7 *(Released Fri, 14 Jan 2011)*
12 - Fixed timezone issues when formatting timestamps
13 - Added support for node groups, available via ``gnt-group`` and other
15 - Added out-of-band framework and management, see :doc:`design
16 document <design-oob>`
17 - Removed support for roman numbers from ``gnt-node list`` and
18 ``gnt-instance list``.
19 - Allowed modification of master network interface via ``gnt-cluster
20 modify --master-netdev``
21 - Accept offline secondaries while shutting down instance disks
22 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
23 - Added support for multiple LVM volume groups
24 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
26 - Added commands to list available fields:
28 - ``gnt-node list-fields``
29 - ``gnt-group list-fields``
30 - ``gnt-instance list-fields``
32 - Updated documentation and man pages
37 - Moved ``rapi_users`` file into separate directory, now named
38 ``…/ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
40 - Added new tool for running commands on many machines,
41 ``tools/ganeti-listrunner``
42 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
43 improving the ``gnt-instance console`` output
44 - Allowed customisation of disk index separator at ``configure`` time
45 - Export node group allocation policy to :doc:`iallocator <iallocator>`
46 - Added support for non-partitioned md disks in ``lvmstrap``
47 - Added script to gracefully power off KVM instances
48 - Split ``utils`` module into smaller parts
49 - Changed query operations to return more detailed information, e.g.
50 whether an information is unavailable due to an offline node. To use
51 this new functionality, the LUXI call ``Query`` must be used. Field
52 information is now stored by the master daemon and can be retrieved
53 using ``QueryFields``. Instances, nodes and groups can also be queried
54 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
55 exposed via RAPI). The following commands make use of this
56 infrastructure change:
59 - ``gnt-group list-fields``
61 - ``gnt-node list-fields``
62 - ``gnt-instance list``
63 - ``gnt-instance list-fields``
69 - New RAPI resources (see :doc:`rapi`):
73 - ``/2/groups/[group_name]``
74 - ``/2/groups/[group_name]/assign-nodes``
75 - ``/2/groups/[group_name]/modify``
76 - ``/2/groups/[group_name]/rename``
77 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
81 - Implemented ``no_install`` for instance creation
82 - Implemented OS parameters for instance reinstallation, allowing
83 use of special settings on reinstallation (e.g. for preserving data)
88 - Added IPv6 support in import/export
89 - Pause DRBD synchronization while wiping disks on instance creation
90 - Updated unittests and QA scripts
91 - Improved network parameters passed to KVM
92 - Converted man pages from docbook to reStructuredText
98 *(Released Mon, 20 Dec 2010)*
100 Released version 2.3.1~rc1 without any changes.
106 *(Released Wed, 1 Dec 2010)*
108 - impexpd: Disable OpenSSL compression in socat if possible (backport
109 from master, commit e90739d625b, see :doc:`installation guide
110 <install-quick>` for details)
111 - Changed unittest coverage report to exclude test scripts
112 - Added script to check version format
118 *(Released Wed, 1 Dec 2010)*
120 Released version 2.3.0~rc1 without any changes.
126 *(Released Fri, 19 Nov 2010)*
128 A number of bugfixes and documentation updates:
130 - Update ganeti-os-interface documentation
131 - Fixed a bug related to duplicate MACs or similar items which should be
133 - Fix breakage in OS state modify
134 - Reinstall instance: disallow offline secondaries (fixes bug related to
135 OS changing but reinstall failing)
136 - plus all the other fixes between 2.2.1 and 2.2.2
142 *(Released Tue, 2 Nov 2010)*
144 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
145 - Fixed master failover race with watcher
146 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
148 - Accept previously stopped instance for export with instance removal
149 - Simplify and extend the environment variables for instance OS scripts
150 - Added new node flags, ``master_capable`` and ``vm_capable``
151 - Added optional instance disk wiping prior during allocation. This is a
152 cluster-wide option and can be set/modified using
153 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
154 - Added IPv6 support, see :doc:`design document <design-2.3>` and
156 - Added a new watcher option (``--ignore-pause``)
157 - Added option to ignore offline node on instance start/stop
158 (``--ignore-offline``)
159 - Allow overriding OS parameters with ``gnt-instance reinstall``
160 - Added ability to change node's secondary IP address using ``gnt-node
162 - Implemented privilege separation for all daemons except
163 ``ganeti-noded``, see ``configure`` options
164 - Complain if an instance's disk is marked faulty in ``gnt-cluster
166 - Implemented job priorities (see ``ganeti(7)`` manpage)
167 - Ignore failures while shutting down instances during failover from
169 - Exit daemon's bootstrap process only once daemon is ready
170 - Export more information via ``LUQueryInstances``/remote API
171 - Improved documentation, QA and unittests
172 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
173 restart if the file was created or changed
174 - Added LUXI protocol version sent with each request and response,
175 allowing detection of server/client mismatches
176 - Moved the Python scripts among gnt-* and ganeti-* into modules
177 - Moved all code related to setting up SSH to an external script,
179 - Infrastructure changes for node group support in future versions
185 *(Released Fri, 19 Nov 2010)*
187 A few small bugs fixed, and some improvements to the build system:
189 - Fix documentation regarding conversion to drbd
190 - Fix validation of parameters in cluster modify (``gnt-cluster modify
192 - Fix error handling in node modify with multiple changes
193 - Allow remote imports without checked names
199 *(Released Tue, 19 Oct 2010)*
201 - Disable SSL session ID cache in RPC client
207 *(Released Thu, 14 Oct 2010)*
209 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
210 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
212 - Fix problems with interaction between Curl and Python's HTTP server,
213 resulting in increased speed in many RPC calls
214 - Improve our release script to prevent breakage with older aclocal and
221 *(Released Thu, 7 Oct 2010)*
223 - Fixed issue 125, replace hardcoded “xenvg” in ``gnt-cluster`` with
224 value retrieved from master
225 - Added support for blacklisted or hidden OS definitions
226 - Added simple lock monitor (accessible via (``gnt-debug locks``)
227 - Added support for -mem-path in KVM hypervisor abstraction layer
228 - Allow overriding instance parameters in tool for inter-cluster
229 instance moves (``tools/move-instance``)
230 - Improved opcode summaries (e.g. in ``gnt-job list``)
231 - Improve consistency of OS listing by sorting it
232 - Documentation updates
238 *(Released Fri, 8 Oct 2010)*
240 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
246 *(Released Mon, 4 Oct 2010)*
248 - Fixed regression in ``gnt-instance rename``
254 *(Released Wed, 22 Sep 2010)*
256 - Fixed OS_VARIANT variable for OS scripts
257 - Fixed cluster tag operations via RAPI
258 - Made ``setup-ssh`` exit with non-zero code if an error occurred
259 - Disabled RAPI CA checks in watcher
265 *(Released Mon, 23 Aug 2010)*
267 - Support DRBD versions of the format "a.b.c.d"
269 - Re-introduce support for usage from multiple threads in RAPI client
270 - Instance renames and modify via RAPI
271 - Work around race condition between processing and archival in job
273 - Mark opcodes following failed one as failed, too
274 - Job field ``lock_status`` was removed due to difficulties making it
275 work with the changed job queue in Ganeti 2.2; a better way to monitor
276 locks is expected for a later 2.2.x release
277 - Fixed dry-run behaviour with many commands
278 - Support ``ssh-agent`` again when adding nodes
279 - Many additional bugfixes
285 *(Released Fri, 30 Jul 2010)*
287 Important change: the internal RPC mechanism between Ganeti nodes has
288 changed from using a home-grown http library (based on the Python base
289 libraries) to use the PycURL library. This requires that PycURL is
290 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
291 against GnuTLS by default. cURL's support for GnuTLS had known issues
292 before cURL 7.21.0 and we recommend using the latest cURL release or
293 linking against OpenSSL. Most other distributions already link PycURL
294 and cURL against OpenSSL. The command::
296 python -c 'import pycurl; print pycurl.version'
298 can be used to determine the libraries PycURL and cURL are linked
301 Other significant changes:
303 - Rewrote much of the internals of the job queue, in order to achieve
304 better parallelism; this decouples job query operations from the job
305 processing, and it should allow much nicer behaviour of the master
306 daemon under load, and it also has uncovered some long-standing bugs
307 related to the job serialisation (now fixed)
308 - Added a default iallocator setting to the cluster parameters,
309 eliminating the need to always pass nodes or an iallocator for
310 operations that require selection of new node(s)
311 - Added experimental support for the LXC virtualization method
312 - Added support for OS parameters, which allows the installation of
313 instances to pass parameter to OS scripts in order to customise the
315 - Added a hypervisor parameter controlling the migration type (live or
316 non-live), since hypervisors have various levels of reliability; this
317 has renamed the 'live' parameter to 'mode'
318 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
319 logical volumes, meaning that cluster verify will ignore them and not
320 flag their presence as errors
321 - The watcher will now reset the error count for failed instances after
322 8 hours, thus allowing self-healing if the problem that caused the
323 instances to be down/fail to start has cleared in the meantime
324 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
325 check for, and warn, if the drbd module parameter ``usermode_helper``
326 is not consistent with the cluster-wide setting; this is needed to
327 make diagnose easier of failed drbd creations
328 - Started adding base IPv6 support, but this is not yet
329 enabled/available for use
330 - Rename operations (cluster, instance) will now return the new name,
331 which is especially useful if a short name was passed in
332 - Added support for instance migration in RAPI
333 - Added a tool to pre-configure nodes for the SSH setup, before joining
334 them to the cluster; this will allow in the future a simplified model
335 for node joining (but not yet fully enabled in 2.2); this needs the
336 paramiko python library
337 - Fixed handling of name-resolving errors
338 - Fixed consistency of job results on the error path
339 - Fixed master-failover race condition when executed multiple times in
341 - Fixed many bugs related to the job queue (mostly introduced during the
342 2.2 development cycle, so not all are impacting 2.1)
343 - Fixed instance migration with missing disk symlinks
344 - Fixed handling of unknown jobs in ``gnt-job archive``
345 - And many other small fixes/improvements
349 - Enhanced both the unittest and the QA coverage
350 - Switched the opcode validation to a generic model, and extended the
351 validation to all opcode parameters
352 - Changed more parts of the code that write shell scripts to use the
354 - Switched the master daemon to use the asyncore library for the Luxi
361 *(Released Thu, 17 Jun 2010)*
363 - Added tool (``move-instance``) and infrastructure to move instances
364 between separate clusters (see :doc:`separate documentation
365 <move-instance>` and :doc:`design document <design-2.2>`)
366 - Added per-request RPC timeout
367 - RAPI now requires a Content-Type header for requests with a body (e.g.
368 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
369 :rfc:`2616` (HTTP/1.1), section 7.2.1)
370 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
372 - Implemented initial support for running Ganeti daemons as separate
373 users, see configure-time flags ``--with-user-prefix`` and
374 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
376 - Instances can be removed after export (``gnt-backup export
378 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
379 key (instead of 1024 bit)
380 - Added new cluster configuration file for cluster domain secret
381 - Import/export now use SSL instead of SSH
382 - Added support for showing estimated time when exporting an instance,
383 see the ``ganeti-os-interface(7)`` manpage and look for
390 *(Released Tue, 16 Nov 2010)*
392 Some more bugfixes. Unless critical bugs occur, this will be the last
395 - Fix case of MAC special-values
396 - Fix mac checker regex
397 - backend: Fix typo causing “out of range” error
398 - Add missing --units in gnt-instance list man page
404 *(Released Tue, 24 Aug 2010)*
407 - Don't ignore secondary node silently on non-mirrored disk templates
409 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
410 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
411 - Properly document the usb_mouse parameter
412 - Fix path in ganeti-rapi(8) (issue 116)
413 - Adjust error message when the ganeti user's .ssh directory is
415 - Add same-node-check when changing the disk template to drbd
421 *(Released Fri, 16 Jul 2010)*
424 - Add an option to only select some reboot types during qa/burnin.
425 (on some hypervisors consequent reboots are not supported)
426 - Fix infrequent race condition in master failover. Sometimes the old
427 master ip address would be still detected as up for a short time
428 after it was removed, causing failover to fail.
429 - Decrease mlockall warnings when the ctypes module is missing. On
430 Python 2.4 we support running even if no ctypes module is installed,
431 but we were too verbose about this issue.
432 - Fix building on old distributions, on which man doesn't have a
434 - Fix RAPI not to ignore the MAC address on instance creation
435 - Implement the old instance creation format in the RAPI client.
441 *(Released Thu, 01 Jul 2010)*
443 A small bugfix release:
444 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
445 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
446 a non-existing option
447 - Fix instance create: the --force-variant option was ignored
448 - Improve pylint 0.21 compatibility and warnings with Python 2.6
449 - Fix modify node storage with non-FQDN arguments
450 - Fix RAPI client to authenticate under Python 2.6 when used
451 for more than 5 requests needing authentication
452 - Fix gnt-instance modify -t (storage) giving a wrong error message
453 when converting a non-shutdown drbd instance to plain
459 *(Released Fri, 18 Jun 2010)*
461 A small bugfix release:
463 - Fix live migration of KVM instances started with older Ganeti
464 versions which had fewer hypervisor parameters
465 - Fix gnt-instance grow-disk on down instances
466 - Fix an error-reporting bug during instance migration
467 - Better checking of the ``--net`` and ``--disk`` values, to avoid
468 silently ignoring broken ones
469 - Fix an RPC error reporting bug affecting, for example, RAPI client
471 - Fix bug triggered by different API version os-es on different nodes
472 - Fix a bug in instance startup with custom hvparams: OS level
473 parameters would fail to be applied.
474 - Fix the RAPI client under Python 2.6 (but more work is needed to
475 make it work completely well with OpenSSL)
476 - Fix handling of errors when resolving names from DNS
482 *(Released Thu, 3 Jun 2010)*
484 A medium sized development cycle. Some new features, and some
485 fixes/small improvements/cleanups.
490 The node deamon now tries to mlock itself into memory, unless the
491 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
492 its logs, and falls back to console logging. This allows emergency
493 features such as ``gnt-node powercycle`` to work even in the event of a
494 broken node disk (tested offlining the disk hosting the node's
495 filesystem and dropping its memory caches; don't try this at home)
497 KVM: add vhost-net acceleration support. It can be tested with a new
498 enough version of the kernel and of qemu-kvm.
500 KVM: Add instance chrooting feature. If you use privilege dropping for
501 your VMs you can also now force them to chroot to an empty directory,
502 before starting the emulated guest.
504 KVM: Add maximum migration bandwith and maximum downtime tweaking
505 support (requires a new-enough version of qemu-kvm).
507 Cluster verify will now warn if the master node doesn't have the master
510 Add a new (incompatible) instance creation request format to RAPI which
511 supports all parameters (previously only a subset was supported, and it
512 wasn't possible to extend the old format to accomodate all the new
513 features. The old format is still supported, and a client can check for
514 this feature, before using it, by checking for its presence in the
515 ``features`` RAPI resource.
517 Now with ancient latin support. Try it passing the ``--roman`` option to
518 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
519 (requires the python-roman module to be installed, in order to work).
524 As usual many internal code refactorings, documentation updates, and
527 - Lots of improvements and cleanups to the experimental Remote API
528 (RAPI) client library.
529 - A new unit test suite for the core daemon libraries.
530 - A fix to creating missing directories makes sure the umask is not
531 applied anymore. This enforces the same directory permissions
533 - Better handling terminating daemons with ctrl+c (used when running
534 them in debugging mode).
535 - Fix a race condition in live migrating a KVM instance, when stat()
536 on the old proc status file returned EINVAL, which is an unexpected
538 - Fixed manpage checking with newer man and utf-8 charachters. But now
539 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
545 *(Released Fri, 7 May 2010)*
547 Fix a bug which prevented untagged KVM instances from starting.
553 *(Released Fri, 7 May 2010)*
555 Another release with a long development cycle, during which many
556 different features were added.
561 The KVM hypervisor now can run the individual instances as non-root, to
562 reduce the impact of a VM being hijacked due to bugs in the
563 hypervisor. It is possible to run all instances as a single (non-root)
564 user, to manually specify a user for each instance, or to dynamically
565 allocate a user out of a cluster-wide pool to each instance, with the
566 guarantee that no two instances will run under the same user ID on any
569 An experimental RAPI client library, that can be used standalone
570 (without the other Ganeti libraries), is provided in the source tree as
571 ``lib/rapi/client.py``. Note this client might change its interface in
572 the future, as we iterate on its capabilities.
574 A new command, ``gnt-cluster renew-crypto`` has been added to easily
575 replace the cluster's certificates and crypto keys. This might help in
576 case they have been compromised, or have simply expired.
578 A new disk option for instance creation has been added that allows one
579 to "adopt" currently existing logical volumes, with data
580 preservation. This should allow easier migration to Ganeti from
581 unmanaged (or managed via other software) instances.
583 Another disk improvement is the possibility to convert between redundant
584 (DRBD) and plain (LVM) disk configuration for an instance. This should
585 allow better scalability (starting with one node and growing the
586 cluster, or shrinking a two-node cluster to one node).
588 A new feature that could help with automated node failovers has been
589 implemented: if a node sees itself as offline (by querying the master
590 candidates), it will try to shutdown (hard) all instances and any active
591 DRBD devices. This reduces the risk of duplicate instances if an
592 external script automatically failovers the instances on such nodes. To
593 enable this, the cluster parameter ``maintain_node_health`` should be
594 enabled; in the future this option (per the name) will enable other
595 automatic maintenance features.
597 Instance export/import now will reuse the original instance
598 specifications for all parameters; that means exporting an instance,
599 deleting it and the importing it back should give an almost identical
600 instance. Note that the default import behaviour has changed from
601 before, where it created only one NIC; now it recreates the original
604 Cluster verify has added a few new checks: SSL certificates validity,
605 /etc/hosts consistency across the cluster, etc.
610 As usual, many internal changes were done, documentation fixes,
613 - Fixed cluster initialization with disabled cluster storage (regression
615 - File-based storage supports growing the disks
616 - Fixed behaviour of node role changes
617 - Fixed cluster verify for some corner cases, plus a general rewrite of
618 cluster verify to allow future extension with more checks
619 - Fixed log spamming by watcher and node daemon (regression introduced
621 - Fixed possible validation issues when changing the list of enabled
623 - Fixed cleanup of /etc/hosts during node removal
624 - Fixed RAPI response for invalid methods
625 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
626 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
627 from ide disks, etc.)
628 - Allow OS changes without re-installation (to record a changed OS
629 outside of Ganeti, or to allow OS renames)
630 - Allow instance creation without OS installation (useful for example if
631 the OS will be installed manually, or restored from a backup not in
633 - Implemented option to make cluster ``copyfile`` use the replication
635 - Added list of enabled hypervisors to ssconf (possibly useful for
637 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
639 - A partial form of node re-IP is possible via node readd, which now
640 allows changed node primary IP
641 - Command line utilities now show an informational message if the job is
643 - The logs of the master daemon now show the PID/UID/GID of the
650 *(Released Fri, 12 Mar 2010)*
652 During the 2.1.0 long release candidate cycle, a lot of improvements and
653 changes have accumulated with were released later as 2.1.1.
658 The node evacuate command (``gnt-node evacuate``) was significantly
659 rewritten, and as such the IAllocator protocol was changed - a new
660 request type has been added. This unfortunate change during a stable
661 series is designed to improve performance of node evacuations; on
662 clusters with more than about five nodes and which are well-balanced,
663 evacuation should proceed in parallel for all instances of the node
664 being evacuated. As such, any existing IAllocator scripts need to be
665 updated, otherwise the above command will fail due to the unknown
666 request. The provided "dumb" allocator has not been updated; but the
667 ganeti-htools package supports the new protocol since version 0.2.4.
669 Another important change is increased validation of node and instance
670 names. This might create problems in special cases, if invalid host
671 names are being used.
673 Also, a new layer of hypervisor parameters has been added, that sits at
674 OS level between the cluster defaults and the instance ones. This allows
675 customisation of virtualization parameters depending on the installed
676 OS. For example instances with OS 'X' may have a different KVM kernel
677 (or any other parameter) than the cluster defaults. This is intended to
678 help managing a multiple OSes on the same cluster, without manual
679 modification of each instance's parameters.
681 A tool for merging clusters, ``cluster-merge``, has been added in the
687 - Improved the int/float conversions that should make the code more
688 robust in face of errors from the node daemons
689 - Fixed the remove node code in case of internal configuration errors
690 - Fixed the node daemon behaviour in face of inconsistent queue
691 directory (e.g. read-only file-system where we can't open the files
693 - Fixed the behaviour of gnt-node modify for master candidate demotion;
694 now it either aborts cleanly or, if given the new “auto_promote”
695 parameter, will automatically promote other nodes as needed
696 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
697 completely prevent Ganeti from working
698 - Fixed bug for instance export when not all disks were successfully
700 - Fixed behaviour of node add when the new node is slow in starting up
702 - Fixed handling of signals in the LUXI client, which should improve
703 behaviour of command-line scripts
704 - Added checks for invalid node/instance names in the configuration (now
705 flagged during cluster verify)
706 - Fixed watcher behaviour for disk activation errors
707 - Fixed two potentially endless loops in http library, which led to the
708 RAPI daemon hanging and consuming 100% CPU in some cases
709 - Fixed bug in RAPI daemon related to hashed passwords
710 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
712 - Enhanced compatibility with non-Debian OSes, but not using absolute
713 path in some commands and allowing customisation of the ssh
714 configuration directory
715 - Fixed possible future issue with new Python versions by abiding to the
716 proper use of ``__slots__`` attribute on classes
717 - Added checks that should prevent directory traversal attacks
718 - Many documentation fixes based on feedback from users
723 - Added an “early_release” more for instance replace disks and node
724 evacuate, where we release locks earlier and thus allow higher
725 parallelism within the cluster
726 - Added watcher hooks, intended to allow the watcher to restart other
727 daemons (e.g. from the ganeti-nbma project), but they can be used of
728 course for any other purpose
729 - Added a compile-time disable for DRBD barriers, to increase
730 performance if the administrator trusts the power supply or the
731 storage system to not lose writes
732 - Added the option of using syslog for logging instead of, or in
733 addition to, Ganeti's own log files
734 - Removed boot restriction for paravirtual NICs for KVM, recent versions
735 can indeed boot from a paravirtual NIC
736 - Added a generic debug level for many operations; while this is not
737 used widely yet, it allows one to pass the debug value all the way to
739 - Enhanced the hooks environment for instance moves (failovers,
740 migrations) where the primary/secondary nodes changed during the
741 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
742 - Enhanced data validations for many user-supplied values; one important
743 item is the restrictions imposed on instance and node names, which
744 might reject some (invalid) host names
745 - Add a configure-time option to disable file-based storage, if it's not
746 needed; this allows greater security separation between the master
747 node and the other nodes from the point of view of the inter-node RPC
749 - Added user notification in interactive tools if job is waiting in the
750 job queue or trying to acquire locks
751 - Added log messages when a job is waiting for locks
752 - Added filtering by node tags in instance operations which admit
753 multiple instances (start, stop, reboot, reinstall)
754 - Added a new tool for cluster mergers, ``cluster-merge``
755 - Parameters from command line which are of the form ``a=b,c=d`` can now
756 use backslash escapes to pass in values which contain commas,
757 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
759 - For KVM, the instance name is the first parameter passed to KVM, so
760 that it's more visible in the process list
766 *(Released Tue, 2 Mar 2010)*
768 Ganeti 2.1 brings many improvements with it. Major changes:
770 - Added infrastructure to ease automated disk repairs
771 - Added new daemon to export configuration data in a cheaper way than
773 - Instance NICs can now be routed instead of being associated with a
775 - Improved job locking logic to reduce impact of jobs acquiring multiple
776 locks waiting for other long-running jobs
778 In-depth implementation details can be found in the Ganeti 2.1 design
784 - Added chroot hypervisor
785 - Added more options to xen-hvm hypervisor (``kernel_path`` and
787 - Added more options to xen-pvm hypervisor (``use_bootloader``,
788 ``bootloader_path`` and ``bootloader_args``)
789 - Added the ``use_localtime`` option for the xen-hvm and kvm
790 hypervisors, and the default value for this has changed to false (in
791 2.0 xen-hvm always enabled it)
792 - Added luxi call to submit multiple jobs in one go
793 - Added cluster initialization option to not modify ``/etc/hosts``
795 - Added network interface parameters
796 - Added dry run mode to some LUs
797 - Added RAPI resources:
799 - ``/2/instances/[instance_name]/info``
800 - ``/2/instances/[instance_name]/replace-disks``
801 - ``/2/nodes/[node_name]/evacuate``
802 - ``/2/nodes/[node_name]/migrate``
803 - ``/2/nodes/[node_name]/role``
804 - ``/2/nodes/[node_name]/storage``
805 - ``/2/nodes/[node_name]/storage/modify``
806 - ``/2/nodes/[node_name]/storage/repair``
808 - Added OpCodes to evacuate or migrate all instances on a node
809 - Added new command to list storage elements on nodes (``gnt-node
810 list-storage``) and modify them (``gnt-node modify-storage``)
811 - Added new ssconf files with master candidate IP address
812 (``ssconf_master_candidates_ips``), node primary IP address
813 (``ssconf_node_primary_ips``) and node secondary IP address
814 (``ssconf_node_secondary_ips``)
815 - Added ``ganeti-confd`` and a client library to query the Ganeti
816 configuration via UDP
817 - Added ability to run hooks after cluster initialization and before
819 - Added automatic mode for disk replace (``gnt-instance replace-disks
821 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
822 after catastrophic data-loss
823 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
825 - Added ``gnt-instance move`` command to move instances
826 - Added ``gnt-cluster watcher`` command to control watcher
827 - Added ``gnt-node powercycle`` command to powercycle nodes
828 - Added new job status field ``lock_status``
829 - Added parseable error codes to cluster verification (``gnt-cluster
830 verify --error-codes``) and made output less verbose (use
831 ``--verbose`` to restore previous behaviour)
832 - Added UUIDs to the main config entities (cluster, nodes, instances)
833 - Added support for OS variants
834 - Added support for hashed passwords in the Ganeti remote API users file
836 - Added option to specify maximum timeout on instance shutdown
837 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
838 - Added new helper script to start and stop Ganeti daemons
839 (``daemon-util``), with the intent to reduce the work necessary to
840 adjust Ganeti for non-Debian distributions and to start/stop daemons
842 - Added more unittests
843 - Fixed critical bug in ganeti-masterd startup
844 - Removed the configure-time ``kvm-migration-port`` parameter, this is
845 now customisable at the cluster level for both the KVM and Xen
846 hypervisors using the new ``migration_port`` parameter
847 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
848 reinstalling an instance
849 - Allowed ``@`` in tag names
850 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
851 - Many documentation updates
852 - Distribute hypervisor files on ``gnt-cluster redist-conf``
853 - ``gnt-instance reinstall`` can now reinstall multiple instances
854 - Updated many command line parameters
855 - Introduced new OS API version 15
856 - No longer support a default hypervisor
857 - Treat virtual LVs as inexistent
858 - Improved job locking logic to reduce lock contention
859 - Match instance and node names case insensitively
860 - Reimplemented bash completion script to be more complete
867 *(Released Thu, 4 Feb 2010)*
869 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
870 - Fix a string formatting bug
871 - Improve safety of the code in some error paths
872 - Improve data validation in the master of values returned from nodes
878 *(Released Thu, 17 Dec 2009)*
880 - Fix security issue due to missing validation of iallocator names; this
881 allows local and remote execution of arbitrary executables
882 - Fix failure of gnt-node list during instance removal
883 - Ship the RAPI documentation in the archive
889 *(Released Wed, 30 Sep 2009)*
891 - Fixed many wrong messages
892 - Fixed a few bugs related to the locking library
893 - Fixed MAC checking at instance creation time
894 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
895 - Fixed a few issues related to signal handling in both daemons and
897 - Fixed the example startup script provided
898 - Fixed insserv dependencies in the example startup script (patch from
900 - Fixed handling of drained nodes in the iallocator framework
901 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
903 - Fixed error related to invalid job IDs in job polling
904 - Fixed job/opcode persistence on unclean master shutdown
905 - Fixed handling of partial job processing after unclean master
907 - Fixed error reporting from LUs, previously all errors were converted
908 into execution errors
909 - Fixed error reporting from burnin
910 - Decreased significantly the memory usage of the job queue
911 - Optimised slightly multi-job submission
912 - Optimised slightly opcode loading
913 - Backported the multi-job submit framework from the development
914 branch; multi-instance start and stop should be faster
915 - Added script to clean archived jobs after 21 days; this will reduce
916 the size of the queue directory
917 - Added some extra checks in disk size tracking
918 - Added an example ethers hook script
919 - Added a cluster parameter that prevents Ganeti from modifying of
921 - Added more node information to RAPI responses
922 - Added a ``gnt-job watch`` command that allows following the ouput of a
924 - Added a bind-address option to ganeti-rapi
925 - Added more checks to the configuration verify
926 - Enhanced the burnin script such that some operations can be retried
928 - Converted instance reinstall to multi-instance model
934 *(Released Fri, 7 Aug 2009)*
936 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
937 to allow using the pre-2.0.2 behaviour in activation, if any existing
938 instances have mismatched disk sizes in the configuration
939 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
940 any configuration mismatches for disk sizes
941 - Added ``gnt-master cluste-failover --no-voting`` to allow master
942 failover to work on two-node clusters
943 - Fixed the ``--net`` option of ``gnt-backup import``, which was
945 - Fixed detection of OS script errors in ``gnt-backup export``
946 - Fixed exit code of ``gnt-backup export``
952 *(Released Fri, 17 Jul 2009)*
954 - Added experimental support for stripped logical volumes; this should
955 enhance performance but comes with a higher complexity in the block
956 device handling; stripping is only enabled when passing
957 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
958 affected even in the non-stripped mode
959 - Improved resiliency against transient failures at the end of DRBD
960 resyncs, and in general of DRBD resync checks
961 - Fixed a couple of issues with exports and snapshot errors
962 - Fixed a couple of issues in instance listing
963 - Added display of the disk size in ``gnt-instance info``
964 - Fixed checking for valid OSes in instance creation
965 - Fixed handling of the "vcpus" parameter in instance listing and in
966 general of invalid parameters
967 - Fixed http server library, and thus RAPI, to handle invalid
968 username/password combinations correctly; this means that now they
969 report unauthorized for queries too, not only for modifications,
970 allowing earlier detect of configuration problems
971 - Added a new "role" node list field, equivalent to the master/master
972 candidate/drained/offline flags combinations
973 - Fixed cluster modify and changes of candidate pool size
974 - Fixed cluster verify error messages for wrong files on regular nodes
975 - Fixed a couple of issues with node demotion from master candidate role
976 - Fixed node readd issues
977 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
978 - Added a new ``--no-voting`` option for masterfailover to fix failover
979 on two-nodes clusters when the former master node is unreachable
980 - Added instance reinstall over RAPI
986 *(Released Tue, 16 Jun 2009)*
988 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
989 allow re-adding the start in single-user option (regression from 1.2)
990 - the watcher writes the instance status to a file, to allow monitoring
991 to report the instance status (from the master) based on cached
992 results of the watcher's queries; while this can get stale if the
993 watcher is being locked due to other work on the cluster, this is
995 - the watcher now also restarts the node daemon and the rapi daemon if
997 - fixed the watcher to handle full and drained queue cases
998 - hooks export more instance data in the environment, which helps if
999 hook scripts need to take action based on the instance's properties
1000 (no longer need to query back into ganeti)
1001 - instance failovers when the instance is stopped do not check for free
1002 RAM, so that failing over a stopped instance is possible in low memory
1004 - rapi uses queries for tags instead of jobs (for less job traffic), and
1005 for cluster tags it won't talk to masterd at all but read them from
1007 - a couple of error handling fixes in RAPI
1008 - drbd handling: improved the error handling of inconsistent disks after
1009 resync to reduce the frequency of "there are some degraded disks for
1010 this instance" messages
1011 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
1012 error handling path called a wrong function name)
1018 *(Released Wed, 27 May 2009)*
1020 - no changes from rc5
1023 Version 2.0 release candidate 5
1024 -------------------------------
1026 *(Released Wed, 20 May 2009)*
1028 - fix a couple of bugs (validation, argument checks)
1029 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
1030 - some small improvements to RAPI and IAllocator
1031 - make watcher automatically start the master daemon if down
1034 Version 2.0 release candidate 4
1035 -------------------------------
1037 *(Released Mon, 27 Apr 2009)*
1039 - change the OS list to not require locks; this helps with big clusters
1040 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
1041 volume group is broken
1042 - ``gnt-instance info``, without any arguments, doesn't run for all
1043 instances anymore; either pass ``--all`` or pass the desired
1044 instances; this helps against mistakes on big clusters where listing
1045 the information for all instances takes a long time
1046 - miscellaneous doc and man pages fixes
1049 Version 2.0 release candidate 3
1050 -------------------------------
1052 *(Released Wed, 8 Apr 2009)*
1054 - Change the internal locking model of some ``gnt-node`` commands, in
1055 order to reduce contention (and blocking of master daemon) when
1056 batching many creation/reinstall jobs
1057 - Fixes to Xen soft reboot
1058 - No longer build documentation at build time, instead distribute it in
1059 the archive, in order to reduce the need for the whole docbook/rst
1063 Version 2.0 release candidate 2
1064 -------------------------------
1066 *(Released Fri, 27 Mar 2009)*
1068 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
1069 - Fix watcher startup sequence, improves the behaviour of busy clusters
1070 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
1071 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
1072 bind address and other places
1073 - Some documentation fixes and updates
1076 Version 2.0 release candidate 1
1077 -------------------------------
1079 *(Released Mon, 2 Mar 2009)*
1081 - More documentation updates, now all docs should be more-or-less
1083 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
1085 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
1087 - Fix an issue related to $libdir/run/ganeti and cluster creation
1093 *(Released Thu, 19 Feb 2009)*
1095 - Xen PVM and KVM have switched the default value for the instance root
1096 disk to the first partition on the first drive, instead of the whole
1097 drive; this means that the OS installation scripts must be changed
1099 - Man pages have been updated
1100 - RAPI has been switched by default to HTTPS, and the exported functions
1101 should all work correctly
1102 - RAPI v1 has been removed
1103 - Many improvements to the KVM hypervisor
1104 - Block device errors are now better reported
1105 - Many other bugfixes and small improvements
1111 *(Released Mon, 26 Jan 2009)*
1113 - Version 2 is a general rewrite of the code and therefore the
1114 differences are too many to list, see the design document for 2.0 in
1115 the ``doc/`` subdirectory for more details
1116 - In this beta version there is not yet a migration path from 1.2 (there
1117 will be one in the final 2.0 release)
1118 - A few significant changes are:
1120 - all commands are executed by a daemon (``ganeti-masterd``) and the
1121 various ``gnt-*`` commands are just front-ends to it
1122 - all the commands are entered into, and executed from a job queue,
1123 see the ``gnt-job(8)`` manpage
1124 - the RAPI daemon supports read-write operations, secured by basic
1125 HTTP authentication on top of HTTPS
1126 - DRBD version 0.7 support has been removed, DRBD 8 is the only
1127 supported version (when migrating from Ganeti 1.2 to 2.0, you need
1128 to migrate to DRBD 8 first while still running Ganeti 1.2)
1129 - DRBD devices are using statically allocated minor numbers, which
1130 will be assigned to existing instances during the migration process
1131 - there is support for both Xen PVM and Xen HVM instances running on
1133 - KVM virtualization is supported too
1134 - file-based storage has been implemented, which means that it is
1135 possible to run the cluster without LVM and DRBD storage, for
1136 example using a shared filesystem exported from shared storage (and
1137 still have live migration)
1143 *(Released Tue, 13 Jan 2009)*
1145 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
1146 - Reuse the old instance mac address by default on instance import, if
1147 the instance name is the same.
1148 - Handle situations in which the node info rpc returns incomplete
1150 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
1151 - Improved version of batcher:
1153 - state file support
1154 - instance mac address support
1155 - support for HVM clusters/instances
1157 - Add an option to show the number of cpu sockets and nodes in
1159 - Support OSes that handle more than one version of the OS api (but do
1160 not change the current API in any other way)
1161 - Fix ``gnt-node migrate``
1162 - ``gnt-debug`` man page
1163 - Fixes various more typos and small issues
1164 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
1170 *(Released Wed, 24 Sep 2008)*
1172 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
1173 type of disk exported to fully virtualized instances.
1174 - provide access to the serial console of HVM instances
1175 - instance auto_balance flag, set by default. If turned off it will
1176 avoid warnings on cluster verify if there is not enough memory to fail
1177 over an instance. in the future it will prevent automatically failing
1178 it over when we will support that.
1179 - batcher tool for instance creation, see ``tools/README.batcher``
1180 - ``gnt-instance reinstall --select-os`` to interactively select a new
1181 operating system when reinstalling an instance.
1182 - when changing the memory amount on instance modify a check has been
1183 added that the instance will be able to start. also warnings are
1184 emitted if the instance will not be able to fail over, if auto_balance
1186 - documentation fixes
1187 - sync fields between ``gnt-instance list/modify/add/import``
1188 - fix a race condition in drbd when the sync speed was set after giving
1189 the device a remote peer.
1195 *(Released Tue, 22 Jul 2008)*
1197 - note: the allowed size and number of tags per object were reduced
1198 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1199 - fixed twisted 8.x compatibility
1200 - fixed ``gnt-instance replace-disks`` with iallocator
1201 - add TCP keepalives on twisted connections to detect restarted nodes
1202 - disk increase support, see ``gnt-instance grow-disk``
1203 - implement bulk node/instance query for RAPI
1204 - add tags in node/instance listing (optional)
1205 - experimental migration (and live migration) support, read the man page
1206 for ``gnt-instance migrate``
1207 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1208 has some small improvements in handling its state file
1214 *(Released Fri, 13 Jun 2008)*
1216 - Experimental readonly, REST-based remote API implementation;
1217 automatically started on master node, TCP port 5080, if enabled by
1218 ``--enable-rapi`` parameter to configure script.
1219 - Instance allocator support. Add and import instance accept a
1220 ``--iallocator`` parameter, and call that instance allocator to decide
1221 which node to use for the instance. The iallocator document describes
1222 what's expected from an allocator script.
1223 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1224 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1225 node is lost there is still enough memory to fail over the instances
1227 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1228 ``gnt-cluster verify``, to check for site-specific compliance. All the
1229 hooks will run, and their output, if any, will be displayed. Any
1230 failing hook will make the verification return an error value.
1231 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1232 primary and secondary interfaces
1233 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1234 that is still declared as part of the cluster and has failed.
1235 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1236 specifying output fields, that just adds the chosen fields to the
1238 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1239 export from the filesystem.
1240 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1241 have been added. Using them you can enable/disable acpi and pae
1242 support, and specify a path for a cd image to be exported to the
1243 instance. These parameters as the name suggest only work on HVM
1245 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1246 PAE support will be set to the previously hardcoded values, but the
1247 (previously hardcoded) path to the CDROM ISO image will be unset and
1248 if required, needs to be set manually with ``gnt-instance modify``
1250 - The address to which an instance's VNC console is bound is now
1251 selectable per-instance, rather than being cluster wide. Of course
1252 this only applies to instances controlled via VNC, so currently just
1253 applies to HVM clusters.
1259 *(Released Mon, 18 Feb 2008)*
1261 - more tweaks to the disk activation code (especially helpful for DRBD)
1262 - change the default ``gnt-instance list`` output format, now there is
1263 one combined status field (see the manpage for the exact values this
1265 - some more fixes for the mac export to hooks change
1266 - make Ganeti not break with DRBD 8.2.x (which changed the version
1267 format in ``/proc/drbd``) (issue 24)
1268 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1269 template, allowing migration from DRBD0.7+MD to DRBD8
1275 *(Released Wed, 30 Jan 2008)*
1277 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1279 - add command aliases infrastructure and a few aliases
1280 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1281 man pages and the ``--help`` option of ``gnt-node
1282 list``/``gnt-instance list``
1283 - fix ``gnt-backup list`` with down nodes (issue 21)
1284 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1285 - fix the dist archive and add a check for including svn/git files in
1287 - some developer-related changes: improve the burnin and the QA suite,
1288 add an upload script for testing during development
1294 *(Released Wed, 16 Jan 2008)*
1296 - experimental HVM support, read the install document, section
1297 "Initializing the cluster"
1298 - allow for the PVM hypervisor per-instance kernel and initrd paths
1299 - add a new command ``gnt-cluster verify-disks`` which uses a new
1300 algorithm to improve the reconnection of the DRBD pairs if the device
1301 on the secondary node has gone away
1302 - make logical volume code auto-activate LVs at disk activation time
1303 - slightly improve the speed of activating disks
1304 - allow specification of the MAC address at instance creation time, and
1305 changing it later via ``gnt-instance modify``
1306 - fix handling of external commands that generate lots of output on
1308 - update documentation with regard to minimum version of DRBD8 supported
1314 *(Released Tue, 4 Dec 2007)*
1316 - Log the ``xm create`` output to the node daemon log on failure (to
1317 help diagnosing the error)
1318 - In debug mode, log all external commands output if failed to the logs
1319 - Change parsing of lvm commands to ignore stderr
1325 *(Released Wed, 28 Nov 2007)*
1327 - Another round of updates to the DRBD 8 code to deal with more failures
1328 in the replace secondary node operation
1329 - Some more logging of failures in disk operations (lvm, drbd)
1330 - A few documentation updates
1337 *(Released Tue, 13 Nov 2007)*
1339 - Change configuration file format from Python's Pickle to JSON.
1340 Upgrading is possible using the cfgupgrade utility.
1341 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1342 faster replace disks and is more stable (DRBD 8 has many improvements
1343 compared to DRBD 0.7)
1344 - Added command line tags support (see man pages for ``gnt-instance``,
1345 ``gnt-node``, ``gnt-cluster``)
1346 - Added instance rename support
1347 - Added multi-instance startup/shutdown
1348 - Added cluster rename support
1349 - Added ``gnt-node evacuate`` to simplify some node operations
1350 - Added instance reboot operation that can speedup reboot as compared to
1352 - Soften the requirement that hostnames are in FQDN format
1353 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1355 - Removed dependency on debian's patched fping that uses the
1356 non-standard ``-S`` option
1357 - Now the OS definitions are searched for in multiple, configurable
1358 paths (easier for distros to package)
1359 - Some changes to the hooks infrastructure (especially the new
1360 post-configuration update hook)
1361 - Other small bugfixes
1363 .. vim: set textwidth=72 syntax=rst :