9 - Adding instance disk wipe prior allocation. This is a cluster wide
10 option and can be set/modified using --prealloc-wipe-disks.
15 *(Released Tue, 19 Oct 2010)*
17 - Disable SSL session ID cache in RPC client
23 *(Released Thu, 14 Oct 2010)*
25 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
26 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
28 - Fix problems with interaction between Curl and Python's HTTP server,
29 resulting in increased speed in many RPC calls
30 - Improve our release script to prevent breakage with older aclocal and
37 *(Released Thu, 7 Oct 2010)*
39 - Fixed issue 125, replace hardcoded “xenvg” in ``gnt-cluster`` with
40 value retrieved from master
41 - Added support for blacklisted or hidden OS definitions
42 - Added simple lock monitor (accessible via (``gnt-debug locks``)
43 - Added support for -mem-path in KVM hypervisor abstraction layer
44 - Allow overriding instance parameters in tool for inter-cluster
45 instance moves (``tools/move-instance``)
46 - Improved opcode summaries (e.g. in ``gnt-job list``)
47 - Improve consistency of OS listing by sorting it
48 - Documentation updates
54 *(Released Fri, 8 Oct 2010)*
56 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
62 *(Released Mon, 4 Oct 2010)*
64 - Fixed regression in ``gnt-instance rename``
70 *(Released Wed, 22 Sep 2010)*
72 - Fixed OS_VARIANT variable for OS scripts
73 - Fixed cluster tag operations via RAPI
74 - Made ``setup-ssh`` exit with non-zero code if an error occurred
75 - Disabled RAPI CA checks in watcher
81 *(Released Mon, 23 Aug 2010)*
83 - Support DRBD versions of the format "a.b.c.d"
85 - Re-introduce support for usage from multiple threads in RAPI client
86 - Instance renames and modify via RAPI
87 - Work around race condition between processing and archival in job
89 - Mark opcodes following failed one as failed, too
90 - Job field ``lock_status`` was removed due to difficulties making it
91 work with the changed job queue in Ganeti 2.2; a better way to monitor
92 locks is expected for a later 2.2.x release
93 - Fixed dry-run behaviour with many commands
94 - Support ``ssh-agent`` again when adding nodes
95 - Many additional bugfixes
101 *(Released Fri, 30 Jul 2010)*
103 Important change: the internal RPC mechanism between Ganeti nodes has
104 changed from using a home-grown http library (based on the Python base
105 libraries) to use the PycURL library. This requires that PycURL is
106 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
107 against GnuTLS by default. cURL's support for GnuTLS had known issues
108 before cURL 7.21.0 and we recommend using the latest cURL release or
109 linking against OpenSSL. Most other distributions already link PycURL
110 and cURL against OpenSSL. The command::
112 python -c 'import pycurl; print pycurl.version'
114 can be used to determine the libraries PycURL and cURL are linked
117 Other significant changes:
119 - Rewrote much of the internals of the job queue, in order to achieve
120 better parallelism; this decouples job query operations from the job
121 processing, and it should allow much nicer behaviour of the master
122 daemon under load, and it also has uncovered some long-standing bugs
123 related to the job serialisation (now fixed)
124 - Added a default iallocator setting to the cluster parameters,
125 eliminating the need to always pass nodes or an iallocator for
126 operations that require selection of new node(s)
127 - Added experimental support for the LXC virtualization method
128 - Added support for OS parameters, which allows the installation of
129 instances to pass parameter to OS scripts in order to customise the
131 - Added a hypervisor parameter controlling the migration type (live or
132 non-live), since hypervisors have various levels of reliability; this
133 has renamed the 'live' parameter to 'mode'
134 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
135 logical volumes, meaning that cluster verify will ignore them and not
136 flag their presence as errors
137 - The watcher will now reset the error count for failed instances after
138 8 hours, thus allowing self-healing if the problem that caused the
139 instances to be down/fail to start has cleared in the meantime
140 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
141 check for, and warn, if the drbd module parameter ``usermode_helper``
142 is not consistent with the cluster-wide setting; this is needed to
143 make diagnose easier of failed drbd creations
144 - Started adding base IPv6 support, but this is not yet
145 enabled/available for use
146 - Rename operations (cluster, instance) will now return the new name,
147 which is especially useful if a short name was passed in
148 - Added support for instance migration in RAPI
149 - Added a tool to pre-configure nodes for the SSH setup, before joining
150 them to the cluster; this will allow in the future a simplified model
151 for node joining (but not yet fully enabled in 2.2); this needs the
152 paramiko python library
153 - Fixed handling of name-resolving errors
154 - Fixed consistency of job results on the error path
155 - Fixed master-failover race condition when executed multiple times in
157 - Fixed many bugs related to the job queue (mostly introduced during the
158 2.2 development cycle, so not all are impacting 2.1)
159 - Fixed instance migration with missing disk symlinks
160 - Fixed handling of unknown jobs in ``gnt-job archive``
161 - And many other small fixes/improvements
165 - Enhanced both the unittest and the QA coverage
166 - Switched the opcode validation to a generic model, and extended the
167 validation to all opcode parameters
168 - Changed more parts of the code that write shell scripts to use the
170 - Switched the master daemon to use the asyncore library for the Luxi
177 *(Released Thu, 17 Jun 2010)*
179 - Added tool (``move-instance``) and infrastructure to move instances
180 between separate clusters (see :doc:`separate documentation
181 <move-instance>` and :doc:`design document <design-2.2>`)
182 - Added per-request RPC timeout
183 - RAPI now requires a Content-Type header for requests with a body (e.g.
184 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
185 :rfc:`2616` (HTTP/1.1), section 7.2.1)
186 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
188 - Implemented initial support for running Ganeti daemons as separate
189 users, see configure-time flags ``--with-user-prefix`` and
190 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
192 - Instances can be removed after export (``gnt-backup export
194 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
195 key (instead of 1024 bit)
196 - Added new cluster configuration file for cluster domain secret
197 - Import/export now use SSL instead of SSH
198 - Added support for showing estimated time when exporting an instance,
199 see the ``ganeti-os-interface(7)`` manpage and look for
205 *(Released Tue, 24 Aug 2010)*
208 - Don't ignore secondary node silently on non-mirrored disk templates
210 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
211 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
212 - Properly document the usb_mouse parameter
213 - Fix path in ganeti-rapi(8) (issue 116)
214 - Adjust error message when the ganeti user's .ssh directory is
216 - Add same-node-check when changing the disk template to drbd
222 *(Released Fri, 16 Jul 2010)*
225 - Add an option to only select some reboot types during qa/burnin.
226 (on some hypervisors consequent reboots are not supported)
227 - Fix infrequent race condition in master failover. Sometimes the old
228 master ip address would be still detected as up for a short time
229 after it was removed, causing failover to fail.
230 - Decrease mlockall warnings when the ctypes module is missing. On
231 Python 2.4 we support running even if no ctypes module is installed,
232 but we were too verbose about this issue.
233 - Fix building on old distributions, on which man doesn't have a
235 - Fix RAPI not to ignore the MAC address on instance creation
236 - Implement the old instance creation format in the RAPI client.
242 *(Released Thu, 01 Jul 2010)*
244 A small bugfix release:
245 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
246 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
247 a non-existing option
248 - Fix instance create: the --force-variant option was ignored
249 - Improve pylint 0.21 compatibility and warnings with Python 2.6
250 - Fix modify node storage with non-FQDN arguments
251 - Fix RAPI client to authenticate under Python 2.6 when used
252 for more than 5 requests needing authentication
253 - Fix gnt-instance modify -t (storage) giving a wrong error message
254 when converting a non-shutdown drbd instance to plain
260 *(Released Fri, 18 Jun 2010)*
262 A small bugfix release:
264 - Fix live migration of KVM instances started with older Ganeti
265 versions which had fewer hypervisor parameters
266 - Fix gnt-instance grow-disk on down instances
267 - Fix an error-reporting bug during instance migration
268 - Better checking of the ``--net`` and ``--disk`` values, to avoid
269 silently ignoring broken ones
270 - Fix an RPC error reporting bug affecting, for example, RAPI client
272 - Fix bug triggered by different API version os-es on different nodes
273 - Fix a bug in instance startup with custom hvparams: OS level
274 parameters would fail to be applied.
275 - Fix the RAPI client under Python 2.6 (but more work is needed to
276 make it work completely well with OpenSSL)
277 - Fix handling of errors when resolving names from DNS
283 *(Released Thu, 3 Jun 2010)*
285 A medium sized development cycle. Some new features, and some
286 fixes/small improvements/cleanups.
291 The node deamon now tries to mlock itself into memory, unless the
292 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
293 its logs, and falls back to console logging. This allows emergency
294 features such as ``gnt-node powercycle`` to work even in the event of a
295 broken node disk (tested offlining the disk hosting the node's
296 filesystem and dropping its memory caches; don't try this at home)
298 KVM: add vhost-net acceleration support. It can be tested with a new
299 enough version of the kernel and of qemu-kvm.
301 KVM: Add instance chrooting feature. If you use privilege dropping for
302 your VMs you can also now force them to chroot to an empty directory,
303 before starting the emulated guest.
305 KVM: Add maximum migration bandwith and maximum downtime tweaking
306 support (requires a new-enough version of qemu-kvm).
308 Cluster verify will now warn if the master node doesn't have the master
311 Add a new (incompatible) instance creation request format to RAPI which
312 supports all parameters (previously only a subset was supported, and it
313 wasn't possible to extend the old format to accomodate all the new
314 features. The old format is still supported, and a client can check for
315 this feature, before using it, by checking for its presence in the
316 ``features`` RAPI resource.
318 Now with ancient latin support. Try it passing the ``--roman`` option to
319 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
320 (requires the python-roman module to be installed, in order to work).
325 As usual many internal code refactorings, documentation updates, and
328 - Lots of improvements and cleanups to the experimental Remote API
329 (RAPI) client library.
330 - A new unit test suite for the core daemon libraries.
331 - A fix to creating missing directories makes sure the umask is not
332 applied anymore. This enforces the same directory permissions
334 - Better handling terminating daemons with ctrl+c (used when running
335 them in debugging mode).
336 - Fix a race condition in live migrating a KVM instance, when stat()
337 on the old proc status file returned EINVAL, which is an unexpected
339 - Fixed manpage checking with newer man and utf-8 charachters. But now
340 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
346 *(Released Fri, 7 May 2010)*
348 Fix a bug which prevented untagged KVM instances from starting.
354 *(Released Fri, 7 May 2010)*
356 Another release with a long development cycle, during which many
357 different features were added.
362 The KVM hypervisor now can run the individual instances as non-root, to
363 reduce the impact of a VM being hijacked due to bugs in the
364 hypervisor. It is possible to run all instances as a single (non-root)
365 user, to manually specify a user for each instance, or to dynamically
366 allocate a user out of a cluster-wide pool to each instance, with the
367 guarantee that no two instances will run under the same user ID on any
370 An experimental RAPI client library, that can be used standalone
371 (without the other Ganeti libraries), is provided in the source tree as
372 ``lib/rapi/client.py``. Note this client might change its interface in
373 the future, as we iterate on its capabilities.
375 A new command, ``gnt-cluster renew-crypto`` has been added to easily
376 replace the cluster's certificates and crypto keys. This might help in
377 case they have been compromised, or have simply expired.
379 A new disk option for instance creation has been added that allows one
380 to "adopt" currently existing logical volumes, with data
381 preservation. This should allow easier migration to Ganeti from
382 unmanaged (or managed via other software) instances.
384 Another disk improvement is the possibility to convert between redundant
385 (DRBD) and plain (LVM) disk configuration for an instance. This should
386 allow better scalability (starting with one node and growing the
387 cluster, or shrinking a two-node cluster to one node).
389 A new feature that could help with automated node failovers has been
390 implemented: if a node sees itself as offline (by querying the master
391 candidates), it will try to shutdown (hard) all instances and any active
392 DRBD devices. This reduces the risk of duplicate instances if an
393 external script automatically failovers the instances on such nodes. To
394 enable this, the cluster parameter ``maintain_node_health`` should be
395 enabled; in the future this option (per the name) will enable other
396 automatic maintenance features.
398 Instance export/import now will reuse the original instance
399 specifications for all parameters; that means exporting an instance,
400 deleting it and the importing it back should give an almost identical
401 instance. Note that the default import behaviour has changed from
402 before, where it created only one NIC; now it recreates the original
405 Cluster verify has added a few new checks: SSL certificates validity,
406 /etc/hosts consistency across the cluster, etc.
411 As usual, many internal changes were done, documentation fixes,
414 - Fixed cluster initialization with disabled cluster storage (regression
416 - File-based storage supports growing the disks
417 - Fixed behaviour of node role changes
418 - Fixed cluster verify for some corner cases, plus a general rewrite of
419 cluster verify to allow future extension with more checks
420 - Fixed log spamming by watcher and node daemon (regression introduced
422 - Fixed possible validation issues when changing the list of enabled
424 - Fixed cleanup of /etc/hosts during node removal
425 - Fixed RAPI response for invalid methods
426 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
427 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
428 from ide disks, etc.)
429 - Allow OS changes without re-installation (to record a changed OS
430 outside of Ganeti, or to allow OS renames)
431 - Allow instance creation without OS installation (useful for example if
432 the OS will be installed manually, or restored from a backup not in
434 - Implemented option to make cluster ``copyfile`` use the replication
436 - Added list of enabled hypervisors to ssconf (possibly useful for
438 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
440 - A partial form of node re-IP is possible via node readd, which now
441 allows changed node primary IP
442 - Command line utilities now show an informational message if the job is
444 - The logs of the master daemon now show the PID/UID/GID of the
451 *(Released Fri, 12 Mar 2010)*
453 During the 2.1.0 long release candidate cycle, a lot of improvements and
454 changes have accumulated with were released later as 2.1.1.
459 The node evacuate command (``gnt-node evacuate``) was significantly
460 rewritten, and as such the IAllocator protocol was changed - a new
461 request type has been added. This unfortunate change during a stable
462 series is designed to improve performance of node evacuations; on
463 clusters with more than about five nodes and which are well-balanced,
464 evacuation should proceed in parallel for all instances of the node
465 being evacuated. As such, any existing IAllocator scripts need to be
466 updated, otherwise the above command will fail due to the unknown
467 request. The provided "dumb" allocator has not been updated; but the
468 ganeti-htools package supports the new protocol since version 0.2.4.
470 Another important change is increased validation of node and instance
471 names. This might create problems in special cases, if invalid host
472 names are being used.
474 Also, a new layer of hypervisor parameters has been added, that sits at
475 OS level between the cluster defaults and the instance ones. This allows
476 customisation of virtualization parameters depending on the installed
477 OS. For example instances with OS 'X' may have a different KVM kernel
478 (or any other parameter) than the cluster defaults. This is intended to
479 help managing a multiple OSes on the same cluster, without manual
480 modification of each instance's parameters.
482 A tool for merging clusters, ``cluster-merge``, has been added in the
488 - Improved the int/float conversions that should make the code more
489 robust in face of errors from the node daemons
490 - Fixed the remove node code in case of internal configuration errors
491 - Fixed the node daemon behaviour in face of inconsistent queue
492 directory (e.g. read-only file-system where we can't open the files
494 - Fixed the behaviour of gnt-node modify for master candidate demotion;
495 now it either aborts cleanly or, if given the new “auto_promote”
496 parameter, will automatically promote other nodes as needed
497 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
498 completely prevent Ganeti from working
499 - Fixed bug for instance export when not all disks were successfully
501 - Fixed behaviour of node add when the new node is slow in starting up
503 - Fixed handling of signals in the LUXI client, which should improve
504 behaviour of command-line scripts
505 - Added checks for invalid node/instance names in the configuration (now
506 flagged during cluster verify)
507 - Fixed watcher behaviour for disk activation errors
508 - Fixed two potentially endless loops in http library, which led to the
509 RAPI daemon hanging and consuming 100% CPU in some cases
510 - Fixed bug in RAPI daemon related to hashed passwords
511 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
513 - Enhanced compatibility with non-Debian OSes, but not using absolute
514 path in some commands and allowing customisation of the ssh
515 configuration directory
516 - Fixed possible future issue with new Python versions by abiding to the
517 proper use of ``__slots__`` attribute on classes
518 - Added checks that should prevent directory traversal attacks
519 - Many documentation fixes based on feedback from users
524 - Added an “early_release” more for instance replace disks and node
525 evacuate, where we release locks earlier and thus allow higher
526 parallelism within the cluster
527 - Added watcher hooks, intended to allow the watcher to restart other
528 daemons (e.g. from the ganeti-nbma project), but they can be used of
529 course for any other purpose
530 - Added a compile-time disable for DRBD barriers, to increase
531 performance if the administrator trusts the power supply or the
532 storage system to not lose writes
533 - Added the option of using syslog for logging instead of, or in
534 addition to, Ganeti's own log files
535 - Removed boot restriction for paravirtual NICs for KVM, recent versions
536 can indeed boot from a paravirtual NIC
537 - Added a generic debug level for many operations; while this is not
538 used widely yet, it allows one to pass the debug value all the way to
540 - Enhanced the hooks environment for instance moves (failovers,
541 migrations) where the primary/secondary nodes changed during the
542 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
543 - Enhanced data validations for many user-supplied values; one important
544 item is the restrictions imposed on instance and node names, which
545 might reject some (invalid) host names
546 - Add a configure-time option to disable file-based storage, if it's not
547 needed; this allows greater security separation between the master
548 node and the other nodes from the point of view of the inter-node RPC
550 - Added user notification in interactive tools if job is waiting in the
551 job queue or trying to acquire locks
552 - Added log messages when a job is waiting for locks
553 - Added filtering by node tags in instance operations which admit
554 multiple instances (start, stop, reboot, reinstall)
555 - Added a new tool for cluster mergers, ``cluster-merge``
556 - Parameters from command line which are of the form ``a=b,c=d`` can now
557 use backslash escapes to pass in values which contain commas,
558 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
560 - For KVM, the instance name is the first parameter passed to KVM, so
561 that it's more visible in the process list
567 *(Released Tue, 2 Mar 2010)*
569 Ganeti 2.1 brings many improvements with it. Major changes:
571 - Added infrastructure to ease automated disk repairs
572 - Added new daemon to export configuration data in a cheaper way than
574 - Instance NICs can now be routed instead of being associated with a
576 - Improved job locking logic to reduce impact of jobs acquiring multiple
577 locks waiting for other long-running jobs
579 In-depth implementation details can be found in the Ganeti 2.1 design
585 - Added chroot hypervisor
586 - Added more options to xen-hvm hypervisor (``kernel_path`` and
588 - Added more options to xen-pvm hypervisor (``use_bootloader``,
589 ``bootloader_path`` and ``bootloader_args``)
590 - Added the ``use_localtime`` option for the xen-hvm and kvm
591 hypervisors, and the default value for this has changed to false (in
592 2.0 xen-hvm always enabled it)
593 - Added luxi call to submit multiple jobs in one go
594 - Added cluster initialization option to not modify ``/etc/hosts``
596 - Added network interface parameters
597 - Added dry run mode to some LUs
598 - Added RAPI resources:
600 - ``/2/instances/[instance_name]/info``
601 - ``/2/instances/[instance_name]/replace-disks``
602 - ``/2/nodes/[node_name]/evacuate``
603 - ``/2/nodes/[node_name]/migrate``
604 - ``/2/nodes/[node_name]/role``
605 - ``/2/nodes/[node_name]/storage``
606 - ``/2/nodes/[node_name]/storage/modify``
607 - ``/2/nodes/[node_name]/storage/repair``
609 - Added OpCodes to evacuate or migrate all instances on a node
610 - Added new command to list storage elements on nodes (``gnt-node
611 list-storage``) and modify them (``gnt-node modify-storage``)
612 - Added new ssconf files with master candidate IP address
613 (``ssconf_master_candidates_ips``), node primary IP address
614 (``ssconf_node_primary_ips``) and node secondary IP address
615 (``ssconf_node_secondary_ips``)
616 - Added ``ganeti-confd`` and a client library to query the Ganeti
617 configuration via UDP
618 - Added ability to run hooks after cluster initialization and before
620 - Added automatic mode for disk replace (``gnt-instance replace-disks
622 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
623 after catastrophic data-loss
624 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
626 - Added ``gnt-instance move`` command to move instances
627 - Added ``gnt-cluster watcher`` command to control watcher
628 - Added ``gnt-node powercycle`` command to powercycle nodes
629 - Added new job status field ``lock_status``
630 - Added parseable error codes to cluster verification (``gnt-cluster
631 verify --error-codes``) and made output less verbose (use
632 ``--verbose`` to restore previous behaviour)
633 - Added UUIDs to the main config entities (cluster, nodes, instances)
634 - Added support for OS variants
635 - Added support for hashed passwords in the Ganeti remote API users file
637 - Added option to specify maximum timeout on instance shutdown
638 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
639 - Added new helper script to start and stop Ganeti daemons
640 (``daemon-util``), with the intent to reduce the work necessary to
641 adjust Ganeti for non-Debian distributions and to start/stop daemons
643 - Added more unittests
644 - Fixed critical bug in ganeti-masterd startup
645 - Removed the configure-time ``kvm-migration-port`` parameter, this is
646 now customisable at the cluster level for both the KVM and Xen
647 hypervisors using the new ``migration_port`` parameter
648 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
649 reinstalling an instance
650 - Allowed ``@`` in tag names
651 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
652 - Many documentation updates
653 - Distribute hypervisor files on ``gnt-cluster redist-conf``
654 - ``gnt-instance reinstall`` can now reinstall multiple instances
655 - Updated many command line parameters
656 - Introduced new OS API version 15
657 - No longer support a default hypervisor
658 - Treat virtual LVs as inexistent
659 - Improved job locking logic to reduce lock contention
660 - Match instance and node names case insensitively
661 - Reimplemented bash completion script to be more complete
668 *(Released Thu, 4 Feb 2010)*
670 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
671 - Fix a string formatting bug
672 - Improve safety of the code in some error paths
673 - Improve data validation in the master of values returned from nodes
679 *(Released Thu, 17 Dec 2009)*
681 - Fix security issue due to missing validation of iallocator names; this
682 allows local and remote execution of arbitrary executables
683 - Fix failure of gnt-node list during instance removal
684 - Ship the RAPI documentation in the archive
690 *(Released Wed, 30 Sep 2009)*
692 - Fixed many wrong messages
693 - Fixed a few bugs related to the locking library
694 - Fixed MAC checking at instance creation time
695 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
696 - Fixed a few issues related to signal handling in both daemons and
698 - Fixed the example startup script provided
699 - Fixed insserv dependencies in the example startup script (patch from
701 - Fixed handling of drained nodes in the iallocator framework
702 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
704 - Fixed error related to invalid job IDs in job polling
705 - Fixed job/opcode persistence on unclean master shutdown
706 - Fixed handling of partial job processing after unclean master
708 - Fixed error reporting from LUs, previously all errors were converted
709 into execution errors
710 - Fixed error reporting from burnin
711 - Decreased significantly the memory usage of the job queue
712 - Optimised slightly multi-job submission
713 - Optimised slightly opcode loading
714 - Backported the multi-job submit framework from the development
715 branch; multi-instance start and stop should be faster
716 - Added script to clean archived jobs after 21 days; this will reduce
717 the size of the queue directory
718 - Added some extra checks in disk size tracking
719 - Added an example ethers hook script
720 - Added a cluster parameter that prevents Ganeti from modifying of
722 - Added more node information to RAPI responses
723 - Added a ``gnt-job watch`` command that allows following the ouput of a
725 - Added a bind-address option to ganeti-rapi
726 - Added more checks to the configuration verify
727 - Enhanced the burnin script such that some operations can be retried
729 - Converted instance reinstall to multi-instance model
735 *(Released Fri, 7 Aug 2009)*
737 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
738 to allow using the pre-2.0.2 behaviour in activation, if any existing
739 instances have mismatched disk sizes in the configuration
740 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
741 any configuration mismatches for disk sizes
742 - Added ``gnt-master cluste-failover --no-voting`` to allow master
743 failover to work on two-node clusters
744 - Fixed the ``--net`` option of ``gnt-backup import``, which was
746 - Fixed detection of OS script errors in ``gnt-backup export``
747 - Fixed exit code of ``gnt-backup export``
753 *(Released Fri, 17 Jul 2009)*
755 - Added experimental support for stripped logical volumes; this should
756 enhance performance but comes with a higher complexity in the block
757 device handling; stripping is only enabled when passing
758 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
759 affected even in the non-stripped mode
760 - Improved resiliency against transient failures at the end of DRBD
761 resyncs, and in general of DRBD resync checks
762 - Fixed a couple of issues with exports and snapshot errors
763 - Fixed a couple of issues in instance listing
764 - Added display of the disk size in ``gnt-instance info``
765 - Fixed checking for valid OSes in instance creation
766 - Fixed handling of the "vcpus" parameter in instance listing and in
767 general of invalid parameters
768 - Fixed http server library, and thus RAPI, to handle invalid
769 username/password combinations correctly; this means that now they
770 report unauthorized for queries too, not only for modifications,
771 allowing earlier detect of configuration problems
772 - Added a new "role" node list field, equivalent to the master/master
773 candidate/drained/offline flags combinations
774 - Fixed cluster modify and changes of candidate pool size
775 - Fixed cluster verify error messages for wrong files on regular nodes
776 - Fixed a couple of issues with node demotion from master candidate role
777 - Fixed node readd issues
778 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
779 - Added a new ``--no-voting`` option for masterfailover to fix failover
780 on two-nodes clusters when the former master node is unreachable
781 - Added instance reinstall over RAPI
787 *(Released Tue, 16 Jun 2009)*
789 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
790 allow re-adding the start in single-user option (regression from 1.2)
791 - the watcher writes the instance status to a file, to allow monitoring
792 to report the instance status (from the master) based on cached
793 results of the watcher's queries; while this can get stale if the
794 watcher is being locked due to other work on the cluster, this is
796 - the watcher now also restarts the node daemon and the rapi daemon if
798 - fixed the watcher to handle full and drained queue cases
799 - hooks export more instance data in the environment, which helps if
800 hook scripts need to take action based on the instance's properties
801 (no longer need to query back into ganeti)
802 - instance failovers when the instance is stopped do not check for free
803 RAM, so that failing over a stopped instance is possible in low memory
805 - rapi uses queries for tags instead of jobs (for less job traffic), and
806 for cluster tags it won't talk to masterd at all but read them from
808 - a couple of error handling fixes in RAPI
809 - drbd handling: improved the error handling of inconsistent disks after
810 resync to reduce the frequency of "there are some degraded disks for
811 this instance" messages
812 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
813 error handling path called a wrong function name)
819 *(Released Wed, 27 May 2009)*
821 - no changes from rc5
824 Version 2.0 release candidate 5
825 -------------------------------
827 *(Released Wed, 20 May 2009)*
829 - fix a couple of bugs (validation, argument checks)
830 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
831 - some small improvements to RAPI and IAllocator
832 - make watcher automatically start the master daemon if down
835 Version 2.0 release candidate 4
836 -------------------------------
838 *(Released Mon, 27 Apr 2009)*
840 - change the OS list to not require locks; this helps with big clusters
841 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
842 volume group is broken
843 - ``gnt-instance info``, without any arguments, doesn't run for all
844 instances anymore; either pass ``--all`` or pass the desired
845 instances; this helps against mistakes on big clusters where listing
846 the information for all instances takes a long time
847 - miscellaneous doc and man pages fixes
850 Version 2.0 release candidate 3
851 -------------------------------
853 *(Released Wed, 8 Apr 2009)*
855 - Change the internal locking model of some ``gnt-node`` commands, in
856 order to reduce contention (and blocking of master daemon) when
857 batching many creation/reinstall jobs
858 - Fixes to Xen soft reboot
859 - No longer build documentation at build time, instead distribute it in
860 the archive, in order to reduce the need for the whole docbook/rst
864 Version 2.0 release candidate 2
865 -------------------------------
867 *(Released Fri, 27 Mar 2009)*
869 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
870 - Fix watcher startup sequence, improves the behaviour of busy clusters
871 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
872 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
873 bind address and other places
874 - Some documentation fixes and updates
877 Version 2.0 release candidate 1
878 -------------------------------
880 *(Released Mon, 2 Mar 2009)*
882 - More documentation updates, now all docs should be more-or-less
884 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
886 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
888 - Fix an issue related to $libdir/run/ganeti and cluster creation
894 *(Released Thu, 19 Feb 2009)*
896 - Xen PVM and KVM have switched the default value for the instance root
897 disk to the first partition on the first drive, instead of the whole
898 drive; this means that the OS installation scripts must be changed
900 - Man pages have been updated
901 - RAPI has been switched by default to HTTPS, and the exported functions
902 should all work correctly
903 - RAPI v1 has been removed
904 - Many improvements to the KVM hypervisor
905 - Block device errors are now better reported
906 - Many other bugfixes and small improvements
912 *(Released Mon, 26 Jan 2009)*
914 - Version 2 is a general rewrite of the code and therefore the
915 differences are too many to list, see the design document for 2.0 in
916 the ``doc/`` subdirectory for more details
917 - In this beta version there is not yet a migration path from 1.2 (there
918 will be one in the final 2.0 release)
919 - A few significant changes are:
921 - all commands are executed by a daemon (``ganeti-masterd``) and the
922 various ``gnt-*`` commands are just front-ends to it
923 - all the commands are entered into, and executed from a job queue,
924 see the ``gnt-job(8)`` manpage
925 - the RAPI daemon supports read-write operations, secured by basic
926 HTTP authentication on top of HTTPS
927 - DRBD version 0.7 support has been removed, DRBD 8 is the only
928 supported version (when migrating from Ganeti 1.2 to 2.0, you need
929 to migrate to DRBD 8 first while still running Ganeti 1.2)
930 - DRBD devices are using statically allocated minor numbers, which
931 will be assigned to existing instances during the migration process
932 - there is support for both Xen PVM and Xen HVM instances running on
934 - KVM virtualization is supported too
935 - file-based storage has been implemented, which means that it is
936 possible to run the cluster without LVM and DRBD storage, for
937 example using a shared filesystem exported from shared storage (and
938 still have live migration)
944 *(Released Tue, 13 Jan 2009)*
946 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
947 - Reuse the old instance mac address by default on instance import, if
948 the instance name is the same.
949 - Handle situations in which the node info rpc returns incomplete
951 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
952 - Improved version of batcher:
955 - instance mac address support
956 - support for HVM clusters/instances
958 - Add an option to show the number of cpu sockets and nodes in
960 - Support OSes that handle more than one version of the OS api (but do
961 not change the current API in any other way)
962 - Fix ``gnt-node migrate``
963 - ``gnt-debug`` man page
964 - Fixes various more typos and small issues
965 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
971 *(Released Wed, 24 Sep 2008)*
973 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
974 type of disk exported to fully virtualized instances.
975 - provide access to the serial console of HVM instances
976 - instance auto_balance flag, set by default. If turned off it will
977 avoid warnings on cluster verify if there is not enough memory to fail
978 over an instance. in the future it will prevent automatically failing
979 it over when we will support that.
980 - batcher tool for instance creation, see ``tools/README.batcher``
981 - ``gnt-instance reinstall --select-os`` to interactively select a new
982 operating system when reinstalling an instance.
983 - when changing the memory amount on instance modify a check has been
984 added that the instance will be able to start. also warnings are
985 emitted if the instance will not be able to fail over, if auto_balance
987 - documentation fixes
988 - sync fields between ``gnt-instance list/modify/add/import``
989 - fix a race condition in drbd when the sync speed was set after giving
990 the device a remote peer.
996 *(Released Tue, 22 Jul 2008)*
998 - note: the allowed size and number of tags per object were reduced
999 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1000 - fixed twisted 8.x compatibility
1001 - fixed ``gnt-instance replace-disks`` with iallocator
1002 - add TCP keepalives on twisted connections to detect restarted nodes
1003 - disk increase support, see ``gnt-instance grow-disk``
1004 - implement bulk node/instance query for RAPI
1005 - add tags in node/instance listing (optional)
1006 - experimental migration (and live migration) support, read the man page
1007 for ``gnt-instance migrate``
1008 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1009 has some small improvements in handling its state file
1015 *(Released Fri, 13 Jun 2008)*
1017 - Experimental readonly, REST-based remote API implementation;
1018 automatically started on master node, TCP port 5080, if enabled by
1019 ``--enable-rapi`` parameter to configure script.
1020 - Instance allocator support. Add and import instance accept a
1021 ``--iallocator`` parameter, and call that instance allocator to decide
1022 which node to use for the instance. The iallocator document describes
1023 what's expected from an allocator script.
1024 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1025 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1026 node is lost there is still enough memory to fail over the instances
1028 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1029 ``gnt-cluster verify``, to check for site-specific compliance. All the
1030 hooks will run, and their output, if any, will be displayed. Any
1031 failing hook will make the verification return an error value.
1032 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1033 primary and secondary interfaces
1034 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1035 that is still declared as part of the cluster and has failed.
1036 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1037 specifying output fields, that just adds the chosen fields to the
1039 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1040 export from the filesystem.
1041 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1042 have been added. Using them you can enable/disable acpi and pae
1043 support, and specify a path for a cd image to be exported to the
1044 instance. These parameters as the name suggest only work on HVM
1046 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1047 PAE support will be set to the previously hardcoded values, but the
1048 (previously hardcoded) path to the CDROM ISO image will be unset and
1049 if required, needs to be set manually with ``gnt-instance modify``
1051 - The address to which an instance's VNC console is bound is now
1052 selectable per-instance, rather than being cluster wide. Of course
1053 this only applies to instances controlled via VNC, so currently just
1054 applies to HVM clusters.
1060 *(Released Mon, 18 Feb 2008)*
1062 - more tweaks to the disk activation code (especially helpful for DRBD)
1063 - change the default ``gnt-instance list`` output format, now there is
1064 one combined status field (see the manpage for the exact values this
1066 - some more fixes for the mac export to hooks change
1067 - make Ganeti not break with DRBD 8.2.x (which changed the version
1068 format in ``/proc/drbd``) (issue 24)
1069 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1070 template, allowing migration from DRBD0.7+MD to DRBD8
1076 *(Released Wed, 30 Jan 2008)*
1078 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1080 - add command aliases infrastructure and a few aliases
1081 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1082 man pages and the ``--help`` option of ``gnt-node
1083 list``/``gnt-instance list``
1084 - fix ``gnt-backup list`` with down nodes (issue 21)
1085 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1086 - fix the dist archive and add a check for including svn/git files in
1088 - some developer-related changes: improve the burnin and the QA suite,
1089 add an upload script for testing during development
1095 *(Released Wed, 16 Jan 2008)*
1097 - experimental HVM support, read the install document, section
1098 "Initializing the cluster"
1099 - allow for the PVM hypervisor per-instance kernel and initrd paths
1100 - add a new command ``gnt-cluster verify-disks`` which uses a new
1101 algorithm to improve the reconnection of the DRBD pairs if the device
1102 on the secondary node has gone away
1103 - make logical volume code auto-activate LVs at disk activation time
1104 - slightly improve the speed of activating disks
1105 - allow specification of the MAC address at instance creation time, and
1106 changing it later via ``gnt-instance modify``
1107 - fix handling of external commands that generate lots of output on
1109 - update documentation with regard to minimum version of DRBD8 supported
1115 *(Released Tue, 4 Dec 2007)*
1117 - Log the ``xm create`` output to the node daemon log on failure (to
1118 help diagnosing the error)
1119 - In debug mode, log all external commands output if failed to the logs
1120 - Change parsing of lvm commands to ignore stderr
1126 *(Released Wed, 28 Nov 2007)*
1128 - Another round of updates to the DRBD 8 code to deal with more failures
1129 in the replace secondary node operation
1130 - Some more logging of failures in disk operations (lvm, drbd)
1131 - A few documentation updates
1138 *(Released Tue, 13 Nov 2007)*
1140 - Change configuration file format from Python's Pickle to JSON.
1141 Upgrading is possible using the cfgupgrade utility.
1142 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1143 faster replace disks and is more stable (DRBD 8 has many improvements
1144 compared to DRBD 0.7)
1145 - Added command line tags support (see man pages for ``gnt-instance``,
1146 ``gnt-node``, ``gnt-cluster``)
1147 - Added instance rename support
1148 - Added multi-instance startup/shutdown
1149 - Added cluster rename support
1150 - Added ``gnt-node evacuate`` to simplify some node operations
1151 - Added instance reboot operation that can speedup reboot as compared to
1153 - Soften the requirement that hostnames are in FQDN format
1154 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1156 - Removed dependency on debian's patched fping that uses the
1157 non-standard ``-S`` option
1158 - Now the OS definitions are searched for in multiple, configurable
1159 paths (easier for distros to package)
1160 - Some changes to the hooks infrastructure (especially the new
1161 post-configuration update hook)
1162 - Other small bugfixes
1164 .. vim: set textwidth=72 :