7 *(Released Tue, 19 Oct 2010)*
9 - Disable SSL session ID cache in RPC client
15 *(Released Thu, 14 Oct 2010)*
17 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
18 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
20 - Fix problems with interaction between Curl and Python's HTTP server,
21 resulting in increased speed in many RPC calls
22 - Improve our release script to prevent breakage with older aclocal and
29 *(Released Thu, 7 Oct 2010)*
31 - Fixed issue 125, replace hardcoded “xenvg” in ``gnt-cluster`` with
32 value retrieved from master
33 - Added support for blacklisted or hidden OS definitions
34 - Added simple lock monitor (accessible via (``gnt-debug locks``)
35 - Added support for -mem-path in KVM hypervisor abstraction layer
36 - Allow overriding instance parameters in tool for inter-cluster
37 instance moves (``tools/move-instance``)
38 - Improved opcode summaries (e.g. in ``gnt-job list``)
39 - Improve consistency of OS listing by sorting it
40 - Documentation updates
46 *(Released Fri, 8 Oct 2010)*
48 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
54 *(Released Mon, 4 Oct 2010)*
56 - Fixed regression in ``gnt-instance rename``
62 *(Released Wed, 22 Sep 2010)*
64 - Fixed OS_VARIANT variable for OS scripts
65 - Fixed cluster tag operations via RAPI
66 - Made ``setup-ssh`` exit with non-zero code if an error occurred
67 - Disabled RAPI CA checks in watcher
73 *(Released Mon, 23 Aug 2010)*
75 - Support DRBD versions of the format "a.b.c.d"
77 - Re-introduce support for usage from multiple threads in RAPI client
78 - Instance renames and modify via RAPI
79 - Work around race condition between processing and archival in job
81 - Mark opcodes following failed one as failed, too
82 - Job field ``lock_status`` was removed due to difficulties making it
83 work with the changed job queue in Ganeti 2.2; a better way to monitor
84 locks is expected for a later 2.2.x release
85 - Fixed dry-run behaviour with many commands
86 - Support ``ssh-agent`` again when adding nodes
87 - Many additional bugfixes
93 *(Released Fri, 30 Jul 2010)*
95 Important change: the internal RPC mechanism between Ganeti nodes has
96 changed from using a home-grown http library (based on the Python base
97 libraries) to use the PycURL library. This requires that PycURL is
98 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
99 against GnuTLS by default. cURL's support for GnuTLS had known issues
100 before cURL 7.21.0 and we recommend using the latest cURL release or
101 linking against OpenSSL. Most other distributions already link PycURL
102 and cURL against OpenSSL. The command::
104 python -c 'import pycurl; print pycurl.version'
106 can be used to determine the libraries PycURL and cURL are linked
109 Other significant changes:
111 - Rewrote much of the internals of the job queue, in order to achieve
112 better parallelism; this decouples job query operations from the job
113 processing, and it should allow much nicer behaviour of the master
114 daemon under load, and it also has uncovered some long-standing bugs
115 related to the job serialisation (now fixed)
116 - Added a default iallocator setting to the cluster parameters,
117 eliminating the need to always pass nodes or an iallocator for
118 operations that require selection of new node(s)
119 - Added experimental support for the LXC virtualization method
120 - Added support for OS parameters, which allows the installation of
121 instances to pass parameter to OS scripts in order to customise the
123 - Added a hypervisor parameter controlling the migration type (live or
124 non-live), since hypervisors have various levels of reliability; this
125 has renamed the 'live' parameter to 'mode'
126 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
127 logical volumes, meaning that cluster verify will ignore them and not
128 flag their presence as errors
129 - The watcher will now reset the error count for failed instances after
130 8 hours, thus allowing self-healing if the problem that caused the
131 instances to be down/fail to start has cleared in the meantime
132 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
133 check for, and warn, if the drbd module parameter ``usermode_helper``
134 is not consistent with the cluster-wide setting; this is needed to
135 make diagnose easier of failed drbd creations
136 - Started adding base IPv6 support, but this is not yet
137 enabled/available for use
138 - Rename operations (cluster, instance) will now return the new name,
139 which is especially useful if a short name was passed in
140 - Added support for instance migration in RAPI
141 - Added a tool to pre-configure nodes for the SSH setup, before joining
142 them to the cluster; this will allow in the future a simplified model
143 for node joining (but not yet fully enabled in 2.2); this needs the
144 paramiko python library
145 - Fixed handling of name-resolving errors
146 - Fixed consistency of job results on the error path
147 - Fixed master-failover race condition when executed multiple times in
149 - Fixed many bugs related to the job queue (mostly introduced during the
150 2.2 development cycle, so not all are impacting 2.1)
151 - Fixed instance migration with missing disk symlinks
152 - Fixed handling of unknown jobs in ``gnt-job archive``
153 - And many other small fixes/improvements
157 - Enhanced both the unittest and the QA coverage
158 - Switched the opcode validation to a generic model, and extended the
159 validation to all opcode parameters
160 - Changed more parts of the code that write shell scripts to use the
162 - Switched the master daemon to use the asyncore library for the Luxi
169 *(Released Thu, 17 Jun 2010)*
171 - Added tool (``move-instance``) and infrastructure to move instances
172 between separate clusters (see :doc:`separate documentation
173 <move-instance>` and :doc:`design document <design-2.2>`)
174 - Added per-request RPC timeout
175 - RAPI now requires a Content-Type header for requests with a body (e.g.
176 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
177 :rfc:`2616` (HTTP/1.1), section 7.2.1)
178 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
180 - Implemented initial support for running Ganeti daemons as separate
181 users, see configure-time flags ``--with-user-prefix`` and
182 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
184 - Instances can be removed after export (``gnt-backup export
186 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
187 key (instead of 1024 bit)
188 - Added new cluster configuration file for cluster domain secret
189 - Import/export now use SSL instead of SSH
190 - Added support for showing estimated time when exporting an instance,
191 see the ``ganeti-os-interface(7)`` manpage and look for
197 *(Released Tue, 24 Aug 2010)*
200 - Don't ignore secondary node silently on non-mirrored disk templates
202 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
203 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
204 - Properly document the usb_mouse parameter
205 - Fix path in ganeti-rapi(8) (issue 116)
206 - Adjust error message when the ganeti user's .ssh directory is
208 - Add same-node-check when changing the disk template to drbd
214 *(Released Fri, 16 Jul 2010)*
217 - Add an option to only select some reboot types during qa/burnin.
218 (on some hypervisors consequent reboots are not supported)
219 - Fix infrequent race condition in master failover. Sometimes the old
220 master ip address would be still detected as up for a short time
221 after it was removed, causing failover to fail.
222 - Decrease mlockall warnings when the ctypes module is missing. On
223 Python 2.4 we support running even if no ctypes module is installed,
224 but we were too verbose about this issue.
225 - Fix building on old distributions, on which man doesn't have a
227 - Fix RAPI not to ignore the MAC address on instance creation
228 - Implement the old instance creation format in the RAPI client.
234 *(Released Thu, 01 Jul 2010)*
236 A small bugfix release:
237 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
238 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
239 a non-existing option
240 - Fix instance create: the --force-variant option was ignored
241 - Improve pylint 0.21 compatibility and warnings with Python 2.6
242 - Fix modify node storage with non-FQDN arguments
243 - Fix RAPI client to authenticate under Python 2.6 when used
244 for more than 5 requests needing authentication
245 - Fix gnt-instance modify -t (storage) giving a wrong error message
246 when converting a non-shutdown drbd instance to plain
252 *(Released Fri, 18 Jun 2010)*
254 A small bugfix release:
256 - Fix live migration of KVM instances started with older Ganeti
257 versions which had fewer hypervisor parameters
258 - Fix gnt-instance grow-disk on down instances
259 - Fix an error-reporting bug during instance migration
260 - Better checking of the ``--net`` and ``--disk`` values, to avoid
261 silently ignoring broken ones
262 - Fix an RPC error reporting bug affecting, for example, RAPI client
264 - Fix bug triggered by different API version os-es on different nodes
265 - Fix a bug in instance startup with custom hvparams: OS level
266 parameters would fail to be applied.
267 - Fix the RAPI client under Python 2.6 (but more work is needed to
268 make it work completely well with OpenSSL)
269 - Fix handling of errors when resolving names from DNS
275 *(Released Thu, 3 Jun 2010)*
277 A medium sized development cycle. Some new features, and some
278 fixes/small improvements/cleanups.
283 The node deamon now tries to mlock itself into memory, unless the
284 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
285 its logs, and falls back to console logging. This allows emergency
286 features such as ``gnt-node powercycle`` to work even in the event of a
287 broken node disk (tested offlining the disk hosting the node's
288 filesystem and dropping its memory caches; don't try this at home)
290 KVM: add vhost-net acceleration support. It can be tested with a new
291 enough version of the kernel and of qemu-kvm.
293 KVM: Add instance chrooting feature. If you use privilege dropping for
294 your VMs you can also now force them to chroot to an empty directory,
295 before starting the emulated guest.
297 KVM: Add maximum migration bandwith and maximum downtime tweaking
298 support (requires a new-enough version of qemu-kvm).
300 Cluster verify will now warn if the master node doesn't have the master
303 Add a new (incompatible) instance creation request format to RAPI which
304 supports all parameters (previously only a subset was supported, and it
305 wasn't possible to extend the old format to accomodate all the new
306 features. The old format is still supported, and a client can check for
307 this feature, before using it, by checking for its presence in the
308 ``features`` RAPI resource.
310 Now with ancient latin support. Try it passing the ``--roman`` option to
311 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
312 (requires the python-roman module to be installed, in order to work).
317 As usual many internal code refactorings, documentation updates, and
320 - Lots of improvements and cleanups to the experimental Remote API
321 (RAPI) client library.
322 - A new unit test suite for the core daemon libraries.
323 - A fix to creating missing directories makes sure the umask is not
324 applied anymore. This enforces the same directory permissions
326 - Better handling terminating daemons with ctrl+c (used when running
327 them in debugging mode).
328 - Fix a race condition in live migrating a KVM instance, when stat()
329 on the old proc status file returned EINVAL, which is an unexpected
331 - Fixed manpage checking with newer man and utf-8 charachters. But now
332 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
338 *(Released Fri, 7 May 2010)*
340 Fix a bug which prevented untagged KVM instances from starting.
346 *(Released Fri, 7 May 2010)*
348 Another release with a long development cycle, during which many
349 different features were added.
354 The KVM hypervisor now can run the individual instances as non-root, to
355 reduce the impact of a VM being hijacked due to bugs in the
356 hypervisor. It is possible to run all instances as a single (non-root)
357 user, to manually specify a user for each instance, or to dynamically
358 allocate a user out of a cluster-wide pool to each instance, with the
359 guarantee that no two instances will run under the same user ID on any
362 An experimental RAPI client library, that can be used standalone
363 (without the other Ganeti libraries), is provided in the source tree as
364 ``lib/rapi/client.py``. Note this client might change its interface in
365 the future, as we iterate on its capabilities.
367 A new command, ``gnt-cluster renew-crypto`` has been added to easily
368 replace the cluster's certificates and crypto keys. This might help in
369 case they have been compromised, or have simply expired.
371 A new disk option for instance creation has been added that allows one
372 to "adopt" currently existing logical volumes, with data
373 preservation. This should allow easier migration to Ganeti from
374 unmanaged (or managed via other software) instances.
376 Another disk improvement is the possibility to convert between redundant
377 (DRBD) and plain (LVM) disk configuration for an instance. This should
378 allow better scalability (starting with one node and growing the
379 cluster, or shrinking a two-node cluster to one node).
381 A new feature that could help with automated node failovers has been
382 implemented: if a node sees itself as offline (by querying the master
383 candidates), it will try to shutdown (hard) all instances and any active
384 DRBD devices. This reduces the risk of duplicate instances if an
385 external script automatically failovers the instances on such nodes. To
386 enable this, the cluster parameter ``maintain_node_health`` should be
387 enabled; in the future this option (per the name) will enable other
388 automatic maintenance features.
390 Instance export/import now will reuse the original instance
391 specifications for all parameters; that means exporting an instance,
392 deleting it and the importing it back should give an almost identical
393 instance. Note that the default import behaviour has changed from
394 before, where it created only one NIC; now it recreates the original
397 Cluster verify has added a few new checks: SSL certificates validity,
398 /etc/hosts consistency across the cluster, etc.
403 As usual, many internal changes were done, documentation fixes,
406 - Fixed cluster initialization with disabled cluster storage (regression
408 - File-based storage supports growing the disks
409 - Fixed behaviour of node role changes
410 - Fixed cluster verify for some corner cases, plus a general rewrite of
411 cluster verify to allow future extension with more checks
412 - Fixed log spamming by watcher and node daemon (regression introduced
414 - Fixed possible validation issues when changing the list of enabled
416 - Fixed cleanup of /etc/hosts during node removal
417 - Fixed RAPI response for invalid methods
418 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
419 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
420 from ide disks, etc.)
421 - Allow OS changes without re-installation (to record a changed OS
422 outside of Ganeti, or to allow OS renames)
423 - Allow instance creation without OS installation (useful for example if
424 the OS will be installed manually, or restored from a backup not in
426 - Implemented option to make cluster ``copyfile`` use the replication
428 - Added list of enabled hypervisors to ssconf (possibly useful for
430 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
432 - A partial form of node re-IP is possible via node readd, which now
433 allows changed node primary IP
434 - Command line utilities now show an informational message if the job is
436 - The logs of the master daemon now show the PID/UID/GID of the
443 *(Released Fri, 12 Mar 2010)*
445 During the 2.1.0 long release candidate cycle, a lot of improvements and
446 changes have accumulated with were released later as 2.1.1.
451 The node evacuate command (``gnt-node evacuate``) was significantly
452 rewritten, and as such the IAllocator protocol was changed - a new
453 request type has been added. This unfortunate change during a stable
454 series is designed to improve performance of node evacuations; on
455 clusters with more than about five nodes and which are well-balanced,
456 evacuation should proceed in parallel for all instances of the node
457 being evacuated. As such, any existing IAllocator scripts need to be
458 updated, otherwise the above command will fail due to the unknown
459 request. The provided "dumb" allocator has not been updated; but the
460 ganeti-htools package supports the new protocol since version 0.2.4.
462 Another important change is increased validation of node and instance
463 names. This might create problems in special cases, if invalid host
464 names are being used.
466 Also, a new layer of hypervisor parameters has been added, that sits at
467 OS level between the cluster defaults and the instance ones. This allows
468 customisation of virtualization parameters depending on the installed
469 OS. For example instances with OS 'X' may have a different KVM kernel
470 (or any other parameter) than the cluster defaults. This is intended to
471 help managing a multiple OSes on the same cluster, without manual
472 modification of each instance's parameters.
474 A tool for merging clusters, ``cluster-merge``, has been added in the
480 - Improved the int/float conversions that should make the code more
481 robust in face of errors from the node daemons
482 - Fixed the remove node code in case of internal configuration errors
483 - Fixed the node daemon behaviour in face of inconsistent queue
484 directory (e.g. read-only file-system where we can't open the files
486 - Fixed the behaviour of gnt-node modify for master candidate demotion;
487 now it either aborts cleanly or, if given the new “auto_promote”
488 parameter, will automatically promote other nodes as needed
489 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
490 completely prevent Ganeti from working
491 - Fixed bug for instance export when not all disks were successfully
493 - Fixed behaviour of node add when the new node is slow in starting up
495 - Fixed handling of signals in the LUXI client, which should improve
496 behaviour of command-line scripts
497 - Added checks for invalid node/instance names in the configuration (now
498 flagged during cluster verify)
499 - Fixed watcher behaviour for disk activation errors
500 - Fixed two potentially endless loops in http library, which led to the
501 RAPI daemon hanging and consuming 100% CPU in some cases
502 - Fixed bug in RAPI daemon related to hashed passwords
503 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
505 - Enhanced compatibility with non-Debian OSes, but not using absolute
506 path in some commands and allowing customisation of the ssh
507 configuration directory
508 - Fixed possible future issue with new Python versions by abiding to the
509 proper use of ``__slots__`` attribute on classes
510 - Added checks that should prevent directory traversal attacks
511 - Many documentation fixes based on feedback from users
516 - Added an “early_release” more for instance replace disks and node
517 evacuate, where we release locks earlier and thus allow higher
518 parallelism within the cluster
519 - Added watcher hooks, intended to allow the watcher to restart other
520 daemons (e.g. from the ganeti-nbma project), but they can be used of
521 course for any other purpose
522 - Added a compile-time disable for DRBD barriers, to increase
523 performance if the administrator trusts the power supply or the
524 storage system to not lose writes
525 - Added the option of using syslog for logging instead of, or in
526 addition to, Ganeti's own log files
527 - Removed boot restriction for paravirtual NICs for KVM, recent versions
528 can indeed boot from a paravirtual NIC
529 - Added a generic debug level for many operations; while this is not
530 used widely yet, it allows one to pass the debug value all the way to
532 - Enhanced the hooks environment for instance moves (failovers,
533 migrations) where the primary/secondary nodes changed during the
534 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
535 - Enhanced data validations for many user-supplied values; one important
536 item is the restrictions imposed on instance and node names, which
537 might reject some (invalid) host names
538 - Add a configure-time option to disable file-based storage, if it's not
539 needed; this allows greater security separation between the master
540 node and the other nodes from the point of view of the inter-node RPC
542 - Added user notification in interactive tools if job is waiting in the
543 job queue or trying to acquire locks
544 - Added log messages when a job is waiting for locks
545 - Added filtering by node tags in instance operations which admit
546 multiple instances (start, stop, reboot, reinstall)
547 - Added a new tool for cluster mergers, ``cluster-merge``
548 - Parameters from command line which are of the form ``a=b,c=d`` can now
549 use backslash escapes to pass in values which contain commas,
550 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
552 - For KVM, the instance name is the first parameter passed to KVM, so
553 that it's more visible in the process list
559 *(Released Tue, 2 Mar 2010)*
561 Ganeti 2.1 brings many improvements with it. Major changes:
563 - Added infrastructure to ease automated disk repairs
564 - Added new daemon to export configuration data in a cheaper way than
566 - Instance NICs can now be routed instead of being associated with a
568 - Improved job locking logic to reduce impact of jobs acquiring multiple
569 locks waiting for other long-running jobs
571 In-depth implementation details can be found in the Ganeti 2.1 design
577 - Added chroot hypervisor
578 - Added more options to xen-hvm hypervisor (``kernel_path`` and
580 - Added more options to xen-pvm hypervisor (``use_bootloader``,
581 ``bootloader_path`` and ``bootloader_args``)
582 - Added the ``use_localtime`` option for the xen-hvm and kvm
583 hypervisors, and the default value for this has changed to false (in
584 2.0 xen-hvm always enabled it)
585 - Added luxi call to submit multiple jobs in one go
586 - Added cluster initialization option to not modify ``/etc/hosts``
588 - Added network interface parameters
589 - Added dry run mode to some LUs
590 - Added RAPI resources:
592 - ``/2/instances/[instance_name]/info``
593 - ``/2/instances/[instance_name]/replace-disks``
594 - ``/2/nodes/[node_name]/evacuate``
595 - ``/2/nodes/[node_name]/migrate``
596 - ``/2/nodes/[node_name]/role``
597 - ``/2/nodes/[node_name]/storage``
598 - ``/2/nodes/[node_name]/storage/modify``
599 - ``/2/nodes/[node_name]/storage/repair``
601 - Added OpCodes to evacuate or migrate all instances on a node
602 - Added new command to list storage elements on nodes (``gnt-node
603 list-storage``) and modify them (``gnt-node modify-storage``)
604 - Added new ssconf files with master candidate IP address
605 (``ssconf_master_candidates_ips``), node primary IP address
606 (``ssconf_node_primary_ips``) and node secondary IP address
607 (``ssconf_node_secondary_ips``)
608 - Added ``ganeti-confd`` and a client library to query the Ganeti
609 configuration via UDP
610 - Added ability to run hooks after cluster initialization and before
612 - Added automatic mode for disk replace (``gnt-instance replace-disks
614 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
615 after catastrophic data-loss
616 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
618 - Added ``gnt-instance move`` command to move instances
619 - Added ``gnt-cluster watcher`` command to control watcher
620 - Added ``gnt-node powercycle`` command to powercycle nodes
621 - Added new job status field ``lock_status``
622 - Added parseable error codes to cluster verification (``gnt-cluster
623 verify --error-codes``) and made output less verbose (use
624 ``--verbose`` to restore previous behaviour)
625 - Added UUIDs to the main config entities (cluster, nodes, instances)
626 - Added support for OS variants
627 - Added support for hashed passwords in the Ganeti remote API users file
629 - Added option to specify maximum timeout on instance shutdown
630 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
631 - Added new helper script to start and stop Ganeti daemons
632 (``daemon-util``), with the intent to reduce the work necessary to
633 adjust Ganeti for non-Debian distributions and to start/stop daemons
635 - Added more unittests
636 - Fixed critical bug in ganeti-masterd startup
637 - Removed the configure-time ``kvm-migration-port`` parameter, this is
638 now customisable at the cluster level for both the KVM and Xen
639 hypervisors using the new ``migration_port`` parameter
640 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
641 reinstalling an instance
642 - Allowed ``@`` in tag names
643 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
644 - Many documentation updates
645 - Distribute hypervisor files on ``gnt-cluster redist-conf``
646 - ``gnt-instance reinstall`` can now reinstall multiple instances
647 - Updated many command line parameters
648 - Introduced new OS API version 15
649 - No longer support a default hypervisor
650 - Treat virtual LVs as inexistent
651 - Improved job locking logic to reduce lock contention
652 - Match instance and node names case insensitively
653 - Reimplemented bash completion script to be more complete
660 *(Released Thu, 4 Feb 2010)*
662 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
663 - Fix a string formatting bug
664 - Improve safety of the code in some error paths
665 - Improve data validation in the master of values returned from nodes
671 *(Released Thu, 17 Dec 2009)*
673 - Fix security issue due to missing validation of iallocator names; this
674 allows local and remote execution of arbitrary executables
675 - Fix failure of gnt-node list during instance removal
676 - Ship the RAPI documentation in the archive
682 *(Released Wed, 30 Sep 2009)*
684 - Fixed many wrong messages
685 - Fixed a few bugs related to the locking library
686 - Fixed MAC checking at instance creation time
687 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
688 - Fixed a few issues related to signal handling in both daemons and
690 - Fixed the example startup script provided
691 - Fixed insserv dependencies in the example startup script (patch from
693 - Fixed handling of drained nodes in the iallocator framework
694 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
696 - Fixed error related to invalid job IDs in job polling
697 - Fixed job/opcode persistence on unclean master shutdown
698 - Fixed handling of partial job processing after unclean master
700 - Fixed error reporting from LUs, previously all errors were converted
701 into execution errors
702 - Fixed error reporting from burnin
703 - Decreased significantly the memory usage of the job queue
704 - Optimised slightly multi-job submission
705 - Optimised slightly opcode loading
706 - Backported the multi-job submit framework from the development
707 branch; multi-instance start and stop should be faster
708 - Added script to clean archived jobs after 21 days; this will reduce
709 the size of the queue directory
710 - Added some extra checks in disk size tracking
711 - Added an example ethers hook script
712 - Added a cluster parameter that prevents Ganeti from modifying of
714 - Added more node information to RAPI responses
715 - Added a ``gnt-job watch`` command that allows following the ouput of a
717 - Added a bind-address option to ganeti-rapi
718 - Added more checks to the configuration verify
719 - Enhanced the burnin script such that some operations can be retried
721 - Converted instance reinstall to multi-instance model
727 *(Released Fri, 7 Aug 2009)*
729 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
730 to allow using the pre-2.0.2 behaviour in activation, if any existing
731 instances have mismatched disk sizes in the configuration
732 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
733 any configuration mismatches for disk sizes
734 - Added ``gnt-master cluste-failover --no-voting`` to allow master
735 failover to work on two-node clusters
736 - Fixed the ``--net`` option of ``gnt-backup import``, which was
738 - Fixed detection of OS script errors in ``gnt-backup export``
739 - Fixed exit code of ``gnt-backup export``
745 *(Released Fri, 17 Jul 2009)*
747 - Added experimental support for stripped logical volumes; this should
748 enhance performance but comes with a higher complexity in the block
749 device handling; stripping is only enabled when passing
750 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
751 affected even in the non-stripped mode
752 - Improved resiliency against transient failures at the end of DRBD
753 resyncs, and in general of DRBD resync checks
754 - Fixed a couple of issues with exports and snapshot errors
755 - Fixed a couple of issues in instance listing
756 - Added display of the disk size in ``gnt-instance info``
757 - Fixed checking for valid OSes in instance creation
758 - Fixed handling of the "vcpus" parameter in instance listing and in
759 general of invalid parameters
760 - Fixed http server library, and thus RAPI, to handle invalid
761 username/password combinations correctly; this means that now they
762 report unauthorized for queries too, not only for modifications,
763 allowing earlier detect of configuration problems
764 - Added a new "role" node list field, equivalent to the master/master
765 candidate/drained/offline flags combinations
766 - Fixed cluster modify and changes of candidate pool size
767 - Fixed cluster verify error messages for wrong files on regular nodes
768 - Fixed a couple of issues with node demotion from master candidate role
769 - Fixed node readd issues
770 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
771 - Added a new ``--no-voting`` option for masterfailover to fix failover
772 on two-nodes clusters when the former master node is unreachable
773 - Added instance reinstall over RAPI
779 *(Released Tue, 16 Jun 2009)*
781 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
782 allow re-adding the start in single-user option (regression from 1.2)
783 - the watcher writes the instance status to a file, to allow monitoring
784 to report the instance status (from the master) based on cached
785 results of the watcher's queries; while this can get stale if the
786 watcher is being locked due to other work on the cluster, this is
788 - the watcher now also restarts the node daemon and the rapi daemon if
790 - fixed the watcher to handle full and drained queue cases
791 - hooks export more instance data in the environment, which helps if
792 hook scripts need to take action based on the instance's properties
793 (no longer need to query back into ganeti)
794 - instance failovers when the instance is stopped do not check for free
795 RAM, so that failing over a stopped instance is possible in low memory
797 - rapi uses queries for tags instead of jobs (for less job traffic), and
798 for cluster tags it won't talk to masterd at all but read them from
800 - a couple of error handling fixes in RAPI
801 - drbd handling: improved the error handling of inconsistent disks after
802 resync to reduce the frequency of "there are some degraded disks for
803 this instance" messages
804 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
805 error handling path called a wrong function name)
811 *(Released Wed, 27 May 2009)*
813 - no changes from rc5
816 Version 2.0 release candidate 5
817 -------------------------------
819 *(Released Wed, 20 May 2009)*
821 - fix a couple of bugs (validation, argument checks)
822 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
823 - some small improvements to RAPI and IAllocator
824 - make watcher automatically start the master daemon if down
827 Version 2.0 release candidate 4
828 -------------------------------
830 *(Released Mon, 27 Apr 2009)*
832 - change the OS list to not require locks; this helps with big clusters
833 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
834 volume group is broken
835 - ``gnt-instance info``, without any arguments, doesn't run for all
836 instances anymore; either pass ``--all`` or pass the desired
837 instances; this helps against mistakes on big clusters where listing
838 the information for all instances takes a long time
839 - miscellaneous doc and man pages fixes
842 Version 2.0 release candidate 3
843 -------------------------------
845 *(Released Wed, 8 Apr 2009)*
847 - Change the internal locking model of some ``gnt-node`` commands, in
848 order to reduce contention (and blocking of master daemon) when
849 batching many creation/reinstall jobs
850 - Fixes to Xen soft reboot
851 - No longer build documentation at build time, instead distribute it in
852 the archive, in order to reduce the need for the whole docbook/rst
856 Version 2.0 release candidate 2
857 -------------------------------
859 *(Released Fri, 27 Mar 2009)*
861 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
862 - Fix watcher startup sequence, improves the behaviour of busy clusters
863 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
864 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
865 bind address and other places
866 - Some documentation fixes and updates
869 Version 2.0 release candidate 1
870 -------------------------------
872 *(Released Mon, 2 Mar 2009)*
874 - More documentation updates, now all docs should be more-or-less
876 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
878 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
880 - Fix an issue related to $libdir/run/ganeti and cluster creation
886 *(Released Thu, 19 Feb 2009)*
888 - Xen PVM and KVM have switched the default value for the instance root
889 disk to the first partition on the first drive, instead of the whole
890 drive; this means that the OS installation scripts must be changed
892 - Man pages have been updated
893 - RAPI has been switched by default to HTTPS, and the exported functions
894 should all work correctly
895 - RAPI v1 has been removed
896 - Many improvements to the KVM hypervisor
897 - Block device errors are now better reported
898 - Many other bugfixes and small improvements
904 *(Released Mon, 26 Jan 2009)*
906 - Version 2 is a general rewrite of the code and therefore the
907 differences are too many to list, see the design document for 2.0 in
908 the ``doc/`` subdirectory for more details
909 - In this beta version there is not yet a migration path from 1.2 (there
910 will be one in the final 2.0 release)
911 - A few significant changes are:
913 - all commands are executed by a daemon (``ganeti-masterd``) and the
914 various ``gnt-*`` commands are just front-ends to it
915 - all the commands are entered into, and executed from a job queue,
916 see the ``gnt-job(8)`` manpage
917 - the RAPI daemon supports read-write operations, secured by basic
918 HTTP authentication on top of HTTPS
919 - DRBD version 0.7 support has been removed, DRBD 8 is the only
920 supported version (when migrating from Ganeti 1.2 to 2.0, you need
921 to migrate to DRBD 8 first while still running Ganeti 1.2)
922 - DRBD devices are using statically allocated minor numbers, which
923 will be assigned to existing instances during the migration process
924 - there is support for both Xen PVM and Xen HVM instances running on
926 - KVM virtualization is supported too
927 - file-based storage has been implemented, which means that it is
928 possible to run the cluster without LVM and DRBD storage, for
929 example using a shared filesystem exported from shared storage (and
930 still have live migration)
936 *(Released Tue, 13 Jan 2009)*
938 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
939 - Reuse the old instance mac address by default on instance import, if
940 the instance name is the same.
941 - Handle situations in which the node info rpc returns incomplete
943 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
944 - Improved version of batcher:
947 - instance mac address support
948 - support for HVM clusters/instances
950 - Add an option to show the number of cpu sockets and nodes in
952 - Support OSes that handle more than one version of the OS api (but do
953 not change the current API in any other way)
954 - Fix ``gnt-node migrate``
955 - ``gnt-debug`` man page
956 - Fixes various more typos and small issues
957 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
963 *(Released Wed, 24 Sep 2008)*
965 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
966 type of disk exported to fully virtualized instances.
967 - provide access to the serial console of HVM instances
968 - instance auto_balance flag, set by default. If turned off it will
969 avoid warnings on cluster verify if there is not enough memory to fail
970 over an instance. in the future it will prevent automatically failing
971 it over when we will support that.
972 - batcher tool for instance creation, see ``tools/README.batcher``
973 - ``gnt-instance reinstall --select-os`` to interactively select a new
974 operating system when reinstalling an instance.
975 - when changing the memory amount on instance modify a check has been
976 added that the instance will be able to start. also warnings are
977 emitted if the instance will not be able to fail over, if auto_balance
979 - documentation fixes
980 - sync fields between ``gnt-instance list/modify/add/import``
981 - fix a race condition in drbd when the sync speed was set after giving
982 the device a remote peer.
988 *(Released Tue, 22 Jul 2008)*
990 - note: the allowed size and number of tags per object were reduced
991 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
992 - fixed twisted 8.x compatibility
993 - fixed ``gnt-instance replace-disks`` with iallocator
994 - add TCP keepalives on twisted connections to detect restarted nodes
995 - disk increase support, see ``gnt-instance grow-disk``
996 - implement bulk node/instance query for RAPI
997 - add tags in node/instance listing (optional)
998 - experimental migration (and live migration) support, read the man page
999 for ``gnt-instance migrate``
1000 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1001 has some small improvements in handling its state file
1007 *(Released Fri, 13 Jun 2008)*
1009 - Experimental readonly, REST-based remote API implementation;
1010 automatically started on master node, TCP port 5080, if enabled by
1011 ``--enable-rapi`` parameter to configure script.
1012 - Instance allocator support. Add and import instance accept a
1013 ``--iallocator`` parameter, and call that instance allocator to decide
1014 which node to use for the instance. The iallocator document describes
1015 what's expected from an allocator script.
1016 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1017 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1018 node is lost there is still enough memory to fail over the instances
1020 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1021 ``gnt-cluster verify``, to check for site-specific compliance. All the
1022 hooks will run, and their output, if any, will be displayed. Any
1023 failing hook will make the verification return an error value.
1024 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1025 primary and secondary interfaces
1026 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1027 that is still declared as part of the cluster and has failed.
1028 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1029 specifying output fields, that just adds the chosen fields to the
1031 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1032 export from the filesystem.
1033 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1034 have been added. Using them you can enable/disable acpi and pae
1035 support, and specify a path for a cd image to be exported to the
1036 instance. These parameters as the name suggest only work on HVM
1038 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1039 PAE support will be set to the previously hardcoded values, but the
1040 (previously hardcoded) path to the CDROM ISO image will be unset and
1041 if required, needs to be set manually with ``gnt-instance modify``
1043 - The address to which an instance's VNC console is bound is now
1044 selectable per-instance, rather than being cluster wide. Of course
1045 this only applies to instances controlled via VNC, so currently just
1046 applies to HVM clusters.
1052 *(Released Mon, 18 Feb 2008)*
1054 - more tweaks to the disk activation code (especially helpful for DRBD)
1055 - change the default ``gnt-instance list`` output format, now there is
1056 one combined status field (see the manpage for the exact values this
1058 - some more fixes for the mac export to hooks change
1059 - make Ganeti not break with DRBD 8.2.x (which changed the version
1060 format in ``/proc/drbd``) (issue 24)
1061 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1062 template, allowing migration from DRBD0.7+MD to DRBD8
1068 *(Released Wed, 30 Jan 2008)*
1070 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1072 - add command aliases infrastructure and a few aliases
1073 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1074 man pages and the ``--help`` option of ``gnt-node
1075 list``/``gnt-instance list``
1076 - fix ``gnt-backup list`` with down nodes (issue 21)
1077 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1078 - fix the dist archive and add a check for including svn/git files in
1080 - some developer-related changes: improve the burnin and the QA suite,
1081 add an upload script for testing during development
1087 *(Released Wed, 16 Jan 2008)*
1089 - experimental HVM support, read the install document, section
1090 "Initializing the cluster"
1091 - allow for the PVM hypervisor per-instance kernel and initrd paths
1092 - add a new command ``gnt-cluster verify-disks`` which uses a new
1093 algorithm to improve the reconnection of the DRBD pairs if the device
1094 on the secondary node has gone away
1095 - make logical volume code auto-activate LVs at disk activation time
1096 - slightly improve the speed of activating disks
1097 - allow specification of the MAC address at instance creation time, and
1098 changing it later via ``gnt-instance modify``
1099 - fix handling of external commands that generate lots of output on
1101 - update documentation with regard to minimum version of DRBD8 supported
1107 *(Released Tue, 4 Dec 2007)*
1109 - Log the ``xm create`` output to the node daemon log on failure (to
1110 help diagnosing the error)
1111 - In debug mode, log all external commands output if failed to the logs
1112 - Change parsing of lvm commands to ignore stderr
1118 *(Released Wed, 28 Nov 2007)*
1120 - Another round of updates to the DRBD 8 code to deal with more failures
1121 in the replace secondary node operation
1122 - Some more logging of failures in disk operations (lvm, drbd)
1123 - A few documentation updates
1130 *(Released Tue, 13 Nov 2007)*
1132 - Change configuration file format from Python's Pickle to JSON.
1133 Upgrading is possible using the cfgupgrade utility.
1134 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1135 faster replace disks and is more stable (DRBD 8 has many improvements
1136 compared to DRBD 0.7)
1137 - Added command line tags support (see man pages for ``gnt-instance``,
1138 ``gnt-node``, ``gnt-cluster``)
1139 - Added instance rename support
1140 - Added multi-instance startup/shutdown
1141 - Added cluster rename support
1142 - Added ``gnt-node evacuate`` to simplify some node operations
1143 - Added instance reboot operation that can speedup reboot as compared to
1145 - Soften the requirement that hostnames are in FQDN format
1146 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1148 - Removed dependency on debian's patched fping that uses the
1149 non-standard ``-S`` option
1150 - Now the OS definitions are searched for in multiple, configurable
1151 paths (easier for distros to package)
1152 - Some changes to the hooks infrastructure (especially the new
1153 post-configuration update hook)
1154 - Other small bugfixes
1156 .. vim: set textwidth=72 :