemon if
727 - fixed the watcher to handle full and drained queue cases
728 - hooks export more instance data in the environment, which helps if
729 hook scripts need to take action based on the instance's properties
730 (no longer need to query back into ganeti)
731 - instance failovers when the instance is stopped do not check for free
732 RAM, so that failing over a stopped instance is possible in low memory
734 - rapi uses queries for tags instead of jobs (for less job traffic), and
735 for cluster tags it won't talk to masterd at all but read them from
737 - a couple of error handling fixes in RAPI
738 - drbd handling: improved the error handling of inconsistent disks after
739 resync to reduce the frequency of "there are some degraded disks for
740 this instance" messages
741 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
742 error handling path called a wrong function name)
748 *(Released Wed, 27 May 2009)*
750 - no changes from rc5
753 Version 2.0 release candidate 5
754 -------------------------------
756 *(Released Wed, 20 May 2009)*
758 - fix a couple of bugs (validation, argument checks)
759 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
760 - some small improvements to RAPI and IAllocator
761 - make watcher automatically start the master daemon if down
764 Version 2.0 release candidate 4
765 -------------------------------
767 *(Released Mon, 27 Apr 2009)*
769 - change the OS list to not require locks; this helps with big clusters
770 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
771 volume group is broken
772 - ``gnt-instance info``, without any arguments, doesn't run for all
773 instances anymore; either pass ``--all`` or pass the desired
774 instances; this helps against mistakes on big clusters where listing
775 the information for all instances takes a long time
776 - miscellaneous doc and man pages fixes
779 Version 2.0 release candidate 3
780 -------------------------------
782 *(Released Wed, 8 Apr 2009)*
784 - Change the internal locking model of some ``gnt-node`` commands, in
785 order to reduce contention (and blocking of master daemon) when
786 batching many creation/reinstall jobs
787 - Fixes to Xen soft reboot
788 - No longer build documentation at build time, instead distribute it in
789 the archive, in order to reduce the need for the whole docbook/rst
793 Version 2.0 release candidate 2
794 -------------------------------
796 *(Released Fri, 27 Mar 2009)*
798 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
799 - Fix watcher startup sequence, improves the behaviour of busy clusters
800 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
801 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
802 bind address and other places
803 - Some documentation fixes and updates
806 Version 2.0 release candidate 1
807 -------------------------------
809 *(Released Mon, 2 Mar 2009)*
811 - More documentation updates, now all docs should be more-or-less
813 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
815 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
817 - Fix an issue related to $libdir/run/ganeti and cluster creation
823 *(Released Thu, 19 Feb 2009)*
825 - Xen PVM and KVM have switched the default value for the instance root
826 disk to the first partition on the first drive, instead of the whole
827 drive; this means that the OS installation scripts must be changed
829 - Man pages have been updated
830 - RAPI has been switched by default to HTTPS, and the exported functions
831 should all work correctly
832 - RAPI v1 has been removed
833 - Many improvements to the KVM hypervisor
834 - Block device errors are now better reported
835 - Many other bugfixes and small improvements
841 *(Released Mon, 26 Jan 2009)*
843 - Version 2 is a general rewrite of the code and therefore the
844 differences are too many to list, see the design document for 2.0 in
845 the ``doc/`` subdirectory for more details
846 - In this beta version there is not yet a migration path from 1.2 (there
847 will be one in the final 2.0 release)
848 - A few significant changes are:
850 - all commands are executed by a daemon (``ganeti-masterd``) and the
851 various ``gnt-*`` commands are just front-ends to it
852 - all the commands are entered into, and executed from a job queue,
853 see the ``gnt-job(8)`` manpage
854 - the RAPI daemon supports read-write operations, secured by basic
855 HTTP authentication on top of HTTPS
856 - DRBD version 0.7 support has been removed, DRBD 8 is the only
857 supported version (when migrating from Ganeti 1.2 to 2.0, you need
858 to migrate to DRBD 8 first while still running Ganeti 1.2)
859 - DRBD devices are using statically allocated minor numbers, which
860 will be assigned to existing instances during the migration process
861 - there is support for both Xen PVM and Xen HVM instances running on
863 - KVM virtualization is supported too
864 - file-based storage has been implemented, which means that it is
865 possible to run the cluster without LVM and DRBD storage, for
866 example using a shared filesystem exported from shared storage (and
867 still have live migration)
873 *(Released Tue, 13 Jan 2009)*
875 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
876 - Reuse the old instance mac address by default on instance import, if
877 the instance name is the same.
878 - Handle situations in which the node info rpc returns incomplete
880 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
881 - Improved version of batcher:
884 - instance mac address support
885 - support for HVM clusters/instances
887 - Add an option to show the number of cpu sockets and nodes in
889 - Support OSes that handle more than one version of the OS api (but do
890 not change the current API in any other way)
891 - Fix ``gnt-node migrate``
892 - ``gnt-debug`` man page
893 - Fixes various more typos and small issues
894 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
900 *(Released Wed, 24 Sep 2008)*
902 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
903 type of disk exported to fully virtualized instances.
904 - provide access to the serial console of HVM instances
905 - instance auto_balance flag, set by default. If turned off it will
906 avoid warnings on cluster verify if there is not enough memory to fail
907 over an instance. in the future it will prevent automatically failing
908 it over when we will support that.
909 - batcher tool for instance creation, see ``tools/README.batcher``
910 - ``gnt-instance reinstall --select-os`` to interactively select a new
911 operating system when reinstalling an instance.
912 - when changing the memory amount on instance modify a check has been
913 added that the instance will be able to start. also warnings are
914 emitted if the instance will not be able to fail over, if auto_balance
916 - documentation fixes
917 - sync fields between ``gnt-instance list/modify/add/import``
918 - fix a race condition in drbd when the sync speed was set after giving
919 the device a remote peer.
925 *(Released Tue, 22 Jul 2008)*
927 - note: the allowed size and number of tags per object were reduced
928 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
929 - fixed twisted 8.x compatibility
930 - fixed ``gnt-instance replace-disks`` with iallocator
931 - add TCP keepalives on twisted connections to detect restarted nodes
932 - disk increase support, see ``gnt-instance grow-disk``
933 - implement bulk node/instance query for RAPI
934 - add tags in node/instance listing (optional)
935 - experimental migration (and live migration) support, read the man page
936 for ``gnt-instance migrate``
937 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
938 has some small improvements in handling its state file
944 *(Released Fri, 13 Jun 2008)*
946 - Experimental readonly, REST-based remote API implementation;
947 automatically started on master node, TCP port 5080, if enabled by
948 ``--enable-rapi`` parameter to configure script.
949 - Instance allocator support. Add and import instance accept a
950 ``--iallocator`` parameter, and call that instance allocator to decide
951 which node to use for the instance. The iallocator document describes
952 what's expected from an allocator script.
953 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
954 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
955 node is lost there is still enough memory to fail over the instances
957 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
958 ``gnt-cluster verify``, to check for site-specific compliance. All the
959 hooks will run, and their output, if any, will be displayed. Any
960 failing hook will make the verification return an error value.
961 - ``gnt-cluster verify`` now checks that its peers are reachable on the
962 primary and secondary interfaces
963 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
964 that is still declared as part of the cluster and has failed.
965 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
966 specifying output fields, that just adds the chosen fields to the
968 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
969 export from the filesystem.
970 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
971 have been added. Using them you can enable/disable acpi and pae
972 support, and specify a path for a cd image to be exported to the
973 instance. These parameters as the name suggest only work on HVM
975 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
976 PAE support will be set to the previously hardcoded values, but the
977 (previously hardcoded) path to the CDROM ISO image will be unset and
978 if required, needs to be set manually with ``gnt-instance modify``
980 - The address to which an instance's VNC console is bound is now
981 selectable per-instance, rather than being cluster wide. Of course
982 this only applies to instances controlled via VNC, so currently just
983 applies to HVM clusters.
989 *(Released Mon, 18 Feb 2008)*
991 - more tweaks to the disk activation code (especially helpful for DRBD)
992 - change the default ``gnt-instance list`` output format, now there is
993 one combined status field (see the manpage for the exact values this
995 - some more fixes for the mac export to hooks change
996 - make Ganeti not break with DRBD 8.2.x (which changed the version
997 format in ``/proc/drbd``) (issue 24)
998 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
999 template, allowing migration from DRBD0.7+MD to DRBD8
1005 *(Released Wed, 30 Jan 2008)*
1007 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1009 - add command aliases infrastructure and a few aliases
1010 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1011 man pages and the ``--help`` option of ``gnt-node
1012 list``/``gnt-instance list``
1013 - fix ``gnt-backup list`` with down nodes (issue 21)
1014 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1015 - fix the dist archive and add a check for including svn/git files in
1017 - some developer-related changes: improve the burnin and the QA suite,
1018 add an upload script for testing during development
1024 *(Released Wed, 16 Jan 2008)*
1026 - experimental HVM support, read the install document, section
1027 "Initializing the cluster"
1028 - allow for the PVM hypervisor per-instance kernel and initrd paths
1029 - add a new command ``gnt-cluster verify-disks`` which uses a new
1030 algorithm to improve the reconnection of the DRBD pairs if the device
1031 on the secondary node has gone away
1032 - make logical volume code auto-activate LVs at disk activation time
1033 - slightly improve the speed of activating disks
1034 - allow specification of the MAC address at instance creation time, and
1035 changing it later via ``gnt-instance modify``
1036 - fix handling of external commands that generate lots of output on
1038 - update documentation with regard to minimum version of DRBD8 supported
1044 *(Released Tue, 4 Dec 2007)*
1046 - Log the ``xm create`` output to the node daemon log on failure (to
1047 help diagnosing the error)
1048 - In debug mode, log all external commands output if failed to the logs
1049 - Change parsing of lvm commands to ignore stderr
1055 *(Released Wed, 28 Nov 2007)*
1057 - Another round of updates to the DRBD 8 code to deal with more failures
1058 in the replace secondary node operation
1059 - Some more logging of failures in disk operations (lvm, drbd)
1060 - A few documentation updates
1067 *(Released Tue, 13 Nov 2007)*
1069 - Change configuration file format from Python's Pickle to JSON.
1070 Upgrading is possible using the cfgupgrade utility.
1071 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1072 faster replace disks and is more stable (DRBD 8 has many improvements
1073 compared to DRBD 0.7)
1074 - Added command line tags support (see man pages for ``gnt-instance``,
1075 ``gnt-node``, ``gnt-cluster``)
1076 - Added instance rename support
1077 - Added multi-instance startup/shutdown
1078 - Added cluster rename support
1079 - Added ``gnt-node evacuate`` to simplify some node operations
1080 - Added instance reboot operation that can speedup reboot as compared to
1082 - Soften the requirement that hostnames are in FQDN format
1083 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1085 - Removed dependency on debian's patched fping that uses the
1086 non-standard ``-S`` option
1087 - Now the OS definitions are searched for in multiple, configurable
1088 paths (easier for distros to package)
1089 - Some changes to the hooks infrastructure (especially the new
1090 post-configuration update hook)
1091 - Other small bugfixes
1093 .. vim: set textwidth=72 :