8 *(Released Wed, 16 Apr 2014)*
10 - Fix filtering of pending jobs with -o id (issue 778)
11 - Make RAPI API calls more symmetric (issue 770)
12 - Make parsing of old cluster configuration more robust (issue 783)
13 - Fix wrong output of gnt-instance info after migrations
14 - Fix reserved PCI slots for KVM hotplugging
15 - Use runtime hypervisor parameters to calculate bockdevice options for KVM
16 - Fix high node daemon load during disk sync if the sync is paused manually
18 - Improve opportunistic locking during instance creation (issue 791)
20 Inherited from the 2.9 branch:
22 - Make watcher submit queries low priority (issue 772)
23 - Add reason parameter to RAPI client functions (issue 776)
24 - Fix failing gnt-node list-drbd command (issue 777)
25 - Properly display fake job locks in gnt-debug.
26 - small fixes in documentation
32 *(Released Mon, 24 Mar 2014)*
34 - Fix conflict between virtio + spice or soundhw (issue 757)
35 - accept relative paths in gnt-cluster copyfile (issue 754)
36 - Introduce shutdown timeout for 'xm shutdown' command
37 - Improve RAPI detection of the watcher (issue 752)
43 *(Released Wed, 5 Mar 2014)*
45 - Fix incorrect invocation of hooks on offline nodes (issue 742)
46 - Fix incorrect exit code of gnt-cluster verify in certain circumstances
49 Inherited from the 2.9 branch:
51 - Fix overflow problem in hbal that caused it to break when waiting for
52 jobs for more than 10 minutes (issue 717)
53 - Make hbal properly handle non-LVM storage
54 - Properly export and import NIC parameters, and do so in a backwards
55 compatible way (issue 716)
56 - Fix net-common script in case of routed mode (issue 728)
57 - Improve documentation (issues 724, 730)
63 *(Released Thu, 20 Feb 2014)*
65 Incompatible/important changes
66 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
68 - Adding disks with 'gnt-instance modify' now waits for the disks to sync per
69 default. Specify --no-wait-for-sync to override this behavior.
70 - The Ganeti python code now adheres to a private-module layout. In particular,
71 the module 'ganeti' is no longer in the python search path.
72 - On instance allocation, the iallocator now considers non-LVM storage
73 properly. In particular, actual file storage space information is used
74 when allocating space for a file/sharedfile instance.
75 - When disabling disk templates cluster-wide, the cluster now first
76 checks whether there are instances still using those templates.
77 - 'gnt-node list-storage' now also reports storage information about
78 file-based storage types.
79 - In case of non drbd instances, export \*_SECONDARY environment variables
80 as empty strings (and not "None") during 'instance-migrate' related hooks.
85 - KVM hypervisors can now access RBD storage directly without having to
86 go through a block device.
87 - A new command 'gnt-cluster upgrade' was added that automates the upgrade
88 procedure between two Ganeti versions that are both 2.10 or higher.
89 - The move-instance command can now change disk templates when moving
90 instances, and does not require any node placement options to be
91 specified if the destination cluster has a default iallocator.
92 - Users can now change the soundhw and cpuid settings for XEN hypervisors.
93 - Hail and hbal now have the (optional) capability of accessing average CPU
94 load information through the monitoring deamon, and to use it to dynamically
95 adapt the allocation of instances.
96 - Hotplug support. Introduce new option '--hotplug' to ``gnt-instance modify``
97 so that disk and NIC modifications take effect without the need of actual
98 reboot. There are a couple of constrains currently for this feature:
100 - only KVM hypervisor (versions >= 1.0) supports it,
101 - one can not (yet) hotplug a disk using userspace access mode for RBD
102 - in case of a downgrade instances should suffer a reboot in order to
103 be migratable (due to core change of runtime files)
104 - ``python-fdsend`` is required for NIC hotplugging.
109 - A new test framework for logical units was introduced and the test
110 coverage for logical units was improved significantly.
111 - Opcodes are entirely generated from Haskell using the tool 'hs2py' and
112 the module 'src/Ganeti/OpCodes.hs'.
113 - Constants are also generated from Haskell using the tool
114 'hs2py-constants' and the module 'src/Ganeti/Constants.hs', with the
115 exception of socket related constants, which require changing the
116 cluster configuration file, and HVS related constants, because they
117 are part of a port of instance queries to Haskell. As a result, these
118 changes will be part of the next release of Ganeti.
123 The following new dependencies have been added/updated.
127 - The version requirements for ``python-mock`` have increased to at least
128 version 1.0.1. It is still used for testing only.
129 - ``python-fdsend`` (https://gitorious.org/python-fdsend) is optional
130 but required for KVM NIC hotplugging to work.
135 - Fix integer overflow problem in hbal
141 *(Released Wed, 12 Feb 2014)*
143 This was the third RC release of the 2.10 series. Since 2.10.0 rc2:
145 - Improved hotplug robustness
146 - Start Ganeti daemons after ensure-dirs during upgrade
147 - Documentation improvements
149 Inherited from the 2.9 branch:
151 - Fix the RAPI instances-multi-alloc call
152 - assign unique filenames to file-based disks
153 - gracefully handle degraded non-diskless instances with 0 disks (issue 697)
154 - noded now runs with its specified group, which is the default group,
155 defaulting to root (issue 707)
156 - make using UUIDs to identify nodes in gnt-node consistently possible
163 *(Released Fri, 31 Jan 2014)*
165 This was the second RC release of the 2.10 series. Since 2.10.0 rc1:
167 - Documentation improvements
168 - Run drbdsetup syncer only on network attach
169 - Include target node in hooks nodes for migration
171 - Support post-upgrade hooks during cluster upgrades
173 Inherited from the 2.9 branch:
175 - Ensure that all the hypervisors exist in the config file (Issue 640)
176 - Correctly recognise the role as master node (Issue 687)
177 - configure: allow detection of Sphinx 1.2+ (Issue 502)
178 - gnt-instance now honors the KVM path correctly (Issue 691)
180 Inherited from the 2.8 branch:
182 - Change the list separator for the usb_devices parameter from comma to space.
183 Commas could not work because they are already the hypervisor option
184 separator (Issue 649)
185 - Add support for blktap2 file-driver (Issue 638)
186 - Add network tag definitions to the haskell codebase (Issue 641)
187 - Fix RAPI network tag handling
188 - Add the network tags to the tags searched by gnt-cluster search-tags
189 - Fix caching bug preventing jobs from being cancelled
190 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
196 *(Released Tue, 17 Dec 2013)*
198 This was the first RC release of the 2.10 series. Since 2.10.0 beta1:
200 - All known issues in 2.10.0 beta1 have been resolved (see changes from
202 - Improve handling of KVM runtime files from earlier Ganeti versions
203 - Documentation fixes
205 Inherited from the 2.9 branch:
207 - use custom KVM path if set for version checking
208 - SingleNotifyPipeCondition: don't share pollers
210 Inherited from the 2.8 branch:
212 - Fixed Luxi daemon socket permissions after master-failover
213 - Improve IP version detection code directly checking for colons rather than
214 passing the family from the cluster object
215 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
216 opportunistically anymore (Issue 622)
217 - Allow link local IPv6 gateways (Issue 624)
218 - Fix error printing (Issue 616)
219 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
220 disk modifications, keep the old one. If name=none then set disk name to
222 - Update build_chroot script to work with the latest hackage packages
223 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
224 - Fix evacuation out of drained node (Issue 615)
225 - Add default file_driver if missing (Issue 571)
226 - Fix job error message after unclean master shutdown (Issue 618)
227 - Lock group(s) when creating instances (Issue 621)
228 - SetDiskID() before accepting an instance (Issue 633)
229 - Allow the ext template disks to receive arbitrary parameters, both at creation
230 time and while being modified
231 - Xen handle domain shutdown (future proofing cherry-pick)
232 - Refactor reading live data in htools (future proofing cherry-pick)
238 *(Released Wed, 27 Nov 2013)*
240 This was the first beta release of the 2.10 series. All important changes
241 are listed in the latest 2.10 entry.
246 The following issues are known to be present in the beta and will be fixed
249 - Issue 477: Wrong permissions for confd LUXI socket
250 - Issue 621: Instance related opcodes do not aquire network/group locks
251 - Issue 622: Assertion Error: Node locks differ from node resource locks
252 - Issue 623: IPv6 Masterd <-> Luxid communication error
258 *(Released Mon, 7 Apr 2014)*
260 - Improve RAPI detection of the watcher (Issue 752)
261 - gnt-cluster copyfile: accept relative paths (Issue 754)
262 - Make watcher submit queries low priority (Issue 772)
263 - Add reason parameter to RAPI client functions (Issue 776)
264 - Fix failing gnt-node list-drbd command (Issue 777)
265 - Properly display fake job locks in gnt-debug.
266 - Enable timeout for instance shutdown
267 - small fixes in documentation
273 *(Released Tue, 25 Feb 2014)*
275 - Fix overflow problem in hbal that caused it to break when waiting for
276 jobs for more than 10 minutes (issue 717)
277 - Make hbal properly handle non-LVM storage
278 - Properly export and import NIC parameters, and do so in a backwards
279 compatible way (issue 716)
280 - Fix net-common script in case of routed mode (issue 728)
281 - Improve documentation (issues 724, 730)
287 *(Released Mon, 10 Feb 2014)*
289 - Fix the RAPI instances-multi-alloc call
290 - assign unique filenames to file-based disks
291 - gracefully handle degraded non-diskless instances with 0 disks (issue 697)
292 - noded now runs with its specified group, which is the default group,
293 defaulting to root (issue 707)
294 - make using UUIDs to identify nodes in gnt-node consistently possible
301 *(Released Mon, 27 Jan 2014)*
303 - Ensure that all the hypervisors exist in the config file (Issue 640)
304 - Correctly recognise the role as master node (Issue 687)
305 - configure: allow detection of Sphinx 1.2+ (Issue 502)
306 - gnt-instance now honors the KVM path correctly (Issue 691)
308 Inherited from the 2.8 branch:
310 - Change the list separator for the usb_devices parameter from comma to space.
311 Commas could not work because they are already the hypervisor option
312 separator (Issue 649)
313 - Add support for blktap2 file-driver (Issue 638)
314 - Add network tag definitions to the haskell codebase (Issue 641)
315 - Fix RAPI network tag handling
316 - Add the network tags to the tags searched by gnt-cluster search-tags
317 - Fix caching bug preventing jobs from being cancelled
318 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
324 *(Released Fri, 13 Dec 2013)*
326 - use custom KVM path if set for version checking
327 - SingleNotifyPipeCondition: don't share pollers
329 Inherited from the 2.8 branch:
331 - Fixed Luxi daemon socket permissions after master-failover
332 - Improve IP version detection code directly checking for colons rather than
333 passing the family from the cluster object
334 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
335 opportunistically anymore (Issue 622)
336 - Allow link local IPv6 gateways (Issue 624)
337 - Fix error printing (Issue 616)
338 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
339 disk modifications, keep the old one. If name=none then set disk name to
341 - Update build_chroot script to work with the latest hackage packages
342 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
343 - Fix evacuation out of drained node (Issue 615)
344 - Add default file_driver if missing (Issue 571)
345 - Fix job error message after unclean master shutdown (Issue 618)
346 - Lock group(s) when creating instances (Issue 621)
347 - SetDiskID() before accepting an instance (Issue 633)
348 - Allow the ext template disks to receive arbitrary parameters, both at creation
349 time and while being modified
350 - Xen handle domain shutdown (future proofing cherry-pick)
351 - Refactor reading live data in htools (future proofing cherry-pick)
357 *(Released Wed, 13 Nov 2013)*
359 - fix bug, that kept nodes offline when readding
360 - when verifying DRBD versions, ignore unavailable nodes
361 - fix bug that made the console unavailable on kvm in split-user
363 - DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
369 *(Released Tue, 5 Nov 2013)*
371 Incompatible/important changes
372 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
374 - hroller now also plans for capacity to move non-redundant instances off
375 any node to be rebooted; the old behavior of completely ignoring any
376 non-redundant instances can be restored by adding the --ignore-non-redundant
378 - The cluster option '--no-lvm-storage' was removed in favor of the new option
379 '--enabled-disk-templates'.
380 - On instance creation, disk templates no longer need to be specified
381 with '-t'. The default disk template will be taken from the list of
382 enabled disk templates.
383 - The monitoring daemon is now running as root, in order to be able to collect
384 information only available to root (such as the state of Xen instances).
385 - The ConfD client is now IPv6 compatible.
386 - File and shared file storage is no longer dis/enabled at configure time,
387 but using the option '--enabled-disk-templates' at cluster initialization and
389 - The default directories for file and shared file storage are not anymore
390 specified at configure time, but taken from the cluster's configuration.
391 They can be set at cluster initialization and modification with
392 '--file-storage-dir' and '--shared-file-storage-dir'.
393 - Cluster verification now includes stricter checks regarding the
394 default file and shared file storage directories. It now checks that
395 the directories are explicitely allowed in the 'file-storage-paths' file and
396 that the directories exist on all nodes.
397 - The list of allowed disk templates in the instance policy and the list
398 of cluster-wide enabled disk templates is now checked for consistency
399 on cluster or group modification. On cluster initialization, the ipolicy
400 disk templates are ensured to be a subset of the cluster-wide enabled
406 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
407 the correct command syntax. It is possible to use different DRBD versions
408 on different nodes as long as they are compatible to each other. This
409 enables rolling upgrades of DRBD with no downtime. As permanent operation
410 of different DRBD versions within a node group is discouraged,
411 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
412 - New "inst-status-xen" data collector for the monitoring daemon, providing
413 information about the state of the xen instances on the nodes.
414 - New "lv" data collector for the monitoring daemon, collecting data about the
415 logical volumes on the nodes, and pairing them with the name of the instances
417 - New "diskstats" data collector, collecting the data from /proc/diskstats and
418 presenting them over the monitoring daemon interface.
419 - The ConfD client is now IPv6 compatible.
423 The following new dependencies have been added.
427 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
428 for the unit tests (and only used for testing).
432 - ``hslogger`` (http://software.complete.org/hslogger) is now always
433 required, even if confd is not enabled.
438 - Correctly start/stop luxid during gnt-cluster master-failover (inherited
440 - Improved error messsages (inherited from stable-2.8)
446 *(Released Tue, 15 Oct 2013)*
448 The third release candidate in the 2.9 series. Since 2.9.0 rc2:
450 - in implicit configuration upgrade, match ipolicy with enabled disk templates
451 - improved harep documentation (inherited from stable-2.8)
457 *(Released Wed, 9 Oct 2013)*
459 The second release candidate in the 2.9 series. Since 2.9.0 rc1:
461 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
462 at least one DRBD instance.
463 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
470 *(Released Tue, 1 Oct 2013)*
472 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
475 - update of the documentation, in particular installation instructions
476 - merging of LD_* constants into DT_* constants
477 - python style changes to be compatible with newer versions of pylint
483 *(Released Thu, 29 Aug 2013)*
485 This was the first beta release of the 2.9 series. All important changes
486 are listed in the latest 2.9 entry.
492 *(Released Thu, 23 Jan 2014)*
494 - Change the list separator for the usb_devices parameter from comma to space.
495 Commas could not work because they are already the hypervisor option
496 separator (Issue 649)
497 - Add support for blktap2 file-driver (Issue 638)
498 - Add network tag definitions to the haskell codebase (Issue 641)
499 - Fix RAPI network tag handling
500 - Add the network tags to the tags searched by gnt-cluster search-tags
501 - Fix caching bug preventing jobs from being cancelled
502 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
508 *(Released Thu, 12 Dec 2013)*
510 - Fixed Luxi daemon socket permissions after master-failover
511 - Improve IP version detection code directly checking for colons rather than
512 passing the family from the cluster object
513 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
514 opportunistically anymore (Issue 622)
515 - Allow link local IPv6 gateways (Issue 624)
516 - Fix error printing (Issue 616)
517 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
518 disk modifications, keep the old one. If name=none then set disk name to
520 - Update build_chroot script to work with the latest hackage packages
521 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
522 - Fix evacuation out of drained node (Issue 615)
523 - Add default file_driver if missing (Issue 571)
524 - Fix job error message after unclean master shutdown (Issue 618)
525 - Lock group(s) when creating instances (Issue 621)
526 - SetDiskID() before accepting an instance (Issue 633)
527 - Allow the ext template disks to receive arbitrary parameters, both at creation
528 time and while being modified
529 - Xen handle domain shutdown (future proofing cherry-pick)
530 - Refactor reading live data in htools (future proofing cherry-pick)
536 *(Released Thu, 07 Nov 2013)*
538 - DRBD: ensure peers are UpToDate for dual-primary
539 - Improve error message for replace-disks
540 - More dependency checks at configure time
541 - Placate warnings on ganeti.outils_unittest.py
547 *(Released Thu, 17 Oct 2013)*
549 - Correctly start/stop luxid during gnt-cluster master-failover
550 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
551 - Fix path for the job queue serial file
552 - Improved harep man page
553 - Minor documentation improvements
559 *(Released Mon, 30 Sep 2013)*
561 Incompatible/important changes
562 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
564 - Instance policy can contain multiple instance specs, as described in
565 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
566 <design-partitioned>`. As a consequence, it's not possible to partially change
567 or override instance specs. Bounding specs (min and max) can be specified as a
568 whole using the new option ``--ipolicy-bounds-specs``, while standard
569 specs use the new option ``--ipolicy-std-specs``.
570 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
571 gnt-instance is a valid YAML object.
572 - hail now honors network restrictions when allocating nodes. This led to an
573 update of the IAllocator protocol. See the IAllocator documentation for
575 - confd now only answers static configuration request over the network. luxid
576 was extracted, listens on the local LUXI socket and responds to live queries.
577 This allows finer grained permissions if using separate users.
582 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
583 to always require authentication, ``--require-authentication``. It can
584 be specified in ``$sysconfdir/default/ganeti``.
585 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
586 be used to manage the disk templates to be used by instances in the cluster.
587 Initially, it will be set to a list that includes plain, drbd, if they were
588 enabled by specifying a volume group name, and file and sharedfile, if those
589 were enabled at configure time. Additionally, it will include all disk
590 templates that are currently used by instances. The order of disk templates
591 will be based on Ganeti's history of supporting them. In the future, the
592 first entry of the list will be used as a default disk template on instance
594 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
595 configuration back to the previous stable version.
596 - Disk templates in group ipolicy can be restored to the default value.
597 - Initial support for diskless instances and virtual clusters in QA.
598 - More QA and unit tests for instance policies.
599 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
600 describing why the opcode itself was executed.
601 - The monitoring daemon is now available. It allows users to query the cluster
602 for obtaining information about the status of the system. The daemon is only
603 responsible for providing the information over the network: the actual data
604 gathering is performed by data collectors (currently, only the DRBD status
605 collector is available).
606 - In order to help developers work on Ganeti, a new script
607 (``devel/build_chroot``) is provided, for building a chroot that contains all
608 the required development libraries and tools for compiling Ganeti on a Debian
610 - A new tool, ``harep``, for performing self-repair and recreation of instances
611 in Ganeti has been added.
612 - Split queries are enabled for tags, network, exports, cluster info, groups,
614 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
615 It prints the command line to set the current policies, to ease
617 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
618 devices for KVM virtio-net interfaces will get created with VNET_HDR
619 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
620 virtio-net interfaces, which prevents host kernel tainting and log
621 flooding, when dealing with broken or malicious virtio-net drivers.
622 It's set to true by default.
623 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
625 - Support 'viridian' parameter in Xen HVM
626 - Support DSA SSH keys in bootstrap
627 - To simplify the work of packaging frameworks that want to add the needed users
628 and groups in a split-user setup themselves, at build time three files in
629 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
630 the groups to be generated, the ``users`` file contains, one per line, the
631 users to be generated, optionally followed by their primary group, where
632 important. The ``groupmemberships`` file contains, one per line, additional
633 user-group membership relations that need to be established. The syntax of
634 these files will remain stable in all future versions.
639 The following new dependencies have been added:
642 - The ``curl`` library is not optional anymore for compiling the Haskell code.
643 - ``snap-server`` library (if monitoring is enabled).
646 - The minimum Python version needed to run Ganeti is now 2.6.
647 - ``yaml`` library (only for running the QA).
651 - Perform proper cleanup on termination of Haskell daemons
652 - Fix corner-case in handling of remaining retry time
658 *(Released Tue, 17 Sep 2013)*
660 - To simplify the work of packaging frameworks that want to add the needed users
661 and groups in a split-user setup themselves, at build time three files in
662 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
663 the groups to be generated, the ``users`` file contains, one per line, the
664 users to be generated, optionally followed by their primary group, where
665 important. The ``groupmemberships`` file contains, one per line, additional
666 user-group membership relations that need to be established. The syntax of
667 these files will remain stable in all future versions.
668 - Add a default to file-driver when unspecified over RAPI (Issue 571)
669 - Mark the DSA host pubkey as optional, and remove it during config downgrade
671 - Some documentation fixes
677 *(Released Tue, 27 Aug 2013)*
679 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
681 - Support 'viridian' parameter in Xen HVM (Issue 233)
682 - Include VCS version in ``gnt-cluster version``
683 - Support DSA SSH keys in bootstrap (Issue 338)
684 - Fix batch creation of instances
685 - Use FQDN to check master node status (Issue 551)
686 - Make the DRBD collector more failure-resilient
692 *(Released Fri, 2 Aug 2013)*
694 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
696 - Fix upgrading/downgrading from 2.7
697 - Increase maximum RAPI message size
698 - Documentation updates
699 - Split ``confd`` between ``luxid`` and ``confd``
700 - Merge 2.7 series up to the 2.7.1 release
701 - Allow the ``modify_etc_hosts`` option to be changed
702 - Add better debugging for ``luxid`` queries
703 - Expose bulk parameter for GetJobs in RAPI client
704 - Expose missing ``network`` fields in RAPI
705 - Add some ``cluster verify`` tests
706 - Some unittest fixes
707 - Fix a malfunction in ``hspace``'s tiered allocation
708 - Fix query compatibility between haskell and python implementations
709 - Add the ``vnet_hdr`` HV parameter for KVM
710 - Add ``--cleanup`` to instance failover
711 - Change the connected groups format in ``gnt-network info`` output; it
712 was previously displayed as a raw list by mistake. (Merged from 2.7)
718 *(Released Mon, 24 Jun 2013)*
720 This was the first beta release of the 2.8 series. All important changes
721 are listed in the latest 2.8 entry.
727 *(Released Thu, 26 Sep 2013)*
729 - Change the connected groups format in ``gnt-network info`` output; it
730 was previously displayed as a raw list by mistake
731 - Check disk template in right dict when copying
732 - Support multi-instance allocs without iallocator
733 - Fix some errors in the documentation
734 - Fix formatting of tuple in an error message
740 *(Released Thu, 25 Jul 2013)*
742 - Add logrotate functionality in daemon-util
743 - Add logrotate example file
744 - Add missing fields to network queries over rapi
745 - Fix network object timestamps
746 - Add support for querying network timestamps
747 - Fix a typo in the example crontab
748 - Fix a documentation typo
754 *(Released Thu, 04 Jul 2013)*
756 Incompatible/important changes
757 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
759 - Instance policies for disk size were documented to be on a per-disk
760 basis, but hail applied them to the sum of all disks. This has been
762 - ``hbal`` will now exit with status 0 if, during job execution over
763 LUXI, early exit has been requested and all jobs are successful;
764 before, exit status 1 was used, which cannot be differentiated from
766 - Compatibility with newer versions of rbd has been fixed
767 - ``gnt-instance batch-create`` has been changed to use the bulk create
768 opcode from Ganeti. This lead to incompatible changes in the format of
769 the JSON file. It's now not a custom dict anymore but a dict
770 compatible with the ``OpInstanceCreate`` opcode.
771 - Parent directories for file storage need to be listed in
772 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
773 write the file automatically based on old configuration values, but it
774 can not distribute it across all nodes and the file contents should be
775 verified. Use ``gnt-cluster copyfile
776 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
777 upgraded. The reason for requiring this list of paths now is that
778 before it would have been possible to inject new paths via RPC,
779 allowing files to be created in arbitrary locations. The RPC protocol
780 is protected using SSL/X.509 certificates, but as a design principle
781 Ganeti does not permit arbitrary paths to be passed.
782 - The parsing of the variants file for OSes (see
783 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
784 lines and comment lines (starting with ``#``) are ignored for better
786 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
787 longer available. ``gnt-node add`` now invokes a new tool on the
788 destination node, named ``prepare-node-join``, to configure the SSH
789 daemon. Paramiko is no longer necessary to configure nodes' SSH
790 daemons via ``gnt-node add``.
791 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
792 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
793 the flag is not reset after a master failover.
794 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
795 is recommended. 2.6 will be mandatory from the 2.8 series.
801 - New network management functionality to support automatic allocation
802 of IP addresses and managing of network parameters. See
803 :manpage:`gnt-network(8)` for more details.
804 - New external storage backend, to allow managing arbitrary storage
805 systems external to the cluster. See
806 :manpage:`ganeti-extstorage-interface(7)`.
807 - New ``exclusive-storage`` node parameter added, restricted to
808 nodegroup level. When it's set to true, physical disks are assigned in
809 an exclusive fashion to instances, as documented in :doc:`Partitioned
810 Ganeti <design-partitioned>`. Currently, only instances using the
811 ``plain`` disk template are supported.
812 - The KVM hypervisor has been updated with many new hypervisor
813 parameters, including a generic one for passing arbitrary command line
814 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
815 compatible up to qemu 1.4.
816 - A new tool, called ``mon-collector``, is the stand-alone executor of
817 the data collectors for a monitoring system. As of this version, it
818 just includes the DRBD data collector, that can be executed by calling
819 ``mon-collector`` using the ``drbd`` parameter. See
820 :manpage:`mon-collector(7)`.
821 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
822 for RAPI users. It allows granting permissions to query for
823 information to a specific user without giving
824 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
825 - A new tool named ``node-cleanup`` has been added. It cleans remains of
826 a cluster from a machine by stopping all daemons, removing
827 certificates and ssconf files. Unless the ``--no-backup`` option is
828 given, copies of the certificates are made.
829 - Instance creations now support the use of opportunistic locking,
830 potentially speeding up the (parallel) creation of multiple instances.
831 This feature is currently only available via the :doc:`RAPI
832 <rapi>` interface and when an instance allocator is used. If the
833 ``opportunistic_locking`` parameter is set the opcode will try to
834 acquire as many locks as possible, but will not wait for any locks
835 held by other opcodes. If not enough resources can be found to
836 allocate the instance, the temporary error code
837 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
838 retried thereafter, with or without opportunistic locking.
839 - New experimental linux-ha resource scripts.
840 - Restricted-commands support: ganeti can now be asked (via command line
841 or rapi) to perform commands on a node. These are passed via ganeti
842 RPC rather than ssh. This functionality is restricted to commands
843 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
844 reasons. The file is not copied automatically.
850 - Diskless instances are now externally mirrored (Issue 237). This for
851 now has only been tested in conjunction with explicit target nodes for
853 - Queries not needing locks or RPC access to the node can now be
854 performed by the confd daemon, making them independent from jobs, and
855 thus faster to execute. This is selectable at configure time.
856 - The functionality for allocating multiple instances at once has been
857 overhauled and is now also available through :doc:`RAPI <rapi>`.
859 There are no significant changes from version 2.7.0~rc3.
865 *(Released Tue, 25 Jun 2013)*
867 - Fix permissions on the confd query socket (Issue 477)
868 - Fix permissions on the job archive dir (Issue 498)
869 - Fix handling of an internal exception in replace-disks (Issue 472)
870 - Fix gnt-node info handling of shortened names (Issue 497)
871 - Fix gnt-instance grow-disk when wiping is enabled
872 - Documentation improvements, and support for newer pandoc
873 - Fix hspace honoring ipolicy for disks (Issue 484)
874 - Improve handling of the ``kvm_extra`` HV parameter
880 *(Released Fri, 24 May 2013)*
882 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
884 - Disks added through ``gnt-instance modify`` or created through
885 ``gnt-instance recreate-disks`` are wiped, if the
886 ``prealloc_wipe_disks`` flag is set.
887 - If wiping newly created disks fails, the disks are removed. Also,
888 partial failures in creating disks through ``gnt-instance modify``
889 triggers a cleanup of the partially-created disks.
890 - Removing the master IP address doesn't fail if the address has been
892 - Fix ownership of the OS log dir
893 - Workaround missing SO_PEERCRED constant (Issue 191)
899 *(Released Fri, 3 May 2013)*
901 This was the first release candidate of the 2.7 series. Since beta3:
903 - Fix kvm compatibility with qemu 1.4 (Issue 389)
904 - Documentation updates (admin guide, upgrade notes, install
905 instructions) (Issue 372)
906 - Fix gnt-group list nodes and instances count (Issue 436)
907 - Fix compilation without non-mandatory libraries (Issue 441)
908 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
909 - Make confd logging more verbose at INFO level (Issue 435)
910 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
911 - Fix failure path for instance storage type conversion (Issue 229)
912 - Update htools text backend documentation
913 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
914 - Disable inter-cluster instance move for file-based instances, because
915 it is dependant on instance export, which is not supported for
916 file-based instances. (Issue 414)
917 - Fix gnt-job crashes on non-ascii characters (Issue 427)
918 - Fix volume group checks on non-vm-capable nodes (Issue 432)
924 *(Released Mon, 22 Apr 2013)*
926 This was the third beta release of the 2.7 series. Since beta2:
928 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
929 - Fix data loss on wrong usage of ``gnt-instance move``
930 - Properly export errors in confd-based job queries
931 - Add ``users-setup`` tool
932 - Fix iallocator protocol to report 0 as a disk size for diskless
933 instances. This avoids hail breaking when a diskless instance is
935 - Fix job queue directory permission problem that made confd job queries
936 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
937 for access to archived jobs (Issue 406).
938 - Limit the sizes of networks supported by ``gnt-network`` to something
939 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
940 - Fix bugs in instance disk template conversion
941 - Fix GHC 7 compatibility
942 - Fix ``burnin`` install path (Issue 426).
943 - Allow very small disk grows (Issue 347).
944 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
945 that noded doesn't import masterd code (Issue 419).
946 - Make sure the default metavg at cluster init is the same as the vg, if
947 unspecified (Issue 358).
948 - Fix cleanup of partially created disks (part of Issue 416)
954 *(Released Tue, 2 Apr 2013)*
956 This was the second beta release of the 2.7 series. Since beta1:
958 - Networks no longer have a "type" slot, since this information was
959 unused in Ganeti: instead of it tags should be used.
960 - The rapi client now has a ``target_node`` option to MigrateInstance.
961 - Fix early exit return code for hbal (Issue 386).
962 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
963 - Fix ``rbd showmapped`` output parsing (Issue 312).
964 - Networks are now referenced indexed by UUID, rather than name. This
965 will require running cfgupgrade, from 2.7.0beta1, if networks are in
967 - The OS environment now includes network information.
968 - Deleting of a network is now disallowed if any instance nic is using
969 it, to prevent dangling references.
970 - External storage is now documented in man pages.
971 - The exclusive_storage flag can now only be set at nodegroup level.
972 - Hbal can now submit an explicit priority with its jobs.
973 - Many network related locking fixes.
974 - Bump up the required pylint version to 0.25.1.
975 - Fix the ``no_remember`` option in RAPI client.
976 - Many ipolicy related tests, qa, and fixes.
977 - Many documentation improvements and fixes.
978 - Fix building with ``--disable-file-storage``.
979 - Fix ``-q`` option in htools, which was broken if passed more than
981 - Some haskell/python interaction improvements and fixes.
982 - Fix iallocator in case of missing LVM storage.
983 - Fix confd config load in case of ``--no-lvm-storage``.
984 - The confd/query functionality is now mentioned in the security
991 *(Released Wed, 6 Feb 2013)*
993 This was the first beta release of the 2.7 series. All important changes
994 are listed in the latest 2.7 entry.
1000 *(Released Fri, 21 Dec 2012)*
1002 Important behaviour change: hbal won't rebalance anymore instances which
1003 have the ``auto_balance`` attribute set to false. This was the intention
1004 all along, but until now it only skipped those from the N+1 memory
1005 reservation (DRBD-specific).
1007 A significant number of bug fixes in this release:
1009 - Fixed disk adoption interaction with ipolicy checks.
1010 - Fixed networking issues when instances are started, stopped or
1011 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
1012 - Fixed the warning in cluster verify for shared storage instances not
1014 - Fixed removal of storage directory on shared file storage (issue 262).
1015 - Fixed validation of LVM volume group name in OpClusterSetParams
1016 (``gnt-cluster modify``) (issue 285).
1017 - Fixed runtime memory increases (``gnt-instance modify -m``).
1018 - Fixed live migration under Xen's ``xl`` mode.
1019 - Fixed ``gnt-instance console`` with ``xl``.
1020 - Fixed building with newer Haskell compiler/libraries.
1021 - Fixed PID file writing in Haskell daemons (confd); this prevents
1022 restart issues if confd was launched manually (outside of
1023 ``daemon-util``) while another copy of it was running
1024 - Fixed a type error when doing live migrations with KVM (issue 297) and
1025 the error messages for failing migrations have been improved.
1026 - Fixed opcode validation for the out-of-band commands (``gnt-node
1028 - Fixed a type error when unsetting OS hypervisor parameters (issue
1029 311); now it's possible to unset all OS-specific hypervisor
1031 - Fixed the ``dry-run`` mode for many operations: verification of
1032 results was over-zealous but didn't take into account the ``dry-run``
1033 operation, resulting in "wrong" failures.
1034 - Fixed bash completion in ``gnt-job list`` when the job queue has
1035 hundreds of entries; especially with older ``bash`` versions, this
1036 results in significant CPU usage.
1038 And lastly, a few other improvements have been made:
1040 - Added option to force master-failover without voting (issue 282).
1041 - Clarified error message on lock conflict (issue 287).
1042 - Logging of newly submitted jobs has been improved (issue 290).
1043 - Hostname checks have been made uniform between instance rename and
1045 - The ``--submit`` option is now supported by ``gnt-debug delay``.
1046 - Shutting down the master daemon by sending SIGTERM now stops it from
1047 processing jobs waiting for locks; instead, those jobs will be started
1048 once again after the master daemon is started the next time (issue
1050 - Support for Xen's ``xl`` program has been improved (besides the fixes
1052 - Reduced logging noise in the Haskell confd daemon (only show one log
1053 entry for each config reload, instead of two).
1054 - Several man page updates and typo fixes.
1060 *(Released Fri, 12 Oct 2012)*
1062 A small bugfix release. Among the bugs fixed:
1064 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
1065 made the command unusable.
1066 - Commands that issue many jobs don't fail anymore just because some jobs
1067 take so long that other jobs are archived.
1068 - Failures during ``gnt-instance reinstall`` are reflected by the exit
1070 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
1072 - When ``always_failover`` is set, ``--allow-failover`` is not required
1073 in migrate commands anymore.
1074 - ``bash_completion`` works even if extglob is disabled.
1075 - Fixed bug with locks that made failover for RDB-based instances fail.
1076 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
1077 a random node instead of one based on the allocator metric.
1078 - Support for newer versions of pylint and pep8.
1079 - Hail doesn't fail anymore when trying to add an instance of type
1080 ``file``, ``sharedfile`` or ``rbd``.
1081 - Added new Makefile target to rebuild the whole distribution, so that
1082 all files are included.
1088 *(Released Fri, 27 Jul 2012)*
1091 .. attention:: The ``LUXI`` protocol has been made more consistent
1092 regarding its handling of command arguments. This, however, leads to
1093 incompatibility issues with previous versions. Please ensure that you
1094 restart Ganeti daemons soon after the upgrade, otherwise most
1095 ``LUXI`` calls (job submission, setting/resetting the drain flag,
1096 pausing/resuming the watcher, cancelling and archiving jobs, querying
1097 the cluster configuration) will fail.
1106 The current ``admin_up`` field, which used to denote whether an instance
1107 should be running or not, has been removed. Instead, ``admin_state`` is
1108 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
1110 The rational behind this is that an instance being “down” can have
1113 - it could be down during a reboot
1114 - it could be temporarily be down for a reinstall
1115 - or it could be down because it is deprecated and kept just for its
1118 The previous Boolean state was making it difficult to do capacity
1119 calculations: should Ganeti reserve memory for a down instance? Now, the
1120 tri-state field makes it clear:
1122 - in ``up`` and ``down`` state, all resources are reserved for the
1123 instance, and it can be at any time brought up if it is down
1124 - in ``offline`` state, only disk space is reserved for it, but not
1127 The field can have an extra use: since the transition between ``up`` and
1128 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
1129 transition between ``offline`` and ``down`` is done via ``gnt-instance
1130 modify``, it is possible to given different rights to users. For
1131 example, owners of an instance could be allowed to start/stop it, but
1132 not transition it out of the offline state.
1134 Instance policies and specs
1135 +++++++++++++++++++++++++++
1137 In previous Ganeti versions, an instance creation request was not
1138 limited on the minimum size and on the maximum size just by the cluster
1139 resources. As such, any policy could be implemented only in third-party
1140 clients (RAPI clients, or shell wrappers over ``gnt-*``
1141 tools). Furthermore, calculating cluster capacity via ``hspace`` again
1142 required external input with regards to instance sizes.
1144 In order to improve these workflows and to allow for example better
1145 per-node group differentiation, we introduced instance specs, which
1148 - minimum instance disk size, disk count, memory size, cpu count
1149 - maximum values for the above metrics
1150 - and “standard” values (used in ``hspace`` to calculate the standard
1153 The minimum/maximum values can be also customised at node-group level,
1154 for example allowing more powerful hardware to support bigger instance
1157 Beside the instance specs, there are a few other settings belonging to
1158 the instance policy framework. It is possible now to customise, per
1159 cluster and node-group:
1161 - the list of allowed disk templates
1162 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
1163 - the maximum ratio of instance to spindles (see below for more
1164 information) for local storage
1166 All these together should allow all tools that talk to Ganeti to know
1167 what are the ranges of allowed values for instances and the
1168 over-subscription that is allowed.
1170 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
1171 instance configuration, and the physical CPU configuration from the
1172 node. For the spindle ratios however, we didn't track before these
1173 values, so new parameters have been added:
1175 - a new node parameter ``spindle_count``, defaults to 1, customisable at
1176 node group or node level
1177 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
1179 Note that spindles in this context doesn't need to mean actual
1180 mechanical hard-drives; it's just a relative number for both the node
1181 I/O capacity and instance I/O consumption.
1183 Instance migration behaviour
1184 ++++++++++++++++++++++++++++
1186 While live-migration is in general desirable over failover, it is
1187 possible that for some workloads it is actually worse, due to the
1188 variable time of the “suspend” phase during live migration.
1190 To allow the tools to work consistently over such instances (without
1191 having to hard-code instance names), a new backend parameter
1192 ``always_failover`` has been added to control the migration/failover
1193 behaviour. When set to True, all migration requests for an instance will
1194 instead fall-back to failover.
1196 Instance memory ballooning
1197 ++++++++++++++++++++++++++
1199 Initial support for memory ballooning has been added. The memory for an
1200 instance is no longer fixed (backend parameter ``memory``), but instead
1201 can vary between minimum and maximum values (backend parameters
1202 ``minmem`` and ``maxmem``). Currently we only change an instance's
1205 - live migrating or failing over and instance and the target node
1206 doesn't have enough memory
1207 - user requests changing the memory via ``gnt-instance modify
1210 Instance CPU pinning
1211 ++++++++++++++++++++
1213 In order to control the use of specific CPUs by instance, support for
1214 controlling CPU pinning has been added for the Xen, HVM and LXC
1215 hypervisors. This is controlled by a new hypervisor parameter
1216 ``cpu_mask``; details about possible values for this are in the
1217 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
1218 VCPU-to-CPU mapping) form will work well only when all nodes in your
1219 cluster have the same amount of CPUs.
1224 Another area in which Ganeti was not customisable were the parameters
1225 used for storage configuration, e.g. how many stripes to use for LVM,
1226 DRBD resync configuration, etc.
1228 To improve this area, we've added disks parameters, which are
1229 customisable at cluster and node group level, and which allow to
1230 specify various parameters for disks (DRBD has the most parameters
1231 currently), for example:
1233 - DRBD resync algorithm and parameters (e.g. speed)
1234 - the default VG for meta-data volumes for DRBD
1235 - number of stripes for LVM (plain disk template)
1238 These parameters can be modified via ``gnt-cluster modify -D …`` and
1239 ``gnt-group modify -D …``, and are used at either instance creation (in
1240 case of LVM stripes, for example) or at disk “activation” time
1241 (e.g. resync speed).
1243 Rados block device support
1244 ++++++++++++++++++++++++++
1246 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
1247 denoted by the ``rbd`` disk template type. This is considered
1248 experimental, feedback is welcome. For details on configuring it, see
1249 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
1254 The existing master IP functionality works well only in simple setups (a
1255 single network shared by all nodes); however, if nodes belong to
1256 different networks, then the ``/32`` setup and lack of routing
1257 information is not enough.
1259 To allow the master IP to function well in more complex cases, the
1260 system was reworked as follows:
1262 - a master IP netmask setting has been added
1263 - the master IP activation/turn-down code was moved from the node daemon
1264 to a separate script
1265 - whether to run the Ganeti-supplied master IP script or a user-supplied
1266 on is a ``gnt-cluster init`` setting
1268 Details about the location of the standard and custom setup scripts are
1269 in the man page :manpage:`gnt-cluster(8)`; for information about the
1270 setup script protocol, look at the Ganeti-supplied script.
1275 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
1278 It is now possible to use TLS-protected connections, and when renewing
1279 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
1280 it is now possible to specify spice or spice CA certificates. Also, it
1281 is possible to configure a password for SPICE sessions via the
1282 hypervisor parameter ``spice_password_file``.
1284 There are also new parameters to control the compression and streaming
1285 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
1286 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
1287 for the spice parameters.
1289 Lastly, it is now possible to see the SPICE connection information via
1290 ``gnt-instance console``.
1295 A new tool (``tools/ovfconverter``) has been added that supports
1296 conversion between Ganeti and the `Open Virtualization Format
1297 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
1300 This relies on the ``qemu-img`` tool to convert the disk formats, so the
1301 actual compatibility with other virtualization solutions depends on it.
1303 Confd daemon changes
1304 ++++++++++++++++++++
1306 The configuration query daemon (``ganeti-confd``) is now optional, and
1307 has been rewritten in Haskell; whether to use the daemon at all, use the
1308 Python (default) or the Haskell version is selectable at configure time
1309 via the ``--enable-confd`` parameter, which can take one of the
1310 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
1311 daemon will result in a smaller footprint; for larger systems, we
1312 welcome feedback on the Haskell version which might become the default
1315 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
1316 daemon running. The Python version doesn't implement the new call.
1319 User interface changes
1320 ~~~~~~~~~~~~~~~~~~~~~~
1322 We have replaced the ``--disks`` option of ``gnt-instance
1323 replace-disks`` with a more flexible ``--disk`` option, which allows
1324 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
1325 disk size and mode can be changed upon recreation (via ``gnt-instance
1326 recreate-disks``, which accepts the same ``--disk`` option).
1328 As many people are used to a ``show`` command, we have added that as an
1329 alias to ``info`` on all ``gnt-*`` commands.
1331 The ``gnt-instance grow-disk`` command has a new mode in which it can
1332 accept the target size of the disk, instead of the delta; this can be
1333 more safe since two runs in absolute mode will be idempotent, and
1334 sometimes it's also easier to specify the desired size directly.
1336 Also the handling of instances with regard to offline secondaries has
1337 been improved. Instance operations should not fail because one of it's
1338 secondary nodes is offline, even though it's safe to proceed.
1340 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
1341 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1342 minors to instance name.
1347 RAPI coverage has improved, with (for example) new resources for
1348 recreate-disks, node power-cycle, etc.
1353 There is partial support for ``xl`` in the Xen hypervisor; feedback is
1356 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1357 whether to still support Python 2.4 or move to Python 2.6 as minimum
1360 Support for Fedora has been slightly improved; the provided example
1361 init.d script should work better on it and the INSTALL file should
1362 document the needed dependencies.
1367 The deprecated ``QueryLocks`` LUXI request has been removed. Use
1368 ``Query(what=QR_LOCK, ...)`` instead.
1370 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1371 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1372 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1373 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1374 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1376 RAPI client: ``CertificateError`` now derives from
1377 ``GanetiApiError``. This should make it more easy to handle Ganeti
1380 Deprecation warnings due to PyCrypto/paramiko import in
1381 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
1382 make sure to run an up-to-date paramiko version, if you use this tool.
1384 The QA scripts now depend on Python 2.5 or above (the main code base
1385 still works with Python 2.4).
1387 The configuration file (``config.data``) is now written without
1388 indentation for performance reasons; if you want to edit it, it can be
1389 re-formatted via ``tools/fmtjson``.
1391 A number of bugs has been fixed in the cluster merge tool.
1393 ``x509`` certification verification (used in import-export) has been
1394 changed to allow the same clock skew as permitted by the cluster
1395 verification. This will remove some rare but hard to diagnose errors in
1402 *(Released Thu, 19 Jul 2012)*
1404 Very few changes from rc4 to the final release, only bugfixes:
1406 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
1407 instance, fix CDROM booting for KVM instances)
1408 - fixed node group modification of node parameters
1409 - fixed issue in LUClusterVerifyGroup with multi-group clusters
1410 - fixed generation of bash completion to ensure a stable ordering
1417 *(Released Fri, 13 Jul 2012)*
1419 Third release candidate for 2.6. The following changes were done from
1422 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1423 - Fixed an inconsistency in the LUXI protocol with the provided
1424 arguments (NOT backwards compatible)
1425 - Fixed a bug with node groups ipolicy where ``min`` was greater than
1426 the cluster ``std`` value
1427 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1428 easier instance debugging on nodes (requires ``hconfd`` to work)
1434 *(Released Tue, 03 Jul 2012)*
1436 Second release candidate for 2.6. The following changes were done from
1439 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1441 - Fixed a RAPI regression on missing beparams/memory
1442 - Fixed redistribution of files on offline nodes
1443 - Added possibility to run activate-disks even though secondaries are
1444 offline. With this change it relaxes also the strictness on some other
1445 commands which use activate disks internally:
1446 * ``gnt-instance start|reboot|rename|backup|export``
1447 - Made it possible to remove safely an instance if its secondaries are
1449 - Made it possible to reinstall even though secondaries are offline
1455 *(Released Mon, 25 Jun 2012)*
1457 First release candidate for 2.6. The following changes were done from
1460 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1462 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1463 behaviour. It supports now the modify operation
1464 - ``hcheck`` implemented to analyze cluster health and possibility of
1465 improving health by rebalance
1466 - ``hbal`` has been improved in dealing with split instances
1472 *(Released Mon, 11 Jun 2012)*
1474 Second beta release of 2.6. The following changes were done from beta2
1477 - Fixed ``daemon-util`` with non-root user models
1478 - Fixed creation of plain instances with ``--no-wait-for-sync``
1479 - Fix wrong iv_names when running ``cfgupgrade``
1480 - Export more information in RAPI group queries
1481 - Fixed bug when changing instance network interfaces
1482 - Extended burnin to do NIC changes
1483 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1484 - Changed default for DRBD barriers
1485 - Fixed DRBD error reporting for syncer rate
1486 - Verify the options on disk parameters
1488 And of course various fixes to documentation and improved unittests and
1495 *(Released Wed, 23 May 2012)*
1497 First beta release of 2.6. The following changes were done from beta1 to
1500 - integrated patch for distributions without ``start-stop-daemon``
1501 - adapted example init.d script to work on Fedora
1502 - fixed log handling in Haskell daemons
1503 - adapted checks in the watcher for pycurl linked against libnss
1504 - add partial support for ``xl`` instead of ``xm`` for Xen
1505 - fixed a type issue in cluster verification
1506 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1509 Plus integrated fixes from the 2.5 branch:
1511 - fixed ``kvm-ifup`` to use ``/bin/bash``
1512 - fixed parallel build failures
1513 - KVM live migration when using a custom keymap
1519 *(Released Tue, 24 Jul 2012)*
1521 A small bugfix release, with no new features:
1523 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1524 different default shell (e.g. Debian, Ubuntu)
1525 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1526 243 and Debian bug #650664)
1527 - fixed compatibility with KVM versions that don't support multiple boot
1528 devices (fixes Issue 230 and Debian bug #624256)
1530 Additionally, a few fixes were done to the build system (fixed parallel
1531 build failures) and to the unittests (fixed race condition in test for
1532 FileID functions, and the default enable/disable mode for QA test is now
1539 *(Released Fri, 11 May 2012)*
1541 A small bugfix release.
1543 The main issues solved are on the topic of compatibility with newer LVM
1546 - fixed parsing of ``lv_attr`` field
1547 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1548 the ``--force`` flag is needed
1550 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1551 to accept kernel 3.x too (was hardcoded to 2.6.*).
1553 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1554 of hook results and that also made display of other errors suboptimal
1555 was fixed; the code behaves now like 2.4 and earlier.
1557 Another change in 2.5, the cleanup of the OS scripts environment, is too
1558 aggressive: it removed even the ``PATH`` variable, which requires the OS
1559 scripts to *always* need to export it. Since this is a bit too strict,
1560 we now export a minimal PATH, the same that we export for hooks.
1562 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1563 integrated into this release.
1565 Finally, a few other miscellaneous changes were done (no new features,
1566 just small improvements):
1568 - Fix ``gnt-group --help`` display
1569 - Fix hardcoded Xen kernel path
1570 - Fix grow-disk handling of invalid units
1571 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1572 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1578 *(Released Thu, 12 Apr 2012)*
1580 Incompatible/important changes and bugfixes
1581 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1583 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1584 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1585 to match the underlying LUXI interface.
1586 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1587 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1588 not maintain backwards-compatibility as the underlying operation
1589 changed in an incompatible way. The RAPI client can talk to old
1590 servers, but it needs to be told so as the return value changed.
1591 - When creating file-based instances via RAPI, the ``file_driver``
1592 parameter no longer defaults to ``loop`` and must be specified.
1593 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1595 - Support for the undocumented and deprecated RAPI instance creation
1596 request format version 0 has been dropped. Use version 1, supported
1597 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1598 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1599 documentation <install>`.
1600 - The "cluster-verify" hooks are now executed per group by the
1601 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1602 if you just run ``gnt-cluster verify``, which generates one opcode per
1604 - The environment as passed to the OS scripts is cleared, and thus no
1605 environment variables defined in the node daemon's environment will be
1606 inherited by the scripts.
1607 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1609 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1610 support operations involving multiple node groups.
1611 - Offline nodes are ignored when failing over an instance.
1612 - Support for KVM version 1.0, which changed the version reporting format
1614 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1616 - ``Makefile`` is now compatible with Automake 1.11.2
1617 - Includes all bugfixes made in the 2.4 series
1622 - The ganeti-htools project has been merged into the ganeti-core source
1623 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1624 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1625 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1626 GPT-style partition tables (requires `parted
1627 <http://www.gnu.org/s/parted/>`_).
1628 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1629 - Allowed adding tags on instance creation.
1630 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1631 - Allow instances to be started in a paused state, enabling the user to
1632 see the complete console output on boot using the console.
1633 - Added new hypervisor flag to control default reboot behaviour
1634 (``reboot_behavior``).
1635 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1636 - Improved out-of-band management support:
1638 - Added ``gnt-node health`` command reporting the health status of
1640 - Added ``gnt-node power`` command to manage power status of nodes.
1641 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1643 - Instance migration can fall back to failover if instance is not
1645 - Filters can be used when listing nodes, instances, groups and locks;
1646 see :manpage:`ganeti(7)` manpage.
1647 - Added post-execution status as variables to :doc:`hooks <hooks>`
1649 - Instance tags are exported/imported together with the instance.
1650 - When given an explicit job ID, ``gnt-job info`` will work for archived
1652 - Jobs can define dependencies on other jobs (not yet supported via
1653 RAPI or command line, but used by internal commands and usable via
1656 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1659 - Instance failover is now available as a RAPI resource
1660 (``/2/instances/[instance_name]/failover``).
1661 - ``gnt-instance info`` defaults to static information if primary node
1663 - Opcodes have a new ``comment`` attribute.
1664 - Added basic SPICE support to KVM hypervisor.
1665 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1667 Node group improvements
1668 ~~~~~~~~~~~~~~~~~~~~~~~
1670 - ``gnt-cluster verify`` has been modified to check groups separately,
1671 thereby improving performance.
1672 - Node group support has been added to ``gnt-cluster verify-disks``,
1673 which now operates per node group.
1674 - Watcher has been changed to work better with node groups.
1676 - One process and state file per node group.
1677 - Slow watcher in one group doesn't block other group's watcher.
1679 - Added new command, ``gnt-group evacuate``, to move all instances in a
1680 node group to other groups.
1681 - Added ``gnt-instance change-group`` to move an instance to another
1683 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1684 per-group operations.
1685 - Node groups can be tagged.
1686 - Some operations switch from an exclusive to a shared lock as soon as
1688 - Instance's primary and secondary nodes' groups are now available as
1689 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1690 and ``snodes.group.uuid``).
1695 - Numerous updates to documentation and manpages.
1697 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1699 - Some opcode/job results are now also documented, see :doc:`RAPI
1702 - A lockset's internal lock is now also visible in lock monitor.
1703 - Log messages from job queue workers now contain information about the
1704 opcode they're processing.
1705 - ``gnt-instance console`` no longer requires the instance lock.
1706 - A short delay when waiting for job changes reduces the number of LUXI
1707 requests significantly.
1708 - DRBD metadata volumes are overwritten with zeros during disk creation.
1709 - Out-of-band commands no longer acquire the cluster lock in exclusive
1711 - ``devel/upload`` now uses correct permissions for directories.
1717 *(Released Fri, 23 Mar 2012)*
1719 This was the sixth release candidate of the 2.5 series.
1725 *(Released Mon, 9 Jan 2012)*
1727 This was the fifth release candidate of the 2.5 series.
1733 *(Released Thu, 27 Oct 2011)*
1735 This was the fourth release candidate of the 2.5 series.
1741 *(Released Wed, 26 Oct 2011)*
1743 This was the third release candidate of the 2.5 series.
1749 *(Released Tue, 18 Oct 2011)*
1751 This was the second release candidate of the 2.5 series.
1757 *(Released Tue, 4 Oct 2011)*
1759 This was the first release candidate of the 2.5 series.
1765 *(Released Wed, 31 Aug 2011)*
1767 This was the third beta release of the 2.5 series.
1773 *(Released Mon, 22 Aug 2011)*
1775 This was the second beta release of the 2.5 series.
1781 *(Released Fri, 12 Aug 2011)*
1783 This was the first beta release of the 2.5 series.
1789 *(Released Thu, 27 Oct 2011)*
1791 - Fixed bug when parsing command line parameter values ending in
1793 - Fixed assertion error after unclean master shutdown
1794 - Disable HTTP client pool for RPC, significantly reducing memory usage
1796 - Fixed queue archive creation with wrong permissions
1802 *(Released Tue, 23 Aug 2011)*
1806 - Fixed documentation for importing with ``--src-dir`` option
1807 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1808 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1814 *(Released Fri, 5 Aug 2011)*
1816 Many bug-fixes and a few small features:
1818 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1819 issues when you tried to add an instance with two MAC addresses in one
1821 - KVM: fixed per-instance stored UID value
1822 - KVM: configure bridged NICs at migration start
1823 - KVM: Fix a bug where instance will not start with never KVM versions
1825 - Added OS search path to ``gnt-cluster info``
1826 - Fixed an issue with ``file_storage_dir`` where you were forced to
1827 provide an absolute path, but the documentation states it is a
1828 relative path, the documentation was right
1829 - Added a new parameter to instance stop/start called ``--no-remember``
1830 that will make the state change to not be remembered
1831 - Implemented ``no_remember`` at RAPI level
1832 - Improved the documentation
1833 - Node evacuation: don't call IAllocator if node is already empty
1834 - Fixed bug in DRBD8 replace disks on current nodes
1835 - Fixed bug in recreate-disks for DRBD instances
1836 - Moved assertion checking locks in ``gnt-instance replace-disks``
1837 causing it to abort with not owning the right locks for some situation
1838 - Job queue: Fixed potential race condition when cancelling queued jobs
1839 - Fixed off-by-one bug in job serial generation
1840 - ``gnt-node volumes``: Fix instance names
1841 - Fixed aliases in bash completion
1842 - Fixed a bug in reopening log files after being sent a SIGHUP
1843 - Added a flag to burnin to allow specifying VCPU count
1844 - Bugfixes to non-root Ganeti configuration
1850 *(Released Thu, 12 May 2011)*
1852 Many bug-fixes and a few new small features:
1854 - Fixed a bug related to log opening failures
1855 - Fixed a bug in instance listing with orphan instances
1856 - Fixed a bug which prevented resetting the cluster-level node parameter
1857 ``oob_program`` to the default
1858 - Many fixes related to the ``cluster-merge`` tool
1859 - Fixed a race condition in the lock monitor, which caused failures
1860 during (at least) creation of many instances in parallel
1861 - Improved output for gnt-job info
1862 - Removed the quiet flag on some ssh calls which prevented debugging
1864 - Improved the N+1 failure messages in cluster verify by actually
1865 showing the memory values (needed and available)
1866 - Increased lock attempt timeouts so that when executing long operations
1867 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1868 too early and thus prevent the use of the 'fair' mechanism
1869 - Changed instance query data (``gnt-instance info``) to not acquire
1870 locks unless needed, thus allowing its use on locked instance if only
1871 static information is asked for
1872 - Improved behaviour with filesystems that do not support rename on an
1874 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1875 kept locks on all nodes during the wipe, which is unneeded
1876 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1877 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1878 10GiB) which caused the wipe to fail right at the end in some cases
1879 - Fixed master IP activation when doing master failover with no-voting
1880 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1881 the master node itself
1882 - Fixed potential data-loss in under disk full conditions, where Ganeti
1883 wouldn't check correctly the return code and would consider
1884 partially-written files 'correct'
1885 - Fixed bug related to multiple VGs and DRBD disk replacing
1886 - Added new disk parameter ``metavg`` that allows placement of the meta
1887 device for DRBD in a different volume group
1888 - Fixed error handling in the node daemon when the system libc doesn't
1889 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1890 - Fixed lock release during replace-disks, which kept cluster-wide locks
1891 when doing disk replaces with an iallocator script
1892 - Added check for missing bridges in cluster verify
1893 - Handle EPIPE errors while writing to the terminal better, so that
1894 piping the output to e.g. ``less`` doesn't cause a backtrace
1895 - Fixed rare case where a ^C during Luxi calls could have been
1896 interpreted as server errors, instead of simply terminating
1897 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1899 - Added a few more parameters to the KVM hypervisor, allowing a second
1900 CDROM, custom disk type for CDROMs and a floppy image
1901 - Removed redundant message in instance rename when the name is given
1903 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1904 disks on new nodes, allowing recreation when the original instance
1905 nodes are completely gone
1906 - Added option when converting disk templates to DRBD to skip waiting
1907 for the resync, in order to make the instance available sooner
1908 - Added two new variables to the OS scripts environment (containing the
1910 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1911 to allow use of ``pvgrub`` as bootloader
1912 - Changed the instance memory modifications to only check out-of-memory
1913 conditions on memory increases, and turned the secondary node warnings
1914 into errors (they can still be overridden via ``--force``)
1915 - Fixed the handling of a corner case when the Python installation gets
1916 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1917 to execute a command that doesn't exist
1918 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1919 node of the instance returned failures during instance shutdown; this
1920 adds the option ``--ignore-consistency`` to gnt-instance move
1922 And as usual, various improvements to the error messages, documentation
1929 *(Released Wed, 09 Mar 2011)*
1931 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1932 the RAPI users file if run twice (even with ``--dry-run``).
1934 The release fixes that bug (nothing else changed).
1940 *(Released Mon, 07 Mar 2011)*
1942 Final 2.4.0 release. Just a few small fixes:
1944 - Fixed RAPI node evacuate
1945 - Fixed the kvm-ifup script
1946 - Fixed internal error handling for special job cases
1947 - Updated man page to specify the escaping feature for options
1953 *(Released Mon, 28 Feb 2011)*
1955 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1956 that this feature wiped the disks of the wrong instance, leading to loss
1961 - Fixed title of query field containing instance name
1962 - Expanded the glossary in the documentation
1963 - Fixed one unittest (internal issue)
1969 *(Released Mon, 21 Feb 2011)*
1971 A number of bug fixes plus just a couple functionality changes.
1973 On the user-visible side, the ``gnt-* list`` command output has changed
1974 with respect to "special" field states. The current rc1 style of display
1975 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1976 the default output mode special fields states are displayed as follows:
1978 - Offline resource: ``*``
1979 - Unavailable/not applicable: ``-``
1980 - Data missing (RPC failure): ``?``
1981 - Unknown field: ``??``
1983 Another user-visible change is the addition of ``--force-join`` to
1988 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1989 - Fixed regression in RAPI/instance reinstall where all parameters were
1990 required (instead of optional)
1991 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1992 - Fixed iallocator usage (offline nodes were not considered offline)
1993 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1994 - Fixed hypervisor and OS parameter validation with respect to
1995 non-vm_capable nodes
1996 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1998 - Fixed ``tools/listrunner`` with respect to agent-based usage
2004 *(Released Fri, 4 Feb 2011)*
2006 Many changes and fixes since the beta1 release. While there were some
2007 internal changes, the code has been mostly stabilised for the RC
2010 Note: the dumb allocator was removed in this release, as it was not kept
2011 up-to-date with the IAllocator protocol changes. It is recommended to
2012 use the ``hail`` command from the ganeti-htools package.
2014 Note: the 2.4 and up versions of Ganeti are not compatible with the
2015 0.2.x branch of ganeti-htools. You need to upgrade to
2016 ganeti-htools-0.3.0 (or later).
2018 Regressions fixed from 2.3
2019 ~~~~~~~~~~~~~~~~~~~~~~~~~~
2021 - Fixed the ``gnt-cluster verify-disks`` command
2022 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
2024 - Fixed disk adoption breakage
2025 - Fixed wrong headers in instance listing for field aliases
2030 - Fixed corner case in KVM handling of NICs
2031 - Fixed many cases of wrong handling of non-vm_capable nodes
2032 - Fixed a bug where a missing instance symlink was not possible to
2033 recreate with any ``gnt-*`` command (now ``gnt-instance
2034 activate-disks`` does it)
2035 - Fixed the volume group name as reported by ``gnt-cluster
2037 - Increased timeouts for the import-export code, hopefully leading to
2038 fewer aborts due network or instance timeouts
2039 - Fixed bug in ``gnt-node list-storage``
2040 - Fixed bug where not all daemons were started on cluster
2041 initialisation, but only at the first watcher run
2042 - Fixed many bugs in the OOB implementation
2043 - Fixed watcher behaviour in presence of instances with offline
2045 - Fixed instance list output for instances running on the wrong node
2046 - a few fixes to the cluster-merge tool, but it still cannot merge
2047 multi-node groups (currently it is not recommended to use this tool)
2053 - Improved network configuration for the KVM hypervisor
2054 - Added e1000 as a supported NIC for Xen-HVM
2055 - Improved the lvmstrap tool to also be able to use partitions, as
2056 opposed to full disks
2057 - Improved speed of disk wiping (the cluster parameter
2058 ``prealloc_wipe_disks``, so that it has a low impact on the total time
2059 of instance creations
2060 - Added documentation for the OS parameters
2061 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
2062 hypervisor is not responding
2063 - Added display of blacklisted and hidden OS information in
2064 ``gnt-cluster info``
2065 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
2066 NIC and node parameters, which might create problems with currently
2067 invalid (but undetected) configuration files, but prevents validation
2068 failures when unrelated parameters are modified
2069 - Changed cluster initialisation to wait for the master daemon to become
2071 - Expanded the RAPI interface:
2073 - Added config redistribution resource
2074 - Added activation/deactivation of instance disks
2075 - Added export of console information
2077 - Implemented log file reopening on SIGHUP, which allows using
2078 logrotate(8) for the Ganeti log files
2079 - Added a basic OOB helper script as an example
2085 *(Released Fri, 14 Jan 2011)*
2090 - Fixed timezone issues when formatting timestamps
2091 - Added support for node groups, available via ``gnt-group`` and other
2093 - Added out-of-band framework and management, see :doc:`design
2094 document <design-oob>`
2095 - Removed support for roman numbers from ``gnt-node list`` and
2096 ``gnt-instance list``.
2097 - Allowed modification of master network interface via ``gnt-cluster
2098 modify --master-netdev``
2099 - Accept offline secondaries while shutting down instance disks
2100 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
2101 - Added support for multiple LVM volume groups
2102 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
2104 - Added commands to list available fields:
2106 - ``gnt-node list-fields``
2107 - ``gnt-group list-fields``
2108 - ``gnt-instance list-fields``
2110 - Updated documentation and man pages
2115 - Moved ``rapi_users`` file into separate directory, now named
2116 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
2118 - Added new tool for running commands on many machines,
2119 ``tools/ganeti-listrunner``
2120 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
2121 improving the ``gnt-instance console`` output
2122 - Allowed customisation of disk index separator at ``configure`` time
2123 - Export node group allocation policy to :doc:`iallocator <iallocator>`
2124 - Added support for non-partitioned md disks in ``lvmstrap``
2125 - Added script to gracefully power off KVM instances
2126 - Split ``utils`` module into smaller parts
2127 - Changed query operations to return more detailed information, e.g.
2128 whether an information is unavailable due to an offline node. To use
2129 this new functionality, the LUXI call ``Query`` must be used. Field
2130 information is now stored by the master daemon and can be retrieved
2131 using ``QueryFields``. Instances, nodes and groups can also be queried
2132 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
2133 exposed via RAPI). The following commands make use of this
2134 infrastructure change:
2136 - ``gnt-group list``
2137 - ``gnt-group list-fields``
2139 - ``gnt-node list-fields``
2140 - ``gnt-instance list``
2141 - ``gnt-instance list-fields``
2142 - ``gnt-debug locks``
2147 - New RAPI resources (see :doc:`rapi`):
2151 - ``/2/groups/[group_name]``
2152 - ``/2/groups/[group_name]/assign-nodes``
2153 - ``/2/groups/[group_name]/modify``
2154 - ``/2/groups/[group_name]/rename``
2155 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
2159 - Implemented ``no_install`` for instance creation
2160 - Implemented OS parameters for instance reinstallation, allowing
2161 use of special settings on reinstallation (e.g. for preserving data)
2166 - Added IPv6 support in import/export
2167 - Pause DRBD synchronization while wiping disks on instance creation
2168 - Updated unittests and QA scripts
2169 - Improved network parameters passed to KVM
2170 - Converted man pages from docbook to reStructuredText
2176 *(Released Mon, 20 Dec 2010)*
2178 Released version 2.3.1~rc1 without any changes.
2184 *(Released Wed, 1 Dec 2010)*
2186 - impexpd: Disable OpenSSL compression in socat if possible (backport
2187 from master, commit e90739d625b, see :doc:`installation guide
2188 <install-quick>` for details)
2189 - Changed unittest coverage report to exclude test scripts
2190 - Added script to check version format
2196 *(Released Wed, 1 Dec 2010)*
2198 Released version 2.3.0~rc1 without any changes.
2204 *(Released Fri, 19 Nov 2010)*
2206 A number of bugfixes and documentation updates:
2208 - Update ganeti-os-interface documentation
2209 - Fixed a bug related to duplicate MACs or similar items which should be
2211 - Fix breakage in OS state modify
2212 - Reinstall instance: disallow offline secondaries (fixes bug related to
2213 OS changing but reinstall failing)
2214 - plus all the other fixes between 2.2.1 and 2.2.2
2220 *(Released Tue, 2 Nov 2010)*
2222 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
2223 - Fixed master failover race with watcher
2224 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
2226 - Accept previously stopped instance for export with instance removal
2227 - Simplify and extend the environment variables for instance OS scripts
2228 - Added new node flags, ``master_capable`` and ``vm_capable``
2229 - Added optional instance disk wiping prior during allocation. This is a
2230 cluster-wide option and can be set/modified using
2231 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
2232 - Added IPv6 support, see :doc:`design document <design-2.3>` and
2233 :doc:`install-quick`
2234 - Added a new watcher option (``--ignore-pause``)
2235 - Added option to ignore offline node on instance start/stop
2236 (``--ignore-offline``)
2237 - Allow overriding OS parameters with ``gnt-instance reinstall``
2238 - Added ability to change node's secondary IP address using ``gnt-node
2240 - Implemented privilege separation for all daemons except
2241 ``ganeti-noded``, see ``configure`` options
2242 - Complain if an instance's disk is marked faulty in ``gnt-cluster
2244 - Implemented job priorities (see ``ganeti(7)`` manpage)
2245 - Ignore failures while shutting down instances during failover from
2247 - Exit daemon's bootstrap process only once daemon is ready
2248 - Export more information via ``LUInstanceQuery``/remote API
2249 - Improved documentation, QA and unittests
2250 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
2251 restart if the file was created or changed
2252 - Added LUXI protocol version sent with each request and response,
2253 allowing detection of server/client mismatches
2254 - Moved the Python scripts among gnt-* and ganeti-* into modules
2255 - Moved all code related to setting up SSH to an external script,
2257 - Infrastructure changes for node group support in future versions
2263 *(Released Fri, 19 Nov 2010)*
2265 A few small bugs fixed, and some improvements to the build system:
2267 - Fix documentation regarding conversion to drbd
2268 - Fix validation of parameters in cluster modify (``gnt-cluster modify
2270 - Fix error handling in node modify with multiple changes
2271 - Allow remote imports without checked names
2277 *(Released Tue, 19 Oct 2010)*
2279 - Disable SSL session ID cache in RPC client
2285 *(Released Thu, 14 Oct 2010)*
2287 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
2288 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
2290 - Fix problems with interaction between Curl and Python's HTTP server,
2291 resulting in increased speed in many RPC calls
2292 - Improve our release script to prevent breakage with older aclocal and
2299 *(Released Thu, 7 Oct 2010)*
2301 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
2302 value retrieved from master
2303 - Added support for blacklisted or hidden OS definitions
2304 - Added simple lock monitor (accessible via (``gnt-debug locks``)
2305 - Added support for -mem-path in KVM hypervisor abstraction layer
2306 - Allow overriding instance parameters in tool for inter-cluster
2307 instance moves (``tools/move-instance``)
2308 - Improved opcode summaries (e.g. in ``gnt-job list``)
2309 - Improve consistency of OS listing by sorting it
2310 - Documentation updates
2316 *(Released Fri, 8 Oct 2010)*
2318 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
2324 *(Released Mon, 4 Oct 2010)*
2326 - Fixed regression in ``gnt-instance rename``
2332 *(Released Wed, 22 Sep 2010)*
2334 - Fixed OS_VARIANT variable for OS scripts
2335 - Fixed cluster tag operations via RAPI
2336 - Made ``setup-ssh`` exit with non-zero code if an error occurred
2337 - Disabled RAPI CA checks in watcher
2343 *(Released Mon, 23 Aug 2010)*
2345 - Support DRBD versions of the format "a.b.c.d"
2347 - Re-introduce support for usage from multiple threads in RAPI client
2348 - Instance renames and modify via RAPI
2349 - Work around race condition between processing and archival in job
2351 - Mark opcodes following failed one as failed, too
2352 - Job field ``lock_status`` was removed due to difficulties making it
2353 work with the changed job queue in Ganeti 2.2; a better way to monitor
2354 locks is expected for a later 2.2.x release
2355 - Fixed dry-run behaviour with many commands
2356 - Support ``ssh-agent`` again when adding nodes
2357 - Many additional bugfixes
2363 *(Released Fri, 30 Jul 2010)*
2365 Important change: the internal RPC mechanism between Ganeti nodes has
2366 changed from using a home-grown http library (based on the Python base
2367 libraries) to use the PycURL library. This requires that PycURL is
2368 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2369 against GnuTLS by default. cURL's support for GnuTLS had known issues
2370 before cURL 7.21.0 and we recommend using the latest cURL release or
2371 linking against OpenSSL. Most other distributions already link PycURL
2372 and cURL against OpenSSL. The command::
2374 python -c 'import pycurl; print pycurl.version'
2376 can be used to determine the libraries PycURL and cURL are linked
2379 Other significant changes:
2381 - Rewrote much of the internals of the job queue, in order to achieve
2382 better parallelism; this decouples job query operations from the job
2383 processing, and it should allow much nicer behaviour of the master
2384 daemon under load, and it also has uncovered some long-standing bugs
2385 related to the job serialisation (now fixed)
2386 - Added a default iallocator setting to the cluster parameters,
2387 eliminating the need to always pass nodes or an iallocator for
2388 operations that require selection of new node(s)
2389 - Added experimental support for the LXC virtualization method
2390 - Added support for OS parameters, which allows the installation of
2391 instances to pass parameter to OS scripts in order to customise the
2393 - Added a hypervisor parameter controlling the migration type (live or
2394 non-live), since hypervisors have various levels of reliability; this
2395 has renamed the 'live' parameter to 'mode'
2396 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
2397 logical volumes, meaning that cluster verify will ignore them and not
2398 flag their presence as errors
2399 - The watcher will now reset the error count for failed instances after
2400 8 hours, thus allowing self-healing if the problem that caused the
2401 instances to be down/fail to start has cleared in the meantime
2402 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2403 check for, and warn, if the drbd module parameter ``usermode_helper``
2404 is not consistent with the cluster-wide setting; this is needed to
2405 make diagnose easier of failed drbd creations
2406 - Started adding base IPv6 support, but this is not yet
2407 enabled/available for use
2408 - Rename operations (cluster, instance) will now return the new name,
2409 which is especially useful if a short name was passed in
2410 - Added support for instance migration in RAPI
2411 - Added a tool to pre-configure nodes for the SSH setup, before joining
2412 them to the cluster; this will allow in the future a simplified model
2413 for node joining (but not yet fully enabled in 2.2); this needs the
2414 paramiko python library
2415 - Fixed handling of name-resolving errors
2416 - Fixed consistency of job results on the error path
2417 - Fixed master-failover race condition when executed multiple times in
2419 - Fixed many bugs related to the job queue (mostly introduced during the
2420 2.2 development cycle, so not all are impacting 2.1)
2421 - Fixed instance migration with missing disk symlinks
2422 - Fixed handling of unknown jobs in ``gnt-job archive``
2423 - And many other small fixes/improvements
2427 - Enhanced both the unittest and the QA coverage
2428 - Switched the opcode validation to a generic model, and extended the
2429 validation to all opcode parameters
2430 - Changed more parts of the code that write shell scripts to use the
2432 - Switched the master daemon to use the asyncore library for the Luxi
2439 *(Released Thu, 17 Jun 2010)*
2441 - Added tool (``move-instance``) and infrastructure to move instances
2442 between separate clusters (see :doc:`separate documentation
2443 <move-instance>` and :doc:`design document <design-2.2>`)
2444 - Added per-request RPC timeout
2445 - RAPI now requires a Content-Type header for requests with a body (e.g.
2446 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2447 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2448 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2450 - Implemented initial support for running Ganeti daemons as separate
2451 users, see configure-time flags ``--with-user-prefix`` and
2452 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2454 - Instances can be removed after export (``gnt-backup export
2455 --remove-instance``)
2456 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2457 key (instead of 1024 bit)
2458 - Added new cluster configuration file for cluster domain secret
2459 - Import/export now use SSL instead of SSH
2460 - Added support for showing estimated time when exporting an instance,
2461 see the ``ganeti-os-interface(7)`` manpage and look for
2468 *(Released Tue, 16 Nov 2010)*
2470 Some more bugfixes. Unless critical bugs occur, this will be the last
2473 - Fix case of MAC special-values
2474 - Fix mac checker regex
2475 - backend: Fix typo causing "out of range" error
2476 - Add missing --units in gnt-instance list man page
2482 *(Released Tue, 24 Aug 2010)*
2485 - Don't ignore secondary node silently on non-mirrored disk templates
2487 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2488 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2489 - Properly document the usb_mouse parameter
2490 - Fix path in ganeti-rapi(8) (issue 116)
2491 - Adjust error message when the ganeti user's .ssh directory is
2493 - Add same-node-check when changing the disk template to drbd
2499 *(Released Fri, 16 Jul 2010)*
2502 - Add an option to only select some reboot types during qa/burnin.
2503 (on some hypervisors consequent reboots are not supported)
2504 - Fix infrequent race condition in master failover. Sometimes the old
2505 master ip address would be still detected as up for a short time
2506 after it was removed, causing failover to fail.
2507 - Decrease mlockall warnings when the ctypes module is missing. On
2508 Python 2.4 we support running even if no ctypes module is installed,
2509 but we were too verbose about this issue.
2510 - Fix building on old distributions, on which man doesn't have a
2512 - Fix RAPI not to ignore the MAC address on instance creation
2513 - Implement the old instance creation format in the RAPI client.
2519 *(Released Thu, 01 Jul 2010)*
2521 A small bugfix release:
2522 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2523 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2524 a non-existing option
2525 - Fix instance create: the --force-variant option was ignored
2526 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2527 - Fix modify node storage with non-FQDN arguments
2528 - Fix RAPI client to authenticate under Python 2.6 when used
2529 for more than 5 requests needing authentication
2530 - Fix gnt-instance modify -t (storage) giving a wrong error message
2531 when converting a non-shutdown drbd instance to plain
2537 *(Released Fri, 18 Jun 2010)*
2539 A small bugfix release:
2541 - Fix live migration of KVM instances started with older Ganeti
2542 versions which had fewer hypervisor parameters
2543 - Fix gnt-instance grow-disk on down instances
2544 - Fix an error-reporting bug during instance migration
2545 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2546 silently ignoring broken ones
2547 - Fix an RPC error reporting bug affecting, for example, RAPI client
2549 - Fix bug triggered by different API version os-es on different nodes
2550 - Fix a bug in instance startup with custom hvparams: OS level
2551 parameters would fail to be applied.
2552 - Fix the RAPI client under Python 2.6 (but more work is needed to
2553 make it work completely well with OpenSSL)
2554 - Fix handling of errors when resolving names from DNS
2560 *(Released Thu, 3 Jun 2010)*
2562 A medium sized development cycle. Some new features, and some
2563 fixes/small improvements/cleanups.
2565 Significant features
2566 ~~~~~~~~~~~~~~~~~~~~
2568 The node deamon now tries to mlock itself into memory, unless the
2569 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2570 its logs, and falls back to console logging. This allows emergency
2571 features such as ``gnt-node powercycle`` to work even in the event of a
2572 broken node disk (tested offlining the disk hosting the node's
2573 filesystem and dropping its memory caches; don't try this at home)
2575 KVM: add vhost-net acceleration support. It can be tested with a new
2576 enough version of the kernel and of qemu-kvm.
2578 KVM: Add instance chrooting feature. If you use privilege dropping for
2579 your VMs you can also now force them to chroot to an empty directory,
2580 before starting the emulated guest.
2582 KVM: Add maximum migration bandwith and maximum downtime tweaking
2583 support (requires a new-enough version of qemu-kvm).
2585 Cluster verify will now warn if the master node doesn't have the master
2586 ip configured on it.
2588 Add a new (incompatible) instance creation request format to RAPI which
2589 supports all parameters (previously only a subset was supported, and it
2590 wasn't possible to extend the old format to accomodate all the new
2591 features. The old format is still supported, and a client can check for
2592 this feature, before using it, by checking for its presence in the
2593 ``features`` RAPI resource.
2595 Now with ancient latin support. Try it passing the ``--roman`` option to
2596 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2597 (requires the python-roman module to be installed, in order to work).
2602 As usual many internal code refactorings, documentation updates, and
2605 - Lots of improvements and cleanups to the experimental Remote API
2606 (RAPI) client library.
2607 - A new unit test suite for the core daemon libraries.
2608 - A fix to creating missing directories makes sure the umask is not
2609 applied anymore. This enforces the same directory permissions
2611 - Better handling terminating daemons with ctrl+c (used when running
2612 them in debugging mode).
2613 - Fix a race condition in live migrating a KVM instance, when stat()
2614 on the old proc status file returned EINVAL, which is an unexpected
2616 - Fixed manpage checking with newer man and utf-8 charachters. But now
2617 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2623 *(Released Fri, 7 May 2010)*
2625 Fix a bug which prevented untagged KVM instances from starting.
2631 *(Released Fri, 7 May 2010)*
2633 Another release with a long development cycle, during which many
2634 different features were added.
2636 Significant features
2637 ~~~~~~~~~~~~~~~~~~~~
2639 The KVM hypervisor now can run the individual instances as non-root, to
2640 reduce the impact of a VM being hijacked due to bugs in the
2641 hypervisor. It is possible to run all instances as a single (non-root)
2642 user, to manually specify a user for each instance, or to dynamically
2643 allocate a user out of a cluster-wide pool to each instance, with the
2644 guarantee that no two instances will run under the same user ID on any
2647 An experimental RAPI client library, that can be used standalone
2648 (without the other Ganeti libraries), is provided in the source tree as
2649 ``lib/rapi/client.py``. Note this client might change its interface in
2650 the future, as we iterate on its capabilities.
2652 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2653 replace the cluster's certificates and crypto keys. This might help in
2654 case they have been compromised, or have simply expired.
2656 A new disk option for instance creation has been added that allows one
2657 to "adopt" currently existing logical volumes, with data
2658 preservation. This should allow easier migration to Ganeti from
2659 unmanaged (or managed via other software) instances.
2661 Another disk improvement is the possibility to convert between redundant
2662 (DRBD) and plain (LVM) disk configuration for an instance. This should
2663 allow better scalability (starting with one node and growing the
2664 cluster, or shrinking a two-node cluster to one node).
2666 A new feature that could help with automated node failovers has been
2667 implemented: if a node sees itself as offline (by querying the master
2668 candidates), it will try to shutdown (hard) all instances and any active
2669 DRBD devices. This reduces the risk of duplicate instances if an
2670 external script automatically failovers the instances on such nodes. To
2671 enable this, the cluster parameter ``maintain_node_health`` should be
2672 enabled; in the future this option (per the name) will enable other
2673 automatic maintenance features.
2675 Instance export/import now will reuse the original instance
2676 specifications for all parameters; that means exporting an instance,
2677 deleting it and the importing it back should give an almost identical
2678 instance. Note that the default import behaviour has changed from
2679 before, where it created only one NIC; now it recreates the original
2682 Cluster verify has added a few new checks: SSL certificates validity,
2683 /etc/hosts consistency across the cluster, etc.
2688 As usual, many internal changes were done, documentation fixes,
2691 - Fixed cluster initialization with disabled cluster storage (regression
2692 introduced in 2.1.1)
2693 - File-based storage supports growing the disks
2694 - Fixed behaviour of node role changes
2695 - Fixed cluster verify for some corner cases, plus a general rewrite of
2696 cluster verify to allow future extension with more checks
2697 - Fixed log spamming by watcher and node daemon (regression introduced
2699 - Fixed possible validation issues when changing the list of enabled
2701 - Fixed cleanup of /etc/hosts during node removal
2702 - Fixed RAPI response for invalid methods
2703 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2704 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2705 from ide disks, etc.)
2706 - Allow OS changes without re-installation (to record a changed OS
2707 outside of Ganeti, or to allow OS renames)
2708 - Allow instance creation without OS installation (useful for example if
2709 the OS will be installed manually, or restored from a backup not in
2711 - Implemented option to make cluster ``copyfile`` use the replication
2713 - Added list of enabled hypervisors to ssconf (possibly useful for
2715 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2717 - A partial form of node re-IP is possible via node readd, which now
2718 allows changed node primary IP
2719 - Command line utilities now show an informational message if the job is
2721 - The logs of the master daemon now show the PID/UID/GID of the
2728 *(Released Fri, 12 Mar 2010)*
2730 During the 2.1.0 long release candidate cycle, a lot of improvements and
2731 changes have accumulated with were released later as 2.1.1.
2736 The node evacuate command (``gnt-node evacuate``) was significantly
2737 rewritten, and as such the IAllocator protocol was changed - a new
2738 request type has been added. This unfortunate change during a stable
2739 series is designed to improve performance of node evacuations; on
2740 clusters with more than about five nodes and which are well-balanced,
2741 evacuation should proceed in parallel for all instances of the node
2742 being evacuated. As such, any existing IAllocator scripts need to be
2743 updated, otherwise the above command will fail due to the unknown
2744 request. The provided "dumb" allocator has not been updated; but the
2745 ganeti-htools package supports the new protocol since version 0.2.4.
2747 Another important change is increased validation of node and instance
2748 names. This might create problems in special cases, if invalid host
2749 names are being used.
2751 Also, a new layer of hypervisor parameters has been added, that sits at
2752 OS level between the cluster defaults and the instance ones. This allows
2753 customisation of virtualization parameters depending on the installed
2754 OS. For example instances with OS 'X' may have a different KVM kernel
2755 (or any other parameter) than the cluster defaults. This is intended to
2756 help managing a multiple OSes on the same cluster, without manual
2757 modification of each instance's parameters.
2759 A tool for merging clusters, ``cluster-merge``, has been added in the
2760 tools sub-directory.
2765 - Improved the int/float conversions that should make the code more
2766 robust in face of errors from the node daemons
2767 - Fixed the remove node code in case of internal configuration errors
2768 - Fixed the node daemon behaviour in face of inconsistent queue
2769 directory (e.g. read-only file-system where we can't open the files
2771 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2772 now it either aborts cleanly or, if given the new "auto_promote"
2773 parameter, will automatically promote other nodes as needed
2774 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2775 completely prevent Ganeti from working
2776 - Fixed bug for instance export when not all disks were successfully
2778 - Fixed behaviour of node add when the new node is slow in starting up
2780 - Fixed handling of signals in the LUXI client, which should improve
2781 behaviour of command-line scripts
2782 - Added checks for invalid node/instance names in the configuration (now
2783 flagged during cluster verify)
2784 - Fixed watcher behaviour for disk activation errors
2785 - Fixed two potentially endless loops in http library, which led to the
2786 RAPI daemon hanging and consuming 100% CPU in some cases
2787 - Fixed bug in RAPI daemon related to hashed passwords
2788 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2790 - Enhanced compatibility with non-Debian OSes, but not using absolute
2791 path in some commands and allowing customisation of the ssh
2792 configuration directory
2793 - Fixed possible future issue with new Python versions by abiding to the
2794 proper use of ``__slots__`` attribute on classes
2795 - Added checks that should prevent directory traversal attacks
2796 - Many documentation fixes based on feedback from users
2801 - Added an "early_release" more for instance replace disks and node
2802 evacuate, where we release locks earlier and thus allow higher
2803 parallelism within the cluster
2804 - Added watcher hooks, intended to allow the watcher to restart other
2805 daemons (e.g. from the ganeti-nbma project), but they can be used of
2806 course for any other purpose
2807 - Added a compile-time disable for DRBD barriers, to increase
2808 performance if the administrator trusts the power supply or the
2809 storage system to not lose writes
2810 - Added the option of using syslog for logging instead of, or in
2811 addition to, Ganeti's own log files
2812 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2813 can indeed boot from a paravirtual NIC
2814 - Added a generic debug level for many operations; while this is not
2815 used widely yet, it allows one to pass the debug value all the way to
2817 - Enhanced the hooks environment for instance moves (failovers,
2818 migrations) where the primary/secondary nodes changed during the
2819 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2820 - Enhanced data validations for many user-supplied values; one important
2821 item is the restrictions imposed on instance and node names, which
2822 might reject some (invalid) host names
2823 - Add a configure-time option to disable file-based storage, if it's not
2824 needed; this allows greater security separation between the master
2825 node and the other nodes from the point of view of the inter-node RPC
2827 - Added user notification in interactive tools if job is waiting in the
2828 job queue or trying to acquire locks
2829 - Added log messages when a job is waiting for locks
2830 - Added filtering by node tags in instance operations which admit
2831 multiple instances (start, stop, reboot, reinstall)
2832 - Added a new tool for cluster mergers, ``cluster-merge``
2833 - Parameters from command line which are of the form ``a=b,c=d`` can now
2834 use backslash escapes to pass in values which contain commas,
2835 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2837 - For KVM, the instance name is the first parameter passed to KVM, so
2838 that it's more visible in the process list
2844 *(Released Tue, 2 Mar 2010)*
2846 Ganeti 2.1 brings many improvements with it. Major changes:
2848 - Added infrastructure to ease automated disk repairs
2849 - Added new daemon to export configuration data in a cheaper way than
2850 using the remote API
2851 - Instance NICs can now be routed instead of being associated with a
2853 - Improved job locking logic to reduce impact of jobs acquiring multiple
2854 locks waiting for other long-running jobs
2856 In-depth implementation details can be found in the Ganeti 2.1 design
2862 - Added chroot hypervisor
2863 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2865 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2866 ``bootloader_path`` and ``bootloader_args``)
2867 - Added the ``use_localtime`` option for the xen-hvm and kvm
2868 hypervisors, and the default value for this has changed to false (in
2869 2.0 xen-hvm always enabled it)
2870 - Added luxi call to submit multiple jobs in one go
2871 - Added cluster initialization option to not modify ``/etc/hosts``
2873 - Added network interface parameters
2874 - Added dry run mode to some LUs
2875 - Added RAPI resources:
2877 - ``/2/instances/[instance_name]/info``
2878 - ``/2/instances/[instance_name]/replace-disks``
2879 - ``/2/nodes/[node_name]/evacuate``
2880 - ``/2/nodes/[node_name]/migrate``
2881 - ``/2/nodes/[node_name]/role``
2882 - ``/2/nodes/[node_name]/storage``
2883 - ``/2/nodes/[node_name]/storage/modify``
2884 - ``/2/nodes/[node_name]/storage/repair``
2886 - Added OpCodes to evacuate or migrate all instances on a node
2887 - Added new command to list storage elements on nodes (``gnt-node
2888 list-storage``) and modify them (``gnt-node modify-storage``)
2889 - Added new ssconf files with master candidate IP address
2890 (``ssconf_master_candidates_ips``), node primary IP address
2891 (``ssconf_node_primary_ips``) and node secondary IP address
2892 (``ssconf_node_secondary_ips``)
2893 - Added ``ganeti-confd`` and a client library to query the Ganeti
2894 configuration via UDP
2895 - Added ability to run hooks after cluster initialization and before
2897 - Added automatic mode for disk replace (``gnt-instance replace-disks
2899 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2900 after catastrophic data-loss
2901 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2903 - Added ``gnt-instance move`` command to move instances
2904 - Added ``gnt-cluster watcher`` command to control watcher
2905 - Added ``gnt-node powercycle`` command to powercycle nodes
2906 - Added new job status field ``lock_status``
2907 - Added parseable error codes to cluster verification (``gnt-cluster
2908 verify --error-codes``) and made output less verbose (use
2909 ``--verbose`` to restore previous behaviour)
2910 - Added UUIDs to the main config entities (cluster, nodes, instances)
2911 - Added support for OS variants
2912 - Added support for hashed passwords in the Ganeti remote API users file
2914 - Added option to specify maximum timeout on instance shutdown
2915 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2916 - Added new helper script to start and stop Ganeti daemons
2917 (``daemon-util``), with the intent to reduce the work necessary to
2918 adjust Ganeti for non-Debian distributions and to start/stop daemons
2920 - Added more unittests
2921 - Fixed critical bug in ganeti-masterd startup
2922 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2923 now customisable at the cluster level for both the KVM and Xen
2924 hypervisors using the new ``migration_port`` parameter
2925 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2926 reinstalling an instance
2927 - Allowed ``@`` in tag names
2928 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2929 - Many documentation updates
2930 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2931 - ``gnt-instance reinstall`` can now reinstall multiple instances
2932 - Updated many command line parameters
2933 - Introduced new OS API version 15
2934 - No longer support a default hypervisor
2935 - Treat virtual LVs as inexistent
2936 - Improved job locking logic to reduce lock contention
2937 - Match instance and node names case insensitively
2938 - Reimplemented bash completion script to be more complete
2945 *(Released Thu, 4 Feb 2010)*
2947 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2948 - Fix a string formatting bug
2949 - Improve safety of the code in some error paths
2950 - Improve data validation in the master of values returned from nodes
2956 *(Released Thu, 17 Dec 2009)*
2958 - Fix security issue due to missing validation of iallocator names; this
2959 allows local and remote execution of arbitrary executables
2960 - Fix failure of gnt-node list during instance removal
2961 - Ship the RAPI documentation in the archive
2967 *(Released Wed, 30 Sep 2009)*
2969 - Fixed many wrong messages
2970 - Fixed a few bugs related to the locking library
2971 - Fixed MAC checking at instance creation time
2972 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2973 - Fixed a few issues related to signal handling in both daemons and
2975 - Fixed the example startup script provided
2976 - Fixed insserv dependencies in the example startup script (patch from
2978 - Fixed handling of drained nodes in the iallocator framework
2979 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2981 - Fixed error related to invalid job IDs in job polling
2982 - Fixed job/opcode persistence on unclean master shutdown
2983 - Fixed handling of partial job processing after unclean master
2985 - Fixed error reporting from LUs, previously all errors were converted
2986 into execution errors
2987 - Fixed error reporting from burnin
2988 - Decreased significantly the memory usage of the job queue
2989 - Optimised slightly multi-job submission
2990 - Optimised slightly opcode loading
2991 - Backported the multi-job submit framework from the development
2992 branch; multi-instance start and stop should be faster
2993 - Added script to clean archived jobs after 21 days; this will reduce
2994 the size of the queue directory
2995 - Added some extra checks in disk size tracking
2996 - Added an example ethers hook script
2997 - Added a cluster parameter that prevents Ganeti from modifying of
2999 - Added more node information to RAPI responses
3000 - Added a ``gnt-job watch`` command that allows following the ouput of a
3002 - Added a bind-address option to ganeti-rapi
3003 - Added more checks to the configuration verify
3004 - Enhanced the burnin script such that some operations can be retried
3006 - Converted instance reinstall to multi-instance model
3012 *(Released Fri, 7 Aug 2009)*
3014 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
3015 to allow using the pre-2.0.2 behaviour in activation, if any existing
3016 instances have mismatched disk sizes in the configuration
3017 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
3018 any configuration mismatches for disk sizes
3019 - Added ``gnt-master cluste-failover --no-voting`` to allow master
3020 failover to work on two-node clusters
3021 - Fixed the ``--net`` option of ``gnt-backup import``, which was
3023 - Fixed detection of OS script errors in ``gnt-backup export``
3024 - Fixed exit code of ``gnt-backup export``
3030 *(Released Fri, 17 Jul 2009)*
3032 - Added experimental support for stripped logical volumes; this should
3033 enhance performance but comes with a higher complexity in the block
3034 device handling; stripping is only enabled when passing
3035 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
3036 affected even in the non-stripped mode
3037 - Improved resiliency against transient failures at the end of DRBD
3038 resyncs, and in general of DRBD resync checks
3039 - Fixed a couple of issues with exports and snapshot errors
3040 - Fixed a couple of issues in instance listing
3041 - Added display of the disk size in ``gnt-instance info``
3042 - Fixed checking for valid OSes in instance creation
3043 - Fixed handling of the "vcpus" parameter in instance listing and in
3044 general of invalid parameters
3045 - Fixed http server library, and thus RAPI, to handle invalid
3046 username/password combinations correctly; this means that now they
3047 report unauthorized for queries too, not only for modifications,
3048 allowing earlier detect of configuration problems
3049 - Added a new "role" node list field, equivalent to the master/master
3050 candidate/drained/offline flags combinations
3051 - Fixed cluster modify and changes of candidate pool size
3052 - Fixed cluster verify error messages for wrong files on regular nodes
3053 - Fixed a couple of issues with node demotion from master candidate role
3054 - Fixed node readd issues
3055 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
3056 - Added a new ``--no-voting`` option for masterfailover to fix failover
3057 on two-nodes clusters when the former master node is unreachable
3058 - Added instance reinstall over RAPI
3064 *(Released Tue, 16 Jun 2009)*
3066 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
3067 allow re-adding the start in single-user option (regression from 1.2)
3068 - the watcher writes the instance status to a file, to allow monitoring
3069 to report the instance status (from the master) based on cached
3070 results of the watcher's queries; while this can get stale if the
3071 watcher is being locked due to other work on the cluster, this is
3072 still an improvement
3073 - the watcher now also restarts the node daemon and the rapi daemon if
3075 - fixed the watcher to handle full and drained queue cases
3076 - hooks export more instance data in the environment, which helps if
3077 hook scripts need to take action based on the instance's properties
3078 (no longer need to query back into ganeti)
3079 - instance failovers when the instance is stopped do not check for free
3080 RAM, so that failing over a stopped instance is possible in low memory
3082 - rapi uses queries for tags instead of jobs (for less job traffic), and
3083 for cluster tags it won't talk to masterd at all but read them from
3085 - a couple of error handling fixes in RAPI
3086 - drbd handling: improved the error handling of inconsistent disks after
3087 resync to reduce the frequency of "there are some degraded disks for
3088 this instance" messages
3089 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
3090 error handling path called a wrong function name)
3096 *(Released Wed, 27 May 2009)*
3098 - no changes from rc5
3104 *(Released Wed, 20 May 2009)*
3106 - fix a couple of bugs (validation, argument checks)
3107 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
3108 - some small improvements to RAPI and IAllocator
3109 - make watcher automatically start the master daemon if down
3115 *(Released Mon, 27 Apr 2009)*
3117 - change the OS list to not require locks; this helps with big clusters
3118 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
3119 volume group is broken
3120 - ``gnt-instance info``, without any arguments, doesn't run for all
3121 instances anymore; either pass ``--all`` or pass the desired
3122 instances; this helps against mistakes on big clusters where listing
3123 the information for all instances takes a long time
3124 - miscellaneous doc and man pages fixes
3130 *(Released Wed, 8 Apr 2009)*
3132 - Change the internal locking model of some ``gnt-node`` commands, in
3133 order to reduce contention (and blocking of master daemon) when
3134 batching many creation/reinstall jobs
3135 - Fixes to Xen soft reboot
3136 - No longer build documentation at build time, instead distribute it in
3137 the archive, in order to reduce the need for the whole docbook/rst
3144 *(Released Fri, 27 Mar 2009)*
3146 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
3147 - Fix watcher startup sequence, improves the behaviour of busy clusters
3148 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
3149 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
3150 bind address and other places
3151 - Some documentation fixes and updates
3157 *(Released Mon, 2 Mar 2009)*
3159 - More documentation updates, now all docs should be more-or-less
3161 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
3163 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
3165 - Fix an issue related to $libdir/run/ganeti and cluster creation
3171 *(Released Thu, 19 Feb 2009)*
3173 - Xen PVM and KVM have switched the default value for the instance root
3174 disk to the first partition on the first drive, instead of the whole
3175 drive; this means that the OS installation scripts must be changed
3177 - Man pages have been updated
3178 - RAPI has been switched by default to HTTPS, and the exported functions
3179 should all work correctly
3180 - RAPI v1 has been removed
3181 - Many improvements to the KVM hypervisor
3182 - Block device errors are now better reported
3183 - Many other bugfixes and small improvements
3189 *(Released Mon, 26 Jan 2009)*
3191 - Version 2 is a general rewrite of the code and therefore the
3192 differences are too many to list, see the design document for 2.0 in
3193 the ``doc/`` subdirectory for more details
3194 - In this beta version there is not yet a migration path from 1.2 (there
3195 will be one in the final 2.0 release)
3196 - A few significant changes are:
3198 - all commands are executed by a daemon (``ganeti-masterd``) and the
3199 various ``gnt-*`` commands are just front-ends to it
3200 - all the commands are entered into, and executed from a job queue,
3201 see the ``gnt-job(8)`` manpage
3202 - the RAPI daemon supports read-write operations, secured by basic
3203 HTTP authentication on top of HTTPS
3204 - DRBD version 0.7 support has been removed, DRBD 8 is the only
3205 supported version (when migrating from Ganeti 1.2 to 2.0, you need
3206 to migrate to DRBD 8 first while still running Ganeti 1.2)
3207 - DRBD devices are using statically allocated minor numbers, which
3208 will be assigned to existing instances during the migration process
3209 - there is support for both Xen PVM and Xen HVM instances running on
3211 - KVM virtualization is supported too
3212 - file-based storage has been implemented, which means that it is
3213 possible to run the cluster without LVM and DRBD storage, for
3214 example using a shared filesystem exported from shared storage (and
3215 still have live migration)
3221 *(Released Tue, 13 Jan 2009)*
3223 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
3224 - Reuse the old instance mac address by default on instance import, if
3225 the instance name is the same.
3226 - Handle situations in which the node info rpc returns incomplete
3228 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
3229 - Improved version of batcher:
3231 - state file support
3232 - instance mac address support
3233 - support for HVM clusters/instances
3235 - Add an option to show the number of cpu sockets and nodes in
3237 - Support OSes that handle more than one version of the OS api (but do
3238 not change the current API in any other way)
3239 - Fix ``gnt-node migrate``
3240 - ``gnt-debug`` man page
3241 - Fixes various more typos and small issues
3242 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
3248 *(Released Wed, 24 Sep 2008)*
3250 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
3251 type of disk exported to fully virtualized instances.
3252 - provide access to the serial console of HVM instances
3253 - instance auto_balance flag, set by default. If turned off it will
3254 avoid warnings on cluster verify if there is not enough memory to fail
3255 over an instance. in the future it will prevent automatically failing
3256 it over when we will support that.
3257 - batcher tool for instance creation, see ``tools/README.batcher``
3258 - ``gnt-instance reinstall --select-os`` to interactively select a new
3259 operating system when reinstalling an instance.
3260 - when changing the memory amount on instance modify a check has been
3261 added that the instance will be able to start. also warnings are
3262 emitted if the instance will not be able to fail over, if auto_balance
3264 - documentation fixes
3265 - sync fields between ``gnt-instance list/modify/add/import``
3266 - fix a race condition in drbd when the sync speed was set after giving
3267 the device a remote peer.
3273 *(Released Tue, 22 Jul 2008)*
3275 - note: the allowed size and number of tags per object were reduced
3276 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
3277 - fixed twisted 8.x compatibility
3278 - fixed ``gnt-instance replace-disks`` with iallocator
3279 - add TCP keepalives on twisted connections to detect restarted nodes
3280 - disk increase support, see ``gnt-instance grow-disk``
3281 - implement bulk node/instance query for RAPI
3282 - add tags in node/instance listing (optional)
3283 - experimental migration (and live migration) support, read the man page
3284 for ``gnt-instance migrate``
3285 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
3286 has some small improvements in handling its state file
3292 *(Released Fri, 13 Jun 2008)*
3294 - Experimental readonly, REST-based remote API implementation;
3295 automatically started on master node, TCP port 5080, if enabled by
3296 ``--enable-rapi`` parameter to configure script.
3297 - Instance allocator support. Add and import instance accept a
3298 ``--iallocator`` parameter, and call that instance allocator to decide
3299 which node to use for the instance. The iallocator document describes
3300 what's expected from an allocator script.
3301 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
3302 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
3303 node is lost there is still enough memory to fail over the instances
3305 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
3306 ``gnt-cluster verify``, to check for site-specific compliance. All the
3307 hooks will run, and their output, if any, will be displayed. Any
3308 failing hook will make the verification return an error value.
3309 - ``gnt-cluster verify`` now checks that its peers are reachable on the
3310 primary and secondary interfaces
3311 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
3312 that is still declared as part of the cluster and has failed.
3313 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
3314 specifying output fields, that just adds the chosen fields to the
3316 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
3317 export from the filesystem.
3318 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
3319 have been added. Using them you can enable/disable acpi and pae
3320 support, and specify a path for a cd image to be exported to the
3321 instance. These parameters as the name suggest only work on HVM
3323 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
3324 PAE support will be set to the previously hardcoded values, but the
3325 (previously hardcoded) path to the CDROM ISO image will be unset and
3326 if required, needs to be set manually with ``gnt-instance modify``
3328 - The address to which an instance's VNC console is bound is now
3329 selectable per-instance, rather than being cluster wide. Of course
3330 this only applies to instances controlled via VNC, so currently just
3331 applies to HVM clusters.
3337 *(Released Mon, 18 Feb 2008)*
3339 - more tweaks to the disk activation code (especially helpful for DRBD)
3340 - change the default ``gnt-instance list`` output format, now there is
3341 one combined status field (see the manpage for the exact values this
3343 - some more fixes for the mac export to hooks change
3344 - make Ganeti not break with DRBD 8.2.x (which changed the version
3345 format in ``/proc/drbd``) (issue 24)
3346 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3347 template, allowing migration from DRBD0.7+MD to DRBD8
3353 *(Released Wed, 30 Jan 2008)*
3355 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3357 - add command aliases infrastructure and a few aliases
3358 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
3359 man pages and the ``--help`` option of ``gnt-node
3360 list``/``gnt-instance list``
3361 - fix ``gnt-backup list`` with down nodes (issue 21)
3362 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3363 - fix the dist archive and add a check for including svn/git files in
3365 - some developer-related changes: improve the burnin and the QA suite,
3366 add an upload script for testing during development
3372 *(Released Wed, 16 Jan 2008)*
3374 - experimental HVM support, read the install document, section
3375 "Initializing the cluster"
3376 - allow for the PVM hypervisor per-instance kernel and initrd paths
3377 - add a new command ``gnt-cluster verify-disks`` which uses a new
3378 algorithm to improve the reconnection of the DRBD pairs if the device
3379 on the secondary node has gone away
3380 - make logical volume code auto-activate LVs at disk activation time
3381 - slightly improve the speed of activating disks
3382 - allow specification of the MAC address at instance creation time, and
3383 changing it later via ``gnt-instance modify``
3384 - fix handling of external commands that generate lots of output on
3386 - update documentation with regard to minimum version of DRBD8 supported
3392 *(Released Tue, 4 Dec 2007)*
3394 - Log the ``xm create`` output to the node daemon log on failure (to
3395 help diagnosing the error)
3396 - In debug mode, log all external commands output if failed to the logs
3397 - Change parsing of lvm commands to ignore stderr
3403 *(Released Wed, 28 Nov 2007)*
3405 - Another round of updates to the DRBD 8 code to deal with more failures
3406 in the replace secondary node operation
3407 - Some more logging of failures in disk operations (lvm, drbd)
3408 - A few documentation updates
3415 *(Released Tue, 13 Nov 2007)*
3417 - Change configuration file format from Python's Pickle to JSON.
3418 Upgrading is possible using the cfgupgrade utility.
3419 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3420 faster replace disks and is more stable (DRBD 8 has many improvements
3421 compared to DRBD 0.7)
3422 - Added command line tags support (see man pages for ``gnt-instance``,
3423 ``gnt-node``, ``gnt-cluster``)
3424 - Added instance rename support
3425 - Added multi-instance startup/shutdown
3426 - Added cluster rename support
3427 - Added ``gnt-node evacuate`` to simplify some node operations
3428 - Added instance reboot operation that can speedup reboot as compared to
3430 - Soften the requirement that hostnames are in FQDN format
3431 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3433 - Removed dependency on debian's patched fping that uses the
3434 non-standard ``-S`` option
3435 - Now the OS definitions are searched for in multiple, configurable
3436 paths (easier for distros to package)
3437 - Some changes to the hooks infrastructure (especially the new
3438 post-configuration update hook)
3439 - Other small bugfixes
3441 .. vim: set textwidth=72 syntax=rst :