8 *(Released Thu, 15 May 2014)*
10 - Support restricted migration in hbal
11 - Fix for the --shared-file-storage-dir of gnt-cluster modify (issue 811)
12 - Fail in replace-disks if attaching disks fails (issue 814)
13 - Set IFF_ONE_QUEUE on created tap interfaces for KVM
14 - Small fixes and enhancements in the build system
15 - Various documentation fixes (e.g. issue 810)
21 *(Released Wed, 16 Apr 2014)*
23 - Fix filtering of pending jobs with -o id (issue 778)
24 - Make RAPI API calls more symmetric (issue 770)
25 - Make parsing of old cluster configuration more robust (issue 783)
26 - Fix wrong output of gnt-instance info after migrations
27 - Fix reserved PCI slots for KVM hotplugging
28 - Use runtime hypervisor parameters to calculate bockdevice options for KVM
29 - Fix high node daemon load during disk sync if the sync is paused manually
31 - Improve opportunistic locking during instance creation (issue 791)
33 Inherited from the 2.9 branch:
35 - Make watcher submit queries low priority (issue 772)
36 - Add reason parameter to RAPI client functions (issue 776)
37 - Fix failing gnt-node list-drbd command (issue 777)
38 - Properly display fake job locks in gnt-debug.
39 - small fixes in documentation
45 *(Released Mon, 24 Mar 2014)*
47 - Fix conflict between virtio + spice or soundhw (issue 757)
48 - accept relative paths in gnt-cluster copyfile (issue 754)
49 - Introduce shutdown timeout for 'xm shutdown' command
50 - Improve RAPI detection of the watcher (issue 752)
56 *(Released Wed, 5 Mar 2014)*
58 - Fix incorrect invocation of hooks on offline nodes (issue 742)
59 - Fix incorrect exit code of gnt-cluster verify in certain circumstances
62 Inherited from the 2.9 branch:
64 - Fix overflow problem in hbal that caused it to break when waiting for
65 jobs for more than 10 minutes (issue 717)
66 - Make hbal properly handle non-LVM storage
67 - Properly export and import NIC parameters, and do so in a backwards
68 compatible way (issue 716)
69 - Fix net-common script in case of routed mode (issue 728)
70 - Improve documentation (issues 724, 730)
76 *(Released Thu, 20 Feb 2014)*
78 Incompatible/important changes
79 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
81 - Adding disks with 'gnt-instance modify' now waits for the disks to sync per
82 default. Specify --no-wait-for-sync to override this behavior.
83 - The Ganeti python code now adheres to a private-module layout. In particular,
84 the module 'ganeti' is no longer in the python search path.
85 - On instance allocation, the iallocator now considers non-LVM storage
86 properly. In particular, actual file storage space information is used
87 when allocating space for a file/sharedfile instance.
88 - When disabling disk templates cluster-wide, the cluster now first
89 checks whether there are instances still using those templates.
90 - 'gnt-node list-storage' now also reports storage information about
91 file-based storage types.
92 - In case of non drbd instances, export \*_SECONDARY environment variables
93 as empty strings (and not "None") during 'instance-migrate' related hooks.
98 - KVM hypervisors can now access RBD storage directly without having to
99 go through a block device.
100 - A new command 'gnt-cluster upgrade' was added that automates the upgrade
101 procedure between two Ganeti versions that are both 2.10 or higher.
102 - The move-instance command can now change disk templates when moving
103 instances, and does not require any node placement options to be
104 specified if the destination cluster has a default iallocator.
105 - Users can now change the soundhw and cpuid settings for XEN hypervisors.
106 - Hail and hbal now have the (optional) capability of accessing average CPU
107 load information through the monitoring deamon, and to use it to dynamically
108 adapt the allocation of instances.
109 - Hotplug support. Introduce new option '--hotplug' to ``gnt-instance modify``
110 so that disk and NIC modifications take effect without the need of actual
111 reboot. There are a couple of constrains currently for this feature:
113 - only KVM hypervisor (versions >= 1.0) supports it,
114 - one can not (yet) hotplug a disk using userspace access mode for RBD
115 - in case of a downgrade instances should suffer a reboot in order to
116 be migratable (due to core change of runtime files)
117 - ``python-fdsend`` is required for NIC hotplugging.
122 - A new test framework for logical units was introduced and the test
123 coverage for logical units was improved significantly.
124 - Opcodes are entirely generated from Haskell using the tool 'hs2py' and
125 the module 'src/Ganeti/OpCodes.hs'.
126 - Constants are also generated from Haskell using the tool
127 'hs2py-constants' and the module 'src/Ganeti/Constants.hs', with the
128 exception of socket related constants, which require changing the
129 cluster configuration file, and HVS related constants, because they
130 are part of a port of instance queries to Haskell. As a result, these
131 changes will be part of the next release of Ganeti.
136 The following new dependencies have been added/updated.
140 - The version requirements for ``python-mock`` have increased to at least
141 version 1.0.1. It is still used for testing only.
142 - ``python-fdsend`` (https://gitorious.org/python-fdsend) is optional
143 but required for KVM NIC hotplugging to work.
148 - Fix integer overflow problem in hbal
154 *(Released Wed, 12 Feb 2014)*
156 This was the third RC release of the 2.10 series. Since 2.10.0 rc2:
158 - Improved hotplug robustness
159 - Start Ganeti daemons after ensure-dirs during upgrade
160 - Documentation improvements
162 Inherited from the 2.9 branch:
164 - Fix the RAPI instances-multi-alloc call
165 - assign unique filenames to file-based disks
166 - gracefully handle degraded non-diskless instances with 0 disks (issue 697)
167 - noded now runs with its specified group, which is the default group,
168 defaulting to root (issue 707)
169 - make using UUIDs to identify nodes in gnt-node consistently possible
176 *(Released Fri, 31 Jan 2014)*
178 This was the second RC release of the 2.10 series. Since 2.10.0 rc1:
180 - Documentation improvements
181 - Run drbdsetup syncer only on network attach
182 - Include target node in hooks nodes for migration
184 - Support post-upgrade hooks during cluster upgrades
186 Inherited from the 2.9 branch:
188 - Ensure that all the hypervisors exist in the config file (Issue 640)
189 - Correctly recognise the role as master node (Issue 687)
190 - configure: allow detection of Sphinx 1.2+ (Issue 502)
191 - gnt-instance now honors the KVM path correctly (Issue 691)
193 Inherited from the 2.8 branch:
195 - Change the list separator for the usb_devices parameter from comma to space.
196 Commas could not work because they are already the hypervisor option
197 separator (Issue 649)
198 - Add support for blktap2 file-driver (Issue 638)
199 - Add network tag definitions to the haskell codebase (Issue 641)
200 - Fix RAPI network tag handling
201 - Add the network tags to the tags searched by gnt-cluster search-tags
202 - Fix caching bug preventing jobs from being cancelled
203 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
209 *(Released Tue, 17 Dec 2013)*
211 This was the first RC release of the 2.10 series. Since 2.10.0 beta1:
213 - All known issues in 2.10.0 beta1 have been resolved (see changes from
215 - Improve handling of KVM runtime files from earlier Ganeti versions
216 - Documentation fixes
218 Inherited from the 2.9 branch:
220 - use custom KVM path if set for version checking
221 - SingleNotifyPipeCondition: don't share pollers
223 Inherited from the 2.8 branch:
225 - Fixed Luxi daemon socket permissions after master-failover
226 - Improve IP version detection code directly checking for colons rather than
227 passing the family from the cluster object
228 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
229 opportunistically anymore (Issue 622)
230 - Allow link local IPv6 gateways (Issue 624)
231 - Fix error printing (Issue 616)
232 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
233 disk modifications, keep the old one. If name=none then set disk name to
235 - Update build_chroot script to work with the latest hackage packages
236 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
237 - Fix evacuation out of drained node (Issue 615)
238 - Add default file_driver if missing (Issue 571)
239 - Fix job error message after unclean master shutdown (Issue 618)
240 - Lock group(s) when creating instances (Issue 621)
241 - SetDiskID() before accepting an instance (Issue 633)
242 - Allow the ext template disks to receive arbitrary parameters, both at creation
243 time and while being modified
244 - Xen handle domain shutdown (future proofing cherry-pick)
245 - Refactor reading live data in htools (future proofing cherry-pick)
251 *(Released Wed, 27 Nov 2013)*
253 This was the first beta release of the 2.10 series. All important changes
254 are listed in the latest 2.10 entry.
259 The following issues are known to be present in the beta and will be fixed
262 - Issue 477: Wrong permissions for confd LUXI socket
263 - Issue 621: Instance related opcodes do not aquire network/group locks
264 - Issue 622: Assertion Error: Node locks differ from node resource locks
265 - Issue 623: IPv6 Masterd <-> Luxid communication error
271 *(Released Mon, 7 Apr 2014)*
273 - Improve RAPI detection of the watcher (Issue 752)
274 - gnt-cluster copyfile: accept relative paths (Issue 754)
275 - Make watcher submit queries low priority (Issue 772)
276 - Add reason parameter to RAPI client functions (Issue 776)
277 - Fix failing gnt-node list-drbd command (Issue 777)
278 - Properly display fake job locks in gnt-debug.
279 - Enable timeout for instance shutdown
280 - small fixes in documentation
286 *(Released Tue, 25 Feb 2014)*
288 - Fix overflow problem in hbal that caused it to break when waiting for
289 jobs for more than 10 minutes (issue 717)
290 - Make hbal properly handle non-LVM storage
291 - Properly export and import NIC parameters, and do so in a backwards
292 compatible way (issue 716)
293 - Fix net-common script in case of routed mode (issue 728)
294 - Improve documentation (issues 724, 730)
300 *(Released Mon, 10 Feb 2014)*
302 - Fix the RAPI instances-multi-alloc call
303 - assign unique filenames to file-based disks
304 - gracefully handle degraded non-diskless instances with 0 disks (issue 697)
305 - noded now runs with its specified group, which is the default group,
306 defaulting to root (issue 707)
307 - make using UUIDs to identify nodes in gnt-node consistently possible
314 *(Released Mon, 27 Jan 2014)*
316 - Ensure that all the hypervisors exist in the config file (Issue 640)
317 - Correctly recognise the role as master node (Issue 687)
318 - configure: allow detection of Sphinx 1.2+ (Issue 502)
319 - gnt-instance now honors the KVM path correctly (Issue 691)
321 Inherited from the 2.8 branch:
323 - Change the list separator for the usb_devices parameter from comma to space.
324 Commas could not work because they are already the hypervisor option
325 separator (Issue 649)
326 - Add support for blktap2 file-driver (Issue 638)
327 - Add network tag definitions to the haskell codebase (Issue 641)
328 - Fix RAPI network tag handling
329 - Add the network tags to the tags searched by gnt-cluster search-tags
330 - Fix caching bug preventing jobs from being cancelled
331 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
337 *(Released Fri, 13 Dec 2013)*
339 - use custom KVM path if set for version checking
340 - SingleNotifyPipeCondition: don't share pollers
342 Inherited from the 2.8 branch:
344 - Fixed Luxi daemon socket permissions after master-failover
345 - Improve IP version detection code directly checking for colons rather than
346 passing the family from the cluster object
347 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
348 opportunistically anymore (Issue 622)
349 - Allow link local IPv6 gateways (Issue 624)
350 - Fix error printing (Issue 616)
351 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
352 disk modifications, keep the old one. If name=none then set disk name to
354 - Update build_chroot script to work with the latest hackage packages
355 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
356 - Fix evacuation out of drained node (Issue 615)
357 - Add default file_driver if missing (Issue 571)
358 - Fix job error message after unclean master shutdown (Issue 618)
359 - Lock group(s) when creating instances (Issue 621)
360 - SetDiskID() before accepting an instance (Issue 633)
361 - Allow the ext template disks to receive arbitrary parameters, both at creation
362 time and while being modified
363 - Xen handle domain shutdown (future proofing cherry-pick)
364 - Refactor reading live data in htools (future proofing cherry-pick)
370 *(Released Wed, 13 Nov 2013)*
372 - fix bug, that kept nodes offline when readding
373 - when verifying DRBD versions, ignore unavailable nodes
374 - fix bug that made the console unavailable on kvm in split-user
376 - DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
382 *(Released Tue, 5 Nov 2013)*
384 Incompatible/important changes
385 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
387 - hroller now also plans for capacity to move non-redundant instances off
388 any node to be rebooted; the old behavior of completely ignoring any
389 non-redundant instances can be restored by adding the --ignore-non-redundant
391 - The cluster option '--no-lvm-storage' was removed in favor of the new option
392 '--enabled-disk-templates'.
393 - On instance creation, disk templates no longer need to be specified
394 with '-t'. The default disk template will be taken from the list of
395 enabled disk templates.
396 - The monitoring daemon is now running as root, in order to be able to collect
397 information only available to root (such as the state of Xen instances).
398 - The ConfD client is now IPv6 compatible.
399 - File and shared file storage is no longer dis/enabled at configure time,
400 but using the option '--enabled-disk-templates' at cluster initialization and
402 - The default directories for file and shared file storage are not anymore
403 specified at configure time, but taken from the cluster's configuration.
404 They can be set at cluster initialization and modification with
405 '--file-storage-dir' and '--shared-file-storage-dir'.
406 - Cluster verification now includes stricter checks regarding the
407 default file and shared file storage directories. It now checks that
408 the directories are explicitely allowed in the 'file-storage-paths' file and
409 that the directories exist on all nodes.
410 - The list of allowed disk templates in the instance policy and the list
411 of cluster-wide enabled disk templates is now checked for consistency
412 on cluster or group modification. On cluster initialization, the ipolicy
413 disk templates are ensured to be a subset of the cluster-wide enabled
419 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
420 the correct command syntax. It is possible to use different DRBD versions
421 on different nodes as long as they are compatible to each other. This
422 enables rolling upgrades of DRBD with no downtime. As permanent operation
423 of different DRBD versions within a node group is discouraged,
424 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
425 - New "inst-status-xen" data collector for the monitoring daemon, providing
426 information about the state of the xen instances on the nodes.
427 - New "lv" data collector for the monitoring daemon, collecting data about the
428 logical volumes on the nodes, and pairing them with the name of the instances
430 - New "diskstats" data collector, collecting the data from /proc/diskstats and
431 presenting them over the monitoring daemon interface.
432 - The ConfD client is now IPv6 compatible.
436 The following new dependencies have been added.
440 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
441 for the unit tests (and only used for testing).
445 - ``hslogger`` (http://software.complete.org/hslogger) is now always
446 required, even if confd is not enabled.
451 - Correctly start/stop luxid during gnt-cluster master-failover (inherited
453 - Improved error messsages (inherited from stable-2.8)
459 *(Released Tue, 15 Oct 2013)*
461 The third release candidate in the 2.9 series. Since 2.9.0 rc2:
463 - in implicit configuration upgrade, match ipolicy with enabled disk templates
464 - improved harep documentation (inherited from stable-2.8)
470 *(Released Wed, 9 Oct 2013)*
472 The second release candidate in the 2.9 series. Since 2.9.0 rc1:
474 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
475 at least one DRBD instance.
476 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
483 *(Released Tue, 1 Oct 2013)*
485 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
488 - update of the documentation, in particular installation instructions
489 - merging of LD_* constants into DT_* constants
490 - python style changes to be compatible with newer versions of pylint
496 *(Released Thu, 29 Aug 2013)*
498 This was the first beta release of the 2.9 series. All important changes
499 are listed in the latest 2.9 entry.
505 *(Released Thu, 23 Jan 2014)*
507 - Change the list separator for the usb_devices parameter from comma to space.
508 Commas could not work because they are already the hypervisor option
509 separator (Issue 649)
510 - Add support for blktap2 file-driver (Issue 638)
511 - Add network tag definitions to the haskell codebase (Issue 641)
512 - Fix RAPI network tag handling
513 - Add the network tags to the tags searched by gnt-cluster search-tags
514 - Fix caching bug preventing jobs from being cancelled
515 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
521 *(Released Thu, 12 Dec 2013)*
523 - Fixed Luxi daemon socket permissions after master-failover
524 - Improve IP version detection code directly checking for colons rather than
525 passing the family from the cluster object
526 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
527 opportunistically anymore (Issue 622)
528 - Allow link local IPv6 gateways (Issue 624)
529 - Fix error printing (Issue 616)
530 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
531 disk modifications, keep the old one. If name=none then set disk name to
533 - Update build_chroot script to work with the latest hackage packages
534 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
535 - Fix evacuation out of drained node (Issue 615)
536 - Add default file_driver if missing (Issue 571)
537 - Fix job error message after unclean master shutdown (Issue 618)
538 - Lock group(s) when creating instances (Issue 621)
539 - SetDiskID() before accepting an instance (Issue 633)
540 - Allow the ext template disks to receive arbitrary parameters, both at creation
541 time and while being modified
542 - Xen handle domain shutdown (future proofing cherry-pick)
543 - Refactor reading live data in htools (future proofing cherry-pick)
549 *(Released Thu, 07 Nov 2013)*
551 - DRBD: ensure peers are UpToDate for dual-primary
552 - Improve error message for replace-disks
553 - More dependency checks at configure time
554 - Placate warnings on ganeti.outils_unittest.py
560 *(Released Thu, 17 Oct 2013)*
562 - Correctly start/stop luxid during gnt-cluster master-failover
563 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
564 - Fix path for the job queue serial file
565 - Improved harep man page
566 - Minor documentation improvements
572 *(Released Mon, 30 Sep 2013)*
574 Incompatible/important changes
575 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
577 - Instance policy can contain multiple instance specs, as described in
578 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
579 <design-partitioned>`. As a consequence, it's not possible to partially change
580 or override instance specs. Bounding specs (min and max) can be specified as a
581 whole using the new option ``--ipolicy-bounds-specs``, while standard
582 specs use the new option ``--ipolicy-std-specs``.
583 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
584 gnt-instance is a valid YAML object.
585 - hail now honors network restrictions when allocating nodes. This led to an
586 update of the IAllocator protocol. See the IAllocator documentation for
588 - confd now only answers static configuration request over the network. luxid
589 was extracted, listens on the local LUXI socket and responds to live queries.
590 This allows finer grained permissions if using separate users.
595 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
596 to always require authentication, ``--require-authentication``. It can
597 be specified in ``$sysconfdir/default/ganeti``.
598 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
599 be used to manage the disk templates to be used by instances in the cluster.
600 Initially, it will be set to a list that includes plain, drbd, if they were
601 enabled by specifying a volume group name, and file and sharedfile, if those
602 were enabled at configure time. Additionally, it will include all disk
603 templates that are currently used by instances. The order of disk templates
604 will be based on Ganeti's history of supporting them. In the future, the
605 first entry of the list will be used as a default disk template on instance
607 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
608 configuration back to the previous stable version.
609 - Disk templates in group ipolicy can be restored to the default value.
610 - Initial support for diskless instances and virtual clusters in QA.
611 - More QA and unit tests for instance policies.
612 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
613 describing why the opcode itself was executed.
614 - The monitoring daemon is now available. It allows users to query the cluster
615 for obtaining information about the status of the system. The daemon is only
616 responsible for providing the information over the network: the actual data
617 gathering is performed by data collectors (currently, only the DRBD status
618 collector is available).
619 - In order to help developers work on Ganeti, a new script
620 (``devel/build_chroot``) is provided, for building a chroot that contains all
621 the required development libraries and tools for compiling Ganeti on a Debian
623 - A new tool, ``harep``, for performing self-repair and recreation of instances
624 in Ganeti has been added.
625 - Split queries are enabled for tags, network, exports, cluster info, groups,
627 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
628 It prints the command line to set the current policies, to ease
630 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
631 devices for KVM virtio-net interfaces will get created with VNET_HDR
632 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
633 virtio-net interfaces, which prevents host kernel tainting and log
634 flooding, when dealing with broken or malicious virtio-net drivers.
635 It's set to true by default.
636 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
638 - Support 'viridian' parameter in Xen HVM
639 - Support DSA SSH keys in bootstrap
640 - To simplify the work of packaging frameworks that want to add the needed users
641 and groups in a split-user setup themselves, at build time three files in
642 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
643 the groups to be generated, the ``users`` file contains, one per line, the
644 users to be generated, optionally followed by their primary group, where
645 important. The ``groupmemberships`` file contains, one per line, additional
646 user-group membership relations that need to be established. The syntax of
647 these files will remain stable in all future versions.
652 The following new dependencies have been added:
655 - The ``curl`` library is not optional anymore for compiling the Haskell code.
656 - ``snap-server`` library (if monitoring is enabled).
659 - The minimum Python version needed to run Ganeti is now 2.6.
660 - ``yaml`` library (only for running the QA).
664 - Perform proper cleanup on termination of Haskell daemons
665 - Fix corner-case in handling of remaining retry time
671 *(Released Tue, 17 Sep 2013)*
673 - To simplify the work of packaging frameworks that want to add the needed users
674 and groups in a split-user setup themselves, at build time three files in
675 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
676 the groups to be generated, the ``users`` file contains, one per line, the
677 users to be generated, optionally followed by their primary group, where
678 important. The ``groupmemberships`` file contains, one per line, additional
679 user-group membership relations that need to be established. The syntax of
680 these files will remain stable in all future versions.
681 - Add a default to file-driver when unspecified over RAPI (Issue 571)
682 - Mark the DSA host pubkey as optional, and remove it during config downgrade
684 - Some documentation fixes
690 *(Released Tue, 27 Aug 2013)*
692 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
694 - Support 'viridian' parameter in Xen HVM (Issue 233)
695 - Include VCS version in ``gnt-cluster version``
696 - Support DSA SSH keys in bootstrap (Issue 338)
697 - Fix batch creation of instances
698 - Use FQDN to check master node status (Issue 551)
699 - Make the DRBD collector more failure-resilient
705 *(Released Fri, 2 Aug 2013)*
707 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
709 - Fix upgrading/downgrading from 2.7
710 - Increase maximum RAPI message size
711 - Documentation updates
712 - Split ``confd`` between ``luxid`` and ``confd``
713 - Merge 2.7 series up to the 2.7.1 release
714 - Allow the ``modify_etc_hosts`` option to be changed
715 - Add better debugging for ``luxid`` queries
716 - Expose bulk parameter for GetJobs in RAPI client
717 - Expose missing ``network`` fields in RAPI
718 - Add some ``cluster verify`` tests
719 - Some unittest fixes
720 - Fix a malfunction in ``hspace``'s tiered allocation
721 - Fix query compatibility between haskell and python implementations
722 - Add the ``vnet_hdr`` HV parameter for KVM
723 - Add ``--cleanup`` to instance failover
724 - Change the connected groups format in ``gnt-network info`` output; it
725 was previously displayed as a raw list by mistake. (Merged from 2.7)
731 *(Released Mon, 24 Jun 2013)*
733 This was the first beta release of the 2.8 series. All important changes
734 are listed in the latest 2.8 entry.
740 *(Released Thu, 26 Sep 2013)*
742 - Change the connected groups format in ``gnt-network info`` output; it
743 was previously displayed as a raw list by mistake
744 - Check disk template in right dict when copying
745 - Support multi-instance allocs without iallocator
746 - Fix some errors in the documentation
747 - Fix formatting of tuple in an error message
753 *(Released Thu, 25 Jul 2013)*
755 - Add logrotate functionality in daemon-util
756 - Add logrotate example file
757 - Add missing fields to network queries over rapi
758 - Fix network object timestamps
759 - Add support for querying network timestamps
760 - Fix a typo in the example crontab
761 - Fix a documentation typo
767 *(Released Thu, 04 Jul 2013)*
769 Incompatible/important changes
770 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
772 - Instance policies for disk size were documented to be on a per-disk
773 basis, but hail applied them to the sum of all disks. This has been
775 - ``hbal`` will now exit with status 0 if, during job execution over
776 LUXI, early exit has been requested and all jobs are successful;
777 before, exit status 1 was used, which cannot be differentiated from
779 - Compatibility with newer versions of rbd has been fixed
780 - ``gnt-instance batch-create`` has been changed to use the bulk create
781 opcode from Ganeti. This lead to incompatible changes in the format of
782 the JSON file. It's now not a custom dict anymore but a dict
783 compatible with the ``OpInstanceCreate`` opcode.
784 - Parent directories for file storage need to be listed in
785 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
786 write the file automatically based on old configuration values, but it
787 can not distribute it across all nodes and the file contents should be
788 verified. Use ``gnt-cluster copyfile
789 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
790 upgraded. The reason for requiring this list of paths now is that
791 before it would have been possible to inject new paths via RPC,
792 allowing files to be created in arbitrary locations. The RPC protocol
793 is protected using SSL/X.509 certificates, but as a design principle
794 Ganeti does not permit arbitrary paths to be passed.
795 - The parsing of the variants file for OSes (see
796 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
797 lines and comment lines (starting with ``#``) are ignored for better
799 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
800 longer available. ``gnt-node add`` now invokes a new tool on the
801 destination node, named ``prepare-node-join``, to configure the SSH
802 daemon. Paramiko is no longer necessary to configure nodes' SSH
803 daemons via ``gnt-node add``.
804 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
805 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
806 the flag is not reset after a master failover.
807 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
808 is recommended. 2.6 will be mandatory from the 2.8 series.
814 - New network management functionality to support automatic allocation
815 of IP addresses and managing of network parameters. See
816 :manpage:`gnt-network(8)` for more details.
817 - New external storage backend, to allow managing arbitrary storage
818 systems external to the cluster. See
819 :manpage:`ganeti-extstorage-interface(7)`.
820 - New ``exclusive-storage`` node parameter added, restricted to
821 nodegroup level. When it's set to true, physical disks are assigned in
822 an exclusive fashion to instances, as documented in :doc:`Partitioned
823 Ganeti <design-partitioned>`. Currently, only instances using the
824 ``plain`` disk template are supported.
825 - The KVM hypervisor has been updated with many new hypervisor
826 parameters, including a generic one for passing arbitrary command line
827 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
828 compatible up to qemu 1.4.
829 - A new tool, called ``mon-collector``, is the stand-alone executor of
830 the data collectors for a monitoring system. As of this version, it
831 just includes the DRBD data collector, that can be executed by calling
832 ``mon-collector`` using the ``drbd`` parameter. See
833 :manpage:`mon-collector(7)`.
834 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
835 for RAPI users. It allows granting permissions to query for
836 information to a specific user without giving
837 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
838 - A new tool named ``node-cleanup`` has been added. It cleans remains of
839 a cluster from a machine by stopping all daemons, removing
840 certificates and ssconf files. Unless the ``--no-backup`` option is
841 given, copies of the certificates are made.
842 - Instance creations now support the use of opportunistic locking,
843 potentially speeding up the (parallel) creation of multiple instances.
844 This feature is currently only available via the :doc:`RAPI
845 <rapi>` interface and when an instance allocator is used. If the
846 ``opportunistic_locking`` parameter is set the opcode will try to
847 acquire as many locks as possible, but will not wait for any locks
848 held by other opcodes. If not enough resources can be found to
849 allocate the instance, the temporary error code
850 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
851 retried thereafter, with or without opportunistic locking.
852 - New experimental linux-ha resource scripts.
853 - Restricted-commands support: ganeti can now be asked (via command line
854 or rapi) to perform commands on a node. These are passed via ganeti
855 RPC rather than ssh. This functionality is restricted to commands
856 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
857 reasons. The file is not copied automatically.
863 - Diskless instances are now externally mirrored (Issue 237). This for
864 now has only been tested in conjunction with explicit target nodes for
866 - Queries not needing locks or RPC access to the node can now be
867 performed by the confd daemon, making them independent from jobs, and
868 thus faster to execute. This is selectable at configure time.
869 - The functionality for allocating multiple instances at once has been
870 overhauled and is now also available through :doc:`RAPI <rapi>`.
872 There are no significant changes from version 2.7.0~rc3.
878 *(Released Tue, 25 Jun 2013)*
880 - Fix permissions on the confd query socket (Issue 477)
881 - Fix permissions on the job archive dir (Issue 498)
882 - Fix handling of an internal exception in replace-disks (Issue 472)
883 - Fix gnt-node info handling of shortened names (Issue 497)
884 - Fix gnt-instance grow-disk when wiping is enabled
885 - Documentation improvements, and support for newer pandoc
886 - Fix hspace honoring ipolicy for disks (Issue 484)
887 - Improve handling of the ``kvm_extra`` HV parameter
893 *(Released Fri, 24 May 2013)*
895 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
897 - Disks added through ``gnt-instance modify`` or created through
898 ``gnt-instance recreate-disks`` are wiped, if the
899 ``prealloc_wipe_disks`` flag is set.
900 - If wiping newly created disks fails, the disks are removed. Also,
901 partial failures in creating disks through ``gnt-instance modify``
902 triggers a cleanup of the partially-created disks.
903 - Removing the master IP address doesn't fail if the address has been
905 - Fix ownership of the OS log dir
906 - Workaround missing SO_PEERCRED constant (Issue 191)
912 *(Released Fri, 3 May 2013)*
914 This was the first release candidate of the 2.7 series. Since beta3:
916 - Fix kvm compatibility with qemu 1.4 (Issue 389)
917 - Documentation updates (admin guide, upgrade notes, install
918 instructions) (Issue 372)
919 - Fix gnt-group list nodes and instances count (Issue 436)
920 - Fix compilation without non-mandatory libraries (Issue 441)
921 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
922 - Make confd logging more verbose at INFO level (Issue 435)
923 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
924 - Fix failure path for instance storage type conversion (Issue 229)
925 - Update htools text backend documentation
926 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
927 - Disable inter-cluster instance move for file-based instances, because
928 it is dependant on instance export, which is not supported for
929 file-based instances. (Issue 414)
930 - Fix gnt-job crashes on non-ascii characters (Issue 427)
931 - Fix volume group checks on non-vm-capable nodes (Issue 432)
937 *(Released Mon, 22 Apr 2013)*
939 This was the third beta release of the 2.7 series. Since beta2:
941 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
942 - Fix data loss on wrong usage of ``gnt-instance move``
943 - Properly export errors in confd-based job queries
944 - Add ``users-setup`` tool
945 - Fix iallocator protocol to report 0 as a disk size for diskless
946 instances. This avoids hail breaking when a diskless instance is
948 - Fix job queue directory permission problem that made confd job queries
949 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
950 for access to archived jobs (Issue 406).
951 - Limit the sizes of networks supported by ``gnt-network`` to something
952 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
953 - Fix bugs in instance disk template conversion
954 - Fix GHC 7 compatibility
955 - Fix ``burnin`` install path (Issue 426).
956 - Allow very small disk grows (Issue 347).
957 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
958 that noded doesn't import masterd code (Issue 419).
959 - Make sure the default metavg at cluster init is the same as the vg, if
960 unspecified (Issue 358).
961 - Fix cleanup of partially created disks (part of Issue 416)
967 *(Released Tue, 2 Apr 2013)*
969 This was the second beta release of the 2.7 series. Since beta1:
971 - Networks no longer have a "type" slot, since this information was
972 unused in Ganeti: instead of it tags should be used.
973 - The rapi client now has a ``target_node`` option to MigrateInstance.
974 - Fix early exit return code for hbal (Issue 386).
975 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
976 - Fix ``rbd showmapped`` output parsing (Issue 312).
977 - Networks are now referenced indexed by UUID, rather than name. This
978 will require running cfgupgrade, from 2.7.0beta1, if networks are in
980 - The OS environment now includes network information.
981 - Deleting of a network is now disallowed if any instance nic is using
982 it, to prevent dangling references.
983 - External storage is now documented in man pages.
984 - The exclusive_storage flag can now only be set at nodegroup level.
985 - Hbal can now submit an explicit priority with its jobs.
986 - Many network related locking fixes.
987 - Bump up the required pylint version to 0.25.1.
988 - Fix the ``no_remember`` option in RAPI client.
989 - Many ipolicy related tests, qa, and fixes.
990 - Many documentation improvements and fixes.
991 - Fix building with ``--disable-file-storage``.
992 - Fix ``-q`` option in htools, which was broken if passed more than
994 - Some haskell/python interaction improvements and fixes.
995 - Fix iallocator in case of missing LVM storage.
996 - Fix confd config load in case of ``--no-lvm-storage``.
997 - The confd/query functionality is now mentioned in the security
1004 *(Released Wed, 6 Feb 2013)*
1006 This was the first beta release of the 2.7 series. All important changes
1007 are listed in the latest 2.7 entry.
1013 *(Released Fri, 21 Dec 2012)*
1015 Important behaviour change: hbal won't rebalance anymore instances which
1016 have the ``auto_balance`` attribute set to false. This was the intention
1017 all along, but until now it only skipped those from the N+1 memory
1018 reservation (DRBD-specific).
1020 A significant number of bug fixes in this release:
1022 - Fixed disk adoption interaction with ipolicy checks.
1023 - Fixed networking issues when instances are started, stopped or
1024 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
1025 - Fixed the warning in cluster verify for shared storage instances not
1027 - Fixed removal of storage directory on shared file storage (issue 262).
1028 - Fixed validation of LVM volume group name in OpClusterSetParams
1029 (``gnt-cluster modify``) (issue 285).
1030 - Fixed runtime memory increases (``gnt-instance modify -m``).
1031 - Fixed live migration under Xen's ``xl`` mode.
1032 - Fixed ``gnt-instance console`` with ``xl``.
1033 - Fixed building with newer Haskell compiler/libraries.
1034 - Fixed PID file writing in Haskell daemons (confd); this prevents
1035 restart issues if confd was launched manually (outside of
1036 ``daemon-util``) while another copy of it was running
1037 - Fixed a type error when doing live migrations with KVM (issue 297) and
1038 the error messages for failing migrations have been improved.
1039 - Fixed opcode validation for the out-of-band commands (``gnt-node
1041 - Fixed a type error when unsetting OS hypervisor parameters (issue
1042 311); now it's possible to unset all OS-specific hypervisor
1044 - Fixed the ``dry-run`` mode for many operations: verification of
1045 results was over-zealous but didn't take into account the ``dry-run``
1046 operation, resulting in "wrong" failures.
1047 - Fixed bash completion in ``gnt-job list`` when the job queue has
1048 hundreds of entries; especially with older ``bash`` versions, this
1049 results in significant CPU usage.
1051 And lastly, a few other improvements have been made:
1053 - Added option to force master-failover without voting (issue 282).
1054 - Clarified error message on lock conflict (issue 287).
1055 - Logging of newly submitted jobs has been improved (issue 290).
1056 - Hostname checks have been made uniform between instance rename and
1058 - The ``--submit`` option is now supported by ``gnt-debug delay``.
1059 - Shutting down the master daemon by sending SIGTERM now stops it from
1060 processing jobs waiting for locks; instead, those jobs will be started
1061 once again after the master daemon is started the next time (issue
1063 - Support for Xen's ``xl`` program has been improved (besides the fixes
1065 - Reduced logging noise in the Haskell confd daemon (only show one log
1066 entry for each config reload, instead of two).
1067 - Several man page updates and typo fixes.
1073 *(Released Fri, 12 Oct 2012)*
1075 A small bugfix release. Among the bugs fixed:
1077 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
1078 made the command unusable.
1079 - Commands that issue many jobs don't fail anymore just because some jobs
1080 take so long that other jobs are archived.
1081 - Failures during ``gnt-instance reinstall`` are reflected by the exit
1083 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
1085 - When ``always_failover`` is set, ``--allow-failover`` is not required
1086 in migrate commands anymore.
1087 - ``bash_completion`` works even if extglob is disabled.
1088 - Fixed bug with locks that made failover for RDB-based instances fail.
1089 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
1090 a random node instead of one based on the allocator metric.
1091 - Support for newer versions of pylint and pep8.
1092 - Hail doesn't fail anymore when trying to add an instance of type
1093 ``file``, ``sharedfile`` or ``rbd``.
1094 - Added new Makefile target to rebuild the whole distribution, so that
1095 all files are included.
1101 *(Released Fri, 27 Jul 2012)*
1104 .. attention:: The ``LUXI`` protocol has been made more consistent
1105 regarding its handling of command arguments. This, however, leads to
1106 incompatibility issues with previous versions. Please ensure that you
1107 restart Ganeti daemons soon after the upgrade, otherwise most
1108 ``LUXI`` calls (job submission, setting/resetting the drain flag,
1109 pausing/resuming the watcher, cancelling and archiving jobs, querying
1110 the cluster configuration) will fail.
1119 The current ``admin_up`` field, which used to denote whether an instance
1120 should be running or not, has been removed. Instead, ``admin_state`` is
1121 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
1123 The rational behind this is that an instance being “down” can have
1126 - it could be down during a reboot
1127 - it could be temporarily be down for a reinstall
1128 - or it could be down because it is deprecated and kept just for its
1131 The previous Boolean state was making it difficult to do capacity
1132 calculations: should Ganeti reserve memory for a down instance? Now, the
1133 tri-state field makes it clear:
1135 - in ``up`` and ``down`` state, all resources are reserved for the
1136 instance, and it can be at any time brought up if it is down
1137 - in ``offline`` state, only disk space is reserved for it, but not
1140 The field can have an extra use: since the transition between ``up`` and
1141 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
1142 transition between ``offline`` and ``down`` is done via ``gnt-instance
1143 modify``, it is possible to given different rights to users. For
1144 example, owners of an instance could be allowed to start/stop it, but
1145 not transition it out of the offline state.
1147 Instance policies and specs
1148 +++++++++++++++++++++++++++
1150 In previous Ganeti versions, an instance creation request was not
1151 limited on the minimum size and on the maximum size just by the cluster
1152 resources. As such, any policy could be implemented only in third-party
1153 clients (RAPI clients, or shell wrappers over ``gnt-*``
1154 tools). Furthermore, calculating cluster capacity via ``hspace`` again
1155 required external input with regards to instance sizes.
1157 In order to improve these workflows and to allow for example better
1158 per-node group differentiation, we introduced instance specs, which
1161 - minimum instance disk size, disk count, memory size, cpu count
1162 - maximum values for the above metrics
1163 - and “standard” values (used in ``hspace`` to calculate the standard
1166 The minimum/maximum values can be also customised at node-group level,
1167 for example allowing more powerful hardware to support bigger instance
1170 Beside the instance specs, there are a few other settings belonging to
1171 the instance policy framework. It is possible now to customise, per
1172 cluster and node-group:
1174 - the list of allowed disk templates
1175 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
1176 - the maximum ratio of instance to spindles (see below for more
1177 information) for local storage
1179 All these together should allow all tools that talk to Ganeti to know
1180 what are the ranges of allowed values for instances and the
1181 over-subscription that is allowed.
1183 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
1184 instance configuration, and the physical CPU configuration from the
1185 node. For the spindle ratios however, we didn't track before these
1186 values, so new parameters have been added:
1188 - a new node parameter ``spindle_count``, defaults to 1, customisable at
1189 node group or node level
1190 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
1192 Note that spindles in this context doesn't need to mean actual
1193 mechanical hard-drives; it's just a relative number for both the node
1194 I/O capacity and instance I/O consumption.
1196 Instance migration behaviour
1197 ++++++++++++++++++++++++++++
1199 While live-migration is in general desirable over failover, it is
1200 possible that for some workloads it is actually worse, due to the
1201 variable time of the “suspend” phase during live migration.
1203 To allow the tools to work consistently over such instances (without
1204 having to hard-code instance names), a new backend parameter
1205 ``always_failover`` has been added to control the migration/failover
1206 behaviour. When set to True, all migration requests for an instance will
1207 instead fall-back to failover.
1209 Instance memory ballooning
1210 ++++++++++++++++++++++++++
1212 Initial support for memory ballooning has been added. The memory for an
1213 instance is no longer fixed (backend parameter ``memory``), but instead
1214 can vary between minimum and maximum values (backend parameters
1215 ``minmem`` and ``maxmem``). Currently we only change an instance's
1218 - live migrating or failing over and instance and the target node
1219 doesn't have enough memory
1220 - user requests changing the memory via ``gnt-instance modify
1223 Instance CPU pinning
1224 ++++++++++++++++++++
1226 In order to control the use of specific CPUs by instance, support for
1227 controlling CPU pinning has been added for the Xen, HVM and LXC
1228 hypervisors. This is controlled by a new hypervisor parameter
1229 ``cpu_mask``; details about possible values for this are in the
1230 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
1231 VCPU-to-CPU mapping) form will work well only when all nodes in your
1232 cluster have the same amount of CPUs.
1237 Another area in which Ganeti was not customisable were the parameters
1238 used for storage configuration, e.g. how many stripes to use for LVM,
1239 DRBD resync configuration, etc.
1241 To improve this area, we've added disks parameters, which are
1242 customisable at cluster and node group level, and which allow to
1243 specify various parameters for disks (DRBD has the most parameters
1244 currently), for example:
1246 - DRBD resync algorithm and parameters (e.g. speed)
1247 - the default VG for meta-data volumes for DRBD
1248 - number of stripes for LVM (plain disk template)
1251 These parameters can be modified via ``gnt-cluster modify -D …`` and
1252 ``gnt-group modify -D …``, and are used at either instance creation (in
1253 case of LVM stripes, for example) or at disk “activation” time
1254 (e.g. resync speed).
1256 Rados block device support
1257 ++++++++++++++++++++++++++
1259 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
1260 denoted by the ``rbd`` disk template type. This is considered
1261 experimental, feedback is welcome. For details on configuring it, see
1262 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
1267 The existing master IP functionality works well only in simple setups (a
1268 single network shared by all nodes); however, if nodes belong to
1269 different networks, then the ``/32`` setup and lack of routing
1270 information is not enough.
1272 To allow the master IP to function well in more complex cases, the
1273 system was reworked as follows:
1275 - a master IP netmask setting has been added
1276 - the master IP activation/turn-down code was moved from the node daemon
1277 to a separate script
1278 - whether to run the Ganeti-supplied master IP script or a user-supplied
1279 on is a ``gnt-cluster init`` setting
1281 Details about the location of the standard and custom setup scripts are
1282 in the man page :manpage:`gnt-cluster(8)`; for information about the
1283 setup script protocol, look at the Ganeti-supplied script.
1288 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
1291 It is now possible to use TLS-protected connections, and when renewing
1292 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
1293 it is now possible to specify spice or spice CA certificates. Also, it
1294 is possible to configure a password for SPICE sessions via the
1295 hypervisor parameter ``spice_password_file``.
1297 There are also new parameters to control the compression and streaming
1298 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
1299 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
1300 for the spice parameters.
1302 Lastly, it is now possible to see the SPICE connection information via
1303 ``gnt-instance console``.
1308 A new tool (``tools/ovfconverter``) has been added that supports
1309 conversion between Ganeti and the `Open Virtualization Format
1310 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
1313 This relies on the ``qemu-img`` tool to convert the disk formats, so the
1314 actual compatibility with other virtualization solutions depends on it.
1316 Confd daemon changes
1317 ++++++++++++++++++++
1319 The configuration query daemon (``ganeti-confd``) is now optional, and
1320 has been rewritten in Haskell; whether to use the daemon at all, use the
1321 Python (default) or the Haskell version is selectable at configure time
1322 via the ``--enable-confd`` parameter, which can take one of the
1323 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
1324 daemon will result in a smaller footprint; for larger systems, we
1325 welcome feedback on the Haskell version which might become the default
1328 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
1329 daemon running. The Python version doesn't implement the new call.
1332 User interface changes
1333 ~~~~~~~~~~~~~~~~~~~~~~
1335 We have replaced the ``--disks`` option of ``gnt-instance
1336 replace-disks`` with a more flexible ``--disk`` option, which allows
1337 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
1338 disk size and mode can be changed upon recreation (via ``gnt-instance
1339 recreate-disks``, which accepts the same ``--disk`` option).
1341 As many people are used to a ``show`` command, we have added that as an
1342 alias to ``info`` on all ``gnt-*`` commands.
1344 The ``gnt-instance grow-disk`` command has a new mode in which it can
1345 accept the target size of the disk, instead of the delta; this can be
1346 more safe since two runs in absolute mode will be idempotent, and
1347 sometimes it's also easier to specify the desired size directly.
1349 Also the handling of instances with regard to offline secondaries has
1350 been improved. Instance operations should not fail because one of it's
1351 secondary nodes is offline, even though it's safe to proceed.
1353 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
1354 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1355 minors to instance name.
1360 RAPI coverage has improved, with (for example) new resources for
1361 recreate-disks, node power-cycle, etc.
1366 There is partial support for ``xl`` in the Xen hypervisor; feedback is
1369 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1370 whether to still support Python 2.4 or move to Python 2.6 as minimum
1373 Support for Fedora has been slightly improved; the provided example
1374 init.d script should work better on it and the INSTALL file should
1375 document the needed dependencies.
1380 The deprecated ``QueryLocks`` LUXI request has been removed. Use
1381 ``Query(what=QR_LOCK, ...)`` instead.
1383 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1384 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1385 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1386 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1387 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1389 RAPI client: ``CertificateError`` now derives from
1390 ``GanetiApiError``. This should make it more easy to handle Ganeti
1393 Deprecation warnings due to PyCrypto/paramiko import in
1394 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
1395 make sure to run an up-to-date paramiko version, if you use this tool.
1397 The QA scripts now depend on Python 2.5 or above (the main code base
1398 still works with Python 2.4).
1400 The configuration file (``config.data``) is now written without
1401 indentation for performance reasons; if you want to edit it, it can be
1402 re-formatted via ``tools/fmtjson``.
1404 A number of bugs has been fixed in the cluster merge tool.
1406 ``x509`` certification verification (used in import-export) has been
1407 changed to allow the same clock skew as permitted by the cluster
1408 verification. This will remove some rare but hard to diagnose errors in
1415 *(Released Thu, 19 Jul 2012)*
1417 Very few changes from rc4 to the final release, only bugfixes:
1419 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
1420 instance, fix CDROM booting for KVM instances)
1421 - fixed node group modification of node parameters
1422 - fixed issue in LUClusterVerifyGroup with multi-group clusters
1423 - fixed generation of bash completion to ensure a stable ordering
1430 *(Released Fri, 13 Jul 2012)*
1432 Third release candidate for 2.6. The following changes were done from
1435 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1436 - Fixed an inconsistency in the LUXI protocol with the provided
1437 arguments (NOT backwards compatible)
1438 - Fixed a bug with node groups ipolicy where ``min`` was greater than
1439 the cluster ``std`` value
1440 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1441 easier instance debugging on nodes (requires ``hconfd`` to work)
1447 *(Released Tue, 03 Jul 2012)*
1449 Second release candidate for 2.6. The following changes were done from
1452 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1454 - Fixed a RAPI regression on missing beparams/memory
1455 - Fixed redistribution of files on offline nodes
1456 - Added possibility to run activate-disks even though secondaries are
1457 offline. With this change it relaxes also the strictness on some other
1458 commands which use activate disks internally:
1459 * ``gnt-instance start|reboot|rename|backup|export``
1460 - Made it possible to remove safely an instance if its secondaries are
1462 - Made it possible to reinstall even though secondaries are offline
1468 *(Released Mon, 25 Jun 2012)*
1470 First release candidate for 2.6. The following changes were done from
1473 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1475 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1476 behaviour. It supports now the modify operation
1477 - ``hcheck`` implemented to analyze cluster health and possibility of
1478 improving health by rebalance
1479 - ``hbal`` has been improved in dealing with split instances
1485 *(Released Mon, 11 Jun 2012)*
1487 Second beta release of 2.6. The following changes were done from beta2
1490 - Fixed ``daemon-util`` with non-root user models
1491 - Fixed creation of plain instances with ``--no-wait-for-sync``
1492 - Fix wrong iv_names when running ``cfgupgrade``
1493 - Export more information in RAPI group queries
1494 - Fixed bug when changing instance network interfaces
1495 - Extended burnin to do NIC changes
1496 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1497 - Changed default for DRBD barriers
1498 - Fixed DRBD error reporting for syncer rate
1499 - Verify the options on disk parameters
1501 And of course various fixes to documentation and improved unittests and
1508 *(Released Wed, 23 May 2012)*
1510 First beta release of 2.6. The following changes were done from beta1 to
1513 - integrated patch for distributions without ``start-stop-daemon``
1514 - adapted example init.d script to work on Fedora
1515 - fixed log handling in Haskell daemons
1516 - adapted checks in the watcher for pycurl linked against libnss
1517 - add partial support for ``xl`` instead of ``xm`` for Xen
1518 - fixed a type issue in cluster verification
1519 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1522 Plus integrated fixes from the 2.5 branch:
1524 - fixed ``kvm-ifup`` to use ``/bin/bash``
1525 - fixed parallel build failures
1526 - KVM live migration when using a custom keymap
1532 *(Released Tue, 24 Jul 2012)*
1534 A small bugfix release, with no new features:
1536 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1537 different default shell (e.g. Debian, Ubuntu)
1538 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1539 243 and Debian bug #650664)
1540 - fixed compatibility with KVM versions that don't support multiple boot
1541 devices (fixes Issue 230 and Debian bug #624256)
1543 Additionally, a few fixes were done to the build system (fixed parallel
1544 build failures) and to the unittests (fixed race condition in test for
1545 FileID functions, and the default enable/disable mode for QA test is now
1552 *(Released Fri, 11 May 2012)*
1554 A small bugfix release.
1556 The main issues solved are on the topic of compatibility with newer LVM
1559 - fixed parsing of ``lv_attr`` field
1560 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1561 the ``--force`` flag is needed
1563 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1564 to accept kernel 3.x too (was hardcoded to 2.6.*).
1566 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1567 of hook results and that also made display of other errors suboptimal
1568 was fixed; the code behaves now like 2.4 and earlier.
1570 Another change in 2.5, the cleanup of the OS scripts environment, is too
1571 aggressive: it removed even the ``PATH`` variable, which requires the OS
1572 scripts to *always* need to export it. Since this is a bit too strict,
1573 we now export a minimal PATH, the same that we export for hooks.
1575 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1576 integrated into this release.
1578 Finally, a few other miscellaneous changes were done (no new features,
1579 just small improvements):
1581 - Fix ``gnt-group --help`` display
1582 - Fix hardcoded Xen kernel path
1583 - Fix grow-disk handling of invalid units
1584 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1585 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1591 *(Released Thu, 12 Apr 2012)*
1593 Incompatible/important changes and bugfixes
1594 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1596 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1597 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1598 to match the underlying LUXI interface.
1599 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1600 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1601 not maintain backwards-compatibility as the underlying operation
1602 changed in an incompatible way. The RAPI client can talk to old
1603 servers, but it needs to be told so as the return value changed.
1604 - When creating file-based instances via RAPI, the ``file_driver``
1605 parameter no longer defaults to ``loop`` and must be specified.
1606 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1608 - Support for the undocumented and deprecated RAPI instance creation
1609 request format version 0 has been dropped. Use version 1, supported
1610 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1611 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1612 documentation <install>`.
1613 - The "cluster-verify" hooks are now executed per group by the
1614 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1615 if you just run ``gnt-cluster verify``, which generates one opcode per
1617 - The environment as passed to the OS scripts is cleared, and thus no
1618 environment variables defined in the node daemon's environment will be
1619 inherited by the scripts.
1620 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1622 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1623 support operations involving multiple node groups.
1624 - Offline nodes are ignored when failing over an instance.
1625 - Support for KVM version 1.0, which changed the version reporting format
1627 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1629 - ``Makefile`` is now compatible with Automake 1.11.2
1630 - Includes all bugfixes made in the 2.4 series
1635 - The ganeti-htools project has been merged into the ganeti-core source
1636 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1637 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1638 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1639 GPT-style partition tables (requires `parted
1640 <http://www.gnu.org/s/parted/>`_).
1641 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1642 - Allowed adding tags on instance creation.
1643 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1644 - Allow instances to be started in a paused state, enabling the user to
1645 see the complete console output on boot using the console.
1646 - Added new hypervisor flag to control default reboot behaviour
1647 (``reboot_behavior``).
1648 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1649 - Improved out-of-band management support:
1651 - Added ``gnt-node health`` command reporting the health status of
1653 - Added ``gnt-node power`` command to manage power status of nodes.
1654 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1656 - Instance migration can fall back to failover if instance is not
1658 - Filters can be used when listing nodes, instances, groups and locks;
1659 see :manpage:`ganeti(7)` manpage.
1660 - Added post-execution status as variables to :doc:`hooks <hooks>`
1662 - Instance tags are exported/imported together with the instance.
1663 - When given an explicit job ID, ``gnt-job info`` will work for archived
1665 - Jobs can define dependencies on other jobs (not yet supported via
1666 RAPI or command line, but used by internal commands and usable via
1669 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1672 - Instance failover is now available as a RAPI resource
1673 (``/2/instances/[instance_name]/failover``).
1674 - ``gnt-instance info`` defaults to static information if primary node
1676 - Opcodes have a new ``comment`` attribute.
1677 - Added basic SPICE support to KVM hypervisor.
1678 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1680 Node group improvements
1681 ~~~~~~~~~~~~~~~~~~~~~~~
1683 - ``gnt-cluster verify`` has been modified to check groups separately,
1684 thereby improving performance.
1685 - Node group support has been added to ``gnt-cluster verify-disks``,
1686 which now operates per node group.
1687 - Watcher has been changed to work better with node groups.
1689 - One process and state file per node group.
1690 - Slow watcher in one group doesn't block other group's watcher.
1692 - Added new command, ``gnt-group evacuate``, to move all instances in a
1693 node group to other groups.
1694 - Added ``gnt-instance change-group`` to move an instance to another
1696 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1697 per-group operations.
1698 - Node groups can be tagged.
1699 - Some operations switch from an exclusive to a shared lock as soon as
1701 - Instance's primary and secondary nodes' groups are now available as
1702 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1703 and ``snodes.group.uuid``).
1708 - Numerous updates to documentation and manpages.
1710 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1712 - Some opcode/job results are now also documented, see :doc:`RAPI
1715 - A lockset's internal lock is now also visible in lock monitor.
1716 - Log messages from job queue workers now contain information about the
1717 opcode they're processing.
1718 - ``gnt-instance console`` no longer requires the instance lock.
1719 - A short delay when waiting for job changes reduces the number of LUXI
1720 requests significantly.
1721 - DRBD metadata volumes are overwritten with zeros during disk creation.
1722 - Out-of-band commands no longer acquire the cluster lock in exclusive
1724 - ``devel/upload`` now uses correct permissions for directories.
1730 *(Released Fri, 23 Mar 2012)*
1732 This was the sixth release candidate of the 2.5 series.
1738 *(Released Mon, 9 Jan 2012)*
1740 This was the fifth release candidate of the 2.5 series.
1746 *(Released Thu, 27 Oct 2011)*
1748 This was the fourth release candidate of the 2.5 series.
1754 *(Released Wed, 26 Oct 2011)*
1756 This was the third release candidate of the 2.5 series.
1762 *(Released Tue, 18 Oct 2011)*
1764 This was the second release candidate of the 2.5 series.
1770 *(Released Tue, 4 Oct 2011)*
1772 This was the first release candidate of the 2.5 series.
1778 *(Released Wed, 31 Aug 2011)*
1780 This was the third beta release of the 2.5 series.
1786 *(Released Mon, 22 Aug 2011)*
1788 This was the second beta release of the 2.5 series.
1794 *(Released Fri, 12 Aug 2011)*
1796 This was the first beta release of the 2.5 series.
1802 *(Released Thu, 27 Oct 2011)*
1804 - Fixed bug when parsing command line parameter values ending in
1806 - Fixed assertion error after unclean master shutdown
1807 - Disable HTTP client pool for RPC, significantly reducing memory usage
1809 - Fixed queue archive creation with wrong permissions
1815 *(Released Tue, 23 Aug 2011)*
1819 - Fixed documentation for importing with ``--src-dir`` option
1820 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1821 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1827 *(Released Fri, 5 Aug 2011)*
1829 Many bug-fixes and a few small features:
1831 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1832 issues when you tried to add an instance with two MAC addresses in one
1834 - KVM: fixed per-instance stored UID value
1835 - KVM: configure bridged NICs at migration start
1836 - KVM: Fix a bug where instance will not start with never KVM versions
1838 - Added OS search path to ``gnt-cluster info``
1839 - Fixed an issue with ``file_storage_dir`` where you were forced to
1840 provide an absolute path, but the documentation states it is a
1841 relative path, the documentation was right
1842 - Added a new parameter to instance stop/start called ``--no-remember``
1843 that will make the state change to not be remembered
1844 - Implemented ``no_remember`` at RAPI level
1845 - Improved the documentation
1846 - Node evacuation: don't call IAllocator if node is already empty
1847 - Fixed bug in DRBD8 replace disks on current nodes
1848 - Fixed bug in recreate-disks for DRBD instances
1849 - Moved assertion checking locks in ``gnt-instance replace-disks``
1850 causing it to abort with not owning the right locks for some situation
1851 - Job queue: Fixed potential race condition when cancelling queued jobs
1852 - Fixed off-by-one bug in job serial generation
1853 - ``gnt-node volumes``: Fix instance names
1854 - Fixed aliases in bash completion
1855 - Fixed a bug in reopening log files after being sent a SIGHUP
1856 - Added a flag to burnin to allow specifying VCPU count
1857 - Bugfixes to non-root Ganeti configuration
1863 *(Released Thu, 12 May 2011)*
1865 Many bug-fixes and a few new small features:
1867 - Fixed a bug related to log opening failures
1868 - Fixed a bug in instance listing with orphan instances
1869 - Fixed a bug which prevented resetting the cluster-level node parameter
1870 ``oob_program`` to the default
1871 - Many fixes related to the ``cluster-merge`` tool
1872 - Fixed a race condition in the lock monitor, which caused failures
1873 during (at least) creation of many instances in parallel
1874 - Improved output for gnt-job info
1875 - Removed the quiet flag on some ssh calls which prevented debugging
1877 - Improved the N+1 failure messages in cluster verify by actually
1878 showing the memory values (needed and available)
1879 - Increased lock attempt timeouts so that when executing long operations
1880 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1881 too early and thus prevent the use of the 'fair' mechanism
1882 - Changed instance query data (``gnt-instance info``) to not acquire
1883 locks unless needed, thus allowing its use on locked instance if only
1884 static information is asked for
1885 - Improved behaviour with filesystems that do not support rename on an
1887 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1888 kept locks on all nodes during the wipe, which is unneeded
1889 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1890 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1891 10GiB) which caused the wipe to fail right at the end in some cases
1892 - Fixed master IP activation when doing master failover with no-voting
1893 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1894 the master node itself
1895 - Fixed potential data-loss in under disk full conditions, where Ganeti
1896 wouldn't check correctly the return code and would consider
1897 partially-written files 'correct'
1898 - Fixed bug related to multiple VGs and DRBD disk replacing
1899 - Added new disk parameter ``metavg`` that allows placement of the meta
1900 device for DRBD in a different volume group
1901 - Fixed error handling in the node daemon when the system libc doesn't
1902 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1903 - Fixed lock release during replace-disks, which kept cluster-wide locks
1904 when doing disk replaces with an iallocator script
1905 - Added check for missing bridges in cluster verify
1906 - Handle EPIPE errors while writing to the terminal better, so that
1907 piping the output to e.g. ``less`` doesn't cause a backtrace
1908 - Fixed rare case where a ^C during Luxi calls could have been
1909 interpreted as server errors, instead of simply terminating
1910 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1912 - Added a few more parameters to the KVM hypervisor, allowing a second
1913 CDROM, custom disk type for CDROMs and a floppy image
1914 - Removed redundant message in instance rename when the name is given
1916 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1917 disks on new nodes, allowing recreation when the original instance
1918 nodes are completely gone
1919 - Added option when converting disk templates to DRBD to skip waiting
1920 for the resync, in order to make the instance available sooner
1921 - Added two new variables to the OS scripts environment (containing the
1923 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1924 to allow use of ``pvgrub`` as bootloader
1925 - Changed the instance memory modifications to only check out-of-memory
1926 conditions on memory increases, and turned the secondary node warnings
1927 into errors (they can still be overridden via ``--force``)
1928 - Fixed the handling of a corner case when the Python installation gets
1929 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1930 to execute a command that doesn't exist
1931 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1932 node of the instance returned failures during instance shutdown; this
1933 adds the option ``--ignore-consistency`` to gnt-instance move
1935 And as usual, various improvements to the error messages, documentation
1942 *(Released Wed, 09 Mar 2011)*
1944 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1945 the RAPI users file if run twice (even with ``--dry-run``).
1947 The release fixes that bug (nothing else changed).
1953 *(Released Mon, 07 Mar 2011)*
1955 Final 2.4.0 release. Just a few small fixes:
1957 - Fixed RAPI node evacuate
1958 - Fixed the kvm-ifup script
1959 - Fixed internal error handling for special job cases
1960 - Updated man page to specify the escaping feature for options
1966 *(Released Mon, 28 Feb 2011)*
1968 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1969 that this feature wiped the disks of the wrong instance, leading to loss
1974 - Fixed title of query field containing instance name
1975 - Expanded the glossary in the documentation
1976 - Fixed one unittest (internal issue)
1982 *(Released Mon, 21 Feb 2011)*
1984 A number of bug fixes plus just a couple functionality changes.
1986 On the user-visible side, the ``gnt-* list`` command output has changed
1987 with respect to "special" field states. The current rc1 style of display
1988 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1989 the default output mode special fields states are displayed as follows:
1991 - Offline resource: ``*``
1992 - Unavailable/not applicable: ``-``
1993 - Data missing (RPC failure): ``?``
1994 - Unknown field: ``??``
1996 Another user-visible change is the addition of ``--force-join`` to
2001 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
2002 - Fixed regression in RAPI/instance reinstall where all parameters were
2003 required (instead of optional)
2004 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
2005 - Fixed iallocator usage (offline nodes were not considered offline)
2006 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
2007 - Fixed hypervisor and OS parameter validation with respect to
2008 non-vm_capable nodes
2009 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
2011 - Fixed ``tools/listrunner`` with respect to agent-based usage
2017 *(Released Fri, 4 Feb 2011)*
2019 Many changes and fixes since the beta1 release. While there were some
2020 internal changes, the code has been mostly stabilised for the RC
2023 Note: the dumb allocator was removed in this release, as it was not kept
2024 up-to-date with the IAllocator protocol changes. It is recommended to
2025 use the ``hail`` command from the ganeti-htools package.
2027 Note: the 2.4 and up versions of Ganeti are not compatible with the
2028 0.2.x branch of ganeti-htools. You need to upgrade to
2029 ganeti-htools-0.3.0 (or later).
2031 Regressions fixed from 2.3
2032 ~~~~~~~~~~~~~~~~~~~~~~~~~~
2034 - Fixed the ``gnt-cluster verify-disks`` command
2035 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
2037 - Fixed disk adoption breakage
2038 - Fixed wrong headers in instance listing for field aliases
2043 - Fixed corner case in KVM handling of NICs
2044 - Fixed many cases of wrong handling of non-vm_capable nodes
2045 - Fixed a bug where a missing instance symlink was not possible to
2046 recreate with any ``gnt-*`` command (now ``gnt-instance
2047 activate-disks`` does it)
2048 - Fixed the volume group name as reported by ``gnt-cluster
2050 - Increased timeouts for the import-export code, hopefully leading to
2051 fewer aborts due network or instance timeouts
2052 - Fixed bug in ``gnt-node list-storage``
2053 - Fixed bug where not all daemons were started on cluster
2054 initialisation, but only at the first watcher run
2055 - Fixed many bugs in the OOB implementation
2056 - Fixed watcher behaviour in presence of instances with offline
2058 - Fixed instance list output for instances running on the wrong node
2059 - a few fixes to the cluster-merge tool, but it still cannot merge
2060 multi-node groups (currently it is not recommended to use this tool)
2066 - Improved network configuration for the KVM hypervisor
2067 - Added e1000 as a supported NIC for Xen-HVM
2068 - Improved the lvmstrap tool to also be able to use partitions, as
2069 opposed to full disks
2070 - Improved speed of disk wiping (the cluster parameter
2071 ``prealloc_wipe_disks``, so that it has a low impact on the total time
2072 of instance creations
2073 - Added documentation for the OS parameters
2074 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
2075 hypervisor is not responding
2076 - Added display of blacklisted and hidden OS information in
2077 ``gnt-cluster info``
2078 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
2079 NIC and node parameters, which might create problems with currently
2080 invalid (but undetected) configuration files, but prevents validation
2081 failures when unrelated parameters are modified
2082 - Changed cluster initialisation to wait for the master daemon to become
2084 - Expanded the RAPI interface:
2086 - Added config redistribution resource
2087 - Added activation/deactivation of instance disks
2088 - Added export of console information
2090 - Implemented log file reopening on SIGHUP, which allows using
2091 logrotate(8) for the Ganeti log files
2092 - Added a basic OOB helper script as an example
2098 *(Released Fri, 14 Jan 2011)*
2103 - Fixed timezone issues when formatting timestamps
2104 - Added support for node groups, available via ``gnt-group`` and other
2106 - Added out-of-band framework and management, see :doc:`design
2107 document <design-oob>`
2108 - Removed support for roman numbers from ``gnt-node list`` and
2109 ``gnt-instance list``.
2110 - Allowed modification of master network interface via ``gnt-cluster
2111 modify --master-netdev``
2112 - Accept offline secondaries while shutting down instance disks
2113 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
2114 - Added support for multiple LVM volume groups
2115 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
2117 - Added commands to list available fields:
2119 - ``gnt-node list-fields``
2120 - ``gnt-group list-fields``
2121 - ``gnt-instance list-fields``
2123 - Updated documentation and man pages
2128 - Moved ``rapi_users`` file into separate directory, now named
2129 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
2131 - Added new tool for running commands on many machines,
2132 ``tools/ganeti-listrunner``
2133 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
2134 improving the ``gnt-instance console`` output
2135 - Allowed customisation of disk index separator at ``configure`` time
2136 - Export node group allocation policy to :doc:`iallocator <iallocator>`
2137 - Added support for non-partitioned md disks in ``lvmstrap``
2138 - Added script to gracefully power off KVM instances
2139 - Split ``utils`` module into smaller parts
2140 - Changed query operations to return more detailed information, e.g.
2141 whether an information is unavailable due to an offline node. To use
2142 this new functionality, the LUXI call ``Query`` must be used. Field
2143 information is now stored by the master daemon and can be retrieved
2144 using ``QueryFields``. Instances, nodes and groups can also be queried
2145 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
2146 exposed via RAPI). The following commands make use of this
2147 infrastructure change:
2149 - ``gnt-group list``
2150 - ``gnt-group list-fields``
2152 - ``gnt-node list-fields``
2153 - ``gnt-instance list``
2154 - ``gnt-instance list-fields``
2155 - ``gnt-debug locks``
2160 - New RAPI resources (see :doc:`rapi`):
2164 - ``/2/groups/[group_name]``
2165 - ``/2/groups/[group_name]/assign-nodes``
2166 - ``/2/groups/[group_name]/modify``
2167 - ``/2/groups/[group_name]/rename``
2168 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
2172 - Implemented ``no_install`` for instance creation
2173 - Implemented OS parameters for instance reinstallation, allowing
2174 use of special settings on reinstallation (e.g. for preserving data)
2179 - Added IPv6 support in import/export
2180 - Pause DRBD synchronization while wiping disks on instance creation
2181 - Updated unittests and QA scripts
2182 - Improved network parameters passed to KVM
2183 - Converted man pages from docbook to reStructuredText
2189 *(Released Mon, 20 Dec 2010)*
2191 Released version 2.3.1~rc1 without any changes.
2197 *(Released Wed, 1 Dec 2010)*
2199 - impexpd: Disable OpenSSL compression in socat if possible (backport
2200 from master, commit e90739d625b, see :doc:`installation guide
2201 <install-quick>` for details)
2202 - Changed unittest coverage report to exclude test scripts
2203 - Added script to check version format
2209 *(Released Wed, 1 Dec 2010)*
2211 Released version 2.3.0~rc1 without any changes.
2217 *(Released Fri, 19 Nov 2010)*
2219 A number of bugfixes and documentation updates:
2221 - Update ganeti-os-interface documentation
2222 - Fixed a bug related to duplicate MACs or similar items which should be
2224 - Fix breakage in OS state modify
2225 - Reinstall instance: disallow offline secondaries (fixes bug related to
2226 OS changing but reinstall failing)
2227 - plus all the other fixes between 2.2.1 and 2.2.2
2233 *(Released Tue, 2 Nov 2010)*
2235 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
2236 - Fixed master failover race with watcher
2237 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
2239 - Accept previously stopped instance for export with instance removal
2240 - Simplify and extend the environment variables for instance OS scripts
2241 - Added new node flags, ``master_capable`` and ``vm_capable``
2242 - Added optional instance disk wiping prior during allocation. This is a
2243 cluster-wide option and can be set/modified using
2244 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
2245 - Added IPv6 support, see :doc:`design document <design-2.3>` and
2246 :doc:`install-quick`
2247 - Added a new watcher option (``--ignore-pause``)
2248 - Added option to ignore offline node on instance start/stop
2249 (``--ignore-offline``)
2250 - Allow overriding OS parameters with ``gnt-instance reinstall``
2251 - Added ability to change node's secondary IP address using ``gnt-node
2253 - Implemented privilege separation for all daemons except
2254 ``ganeti-noded``, see ``configure`` options
2255 - Complain if an instance's disk is marked faulty in ``gnt-cluster
2257 - Implemented job priorities (see ``ganeti(7)`` manpage)
2258 - Ignore failures while shutting down instances during failover from
2260 - Exit daemon's bootstrap process only once daemon is ready
2261 - Export more information via ``LUInstanceQuery``/remote API
2262 - Improved documentation, QA and unittests
2263 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
2264 restart if the file was created or changed
2265 - Added LUXI protocol version sent with each request and response,
2266 allowing detection of server/client mismatches
2267 - Moved the Python scripts among gnt-* and ganeti-* into modules
2268 - Moved all code related to setting up SSH to an external script,
2270 - Infrastructure changes for node group support in future versions
2276 *(Released Fri, 19 Nov 2010)*
2278 A few small bugs fixed, and some improvements to the build system:
2280 - Fix documentation regarding conversion to drbd
2281 - Fix validation of parameters in cluster modify (``gnt-cluster modify
2283 - Fix error handling in node modify with multiple changes
2284 - Allow remote imports without checked names
2290 *(Released Tue, 19 Oct 2010)*
2292 - Disable SSL session ID cache in RPC client
2298 *(Released Thu, 14 Oct 2010)*
2300 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
2301 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
2303 - Fix problems with interaction between Curl and Python's HTTP server,
2304 resulting in increased speed in many RPC calls
2305 - Improve our release script to prevent breakage with older aclocal and
2312 *(Released Thu, 7 Oct 2010)*
2314 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
2315 value retrieved from master
2316 - Added support for blacklisted or hidden OS definitions
2317 - Added simple lock monitor (accessible via (``gnt-debug locks``)
2318 - Added support for -mem-path in KVM hypervisor abstraction layer
2319 - Allow overriding instance parameters in tool for inter-cluster
2320 instance moves (``tools/move-instance``)
2321 - Improved opcode summaries (e.g. in ``gnt-job list``)
2322 - Improve consistency of OS listing by sorting it
2323 - Documentation updates
2329 *(Released Fri, 8 Oct 2010)*
2331 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
2337 *(Released Mon, 4 Oct 2010)*
2339 - Fixed regression in ``gnt-instance rename``
2345 *(Released Wed, 22 Sep 2010)*
2347 - Fixed OS_VARIANT variable for OS scripts
2348 - Fixed cluster tag operations via RAPI
2349 - Made ``setup-ssh`` exit with non-zero code if an error occurred
2350 - Disabled RAPI CA checks in watcher
2356 *(Released Mon, 23 Aug 2010)*
2358 - Support DRBD versions of the format "a.b.c.d"
2360 - Re-introduce support for usage from multiple threads in RAPI client
2361 - Instance renames and modify via RAPI
2362 - Work around race condition between processing and archival in job
2364 - Mark opcodes following failed one as failed, too
2365 - Job field ``lock_status`` was removed due to difficulties making it
2366 work with the changed job queue in Ganeti 2.2; a better way to monitor
2367 locks is expected for a later 2.2.x release
2368 - Fixed dry-run behaviour with many commands
2369 - Support ``ssh-agent`` again when adding nodes
2370 - Many additional bugfixes
2376 *(Released Fri, 30 Jul 2010)*
2378 Important change: the internal RPC mechanism between Ganeti nodes has
2379 changed from using a home-grown http library (based on the Python base
2380 libraries) to use the PycURL library. This requires that PycURL is
2381 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2382 against GnuTLS by default. cURL's support for GnuTLS had known issues
2383 before cURL 7.21.0 and we recommend using the latest cURL release or
2384 linking against OpenSSL. Most other distributions already link PycURL
2385 and cURL against OpenSSL. The command::
2387 python -c 'import pycurl; print pycurl.version'
2389 can be used to determine the libraries PycURL and cURL are linked
2392 Other significant changes:
2394 - Rewrote much of the internals of the job queue, in order to achieve
2395 better parallelism; this decouples job query operations from the job
2396 processing, and it should allow much nicer behaviour of the master
2397 daemon under load, and it also has uncovered some long-standing bugs
2398 related to the job serialisation (now fixed)
2399 - Added a default iallocator setting to the cluster parameters,
2400 eliminating the need to always pass nodes or an iallocator for
2401 operations that require selection of new node(s)
2402 - Added experimental support for the LXC virtualization method
2403 - Added support for OS parameters, which allows the installation of
2404 instances to pass parameter to OS scripts in order to customise the
2406 - Added a hypervisor parameter controlling the migration type (live or
2407 non-live), since hypervisors have various levels of reliability; this
2408 has renamed the 'live' parameter to 'mode'
2409 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
2410 logical volumes, meaning that cluster verify will ignore them and not
2411 flag their presence as errors
2412 - The watcher will now reset the error count for failed instances after
2413 8 hours, thus allowing self-healing if the problem that caused the
2414 instances to be down/fail to start has cleared in the meantime
2415 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2416 check for, and warn, if the drbd module parameter ``usermode_helper``
2417 is not consistent with the cluster-wide setting; this is needed to
2418 make diagnose easier of failed drbd creations
2419 - Started adding base IPv6 support, but this is not yet
2420 enabled/available for use
2421 - Rename operations (cluster, instance) will now return the new name,
2422 which is especially useful if a short name was passed in
2423 - Added support for instance migration in RAPI
2424 - Added a tool to pre-configure nodes for the SSH setup, before joining
2425 them to the cluster; this will allow in the future a simplified model
2426 for node joining (but not yet fully enabled in 2.2); this needs the
2427 paramiko python library
2428 - Fixed handling of name-resolving errors
2429 - Fixed consistency of job results on the error path
2430 - Fixed master-failover race condition when executed multiple times in
2432 - Fixed many bugs related to the job queue (mostly introduced during the
2433 2.2 development cycle, so not all are impacting 2.1)
2434 - Fixed instance migration with missing disk symlinks
2435 - Fixed handling of unknown jobs in ``gnt-job archive``
2436 - And many other small fixes/improvements
2440 - Enhanced both the unittest and the QA coverage
2441 - Switched the opcode validation to a generic model, and extended the
2442 validation to all opcode parameters
2443 - Changed more parts of the code that write shell scripts to use the
2445 - Switched the master daemon to use the asyncore library for the Luxi
2452 *(Released Thu, 17 Jun 2010)*
2454 - Added tool (``move-instance``) and infrastructure to move instances
2455 between separate clusters (see :doc:`separate documentation
2456 <move-instance>` and :doc:`design document <design-2.2>`)
2457 - Added per-request RPC timeout
2458 - RAPI now requires a Content-Type header for requests with a body (e.g.
2459 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2460 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2461 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2463 - Implemented initial support for running Ganeti daemons as separate
2464 users, see configure-time flags ``--with-user-prefix`` and
2465 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2467 - Instances can be removed after export (``gnt-backup export
2468 --remove-instance``)
2469 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2470 key (instead of 1024 bit)
2471 - Added new cluster configuration file for cluster domain secret
2472 - Import/export now use SSL instead of SSH
2473 - Added support for showing estimated time when exporting an instance,
2474 see the ``ganeti-os-interface(7)`` manpage and look for
2481 *(Released Tue, 16 Nov 2010)*
2483 Some more bugfixes. Unless critical bugs occur, this will be the last
2486 - Fix case of MAC special-values
2487 - Fix mac checker regex
2488 - backend: Fix typo causing "out of range" error
2489 - Add missing --units in gnt-instance list man page
2495 *(Released Tue, 24 Aug 2010)*
2498 - Don't ignore secondary node silently on non-mirrored disk templates
2500 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2501 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2502 - Properly document the usb_mouse parameter
2503 - Fix path in ganeti-rapi(8) (issue 116)
2504 - Adjust error message when the ganeti user's .ssh directory is
2506 - Add same-node-check when changing the disk template to drbd
2512 *(Released Fri, 16 Jul 2010)*
2515 - Add an option to only select some reboot types during qa/burnin.
2516 (on some hypervisors consequent reboots are not supported)
2517 - Fix infrequent race condition in master failover. Sometimes the old
2518 master ip address would be still detected as up for a short time
2519 after it was removed, causing failover to fail.
2520 - Decrease mlockall warnings when the ctypes module is missing. On
2521 Python 2.4 we support running even if no ctypes module is installed,
2522 but we were too verbose about this issue.
2523 - Fix building on old distributions, on which man doesn't have a
2525 - Fix RAPI not to ignore the MAC address on instance creation
2526 - Implement the old instance creation format in the RAPI client.
2532 *(Released Thu, 01 Jul 2010)*
2534 A small bugfix release:
2535 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2536 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2537 a non-existing option
2538 - Fix instance create: the --force-variant option was ignored
2539 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2540 - Fix modify node storage with non-FQDN arguments
2541 - Fix RAPI client to authenticate under Python 2.6 when used
2542 for more than 5 requests needing authentication
2543 - Fix gnt-instance modify -t (storage) giving a wrong error message
2544 when converting a non-shutdown drbd instance to plain
2550 *(Released Fri, 18 Jun 2010)*
2552 A small bugfix release:
2554 - Fix live migration of KVM instances started with older Ganeti
2555 versions which had fewer hypervisor parameters
2556 - Fix gnt-instance grow-disk on down instances
2557 - Fix an error-reporting bug during instance migration
2558 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2559 silently ignoring broken ones
2560 - Fix an RPC error reporting bug affecting, for example, RAPI client
2562 - Fix bug triggered by different API version os-es on different nodes
2563 - Fix a bug in instance startup with custom hvparams: OS level
2564 parameters would fail to be applied.
2565 - Fix the RAPI client under Python 2.6 (but more work is needed to
2566 make it work completely well with OpenSSL)
2567 - Fix handling of errors when resolving names from DNS
2573 *(Released Thu, 3 Jun 2010)*
2575 A medium sized development cycle. Some new features, and some
2576 fixes/small improvements/cleanups.
2578 Significant features
2579 ~~~~~~~~~~~~~~~~~~~~
2581 The node deamon now tries to mlock itself into memory, unless the
2582 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2583 its logs, and falls back to console logging. This allows emergency
2584 features such as ``gnt-node powercycle`` to work even in the event of a
2585 broken node disk (tested offlining the disk hosting the node's
2586 filesystem and dropping its memory caches; don't try this at home)
2588 KVM: add vhost-net acceleration support. It can be tested with a new
2589 enough version of the kernel and of qemu-kvm.
2591 KVM: Add instance chrooting feature. If you use privilege dropping for
2592 your VMs you can also now force them to chroot to an empty directory,
2593 before starting the emulated guest.
2595 KVM: Add maximum migration bandwith and maximum downtime tweaking
2596 support (requires a new-enough version of qemu-kvm).
2598 Cluster verify will now warn if the master node doesn't have the master
2599 ip configured on it.
2601 Add a new (incompatible) instance creation request format to RAPI which
2602 supports all parameters (previously only a subset was supported, and it
2603 wasn't possible to extend the old format to accomodate all the new
2604 features. The old format is still supported, and a client can check for
2605 this feature, before using it, by checking for its presence in the
2606 ``features`` RAPI resource.
2608 Now with ancient latin support. Try it passing the ``--roman`` option to
2609 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2610 (requires the python-roman module to be installed, in order to work).
2615 As usual many internal code refactorings, documentation updates, and
2618 - Lots of improvements and cleanups to the experimental Remote API
2619 (RAPI) client library.
2620 - A new unit test suite for the core daemon libraries.
2621 - A fix to creating missing directories makes sure the umask is not
2622 applied anymore. This enforces the same directory permissions
2624 - Better handling terminating daemons with ctrl+c (used when running
2625 them in debugging mode).
2626 - Fix a race condition in live migrating a KVM instance, when stat()
2627 on the old proc status file returned EINVAL, which is an unexpected
2629 - Fixed manpage checking with newer man and utf-8 charachters. But now
2630 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2636 *(Released Fri, 7 May 2010)*
2638 Fix a bug which prevented untagged KVM instances from starting.
2644 *(Released Fri, 7 May 2010)*
2646 Another release with a long development cycle, during which many
2647 different features were added.
2649 Significant features
2650 ~~~~~~~~~~~~~~~~~~~~
2652 The KVM hypervisor now can run the individual instances as non-root, to
2653 reduce the impact of a VM being hijacked due to bugs in the
2654 hypervisor. It is possible to run all instances as a single (non-root)
2655 user, to manually specify a user for each instance, or to dynamically
2656 allocate a user out of a cluster-wide pool to each instance, with the
2657 guarantee that no two instances will run under the same user ID on any
2660 An experimental RAPI client library, that can be used standalone
2661 (without the other Ganeti libraries), is provided in the source tree as
2662 ``lib/rapi/client.py``. Note this client might change its interface in
2663 the future, as we iterate on its capabilities.
2665 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2666 replace the cluster's certificates and crypto keys. This might help in
2667 case they have been compromised, or have simply expired.
2669 A new disk option for instance creation has been added that allows one
2670 to "adopt" currently existing logical volumes, with data
2671 preservation. This should allow easier migration to Ganeti from
2672 unmanaged (or managed via other software) instances.
2674 Another disk improvement is the possibility to convert between redundant
2675 (DRBD) and plain (LVM) disk configuration for an instance. This should
2676 allow better scalability (starting with one node and growing the
2677 cluster, or shrinking a two-node cluster to one node).
2679 A new feature that could help with automated node failovers has been
2680 implemented: if a node sees itself as offline (by querying the master
2681 candidates), it will try to shutdown (hard) all instances and any active
2682 DRBD devices. This reduces the risk of duplicate instances if an
2683 external script automatically failovers the instances on such nodes. To
2684 enable this, the cluster parameter ``maintain_node_health`` should be
2685 enabled; in the future this option (per the name) will enable other
2686 automatic maintenance features.
2688 Instance export/import now will reuse the original instance
2689 specifications for all parameters; that means exporting an instance,
2690 deleting it and the importing it back should give an almost identical
2691 instance. Note that the default import behaviour has changed from
2692 before, where it created only one NIC; now it recreates the original
2695 Cluster verify has added a few new checks: SSL certificates validity,
2696 /etc/hosts consistency across the cluster, etc.
2701 As usual, many internal changes were done, documentation fixes,
2704 - Fixed cluster initialization with disabled cluster storage (regression
2705 introduced in 2.1.1)
2706 - File-based storage supports growing the disks
2707 - Fixed behaviour of node role changes
2708 - Fixed cluster verify for some corner cases, plus a general rewrite of
2709 cluster verify to allow future extension with more checks
2710 - Fixed log spamming by watcher and node daemon (regression introduced
2712 - Fixed possible validation issues when changing the list of enabled
2714 - Fixed cleanup of /etc/hosts during node removal
2715 - Fixed RAPI response for invalid methods
2716 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2717 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2718 from ide disks, etc.)
2719 - Allow OS changes without re-installation (to record a changed OS
2720 outside of Ganeti, or to allow OS renames)
2721 - Allow instance creation without OS installation (useful for example if
2722 the OS will be installed manually, or restored from a backup not in
2724 - Implemented option to make cluster ``copyfile`` use the replication
2726 - Added list of enabled hypervisors to ssconf (possibly useful for
2728 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2730 - A partial form of node re-IP is possible via node readd, which now
2731 allows changed node primary IP
2732 - Command line utilities now show an informational message if the job is
2734 - The logs of the master daemon now show the PID/UID/GID of the
2741 *(Released Fri, 12 Mar 2010)*
2743 During the 2.1.0 long release candidate cycle, a lot of improvements and
2744 changes have accumulated with were released later as 2.1.1.
2749 The node evacuate command (``gnt-node evacuate``) was significantly
2750 rewritten, and as such the IAllocator protocol was changed - a new
2751 request type has been added. This unfortunate change during a stable
2752 series is designed to improve performance of node evacuations; on
2753 clusters with more than about five nodes and which are well-balanced,
2754 evacuation should proceed in parallel for all instances of the node
2755 being evacuated. As such, any existing IAllocator scripts need to be
2756 updated, otherwise the above command will fail due to the unknown
2757 request. The provided "dumb" allocator has not been updated; but the
2758 ganeti-htools package supports the new protocol since version 0.2.4.
2760 Another important change is increased validation of node and instance
2761 names. This might create problems in special cases, if invalid host
2762 names are being used.
2764 Also, a new layer of hypervisor parameters has been added, that sits at
2765 OS level between the cluster defaults and the instance ones. This allows
2766 customisation of virtualization parameters depending on the installed
2767 OS. For example instances with OS 'X' may have a different KVM kernel
2768 (or any other parameter) than the cluster defaults. This is intended to
2769 help managing a multiple OSes on the same cluster, without manual
2770 modification of each instance's parameters.
2772 A tool for merging clusters, ``cluster-merge``, has been added in the
2773 tools sub-directory.
2778 - Improved the int/float conversions that should make the code more
2779 robust in face of errors from the node daemons
2780 - Fixed the remove node code in case of internal configuration errors
2781 - Fixed the node daemon behaviour in face of inconsistent queue
2782 directory (e.g. read-only file-system where we can't open the files
2784 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2785 now it either aborts cleanly or, if given the new "auto_promote"
2786 parameter, will automatically promote other nodes as needed
2787 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2788 completely prevent Ganeti from working
2789 - Fixed bug for instance export when not all disks were successfully
2791 - Fixed behaviour of node add when the new node is slow in starting up
2793 - Fixed handling of signals in the LUXI client, which should improve
2794 behaviour of command-line scripts
2795 - Added checks for invalid node/instance names in the configuration (now
2796 flagged during cluster verify)
2797 - Fixed watcher behaviour for disk activation errors
2798 - Fixed two potentially endless loops in http library, which led to the
2799 RAPI daemon hanging and consuming 100% CPU in some cases
2800 - Fixed bug in RAPI daemon related to hashed passwords
2801 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2803 - Enhanced compatibility with non-Debian OSes, but not using absolute
2804 path in some commands and allowing customisation of the ssh
2805 configuration directory
2806 - Fixed possible future issue with new Python versions by abiding to the
2807 proper use of ``__slots__`` attribute on classes
2808 - Added checks that should prevent directory traversal attacks
2809 - Many documentation fixes based on feedback from users
2814 - Added an "early_release" more for instance replace disks and node
2815 evacuate, where we release locks earlier and thus allow higher
2816 parallelism within the cluster
2817 - Added watcher hooks, intended to allow the watcher to restart other
2818 daemons (e.g. from the ganeti-nbma project), but they can be used of
2819 course for any other purpose
2820 - Added a compile-time disable for DRBD barriers, to increase
2821 performance if the administrator trusts the power supply or the
2822 storage system to not lose writes
2823 - Added the option of using syslog for logging instead of, or in
2824 addition to, Ganeti's own log files
2825 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2826 can indeed boot from a paravirtual NIC
2827 - Added a generic debug level for many operations; while this is not
2828 used widely yet, it allows one to pass the debug value all the way to
2830 - Enhanced the hooks environment for instance moves (failovers,
2831 migrations) where the primary/secondary nodes changed during the
2832 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2833 - Enhanced data validations for many user-supplied values; one important
2834 item is the restrictions imposed on instance and node names, which
2835 might reject some (invalid) host names
2836 - Add a configure-time option to disable file-based storage, if it's not
2837 needed; this allows greater security separation between the master
2838 node and the other nodes from the point of view of the inter-node RPC
2840 - Added user notification in interactive tools if job is waiting in the
2841 job queue or trying to acquire locks
2842 - Added log messages when a job is waiting for locks
2843 - Added filtering by node tags in instance operations which admit
2844 multiple instances (start, stop, reboot, reinstall)
2845 - Added a new tool for cluster mergers, ``cluster-merge``
2846 - Parameters from command line which are of the form ``a=b,c=d`` can now
2847 use backslash escapes to pass in values which contain commas,
2848 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2850 - For KVM, the instance name is the first parameter passed to KVM, so
2851 that it's more visible in the process list
2857 *(Released Tue, 2 Mar 2010)*
2859 Ganeti 2.1 brings many improvements with it. Major changes:
2861 - Added infrastructure to ease automated disk repairs
2862 - Added new daemon to export configuration data in a cheaper way than
2863 using the remote API
2864 - Instance NICs can now be routed instead of being associated with a
2866 - Improved job locking logic to reduce impact of jobs acquiring multiple
2867 locks waiting for other long-running jobs
2869 In-depth implementation details can be found in the Ganeti 2.1 design
2875 - Added chroot hypervisor
2876 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2878 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2879 ``bootloader_path`` and ``bootloader_args``)
2880 - Added the ``use_localtime`` option for the xen-hvm and kvm
2881 hypervisors, and the default value for this has changed to false (in
2882 2.0 xen-hvm always enabled it)
2883 - Added luxi call to submit multiple jobs in one go
2884 - Added cluster initialization option to not modify ``/etc/hosts``
2886 - Added network interface parameters
2887 - Added dry run mode to some LUs
2888 - Added RAPI resources:
2890 - ``/2/instances/[instance_name]/info``
2891 - ``/2/instances/[instance_name]/replace-disks``
2892 - ``/2/nodes/[node_name]/evacuate``
2893 - ``/2/nodes/[node_name]/migrate``
2894 - ``/2/nodes/[node_name]/role``
2895 - ``/2/nodes/[node_name]/storage``
2896 - ``/2/nodes/[node_name]/storage/modify``
2897 - ``/2/nodes/[node_name]/storage/repair``
2899 - Added OpCodes to evacuate or migrate all instances on a node
2900 - Added new command to list storage elements on nodes (``gnt-node
2901 list-storage``) and modify them (``gnt-node modify-storage``)
2902 - Added new ssconf files with master candidate IP address
2903 (``ssconf_master_candidates_ips``), node primary IP address
2904 (``ssconf_node_primary_ips``) and node secondary IP address
2905 (``ssconf_node_secondary_ips``)
2906 - Added ``ganeti-confd`` and a client library to query the Ganeti
2907 configuration via UDP
2908 - Added ability to run hooks after cluster initialization and before
2910 - Added automatic mode for disk replace (``gnt-instance replace-disks
2912 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2913 after catastrophic data-loss
2914 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2916 - Added ``gnt-instance move`` command to move instances
2917 - Added ``gnt-cluster watcher`` command to control watcher
2918 - Added ``gnt-node powercycle`` command to powercycle nodes
2919 - Added new job status field ``lock_status``
2920 - Added parseable error codes to cluster verification (``gnt-cluster
2921 verify --error-codes``) and made output less verbose (use
2922 ``--verbose`` to restore previous behaviour)
2923 - Added UUIDs to the main config entities (cluster, nodes, instances)
2924 - Added support for OS variants
2925 - Added support for hashed passwords in the Ganeti remote API users file
2927 - Added option to specify maximum timeout on instance shutdown
2928 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2929 - Added new helper script to start and stop Ganeti daemons
2930 (``daemon-util``), with the intent to reduce the work necessary to
2931 adjust Ganeti for non-Debian distributions and to start/stop daemons
2933 - Added more unittests
2934 - Fixed critical bug in ganeti-masterd startup
2935 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2936 now customisable at the cluster level for both the KVM and Xen
2937 hypervisors using the new ``migration_port`` parameter
2938 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2939 reinstalling an instance
2940 - Allowed ``@`` in tag names
2941 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2942 - Many documentation updates
2943 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2944 - ``gnt-instance reinstall`` can now reinstall multiple instances
2945 - Updated many command line parameters
2946 - Introduced new OS API version 15
2947 - No longer support a default hypervisor
2948 - Treat virtual LVs as inexistent
2949 - Improved job locking logic to reduce lock contention
2950 - Match instance and node names case insensitively
2951 - Reimplemented bash completion script to be more complete
2958 *(Released Thu, 4 Feb 2010)*
2960 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2961 - Fix a string formatting bug
2962 - Improve safety of the code in some error paths
2963 - Improve data validation in the master of values returned from nodes
2969 *(Released Thu, 17 Dec 2009)*
2971 - Fix security issue due to missing validation of iallocator names; this
2972 allows local and remote execution of arbitrary executables
2973 - Fix failure of gnt-node list during instance removal
2974 - Ship the RAPI documentation in the archive
2980 *(Released Wed, 30 Sep 2009)*
2982 - Fixed many wrong messages
2983 - Fixed a few bugs related to the locking library
2984 - Fixed MAC checking at instance creation time
2985 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2986 - Fixed a few issues related to signal handling in both daemons and
2988 - Fixed the example startup script provided
2989 - Fixed insserv dependencies in the example startup script (patch from
2991 - Fixed handling of drained nodes in the iallocator framework
2992 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2994 - Fixed error related to invalid job IDs in job polling
2995 - Fixed job/opcode persistence on unclean master shutdown
2996 - Fixed handling of partial job processing after unclean master
2998 - Fixed error reporting from LUs, previously all errors were converted
2999 into execution errors
3000 - Fixed error reporting from burnin
3001 - Decreased significantly the memory usage of the job queue
3002 - Optimised slightly multi-job submission
3003 - Optimised slightly opcode loading
3004 - Backported the multi-job submit framework from the development
3005 branch; multi-instance start and stop should be faster
3006 - Added script to clean archived jobs after 21 days; this will reduce
3007 the size of the queue directory
3008 - Added some extra checks in disk size tracking
3009 - Added an example ethers hook script
3010 - Added a cluster parameter that prevents Ganeti from modifying of
3012 - Added more node information to RAPI responses
3013 - Added a ``gnt-job watch`` command that allows following the ouput of a
3015 - Added a bind-address option to ganeti-rapi
3016 - Added more checks to the configuration verify
3017 - Enhanced the burnin script such that some operations can be retried
3019 - Converted instance reinstall to multi-instance model
3025 *(Released Fri, 7 Aug 2009)*
3027 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
3028 to allow using the pre-2.0.2 behaviour in activation, if any existing
3029 instances have mismatched disk sizes in the configuration
3030 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
3031 any configuration mismatches for disk sizes
3032 - Added ``gnt-master cluste-failover --no-voting`` to allow master
3033 failover to work on two-node clusters
3034 - Fixed the ``--net`` option of ``gnt-backup import``, which was
3036 - Fixed detection of OS script errors in ``gnt-backup export``
3037 - Fixed exit code of ``gnt-backup export``
3043 *(Released Fri, 17 Jul 2009)*
3045 - Added experimental support for stripped logical volumes; this should
3046 enhance performance but comes with a higher complexity in the block
3047 device handling; stripping is only enabled when passing
3048 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
3049 affected even in the non-stripped mode
3050 - Improved resiliency against transient failures at the end of DRBD
3051 resyncs, and in general of DRBD resync checks
3052 - Fixed a couple of issues with exports and snapshot errors
3053 - Fixed a couple of issues in instance listing
3054 - Added display of the disk size in ``gnt-instance info``
3055 - Fixed checking for valid OSes in instance creation
3056 - Fixed handling of the "vcpus" parameter in instance listing and in
3057 general of invalid parameters
3058 - Fixed http server library, and thus RAPI, to handle invalid
3059 username/password combinations correctly; this means that now they
3060 report unauthorized for queries too, not only for modifications,
3061 allowing earlier detect of configuration problems
3062 - Added a new "role" node list field, equivalent to the master/master
3063 candidate/drained/offline flags combinations
3064 - Fixed cluster modify and changes of candidate pool size
3065 - Fixed cluster verify error messages for wrong files on regular nodes
3066 - Fixed a couple of issues with node demotion from master candidate role
3067 - Fixed node readd issues
3068 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
3069 - Added a new ``--no-voting`` option for masterfailover to fix failover
3070 on two-nodes clusters when the former master node is unreachable
3071 - Added instance reinstall over RAPI
3077 *(Released Tue, 16 Jun 2009)*
3079 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
3080 allow re-adding the start in single-user option (regression from 1.2)
3081 - the watcher writes the instance status to a file, to allow monitoring
3082 to report the instance status (from the master) based on cached
3083 results of the watcher's queries; while this can get stale if the
3084 watcher is being locked due to other work on the cluster, this is
3085 still an improvement
3086 - the watcher now also restarts the node daemon and the rapi daemon if
3088 - fixed the watcher to handle full and drained queue cases
3089 - hooks export more instance data in the environment, which helps if
3090 hook scripts need to take action based on the instance's properties
3091 (no longer need to query back into ganeti)
3092 - instance failovers when the instance is stopped do not check for free
3093 RAM, so that failing over a stopped instance is possible in low memory
3095 - rapi uses queries for tags instead of jobs (for less job traffic), and
3096 for cluster tags it won't talk to masterd at all but read them from
3098 - a couple of error handling fixes in RAPI
3099 - drbd handling: improved the error handling of inconsistent disks after
3100 resync to reduce the frequency of "there are some degraded disks for
3101 this instance" messages
3102 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
3103 error handling path called a wrong function name)
3109 *(Released Wed, 27 May 2009)*
3111 - no changes from rc5
3117 *(Released Wed, 20 May 2009)*
3119 - fix a couple of bugs (validation, argument checks)
3120 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
3121 - some small improvements to RAPI and IAllocator
3122 - make watcher automatically start the master daemon if down
3128 *(Released Mon, 27 Apr 2009)*
3130 - change the OS list to not require locks; this helps with big clusters
3131 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
3132 volume group is broken
3133 - ``gnt-instance info``, without any arguments, doesn't run for all
3134 instances anymore; either pass ``--all`` or pass the desired
3135 instances; this helps against mistakes on big clusters where listing
3136 the information for all instances takes a long time
3137 - miscellaneous doc and man pages fixes
3143 *(Released Wed, 8 Apr 2009)*
3145 - Change the internal locking model of some ``gnt-node`` commands, in
3146 order to reduce contention (and blocking of master daemon) when
3147 batching many creation/reinstall jobs
3148 - Fixes to Xen soft reboot
3149 - No longer build documentation at build time, instead distribute it in
3150 the archive, in order to reduce the need for the whole docbook/rst
3157 *(Released Fri, 27 Mar 2009)*
3159 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
3160 - Fix watcher startup sequence, improves the behaviour of busy clusters
3161 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
3162 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
3163 bind address and other places
3164 - Some documentation fixes and updates
3170 *(Released Mon, 2 Mar 2009)*
3172 - More documentation updates, now all docs should be more-or-less
3174 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
3176 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
3178 - Fix an issue related to $libdir/run/ganeti and cluster creation
3184 *(Released Thu, 19 Feb 2009)*
3186 - Xen PVM and KVM have switched the default value for the instance root
3187 disk to the first partition on the first drive, instead of the whole
3188 drive; this means that the OS installation scripts must be changed
3190 - Man pages have been updated
3191 - RAPI has been switched by default to HTTPS, and the exported functions
3192 should all work correctly
3193 - RAPI v1 has been removed
3194 - Many improvements to the KVM hypervisor
3195 - Block device errors are now better reported
3196 - Many other bugfixes and small improvements
3202 *(Released Mon, 26 Jan 2009)*
3204 - Version 2 is a general rewrite of the code and therefore the
3205 differences are too many to list, see the design document for 2.0 in
3206 the ``doc/`` subdirectory for more details
3207 - In this beta version there is not yet a migration path from 1.2 (there
3208 will be one in the final 2.0 release)
3209 - A few significant changes are:
3211 - all commands are executed by a daemon (``ganeti-masterd``) and the
3212 various ``gnt-*`` commands are just front-ends to it
3213 - all the commands are entered into, and executed from a job queue,
3214 see the ``gnt-job(8)`` manpage
3215 - the RAPI daemon supports read-write operations, secured by basic
3216 HTTP authentication on top of HTTPS
3217 - DRBD version 0.7 support has been removed, DRBD 8 is the only
3218 supported version (when migrating from Ganeti 1.2 to 2.0, you need
3219 to migrate to DRBD 8 first while still running Ganeti 1.2)
3220 - DRBD devices are using statically allocated minor numbers, which
3221 will be assigned to existing instances during the migration process
3222 - there is support for both Xen PVM and Xen HVM instances running on
3224 - KVM virtualization is supported too
3225 - file-based storage has been implemented, which means that it is
3226 possible to run the cluster without LVM and DRBD storage, for
3227 example using a shared filesystem exported from shared storage (and
3228 still have live migration)
3234 *(Released Tue, 13 Jan 2009)*
3236 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
3237 - Reuse the old instance mac address by default on instance import, if
3238 the instance name is the same.
3239 - Handle situations in which the node info rpc returns incomplete
3241 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
3242 - Improved version of batcher:
3244 - state file support
3245 - instance mac address support
3246 - support for HVM clusters/instances
3248 - Add an option to show the number of cpu sockets and nodes in
3250 - Support OSes that handle more than one version of the OS api (but do
3251 not change the current API in any other way)
3252 - Fix ``gnt-node migrate``
3253 - ``gnt-debug`` man page
3254 - Fixes various more typos and small issues
3255 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
3261 *(Released Wed, 24 Sep 2008)*
3263 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
3264 type of disk exported to fully virtualized instances.
3265 - provide access to the serial console of HVM instances
3266 - instance auto_balance flag, set by default. If turned off it will
3267 avoid warnings on cluster verify if there is not enough memory to fail
3268 over an instance. in the future it will prevent automatically failing
3269 it over when we will support that.
3270 - batcher tool for instance creation, see ``tools/README.batcher``
3271 - ``gnt-instance reinstall --select-os`` to interactively select a new
3272 operating system when reinstalling an instance.
3273 - when changing the memory amount on instance modify a check has been
3274 added that the instance will be able to start. also warnings are
3275 emitted if the instance will not be able to fail over, if auto_balance
3277 - documentation fixes
3278 - sync fields between ``gnt-instance list/modify/add/import``
3279 - fix a race condition in drbd when the sync speed was set after giving
3280 the device a remote peer.
3286 *(Released Tue, 22 Jul 2008)*
3288 - note: the allowed size and number of tags per object were reduced
3289 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
3290 - fixed twisted 8.x compatibility
3291 - fixed ``gnt-instance replace-disks`` with iallocator
3292 - add TCP keepalives on twisted connections to detect restarted nodes
3293 - disk increase support, see ``gnt-instance grow-disk``
3294 - implement bulk node/instance query for RAPI
3295 - add tags in node/instance listing (optional)
3296 - experimental migration (and live migration) support, read the man page
3297 for ``gnt-instance migrate``
3298 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
3299 has some small improvements in handling its state file
3305 *(Released Fri, 13 Jun 2008)*
3307 - Experimental readonly, REST-based remote API implementation;
3308 automatically started on master node, TCP port 5080, if enabled by
3309 ``--enable-rapi`` parameter to configure script.
3310 - Instance allocator support. Add and import instance accept a
3311 ``--iallocator`` parameter, and call that instance allocator to decide
3312 which node to use for the instance. The iallocator document describes
3313 what's expected from an allocator script.
3314 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
3315 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
3316 node is lost there is still enough memory to fail over the instances
3318 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
3319 ``gnt-cluster verify``, to check for site-specific compliance. All the
3320 hooks will run, and their output, if any, will be displayed. Any
3321 failing hook will make the verification return an error value.
3322 - ``gnt-cluster verify`` now checks that its peers are reachable on the
3323 primary and secondary interfaces
3324 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
3325 that is still declared as part of the cluster and has failed.
3326 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
3327 specifying output fields, that just adds the chosen fields to the
3329 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
3330 export from the filesystem.
3331 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
3332 have been added. Using them you can enable/disable acpi and pae
3333 support, and specify a path for a cd image to be exported to the
3334 instance. These parameters as the name suggest only work on HVM
3336 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
3337 PAE support will be set to the previously hardcoded values, but the
3338 (previously hardcoded) path to the CDROM ISO image will be unset and
3339 if required, needs to be set manually with ``gnt-instance modify``
3341 - The address to which an instance's VNC console is bound is now
3342 selectable per-instance, rather than being cluster wide. Of course
3343 this only applies to instances controlled via VNC, so currently just
3344 applies to HVM clusters.
3350 *(Released Mon, 18 Feb 2008)*
3352 - more tweaks to the disk activation code (especially helpful for DRBD)
3353 - change the default ``gnt-instance list`` output format, now there is
3354 one combined status field (see the manpage for the exact values this
3356 - some more fixes for the mac export to hooks change
3357 - make Ganeti not break with DRBD 8.2.x (which changed the version
3358 format in ``/proc/drbd``) (issue 24)
3359 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3360 template, allowing migration from DRBD0.7+MD to DRBD8
3366 *(Released Wed, 30 Jan 2008)*
3368 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3370 - add command aliases infrastructure and a few aliases
3371 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
3372 man pages and the ``--help`` option of ``gnt-node
3373 list``/``gnt-instance list``
3374 - fix ``gnt-backup list`` with down nodes (issue 21)
3375 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3376 - fix the dist archive and add a check for including svn/git files in
3378 - some developer-related changes: improve the burnin and the QA suite,
3379 add an upload script for testing during development
3385 *(Released Wed, 16 Jan 2008)*
3387 - experimental HVM support, read the install document, section
3388 "Initializing the cluster"
3389 - allow for the PVM hypervisor per-instance kernel and initrd paths
3390 - add a new command ``gnt-cluster verify-disks`` which uses a new
3391 algorithm to improve the reconnection of the DRBD pairs if the device
3392 on the secondary node has gone away
3393 - make logical volume code auto-activate LVs at disk activation time
3394 - slightly improve the speed of activating disks
3395 - allow specification of the MAC address at instance creation time, and
3396 changing it later via ``gnt-instance modify``
3397 - fix handling of external commands that generate lots of output on
3399 - update documentation with regard to minimum version of DRBD8 supported
3405 *(Released Tue, 4 Dec 2007)*
3407 - Log the ``xm create`` output to the node daemon log on failure (to
3408 help diagnosing the error)
3409 - In debug mode, log all external commands output if failed to the logs
3410 - Change parsing of lvm commands to ignore stderr
3416 *(Released Wed, 28 Nov 2007)*
3418 - Another round of updates to the DRBD 8 code to deal with more failures
3419 in the replace secondary node operation
3420 - Some more logging of failures in disk operations (lvm, drbd)
3421 - A few documentation updates
3428 *(Released Tue, 13 Nov 2007)*
3430 - Change configuration file format from Python's Pickle to JSON.
3431 Upgrading is possible using the cfgupgrade utility.
3432 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3433 faster replace disks and is more stable (DRBD 8 has many improvements
3434 compared to DRBD 0.7)
3435 - Added command line tags support (see man pages for ``gnt-instance``,
3436 ``gnt-node``, ``gnt-cluster``)
3437 - Added instance rename support
3438 - Added multi-instance startup/shutdown
3439 - Added cluster rename support
3440 - Added ``gnt-node evacuate`` to simplify some node operations
3441 - Added instance reboot operation that can speedup reboot as compared to
3443 - Soften the requirement that hostnames are in FQDN format
3444 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3446 - Removed dependency on debian's patched fping that uses the
3447 non-standard ``-S`` option
3448 - Now the OS definitions are searched for in multiple, configurable
3449 paths (easier for distros to package)
3450 - Some changes to the hooks infrastructure (especially the new
3451 post-configuration update hook)
3452 - Other small bugfixes
3454 .. vim: set textwidth=72 syntax=rst :