7 *(Released Mon, 23 Aug 2010)*
9 - Support DRBD versions of the format "a.b.c.d"
11 - Re-introduce support for usage from multiple threads in RAPI client
12 - Instance renames and modify via RAPI
13 - Work around race condition between processing and archival in job
15 - Mark opcodes following failed one as failed, too
16 - Job field ``lock_status`` was removed due to difficulties making it
17 work with the changed job queue in Ganeti 2.2; a better way to monitor
18 locks is expected for a later 2.2.x release
19 - Fixed dry-run behaviour with many commands
20 - Support ``ssh-agent`` again when adding nodes
21 - Many additional bugfixes
27 *(Released Fri, 30 Jul 2010)*
29 Important change: the internal RPC mechanism between Ganeti nodes has
30 changed from using a home-grown http library (based on the Python base
31 libraries) to use the PycURL library. This requires that PycURL is
32 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
33 against GnuTLS by default. cURL's support for GnuTLS had known issues
34 before cURL 7.21.0 and we recommend using the latest cURL release or
35 linking against OpenSSL. Most other distributions already link PycURL
36 and cURL against OpenSSL. The command::
38 python -c 'import pycurl; print pycurl.version'
40 can be used to determine the libraries PycURL and cURL are linked
43 Other significant changes:
45 - Rewrote much of the internals of the job queue, in order to achieve
46 better parallelism; this decouples job query operations from the job
47 processing, and it should allow much nicer behaviour of the master
48 daemon under load, and it also has uncovered some long-standing bugs
49 related to the job serialisation (now fixed)
50 - Added a default iallocator setting to the cluster parameters,
51 eliminating the need to always pass nodes or an iallocator for
52 operations that require selection of new node(s)
53 - Added experimental support for the LXC virtualization method
54 - Added support for OS parameters, which allows the installation of
55 instances to pass parameter to OS scripts in order to customise the
57 - Added a hypervisor parameter controlling the migration type (live or
58 non-live), since hypervisors have various levels of reliability; this
59 has renamed the 'live' parameter to 'mode'
60 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
61 logical volumes, meaning that cluster verify will ignore them and not
62 flag their presence as errors
63 - The watcher will now reset the error count for failed instances after
64 8 hours, thus allowing self-healing if the problem that caused the
65 instances to be down/fail to start has cleared in the meantime
66 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
67 check for, and warn, if the drbd module parameter ``usermode_helper``
68 is not consistent with the cluster-wide setting; this is needed to
69 make diagnose easier of failed drbd creations
70 - Started adding base IPv6 support, but this is not yet
71 enabled/available for use
72 - Rename operations (cluster, instance) will now return the new name,
73 which is especially useful if a short name was passed in
74 - Added support for instance migration in RAPI
75 - Added a tool to pre-configure nodes for the SSH setup, before joining
76 them to the cluster; this will allow in the future a simplified model
77 for node joining (but not yet fully enabled in 2.2); this needs the
78 paramiko python library
79 - Fixed handling of name-resolving errors
80 - Fixed consistency of job results on the error path
81 - Fixed master-failover race condition when executed multiple times in
83 - Fixed many bugs related to the job queue (mostly introduced during the
84 2.2 development cycle, so not all are impacting 2.1)
85 - Fixed instance migration with missing disk symlinks
86 - Fixed handling of unknown jobs in ``gnt-job archive``
87 - And many other small fixes/improvements
91 - Enhanced both the unittest and the QA coverage
92 - Switched the opcode validation to a generic model, and extended the
93 validation to all opcode parameters
94 - Changed more parts of the code that write shell scripts to use the
96 - Switched the master daemon to use the asyncore library for the Luxi
103 *(Released Thu, 17 Jun 2010)*
105 - Added tool (``move-instance``) and infrastructure to move instances
106 between separate clusters (see :doc:`separate documentation
107 <move-instance>` and :doc:`design document <design-2.2>`)
108 - Added per-request RPC timeout
109 - RAPI now requires a Content-Type header for requests with a body (e.g.
110 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
111 :rfc:`2616` (HTTP/1.1), section 7.2.1)
112 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
114 - Implemented initial support for running Ganeti daemons as separate
115 users, see configure-time flags ``--with-user-prefix`` and
116 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
118 - Instances can be removed after export (``gnt-backup export
120 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
121 key (instead of 1024 bit)
122 - Added new cluster configuration file for cluster domain secret
123 - Import/export now use SSL instead of SSH
124 - Added support for showing estimated time when exporting an instance,
125 see the ``ganeti-os-interface(7)`` manpage and look for
131 *(Released Tue, 24 Aug 2010)*
134 - Don't ignore secondary node silently on non-mirrored disk templates
136 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
137 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
138 - Properly document the usb_mouse parameter
139 - Fix path in ganeti-rapi(8) (issue 116)
140 - Adjust error message when the ganeti user's .ssh directory is
142 - Add same-node-check when changing the disk template to drbd
148 *(Released Fri, 16 Jul 2010)*
151 - Add an option to only select some reboot types during qa/burnin.
152 (on some hypervisors consequent reboots are not supported)
153 - Fix infrequent race condition in master failover. Sometimes the old
154 master ip address would be still detected as up for a short time
155 after it was removed, causing failover to fail.
156 - Decrease mlockall warnings when the ctypes module is missing. On
157 Python 2.4 we support running even if no ctypes module is installed,
158 but we were too verbose about this issue.
159 - Fix building on old distributions, on which man doesn't have a
161 - Fix RAPI not to ignore the MAC address on instance creation
162 - Implement the old instance creation format in the RAPI client.
168 *(Released Thu, 01 Jul 2010)*
170 A small bugfix release:
171 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
172 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
173 a non-existing option
174 - Fix instance create: the --force-variant option was ignored
175 - Improve pylint 0.21 compatibility and warnings with Python 2.6
176 - Fix modify node storage with non-FQDN arguments
177 - Fix RAPI client to authenticate under Python 2.6 when used
178 for more than 5 requests needing authentication
179 - Fix gnt-instance modify -t (storage) giving a wrong error message
180 when converting a non-shutdown drbd instance to plain
186 *(Released Fri, 18 Jun 2010)*
188 A small bugfix release:
190 - Fix live migration of KVM instances started with older Ganeti
191 versions which had fewer hypervisor parameters
192 - Fix gnt-instance grow-disk on down instances
193 - Fix an error-reporting bug during instance migration
194 - Better checking of the ``--net`` and ``--disk`` values, to avoid
195 silently ignoring broken ones
196 - Fix an RPC error reporting bug affecting, for example, RAPI client
198 - Fix bug triggered by different API version os-es on different nodes
199 - Fix a bug in instance startup with custom hvparams: OS level
200 parameters would fail to be applied.
201 - Fix the RAPI client under Python 2.6 (but more work is needed to
202 make it work completely well with OpenSSL)
203 - Fix handling of errors when resolving names from DNS
209 *(Released Thu, 3 Jun 2010)*
211 A medium sized development cycle. Some new features, and some
212 fixes/small improvements/cleanups.
217 The node deamon now tries to mlock itself into memory, unless the
218 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
219 its logs, and falls back to console logging. This allows emergency
220 features such as ``gnt-node powercycle`` to work even in the event of a
221 broken node disk (tested offlining the disk hosting the node's
222 filesystem and dropping its memory caches; don't try this at home)
224 KVM: add vhost-net acceleration support. It can be tested with a new
225 enough version of the kernel and of qemu-kvm.
227 KVM: Add instance chrooting feature. If you use privilege dropping for
228 your VMs you can also now force them to chroot to an empty directory,
229 before starting the emulated guest.
231 KVM: Add maximum migration bandwith and maximum downtime tweaking
232 support (requires a new-enough version of qemu-kvm).
234 Cluster verify will now warn if the master node doesn't have the master
237 Add a new (incompatible) instance creation request format to RAPI which
238 supports all parameters (previously only a subset was supported, and it
239 wasn't possible to extend the old format to accomodate all the new
240 features. The old format is still supported, and a client can check for
241 this feature, before using it, by checking for its presence in the
242 ``features`` RAPI resource.
244 Now with ancient latin support. Try it passing the ``--roman`` option to
245 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
246 (requires the python-roman module to be installed, in order to work).
251 As usual many internal code refactorings, documentation updates, and
254 - Lots of improvements and cleanups to the experimental Remote API
255 (RAPI) client library.
256 - A new unit test suite for the core daemon libraries.
257 - A fix to creating missing directories makes sure the umask is not
258 applied anymore. This enforces the same directory permissions
260 - Better handling terminating daemons with ctrl+c (used when running
261 them in debugging mode).
262 - Fix a race condition in live migrating a KVM instance, when stat()
263 on the old proc status file returned EINVAL, which is an unexpected
265 - Fixed manpage checking with newer man and utf-8 charachters. But now
266 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
272 *(Released Fri, 7 May 2010)*
274 Fix a bug which prevented untagged KVM instances from starting.
280 *(Released Fri, 7 May 2010)*
282 Another release with a long development cycle, during which many
283 different features were added.
288 The KVM hypervisor now can run the individual instances as non-root, to
289 reduce the impact of a VM being hijacked due to bugs in the
290 hypervisor. It is possible to run all instances as a single (non-root)
291 user, to manually specify a user for each instance, or to dynamically
292 allocate a user out of a cluster-wide pool to each instance, with the
293 guarantee that no two instances will run under the same user ID on any
296 An experimental RAPI client library, that can be used standalone
297 (without the other Ganeti libraries), is provided in the source tree as
298 ``lib/rapi/client.py``. Note this client might change its interface in
299 the future, as we iterate on its capabilities.
301 A new command, ``gnt-cluster renew-crypto`` has been added to easily
302 replace the cluster's certificates and crypto keys. This might help in
303 case they have been compromised, or have simply expired.
305 A new disk option for instance creation has been added that allows one
306 to "adopt" currently existing logical volumes, with data
307 preservation. This should allow easier migration to Ganeti from
308 unmanaged (or managed via other software) instances.
310 Another disk improvement is the possibility to convert between redundant
311 (DRBD) and plain (LVM) disk configuration for an instance. This should
312 allow better scalability (starting with one node and growing the
313 cluster, or shrinking a two-node cluster to one node).
315 A new feature that could help with automated node failovers has been
316 implemented: if a node sees itself as offline (by querying the master
317 candidates), it will try to shutdown (hard) all instances and any active
318 DRBD devices. This reduces the risk of duplicate instances if an
319 external script automatically failovers the instances on such nodes. To
320 enable this, the cluster parameter ``maintain_node_health`` should be
321 enabled; in the future this option (per the name) will enable other
322 automatic maintenance features.
324 Instance export/import now will reuse the original instance
325 specifications for all parameters; that means exporting an instance,
326 deleting it and the importing it back should give an almost identical
327 instance. Note that the default import behaviour has changed from
328 before, where it created only one NIC; now it recreates the original
331 Cluster verify has added a few new checks: SSL certificates validity,
332 /etc/hosts consistency across the cluster, etc.
337 As usual, many internal changes were done, documentation fixes,
340 - Fixed cluster initialization with disabled cluster storage (regression
342 - File-based storage supports growing the disks
343 - Fixed behaviour of node role changes
344 - Fixed cluster verify for some corner cases, plus a general rewrite of
345 cluster verify to allow future extension with more checks
346 - Fixed log spamming by watcher and node daemon (regression introduced
348 - Fixed possible validation issues when changing the list of enabled
350 - Fixed cleanup of /etc/hosts during node removal
351 - Fixed RAPI response for invalid methods
352 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
353 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
354 from ide disks, etc.)
355 - Allow OS changes without re-installation (to record a changed OS
356 outside of Ganeti, or to allow OS renames)
357 - Allow instance creation without OS installation (useful for example if
358 the OS will be installed manually, or restored from a backup not in
360 - Implemented option to make cluster ``copyfile`` use the replication
362 - Added list of enabled hypervisors to ssconf (possibly useful for
364 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
366 - A partial form of node re-IP is possible via node readd, which now
367 allows changed node primary IP
368 - Command line utilities now show an informational message if the job is
370 - The logs of the master daemon now show the PID/UID/GID of the
377 *(Released Fri, 12 Mar 2010)*
379 During the 2.1.0 long release candidate cycle, a lot of improvements and
380 changes have accumulated with were released later as 2.1.1.
385 The node evacuate command (``gnt-node evacuate``) was significantly
386 rewritten, and as such the IAllocator protocol was changed - a new
387 request type has been added. This unfortunate change during a stable
388 series is designed to improve performance of node evacuations; on
389 clusters with more than about five nodes and which are well-balanced,
390 evacuation should proceed in parallel for all instances of the node
391 being evacuated. As such, any existing IAllocator scripts need to be
392 updated, otherwise the above command will fail due to the unknown
393 request. The provided "dumb" allocator has not been updated; but the
394 ganeti-htools package supports the new protocol since version 0.2.4.
396 Another important change is increased validation of node and instance
397 names. This might create problems in special cases, if invalid host
398 names are being used.
400 Also, a new layer of hypervisor parameters has been added, that sits at
401 OS level between the cluster defaults and the instance ones. This allows
402 customisation of virtualization parameters depending on the installed
403 OS. For example instances with OS 'X' may have a different KVM kernel
404 (or any other parameter) than the cluster defaults. This is intended to
405 help managing a multiple OSes on the same cluster, without manual
406 modification of each instance's parameters.
408 A tool for merging clusters, ``cluster-merge``, has been added in the
414 - Improved the int/float conversions that should make the code more
415 robust in face of errors from the node daemons
416 - Fixed the remove node code in case of internal configuration errors
417 - Fixed the node daemon behaviour in face of inconsistent queue
418 directory (e.g. read-only file-system where we can't open the files
420 - Fixed the behaviour of gnt-node modify for master candidate demotion;
421 now it either aborts cleanly or, if given the new “auto_promote”
422 parameter, will automatically promote other nodes as needed
423 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
424 completely prevent Ganeti from working
425 - Fixed bug for instance export when not all disks were successfully
427 - Fixed behaviour of node add when the new node is slow in starting up
429 - Fixed handling of signals in the LUXI client, which should improve
430 behaviour of command-line scripts
431 - Added checks for invalid node/instance names in the configuration (now
432 flagged during cluster verify)
433 - Fixed watcher behaviour for disk activation errors
434 - Fixed two potentially endless loops in http library, which led to the
435 RAPI daemon hanging and consuming 100% CPU in some cases
436 - Fixed bug in RAPI daemon related to hashed passwords
437 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
439 - Enhanced compatibility with non-Debian OSes, but not using absolute
440 path in some commands and allowing customisation of the ssh
441 configuration directory
442 - Fixed possible future issue with new Python versions by abiding to the
443 proper use of ``__slots__`` attribute on classes
444 - Added checks that should prevent directory traversal attacks
445 - Many documentation fixes based on feedback from users
450 - Added an “early_release” more for instance replace disks and node
451 evacuate, where we release locks earlier and thus allow higher
452 parallelism within the cluster
453 - Added watcher hooks, intended to allow the watcher to restart other
454 daemons (e.g. from the ganeti-nbma project), but they can be used of
455 course for any other purpose
456 - Added a compile-time disable for DRBD barriers, to increase
457 performance if the administrator trusts the power supply or the
458 storage system to not lose writes
459 - Added the option of using syslog for logging instead of, or in
460 addition to, Ganeti's own log files
461 - Removed boot restriction for paravirtual NICs for KVM, recent versions
462 can indeed boot from a paravirtual NIC
463 - Added a generic debug level for many operations; while this is not
464 used widely yet, it allows one to pass the debug value all the way to
466 - Enhanced the hooks environment for instance moves (failovers,
467 migrations) where the primary/secondary nodes changed during the
468 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
469 - Enhanced data validations for many user-supplied values; one important
470 item is the restrictions imposed on instance and node names, which
471 might reject some (invalid) host names
472 - Add a configure-time option to disable file-based storage, if it's not
473 needed; this allows greater security separation between the master
474 node and the other nodes from the point of view of the inter-node RPC
476 - Added user notification in interactive tools if job is waiting in the
477 job queue or trying to acquire locks
478 - Added log messages when a job is waiting for locks
479 - Added filtering by node tags in instance operations which admit
480 multiple instances (start, stop, reboot, reinstall)
481 - Added a new tool for cluster mergers, ``cluster-merge``
482 - Parameters from command line which are of the form ``a=b,c=d`` can now
483 use backslash escapes to pass in values which contain commas,
484 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
486 - For KVM, the instance name is the first parameter passed to KVM, so
487 that it's more visible in the process list
493 *(Released Tue, 2 Mar 2010)*
495 Ganeti 2.1 brings many improvements with it. Major changes:
497 - Added infrastructure to ease automated disk repairs
498 - Added new daemon to export configuration data in a cheaper way than
500 - Instance NICs can now be routed instead of being associated with a
502 - Improved job locking logic to reduce impact of jobs acquiring multiple
503 locks waiting for other long-running jobs
505 In-depth implementation details can be found in the Ganeti 2.1 design
511 - Added chroot hypervisor
512 - Added more options to xen-hvm hypervisor (``kernel_path`` and
514 - Added more options to xen-pvm hypervisor (``use_bootloader``,
515 ``bootloader_path`` and ``bootloader_args``)
516 - Added the ``use_localtime`` option for the xen-hvm and kvm
517 hypervisors, and the default value for this has changed to false (in
518 2.0 xen-hvm always enabled it)
519 - Added luxi call to submit multiple jobs in one go
520 - Added cluster initialization option to not modify ``/etc/hosts``
522 - Added network interface parameters
523 - Added dry run mode to some LUs
524 - Added RAPI resources:
526 - ``/2/instances/[instance_name]/info``
527 - ``/2/instances/[instance_name]/replace-disks``
528 - ``/2/nodes/[node_name]/evacuate``
529 - ``/2/nodes/[node_name]/migrate``
530 - ``/2/nodes/[node_name]/role``
531 - ``/2/nodes/[node_name]/storage``
532 - ``/2/nodes/[node_name]/storage/modify``
533 - ``/2/nodes/[node_name]/storage/repair``
535 - Added OpCodes to evacuate or migrate all instances on a node
536 - Added new command to list storage elements on nodes (``gnt-node
537 list-storage``) and modify them (``gnt-node modify-storage``)
538 - Added new ssconf files with master candidate IP address
539 (``ssconf_master_candidates_ips``), node primary IP address
540 (``ssconf_node_primary_ips``) and node secondary IP address
541 (``ssconf_node_secondary_ips``)
542 - Added ``ganeti-confd`` and a client library to query the Ganeti
543 configuration via UDP
544 - Added ability to run hooks after cluster initialization and before
546 - Added automatic mode for disk replace (``gnt-instance replace-disks
548 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
549 after catastrophic data-loss
550 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
552 - Added ``gnt-instance move`` command to move instances
553 - Added ``gnt-cluster watcher`` command to control watcher
554 - Added ``gnt-node powercycle`` command to powercycle nodes
555 - Added new job status field ``lock_status``
556 - Added parseable error codes to cluster verification (``gnt-cluster
557 verify --error-codes``) and made output less verbose (use
558 ``--verbose`` to restore previous behaviour)
559 - Added UUIDs to the main config entities (cluster, nodes, instances)
560 - Added support for OS variants
561 - Added support for hashed passwords in the Ganeti remote API users file
563 - Added option to specify maximum timeout on instance shutdown
564 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
565 - Added new helper script to start and stop Ganeti daemons
566 (``daemon-util``), with the intent to reduce the work necessary to
567 adjust Ganeti for non-Debian distributions and to start/stop daemons
569 - Added more unittests
570 - Fixed critical bug in ganeti-masterd startup
571 - Removed the configure-time ``kvm-migration-port`` parameter, this is
572 now customisable at the cluster level for both the KVM and Xen
573 hypervisors using the new ``migration_port`` parameter
574 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
575 reinstalling an instance
576 - Allowed ``@`` in tag names
577 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
578 - Many documentation updates
579 - Distribute hypervisor files on ``gnt-cluster redist-conf``
580 - ``gnt-instance reinstall`` can now reinstall multiple instances
581 - Updated many command line parameters
582 - Introduced new OS API version 15
583 - No longer support a default hypervisor
584 - Treat virtual LVs as inexistent
585 - Improved job locking logic to reduce lock contention
586 - Match instance and node names case insensitively
587 - Reimplemented bash completion script to be more complete
594 *(Released Thu, 4 Feb 2010)*
596 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
597 - Fix a string formatting bug
598 - Improve safety of the code in some error paths
599 - Improve data validation in the master of values returned from nodes
605 *(Released Thu, 17 Dec 2009)*
607 - Fix security issue due to missing validation of iallocator names; this
608 allows local and remote execution of arbitrary executables
609 - Fix failure of gnt-node list during instance removal
610 - Ship the RAPI documentation in the archive
616 *(Released Wed, 30 Sep 2009)*
618 - Fixed many wrong messages
619 - Fixed a few bugs related to the locking library
620 - Fixed MAC checking at instance creation time
621 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
622 - Fixed a few issues related to signal handling in both daemons and
624 - Fixed the example startup script provided
625 - Fixed insserv dependencies in the example startup script (patch from
627 - Fixed handling of drained nodes in the iallocator framework
628 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
630 - Fixed error related to invalid job IDs in job polling
631 - Fixed job/opcode persistence on unclean master shutdown
632 - Fixed handling of partial job processing after unclean master
634 - Fixed error reporting from LUs, previously all errors were converted
635 into execution errors
636 - Fixed error reporting from burnin
637 - Decreased significantly the memory usage of the job queue
638 - Optimised slightly multi-job submission
639 - Optimised slightly opcode loading
640 - Backported the multi-job submit framework from the development
641 branch; multi-instance start and stop should be faster
642 - Added script to clean archived jobs after 21 days; this will reduce
643 the size of the queue directory
644 - Added some extra checks in disk size tracking
645 - Added an example ethers hook script
646 - Added a cluster parameter that prevents Ganeti from modifying of
648 - Added more node information to RAPI responses
649 - Added a ``gnt-job watch`` command that allows following the ouput of a
651 - Added a bind-address option to ganeti-rapi
652 - Added more checks to the configuration verify
653 - Enhanced the burnin script such that some operations can be retried
655 - Converted instance reinstall to multi-instance model
661 *(Released Fri, 7 Aug 2009)*
663 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
664 to allow using the pre-2.0.2 behaviour in activation, if any existing
665 instances have mismatched disk sizes in the configuration
666 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
667 any configuration mismatches for disk sizes
668 - Added ``gnt-master cluste-failover --no-voting`` to allow master
669 failover to work on two-node clusters
670 - Fixed the ``--net`` option of ``gnt-backup import``, which was
672 - Fixed detection of OS script errors in ``gnt-backup export``
673 - Fixed exit code of ``gnt-backup export``
679 *(Released Fri, 17 Jul 2009)*
681 - Added experimental support for stripped logical volumes; this should
682 enhance performance but comes with a higher complexity in the block
683 device handling; stripping is only enabled when passing
684 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
685 affected even in the non-stripped mode
686 - Improved resiliency against transient failures at the end of DRBD
687 resyncs, and in general of DRBD resync checks
688 - Fixed a couple of issues with exports and snapshot errors
689 - Fixed a couple of issues in instance listing
690 - Added display of the disk size in ``gnt-instance info``
691 - Fixed checking for valid OSes in instance creation
692 - Fixed handling of the "vcpus" parameter in instance listing and in
693 general of invalid parameters
694 - Fixed http server library, and thus RAPI, to handle invalid
695 username/password combinations correctly; this means that now they
696 report unauthorized for queries too, not only for modifications,
697 allowing earlier detect of configuration problems
698 - Added a new "role" node list field, equivalent to the master/master
699 candidate/drained/offline flags combinations
700 - Fixed cluster modify and changes of candidate pool size
701 - Fixed cluster verify error messages for wrong files on regular nodes
702 - Fixed a couple of issues with node demotion from master candidate role
703 - Fixed node readd issues
704 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
705 - Added a new ``--no-voting`` option for masterfailover to fix failover
706 on two-nodes clusters when the former master node is unreachable
707 - Added instance reinstall over RAPI
713 *(Released Tue, 16 Jun 2009)*
715 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
716 allow re-adding the start in single-user option (regression from 1.2)
717 - the watcher writes the instance status to a file, to allow monitoring
718 to report the instance status (from the master) based on cached
719 results of the watcher's queries; while this can get stale if the
720 watcher is being locked due to other work on the cluster, this is
722 - the watcher now also restarts the node daemon and the rapi daemon if
724 - fixed the watcher to handle full and drained queue cases
725 - hooks export more instance data in the environment, which helps if
726 hook scripts need to take action based on the instance's properties
727 (no longer need to query back into ganeti)
728 - instance failovers when the instance is stopped do not check for free
729 RAM, so that failing over a stopped instance is possible in low memory
731 - rapi uses queries for tags instead of jobs (for less job traffic), and
732 for cluster tags it won't talk to masterd at all but read them from
734 - a couple of error handling fixes in RAPI
735 - drbd handling: improved the error handling of inconsistent disks after
736 resync to reduce the frequency of "there are some degraded disks for
737 this instance" messages
738 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
739 error handling path called a wrong function name)
745 *(Released Wed, 27 May 2009)*
747 - no changes from rc5
750 Version 2.0 release candidate 5
751 -------------------------------
753 *(Released Wed, 20 May 2009)*
755 - fix a couple of bugs (validation, argument checks)
756 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
757 - some small improvements to RAPI and IAllocator
758 - make watcher automatically start the master daemon if down
761 Version 2.0 release candidate 4
762 -------------------------------
764 *(Released Mon, 27 Apr 2009)*
766 - change the OS list to not require locks; this helps with big clusters
767 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
768 volume group is broken
769 - ``gnt-instance info``, without any arguments, doesn't run for all
770 instances anymore; either pass ``--all`` or pass the desired
771 instances; this helps against mistakes on big clusters where listing
772 the information for all instances takes a long time
773 - miscellaneous doc and man pages fixes
776 Version 2.0 release candidate 3
777 -------------------------------
779 *(Released Wed, 8 Apr 2009)*
781 - Change the internal locking model of some ``gnt-node`` commands, in
782 order to reduce contention (and blocking of master daemon) when
783 batching many creation/reinstall jobs
784 - Fixes to Xen soft reboot
785 - No longer build documentation at build time, instead distribute it in
786 the archive, in order to reduce the need for the whole docbook/rst
790 Version 2.0 release candidate 2
791 -------------------------------
793 *(Released Fri, 27 Mar 2009)*
795 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
796 - Fix watcher startup sequence, improves the behaviour of busy clusters
797 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
798 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
799 bind address and other places
800 - Some documentation fixes and updates
803 Version 2.0 release candidate 1
804 -------------------------------
806 *(Released Mon, 2 Mar 2009)*
808 - More documentation updates, now all docs should be more-or-less
810 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
812 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
814 - Fix an issue related to $libdir/run/ganeti and cluster creation
820 *(Released Thu, 19 Feb 2009)*
822 - Xen PVM and KVM have switched the default value for the instance root
823 disk to the first partition on the first drive, instead of the whole
824 drive; this means that the OS installation scripts must be changed
826 - Man pages have been updated
827 - RAPI has been switched by default to HTTPS, and the exported functions
828 should all work correctly
829 - RAPI v1 has been removed
830 - Many improvements to the KVM hypervisor
831 - Block device errors are now better reported
832 - Many other bugfixes and small improvements
838 *(Released Mon, 26 Jan 2009)*
840 - Version 2 is a general rewrite of the code and therefore the
841 differences are too many to list, see the design document for 2.0 in
842 the ``doc/`` subdirectory for more details
843 - In this beta version there is not yet a migration path from 1.2 (there
844 will be one in the final 2.0 release)
845 - A few significant changes are:
847 - all commands are executed by a daemon (``ganeti-masterd``) and the
848 various ``gnt-*`` commands are just front-ends to it
849 - all the commands are entered into, and executed from a job queue,
850 see the ``gnt-job(8)`` manpage
851 - the RAPI daemon supports read-write operations, secured by basic
852 HTTP authentication on top of HTTPS
853 - DRBD version 0.7 support has been removed, DRBD 8 is the only
854 supported version (when migrating from Ganeti 1.2 to 2.0, you need
855 to migrate to DRBD 8 first while still running Ganeti 1.2)
856 - DRBD devices are using statically allocated minor numbers, which
857 will be assigned to existing instances during the migration process
858 - there is support for both Xen PVM and Xen HVM instances running on
860 - KVM virtualization is supported too
861 - file-based storage has been implemented, which means that it is
862 possible to run the cluster without LVM and DRBD storage, for
863 example using a shared filesystem exported from shared storage (and
864 still have live migration)
870 *(Released Tue, 13 Jan 2009)*
872 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
873 - Reuse the old instance mac address by default on instance import, if
874 the instance name is the same.
875 - Handle situations in which the node info rpc returns incomplete
877 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
878 - Improved version of batcher:
881 - instance mac address support
882 - support for HVM clusters/instances
884 - Add an option to show the number of cpu sockets and nodes in
886 - Support OSes that handle more than one version of the OS api (but do
887 not change the current API in any other way)
888 - Fix ``gnt-node migrate``
889 - ``gnt-debug`` man page
890 - Fixes various more typos and small issues
891 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
897 *(Released Wed, 24 Sep 2008)*
899 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
900 type of disk exported to fully virtualized instances.
901 - provide access to the serial console of HVM instances
902 - instance auto_balance flag, set by default. If turned off it will
903 avoid warnings on cluster verify if there is not enough memory to fail
904 over an instance. in the future it will prevent automatically failing
905 it over when we will support that.
906 - batcher tool for instance creation, see ``tools/README.batcher``
907 - ``gnt-instance reinstall --select-os`` to interactively select a new
908 operating system when reinstalling an instance.
909 - when changing the memory amount on instance modify a check has been
910 added that the instance will be able to start. also warnings are
911 emitted if the instance will not be able to fail over, if auto_balance
913 - documentation fixes
914 - sync fields between ``gnt-instance list/modify/add/import``
915 - fix a race condition in drbd when the sync speed was set after giving
916 the device a remote peer.
922 *(Released Tue, 22 Jul 2008)*
924 - note: the allowed size and number of tags per object were reduced
925 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
926 - fixed twisted 8.x compatibility
927 - fixed ``gnt-instance replace-disks`` with iallocator
928 - add TCP keepalives on twisted connections to detect restarted nodes
929 - disk increase support, see ``gnt-instance grow-disk``
930 - implement bulk node/instance query for RAPI
931 - add tags in node/instance listing (optional)
932 - experimental migration (and live migration) support, read the man page
933 for ``gnt-instance migrate``
934 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
935 has some small improvements in handling its state file
941 *(Released Fri, 13 Jun 2008)*
943 - Experimental readonly, REST-based remote API implementation;
944 automatically started on master node, TCP port 5080, if enabled by
945 ``--enable-rapi`` parameter to configure script.
946 - Instance allocator support. Add and import instance accept a
947 ``--iallocator`` parameter, and call that instance allocator to decide
948 which node to use for the instance. The iallocator document describes
949 what's expected from an allocator script.
950 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
951 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
952 node is lost there is still enough memory to fail over the instances
954 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
955 ``gnt-cluster verify``, to check for site-specific compliance. All the
956 hooks will run, and their output, if any, will be displayed. Any
957 failing hook will make the verification return an error value.
958 - ``gnt-cluster verify`` now checks that its peers are reachable on the
959 primary and secondary interfaces
960 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
961 that is still declared as part of the cluster and has failed.
962 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
963 specifying output fields, that just adds the chosen fields to the
965 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
966 export from the filesystem.
967 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
968 have been added. Using them you can enable/disable acpi and pae
969 support, and specify a path for a cd image to be exported to the
970 instance. These parameters as the name suggest only work on HVM
972 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
973 PAE support will be set to the previously hardcoded values, but the
974 (previously hardcoded) path to the CDROM ISO image will be unset and
975 if required, needs to be set manually with ``gnt-instance modify``
977 - The address to which an instance's VNC console is bound is now
978 selectable per-instance, rather than being cluster wide. Of course
979 this only applies to instances controlled via VNC, so currently just
980 applies to HVM clusters.
986 *(Released Mon, 18 Feb 2008)*
988 - more tweaks to the disk activation code (especially helpful for DRBD)
989 - change the default ``gnt-instance list`` output format, now there is
990 one combined status field (see the manpage for the exact values this
992 - some more fixes for the mac export to hooks change
993 - make Ganeti not break with DRBD 8.2.x (which changed the version
994 format in ``/proc/drbd``) (issue 24)
995 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
996 template, allowing migration from DRBD0.7+MD to DRBD8
1002 *(Released Wed, 30 Jan 2008)*
1004 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1006 - add command aliases infrastructure and a few aliases
1007 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1008 man pages and the ``--help`` option of ``gnt-node
1009 list``/``gnt-instance list``
1010 - fix ``gnt-backup list`` with down nodes (issue 21)
1011 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1012 - fix the dist archive and add a check for including svn/git files in
1014 - some developer-related changes: improve the burnin and the QA suite,
1015 add an upload script for testing during development
1021 *(Released Wed, 16 Jan 2008)*
1023 - experimental HVM support, read the install document, section
1024 "Initializing the cluster"
1025 - allow for the PVM hypervisor per-instance kernel and initrd paths
1026 - add a new command ``gnt-cluster verify-disks`` which uses a new
1027 algorithm to improve the reconnection of the DRBD pairs if the device
1028 on the secondary node has gone away
1029 - make logical volume code auto-activate LVs at disk activation time
1030 - slightly improve the speed of activating disks
1031 - allow specification of the MAC address at instance creation time, and
1032 changing it later via ``gnt-instance modify``
1033 - fix handling of external commands that generate lots of output on
1035 - update documentation with regard to minimum version of DRBD8 supported
1041 *(Released Tue, 4 Dec 2007)*
1043 - Log the ``xm create`` output to the node daemon log on failure (to
1044 help diagnosing the error)
1045 - In debug mode, log all external commands output if failed to the logs
1046 - Change parsing of lvm commands to ignore stderr
1052 *(Released Wed, 28 Nov 2007)*
1054 - Another round of updates to the DRBD 8 code to deal with more failures
1055 in the replace secondary node operation
1056 - Some more logging of failures in disk operations (lvm, drbd)
1057 - A few documentation updates
1064 *(Released Tue, 13 Nov 2007)*
1066 - Change configuration file format from Python's Pickle to JSON.
1067 Upgrading is possible using the cfgupgrade utility.
1068 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1069 faster replace disks and is more stable (DRBD 8 has many improvements
1070 compared to DRBD 0.7)
1071 - Added command line tags support (see man pages for ``gnt-instance``,
1072 ``gnt-node``, ``gnt-cluster``)
1073 - Added instance rename support
1074 - Added multi-instance startup/shutdown
1075 - Added cluster rename support
1076 - Added ``gnt-node evacuate`` to simplify some node operations
1077 - Added instance reboot operation that can speedup reboot as compared to
1079 - Soften the requirement that hostnames are in FQDN format
1080 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1082 - Removed dependency on debian's patched fping that uses the
1083 non-standard ``-S`` option
1084 - Now the OS definitions are searched for in multiple, configurable
1085 paths (easier for distros to package)
1086 - Some changes to the hooks infrastructure (especially the new
1087 post-configuration update hook)
1088 - Other small bugfixes
1090 .. vim: set textwidth=72 :